player.c

/*
 * Slave-clocked ALAC stream player. This file is part of Shairport.
 * Copyright (c) James Laird 2011, 2013
 * All rights reserved.
 *
 * Modifications for audio synchronisation
 * and related work, copyright (c) Mike Brady 2014 -- 2019
 * All rights reserved.
 *
 * Permission is hereby granted, free of charge, to any person
 * obtaining a copy of this software and associated documentation
 * files (the "Software"), to deal in the Software without
 * restriction, including without limitation the rights to use,
 * copy, modify, merge, publish, distribute, sublicense, and/or
 * sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be
 * included in all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
 * OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
 * HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
 * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
 * OTHER DEALINGS IN THE SOFTWARE.
 */

#include <assert.h>
#include <errno.h>
#include <fcntl.h>
#include <inttypes.h>
#include <limits.h>
#include <math.h>
#include <pthread.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
#include <sys/syslog.h>
#include <sys/types.h>
#include <unistd.h>

#include "config.h"

#ifdef CONFIG_MBEDTLS
#include <mbedtls/aes.h>
#include <mbedtls/havege.h>
#endif

#ifdef CONFIG_POLARSSL
#include <polarssl/aes.h>
#include <polarssl/havege.h>
#endif

#ifdef CONFIG_OPENSSL
#include <openssl/aes.h>
#endif

#ifdef CONFIG_SOXR
#include <soxr.h>
#endif

#ifdef CONFIG_CONVOLUTION
#include <FFTConvolver/convolver.h>
#endif

#ifdef CONFIG_METADATA_HUB
#include "metadata_hub.h"
#endif

#ifdef CONFIG_DACP_CLIENT
#include "dacp.h"
#endif

#include "common.h"
#include "mdns.h"
#include "player.h"
#include "rtp.h"
#include "rtsp.h"

#include "alac.h"

#ifdef CONFIG_APPLE_ALAC
#include "apple_alac.h"
#endif

#include "loudness.h"

#include "activity_monitor.h"

// default buffer size
// needs to be a power of 2 because of the way BUFIDX(seqno) works
//#define BUFFER_FRAMES 512
#define MAX_PACKET 2048

// DAC buffer occupancy stuff
#define DAC_BUFFER_QUEUE_MINIMUM_LENGTH 2500

// static abuf_t audio_buffer[BUFFER_FRAMES];
#define BUFIDX(seqno) ((seq_t)(seqno) % BUFFER_FRAMES)

uint32_t modulo_32_offset(uint32_t from, uint32_t to) {
  if (from <= to)
    return to - from;
  else
    return UINT32_MAX - from + to + 1;
}

uint64_t modulo_64_offset(uint64_t from, uint64_t to) {
  if (from <= to)
    return to - from;
  else
    return UINT64_MAX - from + to + 1;
}

void do_flush(uint32_t timestamp, rtsp_conn_info *conn);

static void ab_resync(rtsp_conn_info *conn) {
  int i;
  for (i = 0; i < BUFFER_FRAMES; i++) {
    conn->audio_buffer[i].ready = 0;
    conn->audio_buffer[i].resend_level = 0;
    conn->audio_buffer[i].sequence_number = 0;
  }
  conn->ab_synced = 0;
  conn->last_seqno_read = -1;
  conn->ab_buffering = 1;
}

// given starting and ending points as unsigned 32-bit integers running modulo 2^32, returns the
// position of x in the interval in *pos
// returns true if x is actually within the buffer

int position_in_modulo_uint32_t_buffer(uint32_t x, uint32_t start, uint32_t end, uint32_t *pos) {
  int response = 0; // not in the buffer
  if (start <= end) {
    if (x < start) {
      if (pos)
        *pos = UINT32_MAX - start + 1 + x;
    } else {
      if (pos)
        *pos = x - start;
      if (x < end)
        response = 1;
    }
  } else if ((x >= start) && (x <= UINT32_MAX)) {
    response = 1;
    if (pos)
      *pos = x - start;
  } else {
    if (pos)
      *pos = UINT32_MAX - start + 1 + x;
    if (x < end) {
      response = 1;
    }
  }
  return response;
}

// this is used.
static inline seq_t SUCCESSOR(seq_t x) {
  uint32_t p = x & 0xffff;
  p += 1;
  p = p & 0xffff;
  return p;
}

// used in seq_diff and seq_order

// anything with ORDINATE in it must be protected by the ab_mutex
static inline int32_t ORDINATE(seq_t x, seq_t base) {
  int32_t p = x;    // int32_t from seq_t, i.e. uint16_t, so okay
  int32_t q = base; // int32_t from seq_t, i.e. uint16_t, so okay
  int32_t t = (p + 0x10000 - q) & 0xffff;
  // we definitely will get a positive number in t at this point, but it might be a
  // positive alias of a negative number, i.e. x might actually be "before" ab_read
  // So, if the result is greater than 32767, we will assume its an
  // alias and subtract 65536 from it
  if (t >= 32767) {
    // debug(1,"OOB: %u, ab_r: %u, ab_w: %u",x,ab_read,ab_write);
    t -= 65536;
  }
  return t;
}

// wrapped number between two seq_t.
int32_t seq_diff(seq_t a, seq_t b, seq_t base) {
  int32_t diff = ORDINATE(b, base) - ORDINATE(a, base);
  return diff;
}

// the sequence numbers will wrap pretty often.
// this returns true if the second arg is after the first
static inline int seq_order(seq_t a, seq_t b, seq_t base) {
  int32_t d = ORDINATE(b, base) - ORDINATE(a, base);
  return d > 0;
}

static inline seq_t seq_sum(seq_t a, seq_t b) {
  //  uint32_t p = a & 0xffff;
  //  uint32_t q = b & 0x0ffff;
  uint32_t r = (a + b) & 0xffff;
  return r;
}

void reset_input_flow_metrics(rtsp_conn_info *conn) {
  conn->play_number_after_flush = 0;
  conn->packet_count_since_flush = 0;
  conn->input_frame_rate_starting_point_is_valid = 0;
  conn->initial_reference_time = 0;
  conn->initial_reference_timestamp = 0;
}

void unencrypted_packet_decode(unsigned char *packet, int length, short *dest, int *outsize,
                               int size_limit, rtsp_conn_info *conn) {
  if (conn->stream.type == ast_apple_lossless) {
#ifdef CONFIG_APPLE_ALAC
    if (config.use_apple_decoder) {
      if (conn->decoder_in_use != 1 << decoder_apple_alac) {
        debug(2, "Apple ALAC Decoder used on encrypted audio.");
        conn->decoder_in_use = 1 << decoder_apple_alac;
      }
      apple_alac_decode_frame(packet, length, (unsigned char *)dest, outsize);
      *outsize = *outsize * 4; // bring the size to bytes
    } else
#endif
    {
      if (conn->decoder_in_use != 1 << decoder_hammerton) {
        debug(2, "Hammerton Decoder used on encrypted audio.");
        conn->decoder_in_use = 1 << decoder_hammerton;
      }
      alac_decode_frame(conn->decoder_info, packet, (unsigned char *)dest, outsize);
    }
  } else if (conn->stream.type == ast_uncompressed) {
    int length_to_use = length;
    if (length_to_use > size_limit) {
      warn("unencrypted_packet_decode: uncompressed audio packet too long (size: %d bytes) to "
           "process -- truncated",
           length);
      length_to_use = size_limit;
    }
    int i;
    short *source = (short *)packet;
    for (i = 0; i < (length_to_use / 2); i++) {
      *dest = ntohs(*source);
      dest++;
      source++;
    }
    *outsize = length_to_use;
  }
}

int audio_packet_decode(short *dest, int *destlen, uint8_t *buf, int len, rtsp_conn_info *conn) {
  // parameters: where the decoded stuff goes, its length in samples,
  // the incoming packet, the length of the incoming packet in bytes
  // destlen should contain the allowed max number of samples on entry

  if (len > MAX_PACKET) {
    warn("Incoming audio packet size is too large at %d; it should not exceed %d.", len,
         MAX_PACKET);
    return -1;
  }
  unsigned char packet[MAX_PACKET];
  // unsigned char packetp[MAX_PACKET];
  assert(len <= MAX_PACKET);
  int reply = 0;                                          // everything okay
  int outsize = conn->input_bytes_per_frame * (*destlen); // the size the output should be, in bytes
  int maximum_possible_outsize = outsize;

  if (conn->stream.encrypted) {
    unsigned char iv[16];
    int aeslen = len & ~0xf;
    memcpy(iv, conn->stream.aesiv, sizeof(iv));
#ifdef CONFIG_MBEDTLS
    mbedtls_aes_crypt_cbc(&conn->dctx, MBEDTLS_AES_DECRYPT, aeslen, iv, buf, packet);
#endif
#ifdef CONFIG_POLARSSL
    aes_crypt_cbc(&conn->dctx, AES_DECRYPT, aeslen, iv, buf, packet);
#endif
#ifdef CONFIG_OPENSSL
    AES_cbc_encrypt(buf, packet, aeslen, &conn->aes, iv, AES_DECRYPT);
#endif
    memcpy(packet + aeslen, buf + aeslen, len - aeslen);
    unencrypted_packet_decode(packet, len, dest, &outsize, maximum_possible_outsize, conn);
  } else {
    // not encrypted
    unencrypted_packet_decode(buf, len, dest, &outsize, maximum_possible_outsize, conn);
  }

  if (outsize > maximum_possible_outsize) {
    debug(2, "Output from alac_decode larger (%d bytes, not frames) than expected (%d bytes) -- "
             "truncated, but buffer overflow possible! Encrypted = %d.",
          outsize, maximum_possible_outsize, conn->stream.encrypted);
    reply = -1; // output packet is the wrong size
  }

  *destlen = outsize / conn->input_bytes_per_frame;
  if ((outsize % conn->input_bytes_per_frame) != 0)
    debug(1, "Number of audio frames (%d) does not correspond exactly to the number of bytes (%d) "
             "and the audio frame size (%d).",
          *destlen, outsize, conn->input_bytes_per_frame);
  return reply;
}

static int init_alac_decoder(int32_t fmtp[12], rtsp_conn_info *conn) {

  // This is a guess, but the format of the fmtp looks identical to the format of an
  // ALACSpecificCOnfig
  // which is detailed in the file ALACMagicCookieDescription.txt in the Apple ALAC sample
  // implementation
  // Here it is:

  /*
      struct	ALACSpecificConfig (defined in ALACAudioTypes.h)
      abstract   	This struct is used to describe codec provided information about the encoded
     Apple Lossless bitstream.
                  It must accompany the encoded stream in the containing audio file and be provided
     to the decoder.

      field      	frameLength 		uint32_t	indicating the frames per packet
     when
     no
     explicit
     frames per packet setting is
                                                            present in the packet header. The
     encoder frames per packet can be explicitly set
                                                            but for maximum compatibility, the
     default encoder setting of 4096 should be used.

      field      	compatibleVersion 	uint8_t 	indicating compatible version,
                                                            value must be set to 0

      field      	bitDepth 		uint8_t 	describes the bit depth of the
     source
     PCM
     data
     (maximum
     value = 32)

      field      	pb 			uint8_t 	currently unused tuning
     parametetbugr.
                                                            value should be set to 40

      field      	mb 			uint8_t 	currently unused tuning parameter.
                                                            value should be set to 14

      field      	kb			uint8_t 	currently unused tuning parameter.
                                                            value should be set to 10

      field      	numChannels 		uint8_t 	describes the channel count (1 =
     mono,
     2
     =
     stereo,
     etc...)
                                                            when channel layout info is not provided
     in the 'magic cookie', a channel count > 2
                                                            describes a set of discreet channels
     with no specific ordering

      field      	maxRun			uint16_t 	currently unused.
                                                    value should be set to 255

      field      	maxFrameBytes 		uint32_t 	the maximum size of an Apple
     Lossless
     packet
     within
     the encoded stream.
                                                          value of 0 indicates unknown

      field      	avgBitRate 		uint32_t 	the average bit rate in bits per
     second
     of
     the
     Apple
     Lossless stream.
                                                          value of 0 indicates unknown

      field      	sampleRate 		uint32_t 	sample rate of the encoded stream
   */

  // We are going to go on that basis

  alac_file *alac;

  alac = alac_create(conn->input_bit_depth,
                     conn->input_num_channels); // no pthread cancellation point in here
  if (!alac)
    return 1;
  conn->decoder_info = alac;

  alac->setinfo_max_samples_per_frame = conn->max_frames_per_packet;
  alac->setinfo_7a = fmtp[2];
  alac->setinfo_sample_size = conn->input_bit_depth;
  alac->setinfo_rice_historymult = fmtp[4];
  alac->setinfo_rice_initialhistory = fmtp[5];
  alac->setinfo_rice_kmodifier = fmtp[6];
  alac->setinfo_7f = fmtp[7];
  alac->setinfo_80 = fmtp[8];
  alac->setinfo_82 = fmtp[9];
  alac->setinfo_86 = fmtp[10];
  alac->setinfo_8a_rate = fmtp[11];
  alac_allocate_buffers(alac); // no pthread cancellation point in here

#ifdef CONFIG_APPLE_ALAC
  apple_alac_init(fmtp); // no pthread cancellation point in here
#endif

  return 0;
}

static void terminate_decoders(rtsp_conn_info *conn) {
  alac_free(conn->decoder_info);
#ifdef CONFIG_APPLE_ALAC
  apple_alac_terminate();
#endif
}

static void init_buffer(rtsp_conn_info *conn) {
  int i;
  for (i = 0; i < BUFFER_FRAMES; i++)
    conn->audio_buffer[i].data = malloc(conn->input_bytes_per_frame * conn->max_frames_per_packet);
  ab_resync(conn);
}

static void free_audio_buffers(rtsp_conn_info *conn) {
  int i;
  for (i = 0; i < BUFFER_FRAMES; i++)
    free(conn->audio_buffer[i].data);
}

void player_put_packet(seq_t seqno, uint32_t actual_timestamp, uint8_t *data, int len,
                       rtsp_conn_info *conn) {

  // ignore a request to flush that has been made before the first packet...
  if (conn->packet_count == 0) {
    debug_mutex_lock(&conn->flush_mutex, 1000, 1);
    conn->flush_requested = 0;
    conn->flush_rtp_timestamp = 0;
    debug_mutex_unlock(&conn->flush_mutex, 3);
  }

  debug_mutex_lock(&conn->ab_mutex, 30000, 0);
  conn->packet_count++;
  conn->packet_count_since_flush++;
  conn->time_of_last_audio_packet = get_absolute_time_in_fp();
  if (conn->connection_state_to_output) { // if we are supposed to be processing these packets

    if ((conn->flush_rtp_timestamp != 0) && (actual_timestamp != conn->flush_rtp_timestamp) &&
        (modulo_32_offset(actual_timestamp, conn->flush_rtp_timestamp) <
         conn->input_rate * 10)) { // if it's less than 10 seconds
      debug(3, "Dropping flushed packet in player_put_packet, seqno %u, timestamp %" PRIu32
               ", flushing to "
               "timestamp: %" PRIu32 ".",
            seqno, actual_timestamp, conn->flush_rtp_timestamp);
      conn->initial_reference_time = 0;
      conn->initial_reference_timestamp = 0;
    } else {
      /*
        if ((conn->flush_rtp_timestamp != 0) &&
            (modulo_32_offset(conn->flush_rtp_timestamp, actual_timestamp) > conn->input_rate / 5)
        &&
            (modulo_32_offset(conn->flush_rtp_timestamp, actual_timestamp) < conn->input_rate)) {
          // between 0.2 and 1 second
          debug(2, "Dropping flush request in player_put_packet");
          conn->flush_rtp_timestamp = 0;
        }
      */

      abuf_t *abuf = 0;

      if (!conn->ab_synced) {
        // if this is the first packet...
        debug(3, "syncing to seqno %u.", seqno);
        conn->ab_write = seqno;
        conn->ab_read = seqno;
        conn->ab_synced = 1;
      }

      // here, we should check for missing frames
      int resend_interval = (((250 * 44100) / 352) / 1000); // approximately 250 ms intervals
      const int number_of_resend_attempts = 8;
      int latency_based_resend_interval =
          (conn->latency) / (number_of_resend_attempts * conn->max_frames_per_packet);
      if (latency_based_resend_interval > resend_interval)
        resend_interval = latency_based_resend_interval;

      if (conn->resend_interval != resend_interval) {
        debug(2, "Resend interval for latency of %u frames is %d frames.", conn->latency,
              resend_interval);
        conn->resend_interval = resend_interval;
      }

      if (conn->ab_write ==
          seqno) { // if this is the expected packet (which could be the first packet...)
        uint64_t reception_time = get_absolute_time_in_fp();
        if (conn->input_frame_rate_starting_point_is_valid == 0) {
          if ((conn->packet_count_since_flush >= 500) && (conn->packet_count_since_flush <= 510)) {
            conn->frames_inward_measurement_start_time = reception_time;
            conn->frames_inward_frames_received_at_measurement_start_time = actual_timestamp;
            conn->input_frame_rate_starting_point_is_valid = 1; // valid now
          }
        }

        conn->frames_inward_measurement_time = reception_time;
        conn->frames_inward_frames_received_at_measurement_time = actual_timestamp;

        abuf = conn->audio_buffer + BUFIDX(seqno);
        conn->ab_write = SUCCESSOR(seqno); // move the write pointer to the next free space
      } else if (seq_order(conn->ab_write, seqno, conn->ab_read)) { // newer than expected
        // if (ORDINATE(seqno)>(BUFFER_FRAMES*7)/8)
        // debug(1,"An interval of %u frames has opened, with ab_read: %u, ab_write: %u and
        // seqno:
        // %u.",seq_diff(ab_read,seqno),ab_read,ab_write,seqno);
        int32_t gap = seq_diff(conn->ab_write, seqno, conn->ab_read);
        if (gap <= 0)
          debug(1, "Unexpected gap size: %d.", gap);
        int i;
        for (i = 0; i < gap; i++) {
          abuf = conn->audio_buffer + BUFIDX(seq_sum(conn->ab_write, i));
          abuf->ready = 0; // to be sure, to be sure
          abuf->resend_level = 0;
          // abuf->timestamp = 0;
          abuf->given_timestamp = 0;
          abuf->sequence_number = 0;
        }
        // debug(1,"N %d s %u.",seq_diff(ab_write,PREDECESSOR(seqno))+1,ab_write);
        abuf = conn->audio_buffer + BUFIDX(seqno);
        //        rtp_request_resend(ab_write, gap);
        //        resend_requests++;
        conn->ab_write = SUCCESSOR(seqno);
      } else if (seq_order(conn->ab_read, seqno, conn->ab_read)) { // late but not yet played
        conn->late_packets++;
        abuf = conn->audio_buffer + BUFIDX(seqno);
        /*
        if (abuf->ready)
                debug(1,"Late apparently duplicate packet received that is %d packets
        late.",seq_diff(seqno, conn->ab_write, conn->ab_read));
        else
                debug(1,"Late packet received that is %d packets late.",seq_diff(seqno,
        conn->ab_write, conn->ab_read));
              */
      } else { // too late.

        // debug(1,"Too late packet received that is %d packets late.",seq_diff(seqno,
        // conn->ab_write, conn->ab_read));
        conn->too_late_packets++;
      }
      // pthread_mutex_unlock(&ab_mutex);

      if (abuf) {
        int datalen = conn->max_frames_per_packet;
        if (audio_packet_decode(abuf->data, &datalen, data, len, conn) == 0) {
          abuf->ready = 1;
          abuf->length = datalen;
          // abuf->timestamp = ltimestamp;
          abuf->given_timestamp = actual_timestamp;
          abuf->sequence_number = seqno;
        } else {
          debug(1, "Bad audio packet detected and discarded.");
          abuf->ready = 0;
          abuf->resend_level = 0;
          // abuf->timestamp = 0;
          abuf->given_timestamp = 0;
          abuf->sequence_number = 0;
        }
      }

      // pthread_mutex_lock(&ab_mutex);
      int rc = pthread_cond_signal(&conn->flowcontrol);
      if (rc)
        debug(1, "Error signalling flowcontrol.");

      // if it's at the expected time, do a look back for missing packets
      // but release the ab_mutex when doing a resend
      if (!conn->ab_buffering) {
        int j;
        for (j = 1; j <= number_of_resend_attempts; j++) {
          // check j times, after a short period of has elapsed, assuming 352 frames per packet
          // the higher the step_exponent, the less it will try. 1 means it will try very
          // hard. 2.0 seems good.
          float step_exponent = 2.0;
          int back_step = (int)(resend_interval * pow(j, step_exponent));
          int k;
          for (k = -1; k <= 1; k++) {
            if ((back_step + k) <
                seq_diff(conn->ab_read, conn->ab_write,
                         conn->ab_read)) { // if it's within the range of frames in use...
              int item_to_check = (conn->ab_write - (back_step + k)) & 0xffff;
              seq_t next = item_to_check;
              abuf_t *check_buf = conn->audio_buffer + BUFIDX(next);
              if ((!check_buf->ready) &&
                  (check_buf->resend_level <
                   j)) { // prevent multiple requests from the same level of lookback
                check_buf->resend_level = j;
                if (config.disable_resend_requests == 0) {
                  debug_mutex_unlock(&conn->ab_mutex, 3);
                  rtp_request_resend(next, 1, conn);
                  conn->resend_requests++;
                  debug_mutex_lock(&conn->ab_mutex, 20000, 1);
                }
              }
            }
          }
        }
      }
    }
  }
  debug_mutex_unlock(&conn->ab_mutex, 0);
}

int32_t rand_in_range(int32_t exclusive_range_limit) {
  static uint32_t lcg_prev = 12345;
  // returns a pseudo random integer in the range 0 to (exclusive_range_limit-1) inclusive
  int64_t sp = lcg_prev;
  int64_t rl = exclusive_range_limit;
  lcg_prev = lcg_prev * 69069 + 3; // crappy psrg
  sp = sp * rl; // 64 bit calculation. Interesting part is above the 32 rightmost bits;
  return sp >> 32;
}

static inline void process_sample(int32_t sample, char **outp, enum sps_format_t format, int volume,
                                  int dither, rtsp_conn_info *conn) {
  /*
  {
                static int old_volume = 0;
                if (volume != old_volume) {
                        debug(1,"Volume is now %d.",volume);
                        old_volume = volume;
                }
  }
  */

  int64_t hyper_sample = sample;
  int result = 0;

  if (config.loudness) {
    hyper_sample <<=
        32; // Do not apply volume as it has already been done with the Loudness DSP filter
  } else {
    int64_t hyper_volume = (int64_t)volume << 16;
    hyper_sample = hyper_sample * hyper_volume; // this is 64 bit bit multiplication -- we may need
                                                // to dither it down to its target resolution
  }

  // next, do dither, if necessary
  if (dither) {

    // add a TPDF dither -- see
    // http://educypedia.karadimov.info/library/DitherExplained.pdf
    // and the discussion around https://www.hydrogenaud.io/forums/index.php?showtopic=16963&st=25

    // I think, for a 32 --> 16 bits, the range of
    // random numbers needs to be from -2^16 to 2^16, i.e. from -65536 to 65536 inclusive, not from
    // -32768 to +32767

    // Actually, what would be generated here is from -65535 to 65535, i.e. one less on the limits.

    // See the original paper at
    // http://www.ece.rochester.edu/courses/ECE472/resources/Papers/Lipshitz_1992.pdf
    // by Lipshitz, Wannamaker and Vanderkooy, 1992.

    int64_t dither_mask = 0;
    switch (format) {
    case SPS_FORMAT_S32:
    case SPS_FORMAT_S32_LE:
    case SPS_FORMAT_S32_BE:
      dither_mask = (int64_t)1 << (64 - 32);
      break;
    case SPS_FORMAT_S24:
    case SPS_FORMAT_S24_LE:
    case SPS_FORMAT_S24_BE:
    case SPS_FORMAT_S24_3LE:
    case SPS_FORMAT_S24_3BE:
      dither_mask = (int64_t)1 << (64 - 24);
      break;
    case SPS_FORMAT_S16:
    case SPS_FORMAT_S16_LE:
    case SPS_FORMAT_S16_BE:
      dither_mask = (int64_t)1 << (64 - 16);
      break;
    case SPS_FORMAT_S8:
    case SPS_FORMAT_U8:
      dither_mask = (int64_t)1 << (64 - 8);
      break;
    case SPS_FORMAT_UNKNOWN:
      die("Unexpected SPS_FORMAT_UNKNOWN while calculating dither mask.");
      break;
    case SPS_FORMAT_AUTO:
      die("Unexpected SPS_FORMAT_AUTO while calculating dither mask.");
      break;
    case SPS_FORMAT_INVALID:
      die("Unexpected SPS_FORMAT_INVALID while calculating dither mask.");
      break;
    }
    dither_mask -= 1;
    // int64_t r = r64i();
    int64_t r = ranarray64i(); // use an array of precalculated pseudorandom numbers rather than
                               // calculating them on the fly. Should be easier on low-powered
                               // processors

    int64_t tpdf = (r & dither_mask) - (conn->previous_random_number & dither_mask);
    conn->previous_random_number = r;
    // add dither, allowing for clipping
    if (tpdf >= 0) {
      if (INT64_MAX - tpdf >= hyper_sample)
        hyper_sample += tpdf;
      else
        hyper_sample = INT64_MAX;
    } else {
      if (INT64_MIN - tpdf <= hyper_sample)
        hyper_sample += tpdf;
      else
        hyper_sample = INT64_MIN;
    }
    // dither is complete here
  }

  // move the result to the desired position in the int64_t
  char *op = *outp;
  uint8_t byt;
  switch (format) {
  case SPS_FORMAT_S32_LE:
    hyper_sample >>= (64 - 32);
    byt = (uint8_t)hyper_sample;
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 8);
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 16);
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 24);
    *op++ = byt;
    result = 4;
    break;
  case SPS_FORMAT_S32_BE:
    hyper_sample >>= (64 - 32);
    byt = (uint8_t)(hyper_sample >> 24);
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 16);
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 8);
    *op++ = byt;
    byt = (uint8_t)hyper_sample;
    *op++ = byt;
    result = 4;
    break;
  case SPS_FORMAT_S32:
    hyper_sample >>= (64 - 32);
    *(int32_t *)op = hyper_sample;
    result = 4;
    break;
  case SPS_FORMAT_S24_3LE:
    hyper_sample >>= (64 - 24);
    byt = (uint8_t)hyper_sample;
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 8);
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 16);
    *op++ = byt;
    result = 3;
    break;
  case SPS_FORMAT_S24_3BE:
    hyper_sample >>= (64 - 24);
    byt = (uint8_t)(hyper_sample >> 16);
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 8);
    *op++ = byt;
    byt = (uint8_t)hyper_sample;
    *op++ = byt;
    result = 3;
    break;
  case SPS_FORMAT_S24_LE:
    hyper_sample >>= (64 - 24);
    byt = (uint8_t)hyper_sample;
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 8);
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 16);
    *op++ = byt;
    *op++ = 0;
    result = 4;
    break;
  case SPS_FORMAT_S24_BE:
    hyper_sample >>= (64 - 24);
    *op++ = 0;
    byt = (uint8_t)(hyper_sample >> 16);
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 8);
    *op++ = byt;
    byt = (uint8_t)hyper_sample;
    *op++ = byt;
    result = 4;
    break;
  case SPS_FORMAT_S24:
    hyper_sample >>= (64 - 24);
    *(int32_t *)op = hyper_sample;
    result = 4;
    break;
  case SPS_FORMAT_S16_LE:
    hyper_sample >>= (64 - 16);
    byt = (uint8_t)hyper_sample;
    *op++ = byt;
    byt = (uint8_t)(hyper_sample >> 8);
    *op++ = byt;
    result = 2;
    break;
  case SPS_FORMAT_S16_BE:
    hyper_sample >>= (64 - 16);
    byt = (uint8_t)(hyper_sample >> 8);
    *op++ = byt;
    byt = (uint8_t)hyper_sample;
    *op++ = byt;
    result = 2;
    break;
  case SPS_FORMAT_S16:
    hyper_sample >>= (64 - 16);
    *(int16_t *)op = (int16_t)hyper_sample;
    result = 2;
    break;
  case SPS_FORMAT_S8:
    hyper_sample >>= (int8_t)(64 - 8);
    *op = hyper_sample;
    result = 1;
    break;
  case SPS_FORMAT_U8:
    hyper_sample >>= (uint8_t)(64 - 8);
    hyper_sample += 128;
    *op = hyper_sample;
    result = 1;
    break;
  case SPS_FORMAT_UNKNOWN:
    die("Unexpected SPS_FORMAT_UNKNOWN while outputting samples");
    break;
  case SPS_FORMAT_AUTO:
    die("Unexpected SPS_FORMAT_AUTO while outputting samples");
    break;
  case SPS_FORMAT_INVALID:
    die("Unexpected SPS_FORMAT_INVALID while outputting samples");
    break;
  }

  *outp += result;
}

void buffer_get_frame_cleanup_handler(void *arg) {
  rtsp_conn_info *conn = (rtsp_conn_info *)arg;
  debug_mutex_unlock(&conn->ab_mutex, 0);
}

// get the next frame, when available. return 0 if underrun/stream reset.
static abuf_t *buffer_get_frame(rtsp_conn_info *conn) {
  // int16_t buf_fill;
  uint64_t local_time_now;
  // struct timespec tn;
  abuf_t *curframe = NULL;
  int notified_buffer_empty = 0; // diagnostic only

  debug_mutex_lock(&conn->ab_mutex, 30000, 0);

  int wait;
  long dac_delay = 0; // long because alsa returns a long

  pthread_cleanup_push(buffer_get_frame_cleanup_handler,
                       (void *)conn); // undo what's been done so far
  do {
    // get the time
    local_time_now = get_absolute_time_in_fp(); // type okay
    // debug(3, "buffer_get_frame is iterating");

    int rco = get_requested_connection_state_to_output();

    if (conn->connection_state_to_output != rco) {
      conn->connection_state_to_output = rco;
      // change happening
      if (conn->connection_state_to_output == 0) { // going off
        debug_mutex_lock(&conn->flush_mutex, 1000, 1);
        conn->flush_requested = 1;
        debug_mutex_unlock(&conn->flush_mutex, 3);
      }
    }

    if (config.output->is_running)
      if (config.output->is_running() != 0) { // if the back end isn't running for any reason
        debug(3, "not running");
        debug_mutex_lock(&conn->flush_mutex, 1000, 0);
        conn->flush_requested = 1;
        debug_mutex_unlock(&conn->flush_mutex, 0);
      }

    debug_mutex_lock(&conn->flush_mutex, 1000, 0);
    if (conn->flush_requested == 1) {
      if (config.output->flush)
        config.output->flush(); // no cancellation points
      ab_resync(conn);          // no cancellation points
      conn->first_packet_timestamp = 0;
      conn->first_packet_time_to_play = 0;
      conn->time_since_play_started = 0;
      conn->flush_requested = 0;
    }
    debug_mutex_unlock(&conn->flush_mutex, 0);

    if (conn->ab_synced) {
      curframe = conn->audio_buffer + BUFIDX(conn->ab_read);

      if ((conn->ab_read != conn->ab_write) &&
          (curframe->ready)) { // it could be synced and empty, under
                               // exceptional circumstances, with the
                               // frame unused, thus apparently ready

        if (curframe->sequence_number != conn->ab_read) {
          // some kind of sync problem has occurred.
          if (BUFIDX(curframe->sequence_number) == BUFIDX(conn->ab_read)) {
            // it looks like some kind of aliasing has happened
            if (seq_order(conn->ab_read, curframe->sequence_number, conn->ab_read)) {
              conn->ab_read = curframe->sequence_number;
              debug(1, "Aliasing of buffer index -- reset.");
            }
          } else {
            debug(1, "Inconsistent sequence numbers detected");
          }
        }

        // if (conn->flush_rtp_timestamp != 0)
        //  debug(2,"flush_rtp_timestamp is %" PRIx32 " and curframe->given_timestamp is %" PRIx32
        //  ".", conn->flush_rtp_timestamp , curframe->given_timestamp);

        if ((conn->flush_rtp_timestamp != 0) &&
            (curframe->given_timestamp != conn->flush_rtp_timestamp) &&
            (modulo_32_offset(curframe->given_timestamp, conn->flush_rtp_timestamp) <
             conn->input_rate * 10)) { // if it's less than ten seconds
          debug(3, "Dropping flushed packet in buffer_get_frame, seqno %u, timestamp %" PRIu32
                   ", flushing to "
                   "timestamp: %" PRIu32 ".",
                curframe->sequence_number, curframe->given_timestamp, conn->flush_rtp_timestamp);
          curframe->ready = 0;
          curframe->resend_level = 0;
          curframe = NULL; // this will be returned and will cause the loop to go around again
          conn->initial_reference_time = 0;
          conn->initial_reference_timestamp = 0;
        } else if ((conn->flush_rtp_timestamp != 0) &&
                   (modulo_32_offset(conn->flush_rtp_timestamp, curframe->given_timestamp) >
                    conn->input_rate / 5) &&
                   (modulo_32_offset(conn->flush_rtp_timestamp, curframe->given_timestamp) <
                    conn->input_rate * 10)) {
          debug(3, "Dropping flush request in buffer_get_frame");
          conn->flush_rtp_timestamp = 0;
        }
      }

      if ((curframe) && (curframe->ready)) {
        notified_buffer_empty = 0; // at least one buffer now -- diagnostic only.
        if (conn->ab_buffering) {  // if we are getting packets but not yet forwarding them to the
                                   // player
          int have_sent_prefiller_silence = 1; // set true when we have sent some silent frames to
                                               // the DAC
          /*
          int64_t reference_timestamp;
          uint64_t reference_timestamp_time, remote_reference_timestamp_time;
          get_reference_timestamp_stuff(&reference_timestamp, &reference_timestamp_time,
                                        &remote_reference_timestamp_time, conn);
          reference_timestamp *= conn->output_sample_ratio;
          */
          if (conn->first_packet_timestamp == 0) { // if this is the very first packet
                                                   // debug(1,"First frame seen, time %u, with %d
            // frames...",curframe->timestamp,seq_diff(ab_read, ab_write));

            if (have_timestamp_timing_information(conn)) { // if we have a reference time
              // debug(1,"First frame seen with timestamp...");
              conn->first_packet_timestamp =
                  curframe->given_timestamp; // we will keep buffering until we are
                                             // supposed to start playing this
              have_sent_prefiller_silence = 0;

// debug(1, "First packet timestamp is %" PRId64 ".", conn->first_packet_timestamp);

// say we have started playing here
#ifdef CONFIG_METADATA
              debug(2, "pffr");
              send_ssnc_metadata(
                  'pffr', NULL, 0,
                  0); // "first frame received", but don't wait if the queue is locked
#endif
              // Here, calculate when we should start playing. We need to know when to allow the
              // packets to be sent to the player.
              // We will send packets of silence from now until that time and then we will send the
              // first packet, which will be followed by the subsequent packets.

              // we will get a fix every second or so, which will be stored as a pair consisting of
              // the time when the packet with a particular timestamp should be played, neglecting
              // latencies, etc.

              // It probably won't  be the timestamp of our first packet, however, so we might have
              // to do some calculations.

              // To calculate when the first packet will be played, we figure out the exact time the
              // packet should be played according to its timestamp and the reference time.
              // We then need to add the desired latency, typically 88200 frames.

              // Then we need to offset this by the backend latency offset. For example, if we knew
              // that the audio back end has a latency of 100 ms, we would
              // ask for the first packet to be emitted 100 ms earlier than it should, i.e. -4410
              // frames, so that when it got through the audio back end,
              // if would be in sync. To do this, we would give it a latency offset of -100 ms, i.e.
              // -4410 frames.

              // debug(1, "Output sample ratio is %d", conn->output_sample_ratio);

              // what we are asking for here is "what is the local time at which time the calculated
              // frame should be played"

              uint64_t should_be_time;
              frame_to_local_time(conn->first_packet_timestamp + conn->latency +
                                      (uint32_t)(config.audio_backend_latency_offset *
                                                 conn->input_rate), // this will go modulo 2^32
                                  &should_be_time,
                                  conn);

              conn->first_packet_time_to_play = should_be_time;

              if (local_time_now > conn->first_packet_time_to_play) {
                uint64_t lateness = local_time_now - conn->first_packet_time_to_play;
                lateness = (lateness * 1000000) >> 32; // microseconds
                debug(3, "First packet is %" PRIu64 " microseconds late! Flushing 0.5 seconds",
                      lateness);
                do_flush(conn->first_packet_timestamp + 5 * 4410, conn);
              }
            }
          }

          if (conn->first_packet_time_to_play != 0) {
            // recalculate conn->first_packet_time_to_play -- the latency might change

            uint64_t should_be_time;
            frame_to_local_time(conn->first_packet_timestamp + conn->latency +
                                    (uint32_t)(config.audio_backend_latency_offset *
                                               conn->input_rate), // this should go modulo 2^32
                                &should_be_time,
                                conn);

            conn->first_packet_time_to_play = should_be_time;

            // we want the frames of silence sent at the start to be fairly large in case the output
            // device's minimum buffer size is large. But they can't be greater than the silent
            // lead_in time
            // which is either the agreed latency or the silent lead-in time specified by the
            // setting
            // In fact, if should be some fraction of them, to allow for adjustment.

            int64_t max_dac_delay = conn->latency;
            if (config.audio_backend_silent_lead_in_time >= 0)
              max_dac_delay =
                  (int64_t)(config.audio_backend_silent_lead_in_time * conn->input_rate);

            max_dac_delay = max_dac_delay / 4;

            // debug(1,"max_dac_delay is %" PRIu64 " frames.", max_dac_delay);

            int64_t filler_size = max_dac_delay;

            if (local_time_now > conn->first_packet_time_to_play) {
              uint64_t lateness = local_time_now - conn->first_packet_time_to_play;
              lateness = (lateness * 1000000) >> 32; // microseconds
              debug(3, "Gone past starting time by %" PRIu64 " microseconds.", lateness);
              have_sent_prefiller_silence = 1;
              conn->ab_buffering = 0;

              // we've gone past the time...
              // debug(1,"Run past the exact start time by %llu frames, with time now of %llx, fpttp
              // of %llx and dac_delay of %d and %d packets;
              // flush.",(((tn-conn->first_packet_time_to_play)*config.output_rate)>>32)+dac_delay,tn,conn->first_packet_time_to_play,dac_delay,seq_diff(ab_read,
              // ab_write));

              /*
                            if (config.output->flush)
                              config.output->flush();
                            ab_resync(conn);
                            conn->first_packet_timestamp = 0;
                            conn->first_packet_time_to_play = 0;
                            conn->time_since_play_started = 0;
              */
            } else {
              // do some calculations
              int64_t lead_time = conn->first_packet_time_to_play - local_time_now;
              // an audio_backend_silent_lead_in_time of less than zero means start filling ASAP
              int64_t lead_in_time = -1;
              if (config.audio_backend_silent_lead_in_time >= 0)
                lead_in_time =
                    (int64_t)(config.audio_backend_silent_lead_in_time * (int64_t)0x100000000);
              // debug(1,"Lead time is %llx at fpttp
              // %llx.",lead_time,conn->first_packet_time_to_play);
              if ((lead_in_time < 0) || (lead_time <= lead_in_time)) {
                // debug(1,"Lead time is %" PRIx64 ", lead-in time is %" PRIx64 " at fpttp
                // %llx.",lead_time,conn->first_packet_time_to_play);

                // debug(1,"Checking");
                if (config.output->delay) {
                  // conn->first_packet_time_to_play is definitely later than local_time_now
                  int resp = 0;
                  dac_delay = 0;
                  if (have_sent_prefiller_silence != 0)
                    resp = config.output->delay(&dac_delay);
                  if (resp == 0) {
                    int64_t gross_frame_gap =
                        ((conn->first_packet_time_to_play - local_time_now) * config.output_rate) >>
                        32;
                    int64_t exact_frame_gap = gross_frame_gap - dac_delay;
                    // debug(1,"Exact and gross frame gaps are %" PRId64 " and %" PRId64 " frames,
                    // and the dac delay is %ld.", exact_frame_gap, gross_frame_gap, dac_delay);
                    if (exact_frame_gap < 0) {
                      // we've gone past the time...
                      // debug(1,"Run past time.");

                      // this might happen if a big clock adjustment was made at just the wrong
                      // time.

                      debug(1, "Run a bit past the exact start time by %" PRId64
                               " frames with a DAC delay of %ld frames.",
                            -exact_frame_gap, dac_delay);
                      if (config.output->flush)
                        config.output->flush();
                      ab_resync(conn);
                      conn->first_packet_timestamp = 0;
                      conn->first_packet_time_to_play = 0;
                    } else {
                      int64_t fs = filler_size;
                      if (fs > (max_dac_delay - dac_delay))
                        fs = max_dac_delay - dac_delay;
                      if (fs < 0) {
                        // this could happen if the dac delay mysteriously grows between samples,
                        // which could happen in a transition between having no interpolation and
                        // having interpolated buffer numbers.

                        // this will happen benignly if standby is being prevented, because a
                        // thread in the alsa back end will be stuffing frames of silence in there
                        // to keep it busy

                        debug(3,
                              "frame size (fs) < 0 with max_dac_delay of %lld and dac_delay of %ld",
                              max_dac_delay, dac_delay);
                        fs = 0;
                      }
                      if ((exact_frame_gap <= fs) ||
                          (exact_frame_gap <= conn->max_frames_per_packet * 2)) {
                        fs = exact_frame_gap;
                        // debug(1,"Exact frame gap is %llu; play %d frames of silence. Dac_delay is
                        // %d,
                        // with %d packets, ab_read is %04x, ab_write is
                        // %04x.",exact_frame_gap,fs,dac_delay,seq_diff(ab_read,
                        // ab_write),ab_read,ab_write);
                        conn->ab_buffering = 0;
                      }
                      void *silence;
                      // if (fs==0)
                      //  debug(2,"Zero length silence buffer needed with gross_frame_gap of %lld
                      //  and
                      //  dac_delay of %lld.",gross_frame_gap,dac_delay);
                      // the fs (number of frames of silence to play) can be zero in the DAC doesn't
                      // start
                      // outputting frames for a while -- it could get loaded up but not start
                      // responding
                      // for many milliseconds.
                      if (fs > 0) {
                        silence = malloc(conn->output_bytes_per_frame * fs);
                        if (silence == NULL)
                          debug(1, "Failed to allocate %d byte silence buffer.", fs);
                        else {

                          conn->previous_random_number = generate_zero_frames(
                              silence, fs, config.output_format, conn->enable_dither,
                              conn->previous_random_number);

                          // debug(1,"Frames to start: %llu, DAC delay %d, buffer: %d
                          // packets.",exact_frame_gap,dac_delay,seq_diff(conn->ab_read,
                          // conn->ab_write, conn->ab_read));
                          config.output->play(silence, fs);
                          // debug(1,"Sent %" PRId64 " frames of silence",fs);
                          free(silence);
                        }
                      }
                      have_sent_prefiller_silence =
                          1; // even if we haven't sent silence because it's zero frames long...
                    }
                  } else {
                    if ((resp == sps_extra_code_output_stalled) &&
                        (conn->unfixable_error_reported == 0)) {
                      conn->unfixable_error_reported = 1;
                      if (config.cmd_unfixable) {
                        command_execute(config.cmd_unfixable, "output_device_stalled", 1);
                      } else {
                        warn(
                            "an unrecoverable error, \"output_device_stalled\", has been detected.",
                            conn->connection_number);
                      }
                    }
                  }
                } else {
                  // no delay function on back end -- just send the prefiller silence
                  // debug(1,"Back end has no delay function.");
                  // send the appropriate prefiller here...

                  void *silence;
                  if (lead_time != 0) {
                    int64_t frame_gap = (lead_time * config.output_rate) >> 32;
                    // debug(1,"%d frames needed.",frame_gap);
                    while (frame_gap > 0) {
                      ssize_t fs = config.output_rate / 10;
                      if (fs > frame_gap)
                        fs = frame_gap;

                      silence = malloc(conn->output_bytes_per_frame * fs);
                      if (silence == NULL)
                        debug(1, "Failed to allocate %d frame silence buffer.", fs);
                      else {
                        // debug(1, "No delay function -- outputting %d frames of silence.", fs);
                        conn->previous_random_number =
                            generate_zero_frames(silence, fs, config.output_format,
                                                 conn->enable_dither, conn->previous_random_number);
                        config.output->play(silence, fs);
                        free(silence);
                      }
                      frame_gap -= fs;
                    }
                  }
                  have_sent_prefiller_silence = 1;
                  conn->ab_buffering = 0;
                }
              }
            }
          }
          if (conn->ab_buffering == 0) {
/*
  // note the time of the playing of the first frame
  uint64_t reference_timestamp_time; // don't need this...
  get_reference_timestamp_stuff(&conn->play_segment_reference_frame,
                                &reference_timestamp_time,
                                &conn->play_segment_reference_frame_remote_time, conn);
  conn->play_segment_reference_frame *= conn->output_sample_ratio;
*/
#ifdef CONFIG_METADATA
            debug(2, "prsm");
            send_ssnc_metadata('prsm', NULL, 0,
                               0); // "resume", but don't wait if the queue is locked
#endif
          }
        }
      }
    }

    // Here, we work out whether to release a packet or wait
    // We release a buffer when the time is right.

    // To work out when the time is right, we need to take account of (1) the actual time the packet
    // should be released,
    // (2) the latency requested, (3) the audio backend latency offset and (4) the desired length of
    // the audio backend's buffer

    // The time is right if the current time is later or the same as
    // The packet time + (latency + latency offset - backend_buffer_length).
    // Note: the last three items are expressed in frames and must be converted to time.

    int do_wait = 0; // don't wait unless we can really prove we must
    if ((conn->ab_synced) && (curframe) && (curframe->ready) && (curframe->given_timestamp)) {
      do_wait =
          1; // if the current frame exists and is ready, then wait unless it's time to let it go...

      // here, get the time to play the current frame.

      if (have_timestamp_timing_information(conn)) { // if we have a reference time

        uint64_t time_to_play;
        frame_to_local_time(curframe->given_timestamp + conn->latency +
                                (uint32_t)(config.audio_backend_latency_offset * conn->input_rate) -
                                (uint32_t)(config.audio_backend_buffer_desired_length *
                                           conn->input_rate), // this will go modulo 2^32
                            &time_to_play,
                            conn);

        if (local_time_now >= time_to_play) {
          do_wait = 0;
        }
      }
    }
    if (do_wait == 0)
      if ((conn->ab_synced != 0) && (conn->ab_read == conn->ab_write)) { // the buffer is empty!
        if (notified_buffer_empty == 0) {
          debug(3, "Buffers exhausted.");
          notified_buffer_empty = 1;
          reset_input_flow_metrics(conn);
        }
        do_wait = 1;
      }
    wait = (conn->ab_buffering || (do_wait != 0) || (!conn->ab_synced));

    if (wait) {
      uint64_t time_to_wait_for_wakeup_fp =
          ((uint64_t)1 << 32) / conn->input_rate; // this is time period of one frame
      time_to_wait_for_wakeup_fp *= 2 * 352;      // two full 352-frame packets
      time_to_wait_for_wakeup_fp /= 3;            // two thirds of a packet time

#ifdef COMPILE_FOR_LINUX_AND_FREEBSD_AND_CYGWIN_AND_OPENBSD
      uint64_t time_of_wakeup_fp = local_time_now + time_to_wait_for_wakeup_fp;
      uint64_t sec = time_of_wakeup_fp >> 32;
      uint64_t nsec = ((time_of_wakeup_fp & 0xffffffff) * 1000000000) >> 32;

      struct timespec time_of_wakeup;
      time_of_wakeup.tv_sec = sec;
      time_of_wakeup.tv_nsec = nsec;
      //      pthread_cond_timedwait(&conn->flowcontrol, &conn->ab_mutex, &time_of_wakeup);
      int rc = pthread_cond_timedwait(&conn->flowcontrol, &conn->ab_mutex,
                                      &time_of_wakeup); // this is a pthread cancellation point
      if ((rc != 0) && (rc != ETIMEDOUT))
        debug(3, "pthread_cond_timedwait returned error code %d.", rc);
#endif
#ifdef COMPILE_FOR_OSX
      uint64_t sec = time_to_wait_for_wakeup_fp >> 32;
      uint64_t nsec = ((time_to_wait_for_wakeup_fp & 0xffffffff) * 1000000000) >> 32;
      struct timespec time_to_wait;
      time_to_wait.tv_sec = sec;
      time_to_wait.tv_nsec = nsec;
      pthread_cond_timedwait_relative_np(&conn->flowcontrol, &conn->ab_mutex, &time_to_wait);
#endif
    }
  } while (wait);

  // seq_t read = conn->ab_read;
  if (curframe) {
    if (!curframe->ready) {
      // debug(1, "Supplying a silent frame for frame %u", read);
      conn->missing_packets++;
      curframe->given_timestamp = 0; // indicate a silent frame should be substituted
    }
    curframe->ready = 0;
    curframe->resend_level = 0;
  }
  conn->ab_read = SUCCESSOR(conn->ab_read);
  pthread_cleanup_pop(1);
  return curframe;
}

static inline int32_t mean_32(int32_t a, int32_t b) {
  int64_t al = a;
  int64_t bl = b;
  int64_t mean = (al + bl) / 2;
  int32_t r = (int32_t)mean;
  if (r != mean)
    debug(1, "Error calculating average of two int32_ts");
  return r;
}

// this takes an array of signed 32-bit integers and (a) removes or inserts a frame as specified in
// stuff,
// (b) multiplies each sample by the fixedvolume (a 16-bit quantity)
// (c) dithers the result to the output size 32/24/16/8 bits
// (d) outputs the result in the approprate format
// formats accepted so far include U8, S8, S16, S24, S24_3LE, S24_3BE and S32

// stuff: 1 means add 1; 0 means do nothing; -1 means remove 1
static int stuff_buffer_basic_32(int32_t *inptr, int length, enum sps_format_t l_output_format,
                                 char *outptr, int stuff, int dither, rtsp_conn_info *conn) {
  int tstuff = stuff;
  char *l_outptr = outptr;
  if ((stuff > 1) || (stuff < -1) || (length < 100)) {
    // debug(1, "Stuff argument to stuff_buffer must be from -1 to +1 and length >100.");
    tstuff = 0; // if any of these conditions hold, don't stuff anything/
  }

  int i;
  int stuffsamp = length;
  if (tstuff)
    //      stuffsamp = rand() % (length - 1);
    stuffsamp =
        (rand() % (length - 2)) + 1; // ensure there's always a sample before and after the item

  for (i = 0; i < stuffsamp; i++) { // the whole frame, if no stuffing
    process_sample(*inptr++, &l_outptr, l_output_format, conn->fix_volume, dither, conn);
    process_sample(*inptr++, &l_outptr, l_output_format, conn->fix_volume, dither, conn);
  };
  if (tstuff) {
    if (tstuff == 1) {
      // debug(3, "+++++++++");
      // interpolate one sample
      process_sample(mean_32(inptr[-2], inptr[0]), &l_outptr, l_output_format, conn->fix_volume,
                     dither, conn);
      process_sample(mean_32(inptr[-1], inptr[1]), &l_outptr, l_output_format, conn->fix_volume,
                     dither, conn);
    } else if (stuff == -1) {
      // debug(3, "---------");
      inptr++;
      inptr++;
    }

    // if you're removing, i.e. stuff < 0, copy that much less over. If you're adding, do all the
    // rest.
    int remainder = length;
    if (tstuff < 0)
      remainder = remainder + tstuff; // don't run over the correct end of the output buffer

    for (i = stuffsamp; i < remainder; i++) {
      process_sample(*inptr++, &l_outptr, l_output_format, conn->fix_volume, dither, conn);
      process_sample(*inptr++, &l_outptr, l_output_format, conn->fix_volume, dither, conn);
    }
  }
  conn->amountStuffed = tstuff;
  return length + tstuff;
}

#ifdef CONFIG_SOXR
// this takes an array of signed 32-bit integers and
// (a) uses libsoxr to
// resample the array to have one more or one less frame, as specified in
// stuff,
// (b) multiplies each sample by the fixedvolume (a 16-bit quantity)
// (c) dithers the result to the output size 32/24/16/8 bits
// (d) outputs the result in the approprate format
// formats accepted so far include U8, S8, S16, S24, S24_3LE, S24_3BE and S32

int32_t stat_n = 0;
double stat_mean = 0.0;
double stat_M2 = 0.0;
double longest_soxr_execution_time_us = 0.0;
int64_t packets_processed = 0;

int stuff_buffer_soxr_32(int32_t *inptr, int32_t *scratchBuffer, int length,
                         enum sps_format_t l_output_format, char *outptr, int stuff, int dither,
                         rtsp_conn_info *conn) {
  if (scratchBuffer == NULL) {
    die("soxr scratchBuffer not initialised.");
  }
  packets_processed++;
  int tstuff = stuff;
  if ((stuff > 1) || (stuff < -1) || (length < 100)) {
    // debug(1, "Stuff argument to stuff_buffer must be from -1 to +1 and length >100.");
    tstuff = 0; // if any of these conditions hold, don't stuff anything/
  }

  if (tstuff) {
    // debug(1,"Stuff %d.",stuff);

    soxr_io_spec_t io_spec;
    io_spec.itype = SOXR_INT32_I;
    io_spec.otype = SOXR_INT32_I;
    io_spec.scale = 1.0; // this seems to crash if not = 1.0
    io_spec.e = NULL;
    io_spec.flags = 0;

    size_t odone;

    uint64_t soxr_start_time = get_absolute_time_in_fp();

    soxr_error_t error = soxr_oneshot(length, length + tstuff, 2, // Rates and # of chans.
                                      inptr, length, NULL,        // Input.
                                      scratchBuffer, length + tstuff, &odone, // Output.
                                      &io_spec,    // Input, output and transfer spec.
                                      NULL, NULL); // Default configuration.

    if (error)
      die("soxr error: %s\n", "error: %s\n", soxr_strerror(error));

    if (odone > (size_t)(length + 1))
      die("odone = %u!\n", odone);

    // mean and variance calculations from "online_variance" algorithm at
    // https://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Online_algorithm

    double soxr_execution_time_us =
        (((get_absolute_time_in_fp() - soxr_start_time) * 1000000) >> 32) * 1.0;
    // debug(1,"soxr_execution_time_us: %10.1f",soxr_execution_time_us);
    if (soxr_execution_time_us > longest_soxr_execution_time_us)
      longest_soxr_execution_time_us = soxr_execution_time_us;
    stat_n += 1;
    double stat_delta = soxr_execution_time_us - stat_mean;
    stat_mean += stat_delta / stat_n;
    stat_M2 += stat_delta * (soxr_execution_time_us - stat_mean);

    int i;
    int32_t *ip, *op;
    ip = inptr;
    op = scratchBuffer;

    const int gpm = 5;
    // keep the first (dpm) samples, to mitigate the Gibbs phenomenon
    for (i = 0; i < gpm; i++) {
      *op++ = *ip++;
      *op++ = *ip++;
    }

    // keep the last (dpm) samples, to mitigate the Gibbs phenomenon

    // pointer arithmetic, baby -- it's da bomb.
    op = scratchBuffer + (length + tstuff - gpm) * 2;
    ip = inptr + (length - gpm) * 2;
    for (i = 0; i < gpm; i++) {
      *op++ = *ip++;
      *op++ = *ip++;
    }

    // now, do the volume, dither and formatting processing
    ip = scratchBuffer;
    char *l_outptr = outptr;
    for (i = 0; i < length + tstuff; i++) {
      process_sample(*ip++, &l_outptr, l_output_format, conn->fix_volume, dither, conn);
      process_sample(*ip++, &l_outptr, l_output_format, conn->fix_volume, dither, conn);
    };

  } else { // the whole frame, if no stuffing

    // now, do the volume, dither and formatting processing
    int32_t *ip = inptr;
    char *l_outptr = outptr;
    int i;

    for (i = 0; i < length; i++) {
      process_sample(*ip++, &l_outptr, l_output_format, conn->fix_volume, dither, conn);
      process_sample(*ip++, &l_outptr, l_output_format, conn->fix_volume, dither, conn);
    };
  }

  if (packets_processed % 1250 == 0) {
    debug(3, "soxr_oneshot execution time in microseconds: mean, standard deviation and max "
             "for %" PRId32 " interpolations in the last "
             "1250 packets. %10.1f, %10.1f, %10.1f.",
          stat_n, stat_mean, stat_n <= 1 ? 0.0 : sqrtf(stat_M2 / (stat_n - 1)),
          longest_soxr_execution_time_us);
    stat_n = 0;
    stat_mean = 0.0;
    stat_M2 = 0.0;
    longest_soxr_execution_time_us = 0.0;
  }

  conn->amountStuffed = tstuff;
  return length + tstuff;
}
#endif

typedef struct stats { // statistics for running averages
  int64_t sync_error, correction, drift;
} stats_t;

void player_thread_initial_cleanup_handler(__attribute__((unused)) void *arg) {
  rtsp_conn_info *conn = (rtsp_conn_info *)arg;
  debug(3, "Connection %d: player thread main loop exit via player_thread_initial_cleanup_handler.",
        conn->connection_number);
}

void player_thread_cleanup_handler(void *arg) {
  rtsp_conn_info *conn = (rtsp_conn_info *)arg;
  int oldState;
  pthread_setcancelstate(PTHREAD_CANCEL_DISABLE, &oldState);
  debug(3, "Connection %d: player thread main loop exit via player_thread_cleanup_handler.",
        conn->connection_number);

  if (config.output->stop)
    config.output->stop();

  if (config.statistics_requested) {
    int rawSeconds = (int)difftime(time(NULL), conn->playstart);
    int elapsedHours = rawSeconds / 3600;
    int elapsedMin = (rawSeconds / 60) % 60;
    int elapsedSec = rawSeconds % 60;
    if (conn->frame_rate_status)
      inform("Playback Stopped. Total playing time %02d:%02d:%02d. Input: %0.2f, output: %0.2f "
             "frames per second.",
             elapsedHours, elapsedMin, elapsedSec, conn->input_frame_rate, conn->frame_rate);
    else
      inform("Playback Stopped. Total playing time %02d:%02d:%02d. Input: %0.2f frames per second.",
             elapsedHours, elapsedMin, elapsedSec, conn->input_frame_rate);
  }

#ifdef CONFIG_DACP_CLIENT
  relinquish_dacp_server_information(
      conn); // say it doesn't belong to this conversation thread any more...
#else
  mdns_dacp_monitor_set_id(NULL); // say we're not interested in following that DACP id any more
#endif

  debug(3, "Cancelling timing, control and audio threads...");
  debug(3, "Cancel timing thread.");
  pthread_cancel(conn->rtp_timing_thread);
  debug(3, "Join timing thread.");
  pthread_join(conn->rtp_timing_thread, NULL);
  debug(3, "Timing thread terminated.");
  debug(3, "Cancel control thread.");
  pthread_cancel(conn->rtp_control_thread);
  debug(3, "Join control thread.");
  pthread_join(conn->rtp_control_thread, NULL);
  debug(3, "Control thread terminated.");
  debug(3, "Cancel audio thread.");
  pthread_cancel(conn->rtp_audio_thread);
  debug(3, "Join audio thread.");
  pthread_join(conn->rtp_audio_thread, NULL);
  debug(3, "Audio thread terminated.");

  if (conn->outbuf) {
    free(conn->outbuf);
    conn->outbuf = NULL;
  }
  if (conn->sbuf) {
    free(conn->sbuf);
    conn->sbuf = NULL;
  }
  if (conn->tbuf) {
    free(conn->tbuf);
    conn->tbuf = NULL;
  }
  free_audio_buffers(conn);
  if (conn->stream.type == ast_apple_lossless)
    terminate_decoders(conn);

  clear_reference_timestamp(conn);
  conn->rtp_running = 0;
  pthread_setcancelstate(oldState, NULL);
}

void *player_thread_func(void *arg) {
  rtsp_conn_info *conn = (rtsp_conn_info *)arg;
  // pthread_cleanup_push(player_thread_initial_cleanup_handler, arg);
  conn->packet_count = 0;
  conn->packet_count_since_flush = 0;
  conn->previous_random_number = 0;
  conn->input_bytes_per_frame = 4;
  conn->decoder_in_use = 0;
  conn->ab_buffering = 1;
  conn->ab_synced = 0;
  conn->first_packet_timestamp = 0;
  conn->flush_requested = 0;
  conn->fix_volume = 0x10000;

  if (conn->latency == 0) {
    debug(3, "No latency has (yet) been specified. Setting 88,200 (2 seconds) frames "
             "as a default.");
    conn->latency = 88200;
  }

  if (conn->stream.type == ast_apple_lossless)
    init_alac_decoder((int32_t *)&conn->stream.fmtp,
                      conn); // this sets up incoming rate, bit depth, channels.
                             // No pthread cancellation point in here
  // This must be after init_alac_decoder
  init_buffer(conn); // will need a corresponding deallocation. No cancellation points in here

  if (conn->stream.encrypted) {
#ifdef CONFIG_MBEDTLS
    memset(&conn->dctx, 0, sizeof(mbedtls_aes_context));
    mbedtls_aes_setkey_dec(&conn->dctx, conn->stream.aeskey, 128);
#endif

#ifdef CONFIG_POLARSSL
    memset(&conn->dctx, 0, sizeof(aes_context));
    aes_setkey_dec(&conn->dctx, conn->stream.aeskey, 128);
#endif

#ifdef CONFIG_OPENSSL
    AES_set_decrypt_key(conn->stream.aeskey, 128, &conn->aes);
#endif
  }

  conn->timestamp_epoch = 0; // indicate that the next timestamp will be the first one.
  conn->maximum_timestamp_interval = conn->input_rate * 60; // actually there shouldn't be more than
                                                            // about 13 seconds of a gap between
                                                            // successive rtptimes, at worst

  conn->output_sample_ratio = config.output_rate / conn->input_rate;

  //  debug(1, "Output sample ratio is %d.", conn->output_sample_ratio);

  conn->max_frame_size_change =
      1 * conn->output_sample_ratio; // we add or subtract one frame at the nominal
                                     // rate, multiply it by the frame ratio.
                                     // but, on some occasions, more than one frame could be added

  switch (config.output_format) {
  case SPS_FORMAT_S24_3LE:
  case SPS_FORMAT_S24_3BE:
    conn->output_bytes_per_frame = 6;
    break;

  case SPS_FORMAT_S24:
  case SPS_FORMAT_S24_LE:
  case SPS_FORMAT_S24_BE:
    conn->output_bytes_per_frame = 8;
    break;
  case SPS_FORMAT_S32:
  case SPS_FORMAT_S32_LE:
  case SPS_FORMAT_S32_BE:
    conn->output_bytes_per_frame = 8;
    break;
  default:
    conn->output_bytes_per_frame = 4;
  }

  debug(3, "Output frame bytes is %d.", conn->output_bytes_per_frame);

  conn->dac_buffer_queue_minimum_length = (int64_t)(
      config.audio_backend_buffer_interpolation_threshold_in_seconds * config.output_rate);
  debug(3, "dac_buffer_queue_minimum_length is %" PRId64 " frames.",
        conn->dac_buffer_queue_minimum_length);

  conn->session_corrections = 0;
  // conn->play_segment_reference_frame = 0; // zero signals that we are not in a play segment

  // check that there are enough buffers to accommodate the desired latency and the latency offset

  int maximum_latency =
      conn->latency + (int)(config.audio_backend_latency_offset * config.output_rate);
  if ((maximum_latency + (352 - 1)) / 352 + 10 > BUFFER_FRAMES)
    die("Not enough buffers available for a total latency of %d frames. A maximum of %d 352-frame "
        "packets may be accommodated.",
        maximum_latency, BUFFER_FRAMES);
  conn->connection_state_to_output = get_requested_connection_state_to_output();
// this is about half a minute
//#define trend_interval 3758
#define trend_interval 1003

  stats_t statistics[trend_interval];
  int number_of_statistics, oldest_statistic, newest_statistic;
  int at_least_one_frame_seen = 0;
  int64_t tsum_of_sync_errors, tsum_of_corrections, tsum_of_insertions_and_deletions,
      tsum_of_drifts;
  int64_t previous_sync_error = 0, previous_correction = 0;
  int64_t minimum_dac_queue_size = INT64_MAX;
  int32_t minimum_buffer_occupancy = INT32_MAX;
  int32_t maximum_buffer_occupancy = INT32_MIN;

  conn->playstart = time(NULL);

  conn->frame_rate = 0.0;
  conn->frame_rate_status = 0;

  conn->input_frame_rate = 0.0;
  conn->input_frame_rate_starting_point_is_valid = 0;

  conn->buffer_occupancy = 0;

  int play_samples = 0;
  int64_t current_delay;
  int play_number = 0;
  conn->play_number_after_flush = 0;
  //  int last_timestamp = 0; // for debugging only
  conn->time_of_last_audio_packet = 0;
  // conn->shutdown_requested = 0;
  number_of_statistics = oldest_statistic = newest_statistic = 0;
  tsum_of_sync_errors = tsum_of_corrections = tsum_of_insertions_and_deletions = tsum_of_drifts = 0;

  const int print_interval = trend_interval; // don't ask...
  // I think it's useful to keep this prime to prevent it from falling into a pattern with some
  // other process.

  static char rnstate[256];
  initstate(time(NULL), rnstate, 256);

  signed short *inbuf;
  int inbuflength;

  unsigned int output_bit_depth = 16; // default;

  switch (config.output_format) {
  case SPS_FORMAT_S8:
  case SPS_FORMAT_U8:
    output_bit_depth = 8;
    break;
  case SPS_FORMAT_S16:
  case SPS_FORMAT_S16_LE:
  case SPS_FORMAT_S16_BE:
    output_bit_depth = 16;
    break;
  case SPS_FORMAT_S24:
  case SPS_FORMAT_S24_LE:
  case SPS_FORMAT_S24_BE:
  case SPS_FORMAT_S24_3LE:
  case SPS_FORMAT_S24_3BE:
    output_bit_depth = 24;
    break;
  case SPS_FORMAT_S32:
  case SPS_FORMAT_S32_LE:
  case SPS_FORMAT_S32_BE:
    output_bit_depth = 32;
    break;
  case SPS_FORMAT_UNKNOWN:
    die("Unknown format choosing output bit depth");
    break;
  case SPS_FORMAT_AUTO:
    die("Invalid format -- SPS_FORMAT_AUTO -- choosing output bit depth");
    break;
  case SPS_FORMAT_INVALID:
    die("Invalid format -- SPS_FORMAT_INVALID -- choosing output bit depth");
    break;
  }

  debug(3, "Output bit depth is %d.", output_bit_depth);

  if (conn->input_bit_depth > output_bit_depth) {
    debug(3, "Dithering will be enabled because the input bit depth is greater than the output bit "
             "depth");
  }
  if (config.output->parameters == NULL) {
    debug(3, "Dithering will be enabled because the output volume is being altered in software");
  }

  if ((config.output->parameters == NULL) || (conn->input_bit_depth > output_bit_depth) ||
      (config.playback_mode == ST_mono))
    conn->enable_dither = 1;

  // remember, the output device may never have been initialised prior to this call
  config.output->start(config.output_rate, config.output_format); // will need a corresponding stop

  // we need an intermediate "transition" buffer

  // if ((input_rate!=config.output_rate) || (input_bit_depth!=output_bit_depth)) {
  // debug(1,"Define tbuf of length
  // %d.",output_bytes_per_frame*(max_frames_per_packet*output_sample_ratio+max_frame_size_change));
  conn->tbuf =
      malloc(sizeof(int32_t) * 2 * (conn->max_frames_per_packet * conn->output_sample_ratio +
                                    conn->max_frame_size_change));
  if (conn->tbuf == NULL)
    die("Failed to allocate memory for the transition buffer.");

  // initialise this, because soxr stuffing might be chosen later

  conn->sbuf =
      malloc(sizeof(int32_t) * 2 * (conn->max_frames_per_packet * conn->output_sample_ratio +
                                    conn->max_frame_size_change));
  if (conn->sbuf == NULL)
    die("Failed to allocate memory for the sbuf buffer.");

  // The size of these dependents on the number of frames, the size of each frame and the maximum
  // size change
  conn->outbuf = malloc(
      conn->output_bytes_per_frame *
      (conn->max_frames_per_packet * conn->output_sample_ratio + conn->max_frame_size_change));
  if (conn->outbuf == NULL)
    die("Failed to allocate memory for an output buffer.");
  conn->first_packet_timestamp = 0;
  conn->missing_packets = conn->late_packets = conn->too_late_packets = conn->resend_requests = 0;
  conn->flush_rtp_timestamp = 0; // it seems this number has a special significance -- it seems to
                                 // be used as a null operand, so we'll use it like that too
  int sync_error_out_of_bounds =
      0; // number of times in a row that there's been a serious sync error

  conn->framesProcessedInThisEpoch = 0;
  conn->framesGeneratedInThisEpoch = 0;
  conn->correctionsRequestedInThisEpoch = 0;

  if (config.statistics_requested) {
    if ((config.output->delay)) {
      if (config.no_sync == 0) {
        inform("sync error in milliseconds, "
               "net correction in ppm, "
               "corrections in ppm, "
               "total packets, "
               "missing packets, "
               "late packets, "
               "too late packets, "
               "resend requests, "
               "min DAC queue size, "
               "min buffer occupancy, "
               "max buffer occupancy, "
               "source nominal frames per second, "
               "source actual frames per second, "
               "output frames per second, "
               "source clock drift in ppm, "
               "source clock drift sample count, "
               "rough calculated correction in ppm");
      } else {
        inform("sync error in milliseconds, "
               "total packets, "
               "missing packets, "
               "late packets, "
               "too late packets, "
               "resend requests, "
               "min DAC queue size, "
               "min buffer occupancy, "
               "max buffer occupancy, "
               "source nominal frames per second, "
               "source actual frames per second, "
               "source clock drift in ppm, "
               "source clock drift sample count");
      }
    } else {
      inform("sync error in milliseconds, "
             "total packets, "
             "missing packets, "
             "late packets, "
             "too late packets, "
             "resend requests, "
             "min buffer occupancy, "
             "max buffer occupancy, "
             "source nominal frames per second, "
             "source actual frames per second, "
             "source clock drift in ppm, "
             "source clock drift sample count");
    }
  }

  // create and start the timing, control and audio receiver threads
  pthread_create(&conn->rtp_audio_thread, NULL, &rtp_audio_receiver, (void *)conn);
  pthread_create(&conn->rtp_control_thread, NULL, &rtp_control_receiver, (void *)conn);
  pthread_create(&conn->rtp_timing_thread, NULL, &rtp_timing_receiver, (void *)conn);

  pthread_cleanup_push(player_thread_cleanup_handler, arg); // undo what's been done so far

  // stop looking elsewhere for DACP stuff
  int oldState;
  pthread_setcancelstate(PTHREAD_CANCEL_DISABLE, &oldState);

#ifdef CONFIG_DACP_CLIENT
  set_dacp_server_information(conn);
#else
  mdns_dacp_monitor_set_id(conn->dacp_id);
#endif

  pthread_setcancelstate(oldState, NULL);

  // set the default volume to whatever it was before, as stored in the config airplay_volume
  debug(2, "Set initial volume to %f.", config.airplay_volume);
  player_volume(config.airplay_volume, conn); // will contain a cancellation point if asked to wait

  debug(2, "Play begin");
  while (1) {
    pthread_testcancel();                     // allow a pthread_cancel request to take effect.
    abuf_t *inframe = buffer_get_frame(conn); // this has cancellation point(s), but it's not
                                              // guaranteed that they'll always be executed
    if (inframe) {
      inbuf = inframe->data;
      inbuflength = inframe->length;
      if (inbuf) {
        play_number++;
        //        if (play_number % 100 == 0)
        //          debug(3, "Play frame %d.", play_number);
        conn->play_number_after_flush++;
        if (inframe->given_timestamp == 0) {
          debug(3, "Player has supplied a silent frame, (possibly frame %u) for play number %d.",
                SUCCESSOR(conn->last_seqno_read), play_number);
          conn->last_seqno_read = (SUCCESSOR(conn->last_seqno_read) &
                                   0xffff); // manage the packet out of sequence minder

          void *silence = malloc(conn->output_bytes_per_frame * conn->max_frames_per_packet *
                                 conn->output_sample_ratio);
          if (silence == NULL) {
            debug(1, "Failed to allocate memory for a silent frame silence buffer.");
          } else {
            // the player may change the contents of the buffer, so it has to be zeroed each time;
            // might as well malloc and free it locally
            conn->previous_random_number = generate_zero_frames(
                silence, conn->max_frames_per_packet * conn->output_sample_ratio,
                config.output_format, conn->enable_dither, conn->previous_random_number);
            config.output->play(silence, conn->max_frames_per_packet * conn->output_sample_ratio);
            free(silence);
          }
        } else if (conn->play_number_after_flush < 10) {
          /*
          int64_t difference = 0;
          if (last_timestamp)
            difference = inframe->timestamp - last_timestamp;
          last_timestamp = inframe->timestamp;
          debug(1, "Play number %d, monotonic timestamp %llx, difference
          %lld.",conn->play_number_after_flush,inframe->timestamp,difference);
          */
          void *silence = malloc(conn->output_bytes_per_frame * conn->max_frames_per_packet *
                                 conn->output_sample_ratio);
          if (silence == NULL) {
            debug(1, "Failed to allocate memory for a flush silence buffer.");
          } else {
            // the player may change the contents of the buffer, so it has to be zeroed each time;
            // might as well malloc and free it locally
            conn->previous_random_number = generate_zero_frames(
                silence, conn->max_frames_per_packet * conn->output_sample_ratio,
                config.output_format, conn->enable_dither, conn->previous_random_number);
            config.output->play(silence, conn->max_frames_per_packet * conn->output_sample_ratio);
            free(silence);
          }
        } else {

          if (((config.output->parameters == NULL) && (config.ignore_volume_control == 0) &&
               (config.airplay_volume != 0.0)) ||
              (conn->input_bit_depth > output_bit_depth) || (config.playback_mode == ST_mono))
            conn->enable_dither = 1;
          else
            conn->enable_dither = 0;

          // here, let's transform the frame of data, if necessary

          switch (conn->input_bit_depth) {
          case 16: {
            int i, j;
            int16_t ls, rs;
            int32_t ll = 0, rl = 0;
            int16_t *inps = inbuf;
            // int16_t *outps = tbuf;
            int32_t *outpl = (int32_t *)conn->tbuf;
            for (i = 0; i < inbuflength; i++) {
              ls = *inps++;
              rs = *inps++;

              // here, do the mode stuff -- mono / reverse stereo / leftonly / rightonly
              // also, raise the 16-bit samples to 32 bits.

              switch (config.playback_mode) {
              case ST_mono: {
                int32_t both = ls + rs;
                both = both << (16 - 1); // keep all 17 bits of the sum of the 16bit left and right
                                         // -- the 17th bit will influence dithering later
                ll = both;
                rl = both;
              } break;
              case ST_reverse_stereo: {
                ll = rs;
                rl = ls;
                ll = ll << 16;
                rl = rl << 16;
              } break;
              case ST_left_only:
                rl = ls;
                ll = ls;
                ll = ll << 16;
                rl = rl << 16;
                break;
              case ST_right_only:
                ll = rs;
                rl = rs;
                ll = ll << 16;
                rl = rl << 16;
                break;
              case ST_stereo:
                ll = ls;
                rl = rs;
                ll = ll << 16;
                rl = rl << 16;
                break; // nothing extra to do
              }

              // here, replicate the samples if you're upsampling

              for (j = 0; j < conn->output_sample_ratio; j++) {
                *outpl++ = ll;
                *outpl++ = rl;
              }
            }

          } break;
          default:
            die("Shairport Sync only supports 16 bit input");
          }

          inbuflength *= conn->output_sample_ratio;

          // We have a frame of data. We need to see if we want to add or remove a frame from it to
          // keep in sync.
          // So we calculate the timing error for the first frame in the DAC.
          // If it's ahead of time, we add one audio frame to this frame to delay a subsequent frame
          // If it's late, we remove an audio frame from this frame to bring a subsequent frame
          // forward in time

          at_least_one_frame_seen = 1;

          // now, go back as far as the total latency less, say, 100 ms, and check the presence of
          // frames from then onwards

          uint32_t reference_timestamp;
          uint64_t reference_timestamp_time, remote_reference_timestamp_time;
          get_reference_timestamp_stuff(&reference_timestamp, &reference_timestamp_time,
                                        &remote_reference_timestamp_time, conn); // types okay
          int64_t rt, nt;
          rt = reference_timestamp;      // uint32_t to int64_t
          nt = inframe->given_timestamp; // uint32_t to int64_t
          rt = rt * conn->output_sample_ratio;
          nt = nt * conn->output_sample_ratio;

          uint64_t local_time_now = get_absolute_time_in_fp(); // types okay
          // struct timespec tn;
          // clock_gettime(CLOCK_MONOTONIC,&tn);
          // uint64_t
          // local_time_now=((uint64_t)tn.tv_sec<<32)+((uint64_t)tn.tv_nsec<<32)/1000000000;

          int64_t td = 0; // td is the time difference between the reference timestamp time and the
                          // present time. Only used to calculate td_in_frames
          int64_t td_in_frames = 0; // td_in_frames is the number of frames between between the
                                    // reference timestamp time and the present time

          if (local_time_now >= reference_timestamp_time) {
            td = local_time_now - reference_timestamp_time; // this is the positive value.
                                                            // Conversion is positive uint64_t to
                                                            // int64_t, thus okay
            td_in_frames = (td * config.output_rate) >> 32;
          } else {
            td = reference_timestamp_time - local_time_now; // this is the absolute value, which
                                                            // should be negated. Conversion is
                                                            // positive uint64_t to int64_t, thus
                                                            // okay.
            td_in_frames = (td * config.output_rate) >>
                           32; // use the absolute td value for the present. Types okay
            td_in_frames = -td_in_frames;
            td = -td; // should be okay, as the range of values should be very small w.r.t 64 bits
          }

          // This is the timing error for the next audio frame in the DAC, if applicable
          int64_t sync_error = 0;

          int amount_to_stuff = 0;

          // check sequencing
          if (conn->last_seqno_read == -1)
            conn->last_seqno_read =
                inframe->sequence_number; // int32_t from seq_t, i.e. uint16_t, so okay.
          else {
            conn->last_seqno_read =
                SUCCESSOR(conn->last_seqno_read); // int32_t from seq_t, i.e. uint16_t, so okay.
            if (inframe->sequence_number !=
                conn->last_seqno_read) { // seq_t, ei.e. uint16_t and int32_t, so okay
              debug(2, "Player: packets out of sequence: expected: %u, got: %u, with ab_read: %u "
                       "and ab_write: %u.",
                    conn->last_seqno_read, inframe->sequence_number, conn->ab_read, conn->ab_write);
              conn->last_seqno_read = inframe->sequence_number; // reset warning...
            }
          }

          conn->buffer_occupancy =
              seq_diff(conn->ab_read, conn->ab_write, conn->ab_read); // int32_t from int32

          if (conn->buffer_occupancy < minimum_buffer_occupancy)
            minimum_buffer_occupancy = conn->buffer_occupancy;

          if (conn->buffer_occupancy > maximum_buffer_occupancy)
            maximum_buffer_occupancy = conn->buffer_occupancy;

          // here, we want to check (a) if we are meant to do synchronisation,
          // (b) if we have a delay procedure, (c) if we can get the delay.

          // If any of these are false, we don't do any synchronisation stuff

          int resp = -1; // use this as a flag -- if negative, we can't rely on a real known delay
          current_delay = -1; // use this as a failure flag

          if (config.output->delay) {
            long l_delay;
            resp = config.output->delay(&l_delay);
            if (resp == 0) { // no error
              current_delay = l_delay;
              if (current_delay < 0) {
                debug(2, "Underrun of %lld frames reported, but ignored.", current_delay);
                current_delay =
                    0; // could get a negative value if there was underrun, but ignore it.
              }
              if (current_delay < minimum_dac_queue_size) {
                minimum_dac_queue_size = current_delay; // update for display later
              }
            } else {
              current_delay = 0;
              if ((resp == sps_extra_code_output_stalled) &&
                  (conn->unfixable_error_reported == 0)) {
                conn->unfixable_error_reported = 1;
                if (config.cmd_unfixable) {
                  warn("Connection %d: An unfixable error has been detected -- output device is "
                       "stalled. Executing the "
                       "\"run_this_if_an_unfixable_error_is_detected\" command.",
                       conn->connection_number);
                  command_execute(config.cmd_unfixable, "output_device_stalled", 1);
                } else {
                  warn("Connection %d: An unfixable error has been detected -- output device is "
                       "stalled. \"No "
                       "run_this_if_an_unfixable_error_is_detected\" command provided -- nothing "
                       "done.",
                       conn->connection_number);
                }
              } else
                debug(3, "Delay error %d when checking running latency.", resp);
            }
          }

          if (resp == 0) {

            uint32_t should_be_frame_32;
            local_time_to_frame(local_time_now, &should_be_frame_32, conn);
            int64_t should_be_frame = ((int64_t)should_be_frame_32) * conn->output_sample_ratio;

            // int64_t absolute_difference_in_frames = td_in_frames + rt - should_be_frame;
            // if (absolute_difference_in_frames < 0)
            //  absolute_difference_in_frames = -absolute_difference_in_frames;

            // if (absolute_difference_in_frames > 10 * conn->output_sample_ratio)
            //  debug(1, "Difference between old and new frame number is %" PRId64 " frames.",
            //        td_in_frames + rt - should_be_frame);
            // this is the actual delay, including the latency we actually want, which will
            // fluctuate a good bit about a potentially rising or falling trend.

            //            int64_t delay = td_in_frames + rt - (nt - current_delay); // all int64_t
            // cut over to the new calculation method
            int64_t delay = should_be_frame - (nt - current_delay); // all int64_t

            // td_in_frames + rt is the frame number that should be output at local_time_now.

            // This is the timing error for the next audio frame in the DAC.

            // if positive, it means that the packet will be late -- the delay is longer than
            // requested
            // if negative, the packet will be early -- the delay is less than expected.

            sync_error =
                delay - ((int64_t)conn->latency * conn->output_sample_ratio +
                         (int64_t)(config.audio_backend_latency_offset *
                                   config.output_rate)); // int64_t from int64_t - int32_t, so okay

            // debug(1,"%" PRId64 "",sync_error,inbuflength);

            // not too sure if abs() is implemented for int64_t, so we'll do it manually
            int64_t abs_sync_error = sync_error;
            if (abs_sync_error < 0)
              abs_sync_error = -abs_sync_error;

            if ((config.no_sync == 0) && (inframe->given_timestamp != 0) &&
                (config.resyncthreshold > 0.0) &&
                (abs_sync_error > config.resyncthreshold * config.output_rate)) {
              /*
              if (abs_sync_error > 3 * config.output_rate) {

                warn("Very large sync error: %" PRId64 " frames, with should_be_frame: %" PRId64
                     ",  nt: %" PRId64 ", current_delay: %" PRId64 ", given timestamp %" PRIX32
                     ", reference timestamp %" PRIX32 ", should_be_frame %" PRIX32 ".",
                     sync_error, should_be_frame, nt, current_delay, inframe->given_timestamp,
                     reference_timestamp, should_be_frame_32);
              }
              */
              sync_error_out_of_bounds++;
            } else {
              sync_error_out_of_bounds = 0;
            }

            if (sync_error_out_of_bounds > 3) {
              // debug(1, "New lost sync with source for %d consecutive packets -- flushing and "
              //          "resyncing. Error: %lld.",
              //        sync_error_out_of_bounds, sync_error);
              sync_error_out_of_bounds = 0;

              int64_t filler_length =
                  (int64_t)(config.resyncthreshold * config.output_rate); // number of samples
              if ((sync_error > 0) && (sync_error > filler_length)) {
                debug(2, "Large positive sync error: %" PRId64 ".", sync_error);
                int64_t local_frames_to_drop = sync_error / conn->output_sample_ratio;
                uint32_t frames_to_drop_sized = local_frames_to_drop;

                debug_mutex_lock(&conn->flush_mutex, 1000, 1);
                conn->flush_rtp_timestamp =
                    inframe->given_timestamp +
                    frames_to_drop_sized; // flush all packets up to (and including?) this
                reset_input_flow_metrics(conn);
                debug_mutex_unlock(&conn->flush_mutex, 3);

              } else if ((sync_error < 0) && ((-sync_error) > filler_length)) {
                debug(2,
                      "Large negative sync error: %" PRId64 " with should_be_frame_32 of %" PRIu32
                      ", nt of %" PRId64 " and current_delay of %" PRId64 ".",
                      sync_error, should_be_frame_32, nt, current_delay);
                int64_t silence_length = -sync_error;
                if (silence_length > (filler_length * 5))
                  silence_length = filler_length * 5;
                size_t silence_length_sized = silence_length;
                char *long_silence = malloc(conn->output_bytes_per_frame * silence_length_sized);
                if (long_silence) {

                  conn->previous_random_number =
                      generate_zero_frames(long_silence, silence_length_sized, config.output_format,
                                           conn->enable_dither, conn->previous_random_number);

                  debug(2, "Play a silence of %d frames.", silence_length_sized);
                  config.output->play(long_silence, silence_length_sized);
                  free(long_silence);
                } else {
                  warn("Failed to allocate memory for a long_silence buffer of %d frames for a "
                       "sync error of %d frames.",
                       silence_length_sized, sync_error);
                }
                reset_input_flow_metrics(conn);
              }
            } else {

              /*
              // before we finally commit to this frame, check its sequencing and timing
              // require a certain error before bothering to fix it...
              if (sync_error > config.tolerance * config.output_rate) { // int64_t > int, okay
                amount_to_stuff = -1;
              }
              if (sync_error < -config.tolerance * config.output_rate) {
                amount_to_stuff = 1;
              }
              */

              if (amount_to_stuff == 0) {
                // use a "V" shaped function to decide if stuffing should occur
                int64_t s = r64i();
                s = s >> 31;
                s = s * config.tolerance * config.output_rate;
                s = (s >> 32) + config.tolerance * config.output_rate; // should be a number from 0
                                                                       // to config.tolerance *
                                                                       // config.output_rate;
                if ((sync_error > 0) && (sync_error > s)) {
                  // debug(1,"Extra stuff -1");
                  amount_to_stuff = -1;
                }
                if ((sync_error < 0) && (sync_error < (-s))) {
                  // debug(1,"Extra stuff +1");
                  amount_to_stuff = 1;
                }
              }

              // try to keep the corrections definitely below 1 in 1000 audio frames

              // calculate the time elapsed since the play session started.

              if (amount_to_stuff) {
                if ((local_time_now) && (conn->first_packet_time_to_play) &&
                    (local_time_now >= conn->first_packet_time_to_play)) {

                  int64_t tp = (local_time_now - conn->first_packet_time_to_play) >>
                               32; // seconds int64_t from uint64_t which is always positive, so ok

                  if (tp < 5)
                    amount_to_stuff = 0; // wait at least five seconds
                  /*
                  else if (tp < 30) {
                    if ((random() % 1000) >
                        352) // keep it to about 1:1000 for the first thirty seconds
                      amount_to_stuff = 0;
                  }
                  */
                }
              }

              if (config.no_sync != 0)
                amount_to_stuff = 0; // no stuffing if it's been disabled

              // Apply DSP here
              if (config.loudness
#ifdef CONFIG_CONVOLUTION
                  || config.convolution
#endif
                  ) {
                int32_t *tbuf32 = (int32_t *)conn->tbuf;
                float fbuf_l[inbuflength];
                float fbuf_r[inbuflength];

                // Deinterleave, and convert to float
                int i;
                for (i = 0; i < inbuflength; ++i) {
                  fbuf_l[i] = tbuf32[2 * i];
                  fbuf_r[i] = tbuf32[2 * i + 1];
                }

#ifdef CONFIG_CONVOLUTION
                // Apply convolution
                if (config.convolution) {
                  convolver_process_l(fbuf_l, inbuflength);
                  convolver_process_r(fbuf_r, inbuflength);

                  float gain = pow(10.0, config.convolution_gain / 20.0);
                  for (i = 0; i < inbuflength; ++i) {
                    fbuf_l[i] *= gain;
                    fbuf_r[i] *= gain;
                  }
                }
#endif

                if (config.loudness) {
                  // Apply volume and loudness
                  // Volume must be applied here because the loudness filter will increase the
                  // signal level and it would saturate the int32_t otherwise
                  float gain = conn->fix_volume / 65536.0f;
                  // float gain_db = 20 * log10(gain);
                  // debug(1, "Applying soft volume dB: %f k: %f", gain_db, gain);

                  for (i = 0; i < inbuflength; ++i) {
                    fbuf_l[i] = loudness_process(&loudness_l, fbuf_l[i] * gain);
                    fbuf_r[i] = loudness_process(&loudness_r, fbuf_r[i] * gain);
                  }
                }

                // Interleave and convert back to int32_t
                for (i = 0; i < inbuflength; ++i) {
                  tbuf32[2 * i] = fbuf_l[i];
                  tbuf32[2 * i + 1] = fbuf_r[i];
                }
              }

#ifdef CONFIG_SOXR
              if ((current_delay < conn->dac_buffer_queue_minimum_length) ||
                  (config.packet_stuffing == ST_basic) ||
                  (config.soxr_delay_index == 0) || // not computed yet
                  ((config.packet_stuffing == ST_auto) &&
                   (config.soxr_delay_index >
                    config.soxr_delay_threshold)) // if the CPU is deemed too slow
                  ) {
#endif
                play_samples =
                    stuff_buffer_basic_32((int32_t *)conn->tbuf, inbuflength, config.output_format,
                                          conn->outbuf, amount_to_stuff, conn->enable_dither, conn);
#ifdef CONFIG_SOXR
              } else { // soxr requested or auto requested with the index less or equal to the
                       // threshold
                play_samples = stuff_buffer_soxr_32((int32_t *)conn->tbuf, (int32_t *)conn->sbuf,
                                                    inbuflength, config.output_format, conn->outbuf,
                                                    amount_to_stuff, conn->enable_dither, conn);
              }
#endif

              /*
              {
                int co;
                int is_silent=1;
                short *p = outbuf;
                for (co=0;co<play_samples;co++) {
                  if (*p!=0)
                    is_silent=0;
                  p++;
                }
                if (is_silent)
                  debug(1,"Silence!");
              }
              */

              if (conn->outbuf == NULL)
                debug(1, "NULL outbuf to play -- skipping it.");
              else {
                if (play_samples == 0)
                  debug(1, "play_samples==0 skipping it (1).");
                else {
                  if (conn->software_mute_enabled) {
                    generate_zero_frames(conn->outbuf, play_samples, config.output_format,
                                         conn->enable_dither, conn->previous_random_number);
                  }
                  config.output->play(conn->outbuf, play_samples);
                }
              }

              // check for loss of sync
              // timestamp of zero means an inserted silent frame in place of a missing frame
              /*
              if ((config.no_sync == 0) && (inframe->timestamp != 0) &&
                  && (config.resyncthreshold > 0.0) &&
                  (abs_sync_error > config.resyncthreshold * config.output_rate)) {
                sync_error_out_of_bounds++;
                // debug(1,"Sync error out of bounds: Error: %lld; previous error: %lld; DAC: %lld;
                // timestamp: %llx, time now
                //
              %llx",sync_error,previous_sync_error,current_delay,inframe->timestamp,local_time_now);
                if (sync_error_out_of_bounds > 3) {
                  debug(1, "Lost sync with source for %d consecutive packets -- flushing and "
                           "resyncing. Error: %lld.",
                        sync_error_out_of_bounds, sync_error);
                  sync_error_out_of_bounds = 0;
                  player_flush(nt, conn);
                }
              } else {
                sync_error_out_of_bounds = 0;
              }
              */
            }
          } else {
            // if there is no delay procedure, or it's not working or not allowed, there can be no
            // synchronising
            play_samples =
                stuff_buffer_basic_32((int32_t *)conn->tbuf, inbuflength, config.output_format,
                                      conn->outbuf, 0, conn->enable_dither, conn);
            if (conn->outbuf == NULL)
              debug(1, "NULL outbuf to play -- skipping it.");
            else {
              if (conn->software_mute_enabled) {
                generate_zero_frames(conn->outbuf, play_samples, config.output_format,
                                     conn->enable_dither, conn->previous_random_number);
              }
              config.output->play(conn->outbuf, play_samples); // remove the (short*)!
            }
          }

          // mark the frame as finished
          inframe->given_timestamp = 0;
          inframe->sequence_number = 0;

          // update the watchdog
          if ((config.dont_check_timeout == 0) && (config.timeout != 0)) {
            uint64_t time_now = get_absolute_time_in_fp();
            debug_mutex_lock(&conn->watchdog_mutex, 1000, 0);
            conn->watchdog_bark_time = time_now;
            debug_mutex_unlock(&conn->watchdog_mutex, 0);
          }

          // debug(1,"Sync error %lld frames. Amount to stuff %d." ,sync_error,amount_to_stuff);

          // new stats calculation. We want a running average of sync error, drift, adjustment,
          // number of additions+subtractions

          // this is a misleading hack -- the statistics should include some data on the number of
          // valid samples and the number of times sync wasn't checked due to non availability of a
          // delay figure.
          // for the present, stats are only updated when sync has been checked
          if (sync_error != -1) {
            if (number_of_statistics == trend_interval) {
              // here we remove the oldest statistical data and take it from the summaries as well
              tsum_of_sync_errors -= statistics[oldest_statistic].sync_error;
              tsum_of_drifts -= statistics[oldest_statistic].drift;
              if (statistics[oldest_statistic].correction > 0)
                tsum_of_insertions_and_deletions -= statistics[oldest_statistic].correction;
              else
                tsum_of_insertions_and_deletions += statistics[oldest_statistic].correction;
              tsum_of_corrections -= statistics[oldest_statistic].correction;
              oldest_statistic = (oldest_statistic + 1) % trend_interval;
              number_of_statistics--;
            }

            statistics[newest_statistic].sync_error = sync_error;
            statistics[newest_statistic].correction = conn->amountStuffed;

            if (number_of_statistics == 0)
              statistics[newest_statistic].drift = 0;
            else
              statistics[newest_statistic].drift =
                  sync_error - previous_sync_error - previous_correction;

            previous_sync_error = sync_error;
            previous_correction = conn->amountStuffed;

            tsum_of_sync_errors += sync_error;
            tsum_of_drifts += statistics[newest_statistic].drift;
            if (conn->amountStuffed > 0) {
              tsum_of_insertions_and_deletions += conn->amountStuffed;
            } else {
              tsum_of_insertions_and_deletions -= conn->amountStuffed;
            }
            tsum_of_corrections += conn->amountStuffed;
            conn->session_corrections += conn->amountStuffed;

            newest_statistic = (newest_statistic + 1) % trend_interval;
            number_of_statistics++;
          }
        }
        if (play_number % print_interval == 0) {

          // here, calculate the input and output frame rates, where possible, even if statistics
          // have not been requested
          // this is to calculate them in case they are needed by the D-Bus interface or elsewhere.

          if (conn->input_frame_rate_starting_point_is_valid) {
            uint64_t elapsed_reception_time, frames_received;
            elapsed_reception_time =
                conn->frames_inward_measurement_time - conn->frames_inward_measurement_start_time;
            frames_received = conn->frames_inward_frames_received_at_measurement_time -
                              conn->frames_inward_frames_received_at_measurement_start_time;
            conn->input_frame_rate =
                (1.0 * frames_received) /
                elapsed_reception_time; // an IEEE double calculation with two 64-bit integers
            conn->input_frame_rate =
                conn->input_frame_rate * (uint64_t)0x100000000; // this should just change the
                                                                // [binary] exponent in the IEEE FP
                                                                // representation; the mantissa
                                                                // should be unaffected.
          } else {
            conn->input_frame_rate = 0.0;
          }

          if ((config.output->delay) && (config.no_sync == 0) && (config.output->rate_info)) {
            uint64_t elapsed_play_time, frames_played;
            if (config.output->rate_info(&elapsed_play_time, &frames_played) == 0)
              conn->frame_rate_status = 1;
            else
              conn->frame_rate_status = 0;
            if (conn->frame_rate_status) {
              conn->frame_rate =
                  (1.0 * frames_played) /
                  elapsed_play_time; // an IEEE double calculation with two 64-bit integers
              conn->frame_rate =
                  conn->frame_rate * (uint64_t)0x100000000; // this should just change the [binary]
              // exponent in the IEEE FP representation;
              // the mantissa should be unaffected.
            } else {
              conn->frame_rate = 0.0;
            }
          }

          // we can now calculate running averages for sync error (frames), corrections (ppm),
          // insertions plus deletions (ppm), drift (ppm)
          double moving_average_sync_error = (1.0 * tsum_of_sync_errors) / number_of_statistics;
          double moving_average_correction = (1.0 * tsum_of_corrections) / number_of_statistics;
          double moving_average_insertions_plus_deletions =
              (1.0 * tsum_of_insertions_and_deletions) / number_of_statistics;
          // double moving_average_drift = (1.0 * tsum_of_drifts) / number_of_statistics;
          // if ((play_number/print_interval)%20==0)
          if (config.statistics_requested) {
            if (at_least_one_frame_seen) {

              if ((config.output->delay)) {
                if (config.no_sync == 0) {
                  inform("%*.2f,"        /* Sync error in milliseconds */
                         "%*.1f,"        /* net correction in ppm */
                         "%*.1f,"        /* corrections in ppm */
                         "%*d,"          /* total packets */
                         "%*" PRIu64 "," /* missing packets */
                         "%*" PRIu64 "," /* late packets */
                         "%*" PRIu64 "," /* too late packets */
                         "%*" PRIu64 "," /* resend requests */
                         "%*" PRId64 "," /* min DAC queue size */
                         "%*" PRId32 "," /* min buffer occupancy */
                         "%*" PRId32 "," /* max buffer occupancy */
                         "%*.2f,"        /* source nominal frame rate */
                         "%*.2f,"        /* source actual (average) frame rate */
                         "%*.2f,"        /* output frame rate */
                         "%*.2f,"        /* source clock drift */
                         "%*d,"          /* source clock drift sample count */
                         "%*.2f",        /* rough calculated correction in ppm */
                         10,
                         1000 * moving_average_sync_error / config.output_rate, 10,
                         moving_average_correction * 1000000 / (352 * conn->output_sample_ratio),
                         10, moving_average_insertions_plus_deletions * 1000000 /
                                 (352 * conn->output_sample_ratio),
                         12, play_number, 7, conn->missing_packets, 7, conn->late_packets, 7,
                         conn->too_late_packets, 7, conn->resend_requests, 7,
                         minimum_dac_queue_size, 5, minimum_buffer_occupancy, 5,
                         maximum_buffer_occupancy, 11, conn->remote_frame_rate, 11,
                         conn->input_frame_rate, 11, conn->frame_rate, 10,
                         (conn->local_to_remote_time_gradient - 1.0) * 1000000, 6,
                         conn->local_to_remote_time_gradient_sample_count, 10,
                         (conn->frame_rate > 0.0)
                             ? ((conn->frame_rate -
                                 conn->remote_frame_rate * conn->output_sample_ratio *
                                     conn->local_to_remote_time_gradient) *
                                1000000) /
                                   conn->frame_rate
                             : 0.0);
                } else {
                  inform("%*.2f,"        /* Sync error in milliseconds */
                         "%*d,"          /* total packets */
                         "%*" PRIu64 "," /* missing packets */
                         "%*" PRIu64 "," /* late packets */
                         "%*" PRIu64 "," /* too late packets */
                         "%*" PRIu64 "," /* resend requests */
                         "%*" PRId64 "," /* min DAC queue size */
                         "%*" PRId32 "," /* min buffer occupancy */
                         "%*" PRId32 "," /* max buffer occupancy */
                         "%*.2f,"        /* source nominal frame rate */
                         "%*.2f,"        /* source actual (average) frame rate */
                         "%*.2f,"        /* source clock drift */
                         "%*d",          /* source clock drift sample count */
                         10,
                         1000 * moving_average_sync_error / config.output_rate, 12, play_number, 7,
                         conn->missing_packets, 7, conn->late_packets, 7, conn->too_late_packets, 7,
                         conn->resend_requests, 7, minimum_dac_queue_size, 5,
                         minimum_buffer_occupancy, 5, maximum_buffer_occupancy, 11,
                         conn->remote_frame_rate, 11, conn->input_frame_rate, 10,
                         (conn->local_to_remote_time_gradient - 1.0) * 1000000, 6,
                         conn->local_to_remote_time_gradient_sample_count);
                }
              } else {
                inform("%*.2f,"        /* Sync error in milliseconds */
                       "%*d,"          /* total packets */
                       "%*" PRIu64 "," /* missing packets */
                       "%*" PRIu64 "," /* late packets */
                       "%*" PRIu64 "," /* too late packets */
                       "%*" PRIu64 "," /* resend requests */
                       "%*" PRId32 "," /* min buffer occupancy */
                       "%*" PRId32 "," /* max buffer occupancy */
                       "%*.2f,"        /* source nominal frame rate */
                       "%*.2f,"        /* source actual (average) frame rate */
                       "%*.2f,"        /* source clock drift */
                       "%*d",          /* source clock drift sample count */
                       10,
                       1000 * moving_average_sync_error / config.output_rate, 12, play_number, 7,
                       conn->missing_packets, 7, conn->late_packets, 7, conn->too_late_packets, 7,
                       conn->resend_requests, 5, minimum_buffer_occupancy, 5,
                       maximum_buffer_occupancy, 11, conn->remote_frame_rate, 11,
                       conn->input_frame_rate, 10,
                       (conn->local_to_remote_time_gradient - 1.0) * 1000000, 6,
                       conn->local_to_remote_time_gradient_sample_count);
              }
            } else {
              inform("No frames received in the last sampling interval.");
            }
          }
          minimum_dac_queue_size = INT64_MAX;   // hack reset
          maximum_buffer_occupancy = INT32_MIN; // can't be less than this
          minimum_buffer_occupancy = INT32_MAX; // can't be more than this
          at_least_one_frame_seen = 0;
        }
      }
    }
  }

  debug(1, "This should never be called.");
  pthread_cleanup_pop(1); // pop the cleanup handler
                          //  debug(1, "This should never be called either.");
                          //  pthread_cleanup_pop(1); // pop the initial cleanup handler
  pthread_exit(NULL);
}

void player_volume_without_notification(double airplay_volume, rtsp_conn_info *conn) {
  debug_mutex_lock(&conn->volume_control_mutex, 5000, 1);
  debug(2, "player_volume_without_notification %f", airplay_volume);
  // first, see if we are hw only, sw only, both with hw attenuation on the top or both with sw
  // attenuation on top

  enum volume_mode_type { vol_sw_only, vol_hw_only, vol_both } volume_mode;

  // take account of whether there is a hardware mixer, if a max volume has been specified and if a
  // range has been specified
  // the range might imply that both hw and software mixers are needed, so calculate this

  int32_t hw_max_db = 0, hw_min_db = 0; // zeroed to quieten an incorrect uninitialised warning
  int32_t sw_max_db = 0, sw_min_db = -9630;
  if (config.output->parameters) {
    volume_mode = vol_hw_only;
    audio_parameters audio_information;
    config.output->parameters(&audio_information);
    hw_max_db = audio_information.maximum_volume_dB;
    hw_min_db = audio_information.minimum_volume_dB;
    if (config.volume_max_db_set) {
      if (((config.volume_max_db * 100) <= hw_max_db) &&
          ((config.volume_max_db * 100) >= hw_min_db))
        hw_max_db = (int32_t)config.volume_max_db * 100;
      else if (config.volume_range_db) {
        hw_max_db = hw_min_db;
        sw_max_db = (config.volume_max_db * 100) - hw_min_db;
      } else {
        warn("The maximum output level is outside the range of the hardware mixer -- ignored");
      }
    }

    // here, we have set limits on the hw_max_db and the sw_max_db
    // but we haven't actually decided whether we need both hw and software attenuation
    // only if a range is specified could we need both
    if (config.volume_range_db) {
      // see if the range requested exceeds the hardware range available
      int32_t desired_range_db = (int32_t)trunc(config.volume_range_db * 100);
      if ((desired_range_db) > (hw_max_db - hw_min_db)) {
        volume_mode = vol_both;
        int32_t desired_sw_range = desired_range_db - (hw_max_db - hw_min_db);
        if ((sw_max_db - desired_sw_range) < sw_min_db)
          warn("The range requested is too large to accommodate -- ignored.");
        else
          sw_min_db = (sw_max_db - desired_sw_range);
      }
    }
  } else {
    // debug(1,"has no hardware mixer");
    volume_mode = vol_sw_only;
    if (config.volume_max_db_set) {
      if (((config.volume_max_db * 100) <= sw_max_db) &&
          ((config.volume_max_db * 100) >= sw_min_db))
        sw_max_db = (int32_t)config.volume_max_db * 100;
    }
    if (config.volume_range_db) {
      // see if the range requested exceeds the software range available
      int32_t desired_range_db = (int32_t)trunc(config.volume_range_db * 100);
      if ((desired_range_db) > (sw_max_db - sw_min_db))
        warn("The range requested is too large to accommodate -- ignored.");
      else
        sw_min_db = (sw_max_db - desired_range_db);
    }
  }

  // here, we know whether it's hw volume control only, sw only or both, and we have the hw and sw
  // limits.
  // if it's both, we haven't decided whether hw or sw should be on top
  // we have to consider the settings ignore_volume_control and mute.

  if (config.ignore_volume_control == 0) {
    if (airplay_volume == -144.0) {

      if ((config.output->mute) && (config.output->mute(1) == 0))
        debug(2, "player_volume_without_notification: volume mode is %d, airplay_volume is %f, "
                 "hardware mute is enabled.",
              volume_mode, airplay_volume);
      else {
        conn->software_mute_enabled = 1;
        debug(2, "player_volume_without_notification: volume mode is %d, airplay_volume is %f, "
                 "software mute is enabled.",
              volume_mode, airplay_volume);
      }

    } else {
      int32_t max_db = 0, min_db = 0;
      switch (volume_mode) {
      case vol_hw_only:
        max_db = hw_max_db;
        min_db = hw_min_db;
        break;
      case vol_sw_only:
        max_db = sw_max_db;
        min_db = sw_min_db;
        break;
      case vol_both:
        // debug(1, "dB range passed is hw: %d, sw: %d, total: %d", hw_max_db - hw_min_db,
        //      sw_max_db - sw_min_db, (hw_max_db - hw_min_db) + (sw_max_db - sw_min_db));
        max_db =
            (hw_max_db - hw_min_db) + (sw_max_db - sw_min_db); // this should be the range requested
        min_db = 0;
        break;
      default:
        debug(1, "player_volume_without_notification: error: not in a volume mode");
        break;
      }
      double scaled_attenuation = 0.0;
      if (config.volume_control_profile == VCP_standard)
        scaled_attenuation = vol2attn(airplay_volume, max_db, min_db); // no cancellation points
      else if (config.volume_control_profile == VCP_flat)
        scaled_attenuation =
            flat_vol2attn(airplay_volume, max_db, min_db); // no cancellation points
      else
        debug(1, "player_volume_without_notification: unrecognised volume control profile");

      // so here we have the scaled attenuation. If it's for hw or sw only, it's straightforward.
      double hardware_attenuation = 0.0;
      double software_attenuation = 0.0;

      switch (volume_mode) {
      case vol_hw_only:
        hardware_attenuation = scaled_attenuation;
        break;
      case vol_sw_only:
        software_attenuation = scaled_attenuation;
        break;
      case vol_both:
        // here, we now the attenuation required, so we have to apportion it to the sw and hw mixers
        // if we give the hw priority, that means when lowering the volume, set the hw volume to its
        // lowest
        // before using the sw attenuation.
        // similarly, if we give the sw priority, that means when lowering the volume, set the sw
        // volume to its lowest
        // before using the hw attenuation.
        // one imagines that hw priority is likely to be much better
        // if (config.volume_range_hw_priority) {
        if (config.volume_range_hw_priority != 0) {
          // hw priority
          if ((sw_max_db - sw_min_db) > scaled_attenuation) {
            software_attenuation = sw_min_db + scaled_attenuation;
            hardware_attenuation = hw_min_db;
          } else {
            software_attenuation = sw_max_db;
            hardware_attenuation = hw_min_db + scaled_attenuation - (sw_max_db - sw_min_db);
          }
        } else {
          // sw priority
          if ((hw_max_db - hw_min_db) > scaled_attenuation) {
            hardware_attenuation = hw_min_db + scaled_attenuation;
            software_attenuation = sw_min_db;
          } else {
            hardware_attenuation = hw_max_db;
            software_attenuation = sw_min_db + scaled_attenuation - (hw_max_db - hw_min_db);
          }
        }
        break;
      default:
        debug(1, "player_volume_without_notification: error: not in a volume mode");
        break;
      }

      if (((volume_mode == vol_hw_only) || (volume_mode == vol_both)) && (config.output->volume)) {
        config.output->volume(hardware_attenuation); // otherwise set the output to the lowest value
        // debug(1,"Hardware attenuation set to %f for airplay volume of
        // %f.",hardware_attenuation,airplay_volume);
        if (volume_mode == vol_hw_only)
          conn->fix_volume = 0x10000;
      }

      if ((volume_mode == vol_sw_only) || (volume_mode == vol_both)) {
        double temp_fix_volume = 65536.0 * pow(10, software_attenuation / 2000);
        // debug(1,"Software attenuation set to %f, i.e %f out of 65,536, for airplay volume of
        // %f",software_attenuation,temp_fix_volume,airplay_volume);

        conn->fix_volume = temp_fix_volume;

        if (config.loudness)
          loudness_set_volume(software_attenuation / 100);
      }

      if (config.logOutputLevel) {
        inform("Output Level set to: %.2f dB.", scaled_attenuation / 100.0);
      }

#ifdef CONFIG_METADATA
      char *dv = malloc(128); // will be freed in the metadata thread
      if (dv) {
        memset(dv, 0, 128);
        if (config.ignore_volume_control == 1)
          snprintf(dv, 127, "%.2f,%.2f,%.2f,%.2f", airplay_volume, 0.0, 0.0, 0.0);
        else
          snprintf(dv, 127, "%.2f,%.2f,%.2f,%.2f", airplay_volume, scaled_attenuation / 100.0,
                   min_db / 100.0, max_db / 100.0);
        send_ssnc_metadata('pvol', dv, strlen(dv), 1);
      }
#endif
      // here, store the volume for possible use in the future

      if (config.output->mute)
        config.output->mute(0);
      conn->software_mute_enabled = 0;

      debug(2, "player_volume_without_notification: volume mode is %d, airplay volume is %f, "
               "software_attenuation: %f, hardware_attenuation: %f, muting "
               "is disabled.",
            volume_mode, airplay_volume, software_attenuation, hardware_attenuation);
    }
  }
  config.airplay_volume = airplay_volume;
  debug_mutex_unlock(&conn->volume_control_mutex, 3);
}

void player_volume(double airplay_volume, rtsp_conn_info *conn) {
  command_set_volume(airplay_volume);
  player_volume_without_notification(airplay_volume, conn);
}

void do_flush(uint32_t timestamp, rtsp_conn_info *conn) {

  debug(3, "Flush requested up to %u. It seems as if 0 is special.", timestamp);
  debug_mutex_lock(&conn->flush_mutex, 1000, 1);
  conn->flush_requested = 1;
  // if (timestamp!=0)
  conn->flush_rtp_timestamp = timestamp; // flush all packets up to (and including?) this
  // conn->play_segment_reference_frame = 0;
  reset_input_flow_metrics(conn);
  debug_mutex_unlock(&conn->flush_mutex, 3);
  debug(3, "Flush request made.");
}

void player_flush(uint32_t timestamp, rtsp_conn_info *conn) {
  debug(3, "player_flush");
  do_flush(timestamp, conn);
#ifdef CONFIG_METADATA
  // only send a flush metadata message if the first packet has been seen -- it's a bogus message
  // otherwise
  if (conn->first_packet_timestamp) {
    debug(2, "pfls");
    send_ssnc_metadata('pfls', NULL, 0, 1); // contains cancellation points
  }
#endif
}

int player_play(rtsp_conn_info *conn) {
  // need to use conn in place of stream below. Need to put the stream as a parameter to he
  if (conn->player_thread != NULL)
    die("Trying to create a second player thread for this RTSP session");
  if (config.buffer_start_fill > BUFFER_FRAMES)
    die("specified buffer starting fill %d > buffer size %d", config.buffer_start_fill,
        BUFFER_FRAMES);
  activity_monitor_signify_activity(
      1); // active, and should be before play's command hook, command_start()
  command_start();
  // call on the output device to prepare itself
  if ((config.output) && (config.output->prepare))
    config.output->prepare();

  pthread_t *pt = malloc(sizeof(pthread_t));
  if (pt == NULL)
    die("Couldn't allocate space for pthread_t");
  conn->player_thread = pt;
  size_t size = (PTHREAD_STACK_MIN + 256 * 1024);
  pthread_attr_t tattr;
  pthread_attr_init(&tattr);
  int rc = pthread_attr_setstacksize(&tattr, size);
  if (rc)
    debug(1, "Error setting stack size for player_thread: %s", strerror(errno));
  // finished initialising.
  rc = pthread_create(pt, &tattr, player_thread_func, (void *)conn);
  if (rc)
    debug(1, "Error creating player_thread: %s", strerror(errno));
  pthread_attr_destroy(&tattr);
#ifdef CONFIG_METADATA
  debug(2, "pbeg");
  send_ssnc_metadata('pbeg', NULL, 0, 1); // contains cancellation points
#endif
  return 0;
}

int player_stop(rtsp_conn_info *conn) {
  // note -- this may be called from another connection thread.
  // int dl = debuglev;
  // debuglev = 3;
  debug(3, "player_stop");
  if (conn->player_thread) {
    debug(3, "player_thread cancel...");
    pthread_cancel(*conn->player_thread);
    debug(3, "player_thread join...");
    if (pthread_join(*conn->player_thread, NULL) == -1) {
      char errorstring[1024];
      strerror_r(errno, (char *)errorstring, sizeof(errorstring));
      debug(1, "Connection %d: error %d joining player thread: \"%s\".", conn->connection_number,
            errno, (char *)errorstring);
    } else {
      debug(3, "player_thread joined.");
    }
    free(conn->player_thread);
    conn->player_thread = NULL;
#ifdef CONFIG_METADATA
    debug(2, "pend");
    send_ssnc_metadata('pend', NULL, 0, 1); // contains cancellation points
#endif
    // debuglev = dl;
    command_stop();
    activity_monitor_signify_activity(0); // inactive, and should be after command_stop()
    return 0;
  } else {
    debug(3, "Connection %d: player thread already deleted.", conn->connection_number);
    // debuglev = dl;
    return -1;
  }
}