ChangeLog

openal-soft-1.24.2:

    Implemented the AL_SOFT_bformat_hoa extension.

    Implemented default device change events for the PulseAudio backend.

    Implemented an option for WASAPI exclusive mode playback.

    Fixed reverb being too quiet for sounds from different directions.

    Fixed compiling with certain versions of Clang.

    Fixed compiling for some older macOS versions.

    Fixed building alffplay on systems without pkg-config.

    Improved output format detection for CoreAudio.

    Changed the default resampler back to Cubic Spline.

    Added an SDL3 playback backend. Disabled by default to avoid a runtime
    dependency and for compatibility; a single process can't safely use SDL2
    and SDL3 together on some OSs, so enable with care.

    Converted examples from SDL2 to SDL3.

    Integrated fmtlib into the main library and router for logging and string
    formatting.

openal-soft-1.24.1:

    Fixed compilation on PowerPC.

    Fixed compilation on some targets that lack lock-free 64-bit atomics.

    Fixed a crash when parsing certain option values.

    Fixed applying noexcept in the public headers with MSVC.

    Fixed building for UWP with vcpkg.

    Improved compatibility when compiling as C++20 or later.

    Integrated fmtlib for some examples and utilities.

openal-soft-1.24.0:

    Updated library codebase to C++17.

    Implemented the ALC_SOFT_system_events extension.

    Implemented the AL_EXT_debug extension.

    Implemented the AL_EXT_direct_context extension.

    Implemented speaker configuration and headphones detection on CoreAudio.

    Fixed a potential crash with some extension functions on 32-bit Windows.

    Fixed a crash that can occur when stopping playback with the Oboe backend.

    Fixed calculating the reverb room rolloff.

    Fixed EAX occlusion, obstruction, and exclusion low-pass filter strength.

    Fixed EAX distance factor calculations.

    Fixed querying AL_EFFECTSLOT_EFFECT on auxiliary effect slots.

    Fixed compilation on some macOS systems that lack libdispatch.

    Fixed compilation as a subproject with MinGW.

    Changed the context error state to be thread-local. This is technically out
    of spec, but necessary to avoid race conditions with multi-threaded use.

    Split the cubic resampler into 4-point spline and gaussian variants. The
    latter prioritizing the suppression of aliasing distortion and harmonics,
    the former not reducing high frequencies as much.

    Improved timing precision of starting delayed sources.

    Improved ring modulator quality.

    Improved performance of convolution reverb.

    Improved WASAPI device enumeration performance.

    Added UWP support.

    Added 'noexcept' to functions and function types when compiled as C++. As a
    C API, OpenAL can't be expected to throw C++ exceptions, nor can it handle
    them if they leave a callback.

    Added an experimental config option for using WASAPI spatial audio output.

    Added enumeration support to the PortAudio backend.

    Added compatibility options to override the AL_VENDOR, AL_VERSION, and
    AL_RENDERER strings.

    Added an example to play LAF files.

    Disabled real-time mixing by default for PipeWire playback.

    Disabled the SndIO backend by default on non-BSD targets.

openal-soft-1.23.1:

    Implemented the AL_SOFT_UHJ_ex extension.

    Implemented the AL_SOFT_buffer_length_query extension.

    Implemented the AL_SOFT_source_start_delay extension.

    Implemented the AL_EXT_STATIC_BUFFER extension.

    Fixed compiling with certain older versions of GCC.

    Fixed compiling as a submodule.

    Fixed compiling with newer versions of Oboe.

    Improved EAX effect version switching.

    Improved the quality of the reverb modulator.

    Improved performance of the cubic resampler.

    Added a compatibility option to restore AL_SOFT_buffer_sub_data. The option
    disables AL_EXT_SOURCE_RADIUS due to incompatibility.

    Reduced CPU usage when EAX is initialized and FXSlot0 or FXSlot1 are not
    used.

    Reduced memory usage for ADPCM buffer formats. They're no longer converted
    to 16-bit samples on load.

openal-soft-1.23.0:

    Fixed CoreAudio capture support.

    Fixed handling per-version EAX properties.

    Fixed interpolating changes to the Super Stereo width source property.

    Fixed detection of the update and buffer size from PipeWire.

    Fixed resuming playback devices with OpenSL.

    Fixed support for certain OpenAL implementations with the router.

    Improved reverb environment transitions.

    Improved performance of convolution reverb.

    Improved quality and performance of the pitch shifter effect slightly.

    Improved sub-sample precision for resampled sources.

    Improved blending spatialized multi-channel sources that use the source
    radius property.

    Improved mixing 2D ambisonic sources for higher-order 3D ambisonic mixing.

    Improved quadraphonic and 7.1 surround sound output slightly.

    Added config options for UHJ encoding/decoding quality. Including Super
    Stereo processing.

    Added a config option for specifying the speaker distance.

    Added a compatibility config option for specifying the NFC distance
    scaling.

    Added a config option for mixing on PipeWire's non-real-time thread.

    Added support for virtual source nodes with PipeWire capture.

    Added the ability for the WASAPI backend to use different playback rates.

    Added support for SOFA files that define per-response delays in makemhr.

    Changed the default fallback playback sample rate to 48khz. This doesn't
    affect most backends, which can detect a default rate from the system.

    Changed the default resampler to cubic.

    Changed the default HRTF size from 32 to 64 points.

openal-soft-1.22.2:

    Fixed PipeWire version check.

    Fixed building with PipeWire versions before 0.3.33.

openal-soft-1.22.1:

    Fixed CoreAudio capture.

    Fixed air absorption strength.

    Fixed handling 5.1 devices on Windows that use Rear channels instead of
    Side channels.

    Fixed some compilation issues on MinGW.

    Fixed ALSA not being used on some systems without PipeWire and PulseAudio.

    Fixed OpenSL capturing noise.

    Fixed Oboe capture failing with some buffer sizes.

    Added checks for the runtime PipeWire version. The same or newer version
    than is used for building will be needed at runtime for the backend to
    work.

    Separated 3D7.1 into its own speaker configuration.

openal-soft-1.22.0:

    Implemented the ALC_SOFT_reopen_device extension. This allows for moving
    devices to different outputs without losing object state.

    Implemented the ALC_SOFT_output_mode extension.

    Implemented the AL_SOFT_callback_buffer extension.

    Implemented the AL_SOFT_UHJ extension. This supports native UHJ buffer
    formats and Super Stereo processing.

    Implemented the legacy EAX extensions. Enabled by default only on Windows.

    Improved sound positioning stability when a source is near the listener.

    Improved the default 5.1 output decoder.

    Improved the high frequency response for the HRTF second-order ambisonic
    decoder.

    Improved SoundIO capture behavior.

    Fixed UHJ output on NEON-capable CPUs.

    Fixed redundant effect updates when setting an effect property to the
    current value.

    Fixed WASAPI capture using really low sample rates, and sources with very
    high pitch shifts when using a bsinc resampler.

    Added a PipeWire backend.

    Added enumeration for the JACK and CoreAudio backends.

    Added optional support for RTKit to get real-time priority. Only used as a
    backup when pthread_setschedparam fails.

    Added an option for JACK playback to render directly in the real-time
    processing callback. For lower playback latency, on by default.

    Added an option for custom JACK devices.

    Added utilities to encode and decode UHJ audio files. Files are decoded to
    the .amb format, and are encoded from libsndfile-compatible formats.

    Added an in-progress extension to hold sources in a playing state when a
    device disconnects. Allows devices to be reset or reopened and have sources
    resume from where they left off.

    Lowered the priority of the JACK backend. To avoid it getting picked when
    PipeWire is providing JACK compatibility, since the JACK backend is less
    robust with auto-configuration.

openal-soft-1.21.1:

    Improved alext.h's detection of standard types.

    Improved slightly the local source position when the listener and source
    are near each other.

    Improved click/pop prevention for sounds that stop prematurely.

    Fixed compilation for Windows ARM targets with MSVC.

    Fixed ARM NEON detection on Windows.

    Fixed CoreAudio capture when the requested sample rate doesn't match the
    system configuration.

    Fixed OpenSL capture desyncing from the internal capture buffer.

    Fixed sources missing a batch update when applied after quickly restarting
    the source.

    Fixed missing source stop events when stopping a paused source.

    Added capture support to the experimental Oboe backend.

openal-soft-1.21.0:

    Updated library codebase to C++14.

    Implemented the AL_SOFT_effect_target extension.

    Implemented the AL_SOFT_events extension.

    Implemented the ALC_SOFT_loopback_bformat extension.

    Improved memory use for mixing voices.

    Improved detection of NEON capabilities.

    Improved handling of PulseAudio devices that lack manual start control.

    Improved mixing performance with PulseAudio.

    Improved high-frequency scaling quality for the HRTF B-Format decoder.

    Improved makemhr's HRIR delay calculation.

    Improved WASAPI capture of mono formats with multichannel input.

    Reimplemented the modulation stage for reverb.

    Enabled real-time mixing priority by default, for backends that use the
    setting. It can still be disabled in the config file.

    Enabled dual-band processing for the built-in quad and 7.1 output decoders.

    Fixed a potential crash when deleting an effect slot immediately after the
    last source using it stops.

    Fixed building with the static runtime on MSVC.

    Fixed using source stereo angles outside of -pi...+pi.

    Fixed the buffer processed event count for sources that start with empty
    buffers.

    Fixed trying to open an unopenable WASAPI device causing all devices to
    stop working.

    Fixed stale devices when re-enumerating WASAPI devices.

    Fixed using unicode paths with the log file on Windows.

    Fixed DirectSound capture reporting bad sample counts or erroring when
    reading samples.

    Added an in-progress extension for a callback-driven buffer type.

    Added an in-progress extension for higher-order B-Format buffers.

    Added an in-progress extension for convolution reverb.

    Added an experimental Oboe backend for Android playback. This requires the
    Oboe sources at build time, so that it's built as a static library included
    in libopenal.

    Added an option for auto-connecting JACK ports.

    Added greater-than-stereo support to the SoundIO backend.

    Modified the mixer to be fully asynchronous with the external API, and
    should now be real-time safe. Although alcRenderSamplesSOFT is not due to
    locking to check the device handle validity.

    Modified the UHJ encoder to use an all-pass FIR filter that's less harmful
    to non-filtered signal phase.

    Converted examples from SDL_sound to libsndfile. To avoid issues when
    combining SDL2 and SDL_sound.

    Worked around a 32-bit GCC/MinGW bug with TLS destructors. See:
    https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83562

    Reduced the maximum number of source sends from 16 to 6.

    Removed the QSA backend. It's been broken for who knows how long.

    Got rid of the compile-time native-tools targets, using cmake and global
    initialization instead. This should make cross-compiling less troublesome.

openal-soft-1.20.1:

    Implemented the AL_SOFT_direct_channels_remix extension. This extends
    AL_DIRECT_CHANNELS_SOFT to optionally remix input channels that don't have
    a matching output channel.

    Implemented the AL_SOFT_bformat_ex extension. This extends B-Format buffer
    support for N3D or SN3D scaling, or ACN channel ordering.

    Fixed a potential voice leak when a source is started and stopped or
    restarted in quick succession.

    Fixed a potential device reset failure with JACK.

    Improved handling of unsupported channel configurations with WASAPI. Such
    setups will now try to output at least a stereo mix.

    Improved clarity a bit for the HRTF second-order ambisonic decoder.

    Improved detection of compatible layouts for SOFA files in makemhr and
    sofa-info.

    Added the ability to resample HRTFs on load. MHR files no longer need to
    match the device sample rate to be usable.

    Added an option to limit the HRTF's filter length.

openal-soft-1.20.0:

    Converted the library codebase to C++11. A lot of hacks and custom
    structures have been replaced with standard or cleaner implementations.

    Partially implemented the Vocal Morpher effect.

    Fixed the bsinc SSE resamplers on non-GCC compilers.

    Fixed OpenSL capture.

    Fixed support for extended capture formats with OpenSL.

    Fixed handling of WASAPI not reporting a default device.

    Fixed performance problems relating to semaphores on macOS.

    Modified the bsinc12 resampler's transition band to better avoid aliasing
    noise.

    Modified alcResetDeviceSOFT to attempt recovery of disconnected devices.

    Modified the virtual speaker layout for HRTF B-Format decoding.

    Modified the PulseAudio backend to use a custom processing loop.

    Renamed the makehrtf utility to makemhr.

    Improved the efficiency of the bsinc resamplers when up-sampling.

    Improved the quality of the bsinc resamplers slightly.

    Improved the efficiency of the HRTF filters.

    Improved the HRTF B-Format decoder coefficient generation.

    Improved reverb feedback fading to be more consistent with pan fading.

    Improved handling of sources that end prematurely, avoiding loud clicks.

    Improved the performance of some reverb processing loops.

    Added fast_bsinc12 and 24 resamplers that improve efficiency at the cost of
    some quality. Notably, down-sampling has less smooth pitch ramping.

    Added support for SOFA input files with makemhr.

    Added a build option to use pre-built native tools. For cross-compiling,
    use with caution and ensure the native tools' binaries are kept up-to-date.

    Added an adjust-latency config option for the PulseAudio backend.

    Added basic support for multi-field HRTFs.

    Added an option for mixing first- or second-order B-Format with HRTF
    output. This can improve HRTF performance given a number of sources.

    Added an RC file for proper DLL version information.

    Disabled some old KDE workarounds by default. Specifically, PulseAudio
    streams can now be moved (KDE may try to move them after opening).

openal-soft-1.19.1:

    Implemented capture support for the SoundIO backend.

    Fixed source buffer queues potentially not playing properly when a queue
    entry completes.

    Fixed possible unexpected failures when generating auxiliary effect slots.

    Fixed a crash with certain reverb or device settings.

    Fixed OpenSL capture.

    Improved output limiter response, better ensuring the sample amplitude is
    clamped for output.

openal-soft-1.19.0:

    Implemented the ALC_SOFT_device_clock extension.

    Implemented the Pitch Shifter, Frequency Shifter, and Autowah effects.

    Fixed compiling on FreeBSD systems that use freebsd-lib 9.1.

    Fixed compiling on NetBSD.

    Fixed the reverb effect's density scale and panning parameters.

    Fixed use of the WASAPI backend with certain games, which caused odd COM
    initialization errors.

    Increased the number of virtual channels for decoding Ambisonics to HRTF
    output.

    Changed 32-bit x86 builds to use SSE2 math by default for performance.
    Build-time options are available to use just SSE1 or x87 instead.

    Replaced the 4-point Sinc resampler with a more efficient cubic resampler.

    Renamed the MMDevAPI backend to WASAPI.

    Added support for 24-bit, dual-ear HRTF data sets. The built-in data set
    has been updated to 24-bit.

    Added a 24- to 48-point band-limited Sinc resampler.

    Added an SDL2 playback backend. Disabled by default to avoid a dependency
    on SDL2.

    Improved the performance and quality of the Chorus and Flanger effects.

    Improved the efficiency of the band-limited Sinc resampler.

    Improved the Sinc resampler's transition band to avoid over-attenuating
    higher frequencies.

    Improved the performance of some filter operations.

    Improved the efficiency of object ID lookups.

    Improved the efficienty of internal voice/source synchronization.

    Improved AL call error logging with contextualized messages.

    Removed the reverb effect's modulation stage. Due to the lack of reference
    for its intended behavior and strength.

openal-soft-1.18.2:

    Fixed resetting the FPU rounding mode after certain function calls on
    Windows.

    Fixed use of SSE intrinsics when building with Clang on Windows.

    Fixed a crash with the JACK backend when using JACK1.

    Fixed use of pthread_setnane_np on NetBSD.

    Fixed building on FreeBSD with an older freebsd-lib.

    OSS now links with libossaudio if found at build time (for NetBSD).

openal-soft-1.18.1:

    Fixed an issue where resuming a source might not restart playing it.

    Fixed PulseAudio playback when the configured stream length is much less
    than the requested length.

    Fixed MMDevAPI capture with sample rates not matching the backing device.

    Fixed int32 output for the Wave Writer.

    Fixed enumeration of OSS devices that are missing device files.

    Added correct retrieval of the executable's path on FreeBSD.

    Added a config option to specify the dithering depth.

    Added a 5.1 decoder preset that excludes front-center output.

openal-soft-1.18.0:

    Implemented the AL_EXT_STEREO_ANGLES and AL_EXT_SOURCE_RADIUS extensions.

    Implemented the AL_SOFT_gain_clamp_ex, AL_SOFT_source_resampler,
    AL_SOFT_source_spatialize, and ALC_SOFT_output_limiter extensions.

    Implemented 3D processing for some effects. Currently implemented for
    Reverb, Compressor, Equalizer, and Ring Modulator.

    Implemented 2-channel UHJ output encoding. This needs to be enabled with a
    config option to be used.

    Implemented dual-band processing for high-quality ambisonic decoding.

    Implemented distance-compensation for surround sound output.

    Implemented near-field emulation and compensation with ambisonic rendering.
    Currently only applies when using the high-quality ambisonic decoder or
    ambisonic output, with appropriate config options.

    Implemented an output limiter to reduce the amount of distortion from
    clipping.

    Implemented dithering for 8-bit and 16-bit output.

    Implemented a config option to select a preferred HRTF.

    Implemented a run-time check for NEON extensions using /proc/cpuinfo.

    Implemented experimental capture support for the OpenSL backend.

    Fixed building on compilers with NEON support but don't default to having
    NEON enabled.

    Fixed support for JACK on Windows.

    Fixed starting a source while alcSuspendContext is in effect.

    Fixed detection of headsets as headphones, with MMDevAPI.

    Added support for AmbDec config files, for custom ambisonic decoder
    configurations. Version 3 files only.

    Added backend-specific options to alsoft-config.

    Added first-, second-, and third-order ambisonic output formats. Currently
    only works with backends that don't rely on channel labels, like JACK,
    ALSA, and OSS.

    Added a build option to embed the default HRTFs into the lib.

    Added AmbDec presets to enable high-quality ambisonic decoding.

    Added an AmbDec preset for 3D7.1 speaker setups.

    Added documentation regarding Ambisonics, 3D7.1, AmbDec config files, and
    the provided ambdec presets.

    Added the ability for MMDevAPI to open devices given a Device ID or GUID
    string.

    Added an option to the example apps to open a specific device.

    Increased the maximum auxiliary send limit to 16 (up from 4). Requires
    requesting them with the ALC_MAX_AUXILIARY_SENDS context creation
    attribute.

    Increased the default auxiliary effect slot count to 64 (up from 4).

    Reduced the default period count to 3 (down from 4).

    Slightly improved automatic naming for enumerated HRTFs.

    Improved B-Format decoding with HRTF output.

    Improved internal property handling for better batching behavior.

    Improved performance of certain filter uses.

    Removed support for the AL_SOFT_buffer_samples and AL_SOFT_buffer_sub_data
    extensions. Due to conflicts with AL_EXT_SOURCE_RADIUS.

openal-soft-1.17.2:

    Implemented device enumeration for OSSv4.

    Fixed building on OSX.

    Fixed building on non-Windows systems without POSIX-2008.

    Fixed Dedicated Dialog and Dedicated LFE effect output.

    Added a build option to override the share install dir.

    Added a build option to static-link libgcc for MinGW.

openal-soft-1.17.1:

    Fixed building with JACK and without PulseAudio.

    Fixed building on FreeBSD.

    Fixed the ALSA backend's allow-resampler option.

    Fixed handling of inexact ALSA period counts.

    Altered device naming scheme on Windows backends to better match other
    drivers.

    Updated the CoreAudio backend to use the AudioComponent API. This clears up
    deprecation warnings for OSX 10.11, although requires OSX 10.6 or newer.

openal-soft-1.17.0:

    Implemented a JACK playback backend.

    Implemented the AL_EXT_BFORMAT and AL_EXT_MULAW_BFORMAT extensions.

    Implemented the ALC_SOFT_HRTF extension.

    Implemented C, SSE3, and SSE4.1 based 4- and 8-point Sinc resamplers.

    Implemented a C and SSE based band-limited Sinc resampler. This does 12- to
    24-point Sinc resampling, and performs anti-aliasing.

    Implemented B-Format output support for the wave file writer. This creates
    FuMa-style first-order Ambisonics wave files (AMB format).

    Implemented a stereo-mode config option for treating stereo modes as either
    speakers or headphones.

    Implemented per-device configuration options.

    Fixed handling of PulseAudio and MMDevAPI devices that have identical
    descriptions.

    Fixed a potential lockup when stopping playback of suspended PulseAudio devices.

    Fixed logging of Unicode characters on Windows.

    Fixed 5.1 surround sound channels. By default it will now use the side
    channels for the surround output. A configuration using rear channels is
    still available.

    Fixed the QSA backend potentially altering the capture format.

    Fixed detecting MMDevAPI's default device.

    Fixed returning the default capture device name.

    Fixed mixing property calculations when deferring context updates.

    Altered the behavior of alcSuspendContext and alcProcessContext to better
    match certain Windows drivers.

    Altered the panning algorithm, utilizing Ambisonics for better side and
    back positioning cues with surround sound output.

    Improved support for certain older Windows apps.

    Improved the alffplay example to support surround sound streams.

    Improved support for building as a sub-project.

    Added an HRTF playback example.

    Added a tone generator output test.

    Added a toolchain to help with cross-compiling to Android.

openal-soft-1.16.0:

    Implemented EFX Chorus, Flanger, Distortion, Equalizer, and Compressor
    effects.

    Implemented high-pass and band-pass EFX filters.

    Implemented the high-pass filter for the EAXReverb effect.

    Implemented SSE2 and SSE4.1 linear resamplers.

    Implemented Neon-enhanced non-HRTF mixers.

    Implemented a QSA backend, for QNX.

    Implemented the ALC_SOFT_pause_device, AL_SOFT_deferred_updates,
    AL_SOFT_block_alignment, AL_SOFT_MSADPCM, and AL_SOFT_source_length
    extensions.

    Fixed resetting mmdevapi backend devices.

    Fixed clamping when converting 32-bit float samples to integer.

    Fixed modulation range in the Modulator effect.

    Several fixes for the OpenSL playback backend.

    Fixed device specifier names that have Unicode characters on Windows.

    Added support for filenames and paths with Unicode (UTF-8) characters on
    Windows.

    Added support for alsoft.conf config files found in XDG Base Directory
    Specification locations (XDG_CONFIG_DIRS and XDG_CONFIG_HOME, or their
    defaults) on non-Windows systems.

    Added a GUI configuration utility (requires Qt 4.8).

    Added support for environment variable expansion in config options (not
    keys or section names).

    Added an example that uses SDL2 and ffmpeg.

    Modified examples to use SDL_sound.

    Modified CMake config option names for better sorting.

    HRTF data sets specified in the hrtf_tables config option may now be
    relative or absolute filenames.

    Made the default HRTF data set an external file, and added a data set for
    48khz playback in addition to 44.1khz.

    Added support for C11 atomic methods.

    Improved support for some non-GNU build systems.

openal-soft-1.15.1:

    Fixed a regression with retrieving the source's AL_GAIN property.

openal-soft-1.15:

    Fixed device enumeration with the OSS backend.

    Reorganized internal mixing logic, so unneeded steps can potentially be
    skipped for better performance.

    Removed the lookup table for calculating the mixing pans. The panning is
    now calculated directly for better precision.

    Improved the panning of stereo source channels when using stereo output.

    Improved source filter quality on send paths.

    Added a config option to allow PulseAudio to move streams between devices.

    The PulseAudio backend will now attempt to spawn a server by default.

    Added a workaround for a DirectSound bug relating to float32 output.

    Added SSE-based mixers, for HRTF and non-HRTF mixing.

    Added support for the new AL_SOFT_source_latency extension.

    Improved ALSA capture by avoiding an extra buffer when using sizes
    supported by the underlying device.

    Improved the makehrtf utility to support new options and input formats.

    Modified the CFLAGS declared in the pkg-config file so the "AL/" portion of
    the header includes can optionally be omitted.

    Added a couple example code programs to show how to apply reverb, and
    retrieve latency.

    The configuration sample is now installed into the share/openal/ directory
    instead of /etc/openal.

    The configuration sample now gets installed by default.