forked from grame-cncm/faust
-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
1 changed file
with
146 additions
and
130 deletions.
There are no files selected for viewing
276 changes: 146 additions & 130 deletions
276
tools/physicalModeling/ir2dsp/ir2dsp.py
100755 → 100644
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,130 +1,146 @@ | ||
| #!/usr/bin/env python | ||
| # -*- coding: utf-8 -*- | ||
| # id2dsp.py | ||
| # Copyright Pierre-Amaury Grumiaux, Pierre Jouvelot, Emilio Jesus Gallego Arias, | ||
| # Romain Michon and Jakob Dübel | ||
|
|
||
| from __future__ import division | ||
| import math | ||
| import numpy as np | ||
| import matplotlib.pyplot as plt | ||
| from sys import argv | ||
| import subprocess | ||
| from scipy.io.wavfile import read | ||
| import peakutils | ||
| import peakutils.plot as pkplt | ||
| import operator | ||
| import argparse | ||
|
|
||
| # Help for use of the script | ||
| parser = argparse.ArgumentParser(description = "IR2dsp.py is a python script that generates a .dsp file from an impulse response file (a .wav file).The impulse response is analyzed in order to rebuild the sound of the vibration of the object based on this impulse response.", epilog = "See http://faust.grame.fr/images/faust-tutorial2.pdf for more information about Faust") | ||
| parser.add_argument("soundFile", type = str, help = "Path of the sound file") | ||
| parser.add_argument("modelName", help = "The name of the model created in the .dsp file") | ||
| parser.add_argument("peakThreshold", help = "Minimum value of peaks in dB (between -infinity and 0)") | ||
| parser.add_argument("peakDistance", help = "Minimum distance between two peaks in Hertz") | ||
| ars = parser.parse_args () | ||
| # Arguments | ||
| script, soundFile, modelName, peakThreshold, peakDistance = argv | ||
|
|
||
| # Reading file | ||
| (fs, x) = read(soundFile) | ||
| # Normalizing sound | ||
| x = x/max(x) | ||
|
|
||
| # FFT | ||
| X = np.abs(np.fft.fft(x)) | ||
| X = X/(max(X)) | ||
| # computing corresponding frequencies | ||
| time_step = 1/fs | ||
| freqs = np.fft.fftfreq(x.size, time_step) | ||
| idx = np.argsort(freqs) | ||
|
|
||
| # plot for debug | ||
| # plt.plot(freqs[idx],20*np.log(X[idx])) | ||
| # plt.show() | ||
|
|
||
| # detecting peaks | ||
| threshold = math.pow(10, float(peakThreshold)/20) # from dB to X unit | ||
| distance = int(peakDistance)/(fs/x.size) | ||
| indexes = peakutils.indexes(X[idx], thres = threshold, min_dist = distance) | ||
|
|
||
| # Storing frequencies and modes for each bp filters | ||
| peaksFreq = [] | ||
| peaksGains = [] | ||
| nbOfPeaks = 0 | ||
| for i in indexes : | ||
| if freqs[idx][i] > 0: | ||
| peaksFreq.append(freqs[idx][i]) | ||
| peaksGains.append(X[idx][i]) | ||
| nbOfPeaks += 1 | ||
| peaksGains = peaksGains/(max(peaksGains)) | ||
|
|
||
| #Computing t60 values | ||
| peakst60 = [] | ||
| for i in range(0, nbOfPeaks) : | ||
| offset = pow(10, -3/20) # conversion of -3 dB in X unit | ||
| peakIndex = indexes[len(indexes) - nbOfPeaks + i] | ||
|
|
||
| n = peakIndex | ||
| while X[idx][n] > (X[idx][peakIndex]*offset): | ||
| n = n-1 | ||
| a = n | ||
|
|
||
| n = peakIndex | ||
| while X[idx][n] > (X[idx][peakIndex]*offset): | ||
| n = n+1 | ||
| b = n | ||
|
|
||
| bandwidth = (b-a)/(fs/x.size) # bandwidth in Hz | ||
| print(bandwidth) | ||
| peakst60.append(6.91/fs/(1-math.exp(-math.pi*bandwidth/fs))) | ||
|
|
||
| print(" peaks frequencies : ") | ||
| print(peaksFreq) | ||
| print(" corresponding gains : ") | ||
| print(peaksGains) | ||
| print(" corresponding t60 : ") | ||
| print(peakst60) | ||
|
|
||
| # Writing the dsp file # | ||
| # ####################### | ||
| file = open(modelName + ".dsp", "w") | ||
| file.write("// -*-Faust-*-\n") | ||
| file.write("// generated with ir2dsp.py {0} {1} {2}\n".format(modelName, peakThreshold, peakDistance)) | ||
| file.write("declare name \"{0}\";\n".format(modelName)) | ||
| file.write("import(\"stdfaust.lib\");\n") | ||
| file.write("nModes = ") | ||
| file.write(str(len(peaksGains))) | ||
| file.write(";\n") | ||
| file.write("modeFrequencies=("); # writing the frequencies list | ||
| k = 0 | ||
| for i in peaksFreq: | ||
| file.write(str(i)) | ||
| if (k+1<len(peaksGains)): | ||
| file.write(",") | ||
| k += 1 | ||
| file.write(");\n"); | ||
| file.write("massEigenValues=("); # writing the masses list | ||
| k = 0 | ||
| for i in peaksGains: | ||
| file.write(str(i)) | ||
| if (k+1 < len(peaksGains)): | ||
| file.write(",") | ||
| k += 1 | ||
| file.write(");\n"); | ||
| file.write("t60=("); # writing the t60 list | ||
| k = 0 | ||
| for i in peakst60: | ||
| file.write(str(i)) | ||
| if (k+1 < len(peaksGains)): | ||
| file.write(",") | ||
| k += 1 | ||
| file.write(");\n"); | ||
| file.write("modeFreqs=par(i,nModes,ba.take(i+1, modeFrequencies));\n") | ||
| file.write("modeGains=par(i,nModes,ba.take(i+1, massEigenValues));\n") | ||
| file.write("modeT60 = par(i,nModes,ba.take(i+1,t60));\n") | ||
| file.write(modelName) | ||
| file.write("=pm.modalModel(nModes,modeFreqs,modeT60,modeGains);"); | ||
| file.write('\ngate = button("gate");') | ||
| file.write("\nprocess = pm.impulseExcitation(gate) : " + modelName + " <: _,_; ") | ||
| file.close(); | ||
| #!/usr/bin/env python | ||
| # -*- coding: utf-8 -*- | ||
| # id2dsp.py | ||
| # Copyright Pierre-Amaury Grumiaux, Pierre Jouvelot, Emilio Jesus Gallego Arias, | ||
| # Romain Michon and Jakob Dübel | ||
|
|
||
| from __future__ import division | ||
| import math | ||
| import numpy as np | ||
| import matplotlib.pyplot as plt | ||
| from sys import argv | ||
| import subprocess | ||
| from scipy.io.wavfile import read | ||
| import peakutils | ||
| import peakutils.plot as pkplt | ||
| import operator | ||
| import argparse | ||
|
|
||
| # Help for use of the script | ||
| parser = argparse.ArgumentParser(description = "IR2dsp.py is a python script that generates a .dsp file from an impulse response file (a .wav file).The impulse response is analyzed in order to rebuild the sound of the vibration of the object based on this impulse response.", epilog = "See http://faust.grame.fr/images/faust-tutorial2.pdf for more information about Faust") | ||
| parser.add_argument("soundFile", type = str, help = "Path of the sound file") | ||
| parser.add_argument("modelName", help = "The name of the model created in the .dsp file") | ||
| parser.add_argument("peakThreshold", help = "Minimum value of peaks in dB (between -infinity and 0)") | ||
| parser.add_argument("peakDistance", help = "Minimum distance between two peaks in Hertz") | ||
| ars = parser.parse_args () | ||
| # Arguments | ||
| script, soundFile, modelName, peakThreshold, peakDistance = argv | ||
|
|
||
| # Reading file | ||
| print("Reading wav file...") | ||
| (fs, x) = read(soundFile) | ||
| if x.shape[1] == 2: | ||
| x = x[:, 0] | ||
| # Normalizing sound | ||
| x = x/np.max(x) | ||
|
|
||
| # FFT | ||
| print("Computing FFT...") | ||
| X = np.abs(np.fft.fft(x)) | ||
| X = X/(np.max(X)) | ||
| # computing corresponding frequencies | ||
| time_step = 1/fs | ||
| freqs = np.fft.fftfreq(x.size, time_step) | ||
| idx = np.argsort(freqs) | ||
|
|
||
| # plot for debug | ||
| # plt.plot(freqs[idx],20*np.log(X[idx])) | ||
| # plt.show() | ||
|
|
||
| # detecting peaks | ||
| print("Detecting FFT peaks...") | ||
| threshold = math.pow(10, float(peakThreshold)/20) # from dB to X unit | ||
| distance = int(peakDistance)/(fs/x.size) | ||
| indexes = peakutils.indexes(X[idx], thres = threshold, min_dist = distance) | ||
|
|
||
| # Storing frequencies and modes for each bp filters | ||
| peaksFreq = [] | ||
| peaksGains = [] | ||
| nbOfPeaks = 0 | ||
| for i in indexes : | ||
| if freqs[idx][i] > 0: | ||
| peaksFreq.append(freqs[idx][i]) | ||
| peaksGains.append(X[idx][i]) | ||
| nbOfPeaks += 1 | ||
| peaksGains = peaksGains/(np.max(peaksGains)) | ||
|
|
||
| #Computing t60 values | ||
| print("Calculating T60 values...") | ||
| peakst60 = [] | ||
| for i in range(0, nbOfPeaks) : | ||
| offset = pow(10, -3/20) # conversion of -3 dB in X unit | ||
| peakIndex = indexes[len(indexes) - nbOfPeaks + i] | ||
|
|
||
| n = peakIndex | ||
| while X[idx][n] > (X[idx][peakIndex]*offset): | ||
| n = n-1 | ||
| a = n | ||
|
|
||
| n = peakIndex | ||
| while X[idx][n] > (X[idx][peakIndex]*offset): | ||
| n = n+1 | ||
| b = n | ||
|
|
||
| bandwidth = (b-a)/(fs/x.size) # bandwidth in Hz | ||
| peakst60.append(6.91/fs/(1-math.exp(-math.pi*bandwidth/fs))) | ||
|
|
||
| print("Frequency peaks and corresponding gains and T60 has been successfully calculated.") | ||
|
|
||
| # Writing the dsp file # | ||
| # ####################### | ||
| print("Writing the dsp file...") | ||
| file = open(modelName + ".dsp", "w") | ||
| file.write("// -*-Faust-*-\n") | ||
| file.write("// generated with ir2dsp.py {0} {1} {2}\n".format(modelName, peakThreshold, peakDistance)) | ||
| file.write("declare name \"{0}\";\n".format(modelName)) | ||
| file.write("import(\"stdfaust.lib\");\n\n") | ||
| file.write("nModes = ") | ||
| file.write(str(len(peaksGains))) | ||
| file.write(";\n") | ||
| file.write("modeFrequencies = ("); # writing the frequencies list | ||
| k = 0 | ||
| for i in peaksFreq: | ||
| file.write(str(i)) | ||
| if (k+1<len(peaksGains)): | ||
| file.write(",") | ||
| k += 1 | ||
| file.write(");\n"); | ||
| file.write("massEigenValues = ("); # writing the masses list | ||
| k = 0 | ||
| for i in peaksGains: | ||
| file.write(str(i)) | ||
| if (k+1 < len(peaksGains)): | ||
| file.write(",") | ||
| k += 1 | ||
| file.write(");\n"); | ||
| file.write("t60 = ("); # writing the t60 list | ||
| k = 0 | ||
| for i in peakst60: | ||
| file.write(str(i)) | ||
| if (k+1 < len(peaksGains)): | ||
| file.write(",") | ||
| k += 1 | ||
| file.write(");\n"); | ||
| # file.write("modeFreqs=par(i,nModes,ba.take(i+1, modeFrequencies));\n") | ||
| # file.write("modeGains=par(i,nModes,ba.take(i+1, massEigenValues));\n") | ||
| # file.write("modeT60 = par(i,nModes,ba.take(i+1,t60));\n") | ||
| # file.write(modelName) | ||
| # file.write("=pm.modalModel(nModes,modeFreqs,modeT60,modeGains);"); | ||
| # file.write('\ngate = button("gate");') | ||
| # file.write("\nprocess = pm.impulseExcitation(gate) : " + modelName + " <: _,_; ") | ||
|
|
||
| file.write("\nmodalModel(n,modeFreqs,modeRes,modeGains) = _ <: par(i,n,gain(i)*pm.modeFilter(freqs(i),res(i))) :> _\n") | ||
| file.write("with{\n") | ||
| file.write("freqs(i) = ba.take(i+1,modeFreqs);\n") | ||
| file.write("res(i) = ba.take(i+1,modeRes);\n") | ||
| file.write("gain(i) = ba.take(i+1,modeGains);\n") | ||
| file.write("};\n\n") | ||
|
|
||
| file.write(modelName) | ||
| file.write(" = modalModel(nModes,modeFrequencies,t60,massEigenValues);"); | ||
| file.write('\ngate = button("gate");') | ||
| file.write("\nprocess = pm.impulseExcitation(gate) : " + modelName + " <: _,_; ") | ||
|
|
||
| file.close(); | ||
|
|
||
| print(modelName + ".dsp has been successfully created!") |