def __init__(
self,
model,
fs,
outputName,
globalLoudnessFeature=None,
loudnessLevelFunction=None,
loudnessLevel='abs'
):
self.iterator = DynamicLoudnessIterator(
model,
fs,
outputName,
globalLoudnessFeature,
loudnessLevelFunction
)
self.targetLoudnessLevel = loudnessLevel
self.tol = 0.02
self.nIters = 20
self.predictions = None
self.converged = False
self.fs = fs
self.duration = 1
self.alpha = 0.5
if ((type(loudnessLevel) is str) or (loudnessLevel == 2.4)):
self.freqs = freqsISO389
self.sPLs = thresholdsISO389
self.targetLoudnessLevel = 2.4
else:
contour = ISO2262003LoudnessContours()
self.freqs, self.sPLs = contour.phonToSPL(None, loudnessLevel)
self.predictions = np.zeros(self.sPLs.size)
def __init__(self, model, fs, outputName,
globalLoudnessFeature = None,
loudnessLevelFunction = None,
loudnessLevel = 'abs'):
self.iterator = DynamicLoudnessIterator(model, fs, outputName,
globalLoudnessFeature,
loudnessLevelFunction)
self.targetLoudnessLevel = loudnessLevel
self.tol = 0.02
self.nIters = 20
self.predictions = None
self.converged = False
self.fs = fs
self.duration = 1
self.alpha = 0.5
if ((type(loudnessLevel) is str) or (loudnessLevel == 2.4)):
self.freqs = freqsISO389
self.sPLs = thresholdsISO389
self.targetLoudnessLevel = 2.4
else:
contour = ISO2262003LoudnessContours()
self.freqs, self.sPLs = contour.phonToSPL(None, loudnessLevel)
self.predictions = np.zeros(self.sPLs.size)
class DynamicLoudnessContourPredictor():
def __init__(self, model, fs, outputName,
globalLoudnessFeature = None,
loudnessLevelFunction = None,
loudnessLevel = 'abs'):
self.iterator = DynamicLoudnessIterator(model, fs, outputName,
globalLoudnessFeature,
loudnessLevelFunction)
self.targetLoudnessLevel = loudnessLevel
self.tol = 0.02
self.nIters = 20
self.predictions = None
self.converged = False
self.fs = fs
self.duration = 1
self.alpha = 0.5
if ((type(loudnessLevel) is str) or (loudnessLevel == 2.4)):
self.freqs = freqsISO389
self.sPLs = thresholdsISO389
self.targetLoudnessLevel = 2.4
else:
contour = ISO2262003LoudnessContours()
self.freqs, self.sPLs = contour.phonToSPL(None, loudnessLevel)
self.predictions = np.zeros(self.sPLs.size)
def process(self):
self.predictions = np.zeros(self.freqs.size)
self.converged = np.zeros(self.freqs.size, dtype = bool)
for i, freq in enumerate(self.freqs):
print 'Freq: %0.2f, initial guess: %0.2f' % (freq, self.sPLs[i])
s = Sound.tone([freq], dur = self.duration, fs = self.fs)
s.applyRamp(0.1)
s.useDBSPL()
s.normalise(self.sPLs[i], 'RMS')
self.predictions[i] = self.sPLs[i]
self.predictions[i] += self.iterator.process(s.data,
self.targetLoudnessLevel,
self.tol,
self.nIters,
self.alpha)
self.converged[i] = self.iterator.converged
def plotPredictions(self):
plt.semilogx(self.freqs, self.sPLs, label = 'ISO')
plt.semilogx(self.freqs, self.predictions, color = 'r',
linestyle = '--', label = 'Predicted')
plt.legend()
plt.show()
def computeErrors(self):
'''
Returns a tuple of error vector (target - prediction), RMSE and maximum
absolute error.
'''
error = self.sPLs - self.predictions
rMSE = np.sqrt(np.mean(error * error))
maxE = np.max(np.abs(error))
return (error, rMSE, maxE)
def getResults(self):
outputDic = {}
outputDic['Errors'], outputDic['RMSE'], outputDic['MXAE'] = self.computeErrors()
outputDic['Frequencies'] = self.freqs
outputDic['Targets'] = self.sPLs
outputDic['Predictions'] = self.predictions
outputDic['Converged'] = self.converged
return outputDic
class DynamicLoudnessContourPredictor():
def __init__(
self,
model,
fs,
outputName,
globalLoudnessFeature=None,
loudnessLevelFunction=None,
loudnessLevel='abs'
):
self.iterator = DynamicLoudnessIterator(
model,
fs,
outputName,
globalLoudnessFeature,
loudnessLevelFunction
)
self.targetLoudnessLevel = loudnessLevel
self.tol = 0.02
self.nIters = 20
self.predictions = None
self.converged = False
self.fs = fs
self.duration = 1
self.alpha = 0.5
if ((type(loudnessLevel) is str) or (loudnessLevel == 2.4)):
self.freqs = freqsISO389
self.sPLs = thresholdsISO389
self.targetLoudnessLevel = 2.4
else:
contour = ISO2262003LoudnessContours()
self.freqs, self.sPLs = contour.phonToSPL(None, loudnessLevel)
self.predictions = np.zeros(self.sPLs.size)
def process(self):
self.predictions = np.zeros(self.freqs.size)
self.converged = np.zeros(
self.freqs.size, dtype=bool
)
for i, freq in enumerate(self.freqs):
print 'Freq: %0.2f, initial guess: %0.2f' % (freq, self.sPLs[i])
s = Sound.tone([freq], dur=self.duration, fs=self.fs)
s.applyRamp(0.1)
s.useDBSPL()
s.normalise(self.sPLs[i], 'RMS')
self.predictions[i] = self.sPLs[i]
self.predictions[i] += self.iterator.process(
s.data,
self.targetLoudnessLevel,
self.tol,
self.nIters,
self.alpha
)
self.converged[i] = self.iterator.converged
def plotPredictions(self):
plt.semilogx(
self.freqs, self.sPLs, label='ISO'
)
plt.semilogx(
self.freqs, self.predictions, color='r',
linestyle='--', label='Predicted'
)
plt.legend()
plt.show()
def computeErrors(self):
'''
Returns a tuple of error vector (target - prediction), RMSE and maximum
absolute error.
'''
error = self.sPLs - self.predictions
rMSE = np.sqrt(np.mean(error * error))
maxE = np.max(np.abs(error))
return (error, rMSE, maxE)
def getResults(self):
outputDic = {}
outputDic['Errors'], outputDic['RMSE'], outputDic['MXAE'] = \
self.computeErrors()
outputDic['Frequencies'] = self.freqs
outputDic['Targets'] = self.sPLs
outputDic['Predictions'] = self.predictions
outputDic['Converged'] = self.converged
return outputDic
