Add non-HDR SWIMulation code used in experiments

SWIM/axis_swim.py

@@ -0,0 +1,102 @@
+import pyaudio
+import os
+import struct
+import numpy as np
+import matplotlib.pyplot as plt
+import time
+from tkinter import TclError
+import math
+
+# %matplotlib tk
+
+# constants
+# CHUNK = 1024 * 2             # samples per frame
+CHUNK = 802 * 2
+FORMAT = pyaudio.paInt16     # audio format (bytes per sample?)
+CHANNELS = 1                 # single channel for microphone
+RATE = 44100                 # samples per second
+
+
+# create matplotlib figure and axes
+fig, ax = plt.subplots(1, figsize=(15, 7), subplot_kw=dict(polar=True))
+
+# pyaudio class instance
+p = pyaudio.PyAudio()
+
+# stream object to get data from microphone
+stream = p.open(
+    format=FORMAT,
+    channels=CHANNELS,
+    rate=RATE,
+    input=True,
+    output=True,
+    frames_per_buffer=CHUNK
+)
+
+# variable for plotting
+xp = np.arange(0, 2*math.pi, 8*math.pi/(CHUNK), dtype=float)
+# x = np.append(xp)
+x = np.append(xp, xp)
+x = np.append(x, x)
+print(x.shape)
+
+print(x)
+# create a line object with random data
+line, = ax.plot(x, np.random.rand(CHUNK), '-', lw=1, color='#330033')
+
+
+# basic formatting for the axes
+ax.set_title('')
+# ax.set_xlabel('samples')
+# ax.set_ylabel('volume')
+ax.set_ylim(-1, 259)
+ax.set_rticks([])
+# ax.set_yticks([])
+ax.set_xticks(np.arange(0,2.0*np.pi, np.pi/2.0))
+ax.set_yticklabels([])
+ax.set_xticklabels([])
+
+
+
+
+ax.grid(True, color='#000000', linestyle='-', linewidth=1)
+
+
+ax.set_facecolor((1,1,1))
+
+plt.show(block=False)
+
+print('stream started')
+
+# for measuring frame rate
+frame_count = 0
+start_time = time.time()
+
+# smallest = 999;
+
+while True:
+    data = stream.read(CHUNK, exception_on_overflow=False) 
+
+    # convert data to integers, make np array, then offset it by 127
+    # Note that this only works on low-gain signals. If you are interested in making this work on high-gain signals, you can play around with this line.
+    data_int = struct.unpack(str(2 * CHUNK) + 'B', data) 
+
+    # create np array and offset by 128
+    data_np = np.array(data_int, dtype='b')[::2] + 129
+
+    line.set_ydata(data_np)
+
+    # update figure canvas
+    try:
+        fig.canvas.draw()
+        fig.canvas.flush_events()
+        frame_count += 1
+
+    except TclError:
+
+        # calculate average frame rate
+        frame_rate = frame_count / (time.time() - start_time)
+
+        print('stream stopped')
+        print('average frame rate = {:.0f} FPS'.format(frame_rate))
+        break

SWIM/radial_swim.py

@@ -0,0 +1,112 @@
+
+"""
+Notebook for streaming data from a microphone in realtime
+
+audio is captured using pyaudio
+then converted from binary data to ints using struct
+then displayed using matplotlib
+
+if you don't have pyaudio, then run
+
+>>> pip install pyaudio
+
+note: with 2048 samples per chunk, I'm getting 20FPS
+"""
+
+import pyaudio
+import os
+import struct
+import numpy as np
+import matplotlib.pyplot as plt
+import time
+from tkinter import TclError
+import math
+
+# use this backend to display in separate Tk window
+# %matplotlib tk
+
+# constants
+# CHUNK = 1024 * 2             # samples per frame
+CHUNK = 802 * 2
+FORMAT = pyaudio.paInt16     # audio format (bytes per sample?)
+CHANNELS = 1                 # single channel for microphone
+RATE = 44100                 # samples per second
+
+
+# create matplotlib figure and axes
+fig, ax = plt.subplots(1, figsize=(15, 7), subplot_kw=dict(polar=True))
+
+# pyaudio class instance
+p = pyaudio.PyAudio()
+
+# stream object to get data from microphone
+stream = p.open(
+    format=FORMAT,
+    channels=CHANNELS,
+    rate=RATE,
+    input=True,
+    output=True,
+    frames_per_buffer=CHUNK
+)
+
+# variable for plotting
+xp = np.arange(0, 2*math.pi, 8*math.pi/(CHUNK), dtype=float)
+# x = np.append(xp)
+x = np.append(xp, xp)
+x = np.append(x, x)
+print(x.shape)
+
+print(x)
+# create a line object with random data
+line, = ax.plot(x, np.random.rand(CHUNK), '-', lw=2, color='#330033')
+
+# basic formatting for the axes
+ax.set_title('')
+# ax.set_xlabel('samples')
+# ax.set_ylabel('volume')
+ax.set_ylim(-1, 259)
+ax.set_xticks([]) 
+ax.set_yticks([]) 
+
+ax.set_facecolor((1,1,1))
+# ax.set_axis_bgcolor((0,0,0))
+# ax.set_xlim(0, 360)
+# plt.setp(ax, xticks=[0, CHUNK, 2 * CHUNK], yticks=[0, 128, 255])
+
+# show the plot
+plt.show(block=False)
+
+print('stream started')
+
+# for measuring frame rate
+frame_count = 0
+start_time = time.time()
+
+# smallest = 999;
+
+while True:
+    # binary data
+    data = stream.read(CHUNK, exception_on_overflow=False)  
+
+    # convert data to integers, make np array, then offset it by 127
+    data_int = struct.unpack(str(2 * CHUNK) + 'B', data)
+
+    # create np array and offset by 128
+    data_np = np.array(data_int, dtype='b')[::2] + 129
+
+    line.set_ydata(data_np)
+
+    # update figure canvas
+    try:
+        fig.canvas.draw()
+        fig.canvas.flush_events()
+        frame_count += 1
+
+    except TclError:
+
+        # calculate average frame rate
+        frame_rate = frame_count / (time.time() - start_time)
+
+        print('stream stopped')
+        print('average frame rate = {:.0f} FPS'.format(frame_rate))
+        break

SWIM/swimulator.py

@@ -0,0 +1,141 @@
+
+"""
+Notebook for streaming data from a microphone in realtime
+
+audio is captured using pyaudio
+then converted from binary data to ints using struct
+then displayed using matplotlib
+
+if you don't have pyaudio, then run
+
+>>> pip install pyaudio
+
+note: with 2048 samples per chunk, I'm getting 20FPS
+"""
+
+import pyaudio
+import os
+import struct
+import numpy as np
+import matplotlib.pyplot as plt
+import time
+from tkinter import TclError
+import math
+
+class SWIMulator:
+
+    """ 
+    Class to vizualize signal input data in a SWIM style in Python
+
+    dataStream: an audio stream (using pyaudio) from which the vizualisation feeds
+    buffSize: The buffer size
+
+    """
+    FORMAT = pyaudio.paInt16
+
+    def __init__(self, buffSize=802, channelCount=1, rate=44100, axes=False):
+        self.chunk = buffSize * 2
+        self.channels = channelCount
+        self.rate = rate
+        self.crosshair = False
+
+        self.init_plot()
+
+        if axes:
+            self.toggle_axes()
+
+    def open_stream(self):
+        # pyaudio class instance
+        p = pyaudio.PyAudio()
+
+        # stream object to get data from microphone
+        self.stream = p.open(
+            format=SWIMulator.FORMAT,
+            channels=self.channels,
+            rate=self.rate,
+            input=True,
+            output=True,
+            frames_per_buffer=self.chunk
+        )
+
+        return
+
+    def init_plot(self):
+        self.plot, self.axes = plt.subplots(1, figsize=(15, 7), subplot_kw=dict(polar=True))
+
+        # Dummy line info
+        xp = np.arange(0, 2*math.pi, 8*math.pi/(self.chunk), dtype=float)
+        x = np.append(xp, xp)
+        x = np.append(x, x)
+
+        # Arrange line
+        self.line, = self.axes.plot(x, np.random.rand(self.chunk), '-', lw=2, color='#330033')
+
+        # Build axes
+        self.axes.set_title('')
+        self.axes.set_ylim(-1, 259)
+        self.axes.set_xticks([]) 
+        self.axes.set_yticks([])
+        self.axes.set_rticks([])
+
+        # Stylaize
+        self.axes.set_facecolor((1,1,1))
+        # self.axes.set_axis_bgcolor((0,0,0))
+        # self.axes.set_xlim(0, 360)
+        # self.plot.setp(ax, xticks=[0, CHUNK, 2 * CHUNK], yticks=[0, 128, 255])
+
+        return
+
+    def toggle_axes(self):
+        self.crosshair = not self.crosshair
+        if self.crosshair:
+            self.axes.grid(True, color='#000000', linestyle='-', linewidth=1)
+            self.axes.set_xticks(np.arange(0,2.0*np.pi, np.pi/2.0))
+        else:
+            self.axes.grid(False)
+            self.axes.set_xticks([])
+
+        # ax.set_title('')
+        # ax.set_xlabel('samples')
+        # ax.set_ylabel('volume')
+        # ax.set_ylim(-1, 259)
+
+        # ax.set_yticks([])
+
+        return
+
+    def start(self):
+        # Begin
+        plt.show(block=False)
+
+        # Count frames
+        frame_count = 0
+        start_time = time.time()
+
+        # Loop
+        while True: # End this with ctrl+C
+            # Begin reading the data
+            data = self.stream.read(self.chunk, exception_on_overflow=False)  
+
+            # Convert data to integers packed as numpy array, then offset it by 127
+            data_int = struct.unpack(str(2 * self.chunk) + 'B', data)
+
+            # Create numpy array and offset by 128
+            data_np = np.array(data_int, dtype='b')[::2] + 129
+
+            self.line.set_ydata(data_np)
+
+            # update figure canvas
+            try:
+                self.plot.canvas.draw()
+                self.plot.canvas.flush_events()
+                frame_count += 1
+
+            except TclError:
+
+                # calculate average frame rate
+                frame_rate = frame_count / (time.time() - start_time)
+
+                print('Stream stopped')
+                print('Average frame rate = {:.0f} FPS'.format(frame_rate))
+                break