Spectrogram works!

animation.py

@@ -0,0 +1,110 @@
+# from numpy.fft import fft
+# from scipy import signal
+# from scipy.fft import fftshift
+# import matplotlib.pyplot as plt
+# import numpy as np
+import json
+# from scipy.io import wavfile
+# from os import path
+# from pydub import AudioSegment
+# from matplotlib.colors import Normalize 
+# import seaborn as sns
+import numpy as np
+import matplotlib.pyplot as plt
+from matplotlib.animation import FuncAnimation
+
+import threading
+from pydub.playback import play
+from pydub import AudioSegment
+import time
+
+import pandas as pd
+
+data_path = 'ch_wine.txt'
+data = {}
+with open(data_path, 'r') as infile:
+    data = json.load(infile)
+pxx_data = data['Pxx']
+db_data = data['dB']
+freqs = data['freqs']
+
+db_df = pd.DataFrame(np.asarray(pxx_data), index=freqs)
+db_amp_slice = (db_df + min(db_df))[100:300].sum()#.rolling(window=1, min_periods=0).mean().fillna(0)
+
+
+# plt.plot(db_amp_slice)
+# plt.show()
+
+
+# Plot Spec
+# plt.imshow(db_df, aspect='auto', origin='lower')
+# plt.show()
+# exit()
+
+# Getting a single frequency's dB data over the course of audio
+desired_frequency = 1000
+target_index = int((desired_frequency / max(freqs)) * len(freqs))
+# print(len(db_data[target_index]))
+
+# Animation Vars
+size = 10
+
+# Plot Stuff
+fig = plt.figure(figsize=(7, 7))
+ax = fig.add_axes([0, 0, 1, 1], frameon=False)
+ax.set_xlim(0, 1), ax.set_xticks([])
+ax.set_ylim(0, 1), ax.set_yticks([])
+scat = ax.scatter([0.5], [0.5], s=[size], lw=0.5, edgecolors=(0, 0, 0, 1), facecolors='blue')#, facecolors='none')
+
+ax.set_xlim(0, 1), ax.set_xticks([])
+ax.set_ylim(0, 1), ax.set_yticks([])
+
+# Music stuff
+sound = AudioSegment.from_wav("assets/Hozier - Cherry Wine - Cover (Fingerstyle Guitar).wav")
+music_thread = threading.Thread(target=play, args=(sound,))
+song_length = sound.duration_seconds
+
+global music_start
+music_start = 0
+def update(frame_number):
+    global music_start
+
+    if frame_number == 0:
+        music_thread.start()
+        music_start = time.perf_counter()
+
+    print('music_start', music_start)
+    i = (time.perf_counter() - music_start)/song_length
+    print('i', i)
+    print(len(db_amp_slice))
+    i = round(i * len(db_amp_slice))
+    print(frame_number, i)
+
+    size = db_amp_slice[i]
+    # size = db_data[target_index][frame_number]
+    # if size > 1000:
+    #     size = 1000
+    # if size < 0:
+    #     size = 0
+    print('size', size)
+    scat.set_sizes([size])
+
+animation = FuncAnimation(fig, update, interval=5)
+plt.show()
+
+
+print("closing program...")
+
+
+
+
+
+
+
+
+
+
+
+
+
+

main.py

@@ -7,88 +7,82 @@
 from scipy.io import wavfile
 from os import path
 from pydub import AudioSegment
+from matplotlib.colors import Normalize 
+# import seaborn as sns
 
 # import files
-src = "assets/piano2.wav"
-# dst = "test.wav"
-dst = "assets/piano3.wav"
+piano_3 = "assets/piano3.wav"
+audio_tone = "assets/100hz.wav"
+bread = "assets/bread.mp3"
+shepard = "assets/Shepard Tone.wav"
+wine = "assets/Hozier - Cherry Wine - Cover (Fingerstyle Guitar).wav"
 
-print("converting to wav file...")
-# sound = AudioSegment.from_mp3(src)
+
+# Code for converting Mp3 to Wav file. Need to add other conversions and type checking in the future.
+# sound = AudioSegment.from_mp3(shepard)
 # sound = sound.set_channels(1)
+# sound = sound.set_frame_rate(5000) # sample rate in kHz, higher sample rate, longer spectrogram computation takes
 # sound.export(dst, format="wav")
 
-sound = AudioSegment.from_wav(src)
-print(sound.duration_seconds)
-sound = sound.set_channels(1)
-sound = sound.set_frame_rate(5000) # sample rate in kHz
-sound.export(dst, format="wav")
-
-print("importing wav file...")
-sample_rate, samples = wavfile.read(dst)
-print(f"sample_rate:\n{sample_rate}")
-print(f"samples:\n{samples}")
-print(len(samples))
-# sample_rate = len(samples)
-# print(samples.tolist())
-
-
-
-# !
-rng = np.random.default_rng()
-# # Generate a test signal, a 2 Vrms sine wave whose frequency is slowly modulated around 3kHz, corrupted by white noise of exponentially decreasing magnitude sampled at 10 kHz.
-fs = 10e3
-N = 1e5
-amp = 2 * np.sqrt(2)
-noise_power = 0.01 * fs / 2
-time = np.arange(N) / float(fs)
-mod = 500*np.cos(2*np.pi*0.25*time)
-carrier = amp * np.sin(2*np.pi*3e3*time + mod)
-noise = rng.normal(scale=np.sqrt(noise_power), size=time.shape)
-noise *= np.exp(-time/5)
-x = carrier + noise
-
-# Compute and plot the spectrogram.
-f, t, Sxx = signal.spectrogram(x, fs)
-
-plt.pcolormesh(t, f, Sxx, shading='gouraud')
-plt.ylabel('Frequency [Hz]')
-plt.xlabel('Time [sec]')
-plt.show()
-# !
-
-Pxx, freqs, bins, im = plt.specgram(samples, NFFT=1024, Fs=44100, noverlap=900)
-# plt.yscale('log')
+# AudioSegment.frame_rate
+# sound = AudioSegment.from_mp3(bread)
+# sound = sound.set_channels(1)
+# sound = sound.set_frame_rate(5000)
+# sample_rate = AudioSegment.frame_rate
+# samples = sound.raw_data
+
+# Reading Wav file. Must be Mono (single channel)
+sample_rate, samples = wavfile.read(wine)
+print("sample_rate:", sample_rate)
+
+if len(samples[0]) == 2:
+	print("moving from duo to mono...")
+	samples = np.asarray(samples)
+	samples = samples[:,0]
+	print(samples)
+
+Pxx, freqs, bins, im = plt.specgram(samples, NFFT=1024, Fs=sample_rate, noverlap=128, scale="dB")
+# plt.ylim((0, 20000))
+# plt.yscale('symlog')
 # plt.xscale('log')
+# plt.show()
+
+print(np.shape(Pxx))
+print(np.shape(freqs))
+print(np.shape(bins))
+
+Pxx_dB = 10*np.log(Pxx) # Convert to dB scale
+plt.imshow(Pxx_dB, aspect='auto', origin='lower')
 plt.show()
+# print("exporing data...")
+
+data = {}
+data['Pxx'] = Pxx.tolist()
+data['dB'] = Pxx_dB.tolist()
+data['freqs'] = freqs.tolist()
+data['bins'] = bins.tolist()
+
+with open('ch_wine.txt', 'w') as outfile:
+    json.dump(data, outfile)
+
+
+
+
+
+
+
+print("closing program...")
+
+
+
+
+
+
+
+
 
-print("making spectrogram...")
-f, t, Sxx = signal.spectrogram(samples, sample_rate, mode='magnitude')
-print(f"f:\n{f}")
-print(f"t:\n{t}")
-print(f"Sxx:\n{Sxx}")
 
-print("making graph...")
-# plt.pcolormesh(t, f, Sxx, shading='gouraud')
-# plt.ylabel('Frequency [Hz]')
-# plt.xlabel('Time [sec]')
-# plt.show()
 
-plt.pcolormesh(t, f, np.log10(Sxx), shading='gouraud')
-# plt.pcolormesh(t, f, Sxx, shading='gouraud')
-plt.imshow(Sxx)
-plt.ylabel('Frequency [Hz]')
-plt.xlabel('Time [sec]')
-# plt.yscale('log')
-# plt.xscale('log')
-plt.show()
 
-print("exporing data...")
 
-# data = {}
-# data['f'] = f.tolist()
-# data['t'] = t.tolist()
-# data['Sxx'] = Sxx.tolist()
 
-# with open('data.txt', 'w') as outfile:
-#     json.dump(data, outfile)