def smallSampleSize():
try:
filePath = sys.argv[1]
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
print "File is not a wave file"
return
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
timeCodeSampleRate = int(sampleRate * 1.001)
byteWidth = wavefile.numberOfBytesPerSample()
numberOfChannels = wavefile.numberOfChannels()
frameNumber = 0
print datetime.datetime.now()
with open(filePath, 'rb') as f:
startTime = datetime.datetime.now()
f.seek(dataStartPos)
data_size = int(
(sampleRate * 1.001) /
24) #1 frame of audio at 23.98 or 48048 sample rate
dataToReadChunkSize = data_size * byteWidth
#examine 160 seconds of data
dataRemaining = (160 * timeCodeSampleRate
) * byteWidth * numberOfChannels #dataLength
while dataRemaining > 0:
if dataRemaining > dataToReadChunkSize:
data = f.read(dataToReadChunkSize)
else:
data = f.read(dataRemaining)
dataRemaining -= len(data)
print pctimecode.convertUnsignedLongLongToTimeCode(
wavefile.timestamp + (frameNumber * data_size),
timeCodeSampleRate, 24)
frameNumber += 1
for channel in range(0, numberOfChannels):
subdata = ""
for x in range(channel * byteWidth, len(data),
byteWidth * numberOfChannels):
# subdata.extend(data[x:x+3])
subdata += (data[x] + data[x + 1] + data[x + 2])
print ultimateFreqAndPowerForData(
subdata, byteWidth, sampleRate)
print datetime.datetime.now()
else:
print "The file", filename, "doesn't exist"
except IndexError as e:
print "No file was passed in as a variable"
def findTwoPop():
try:
filePath = sys.argv[1]
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
print "File is not a wave file"
return
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
byteWidth = wavefile.numberOfBytesPerSample()
numberOfChannels = wavefile.numberOfChannels()
with open(filePath, 'rb') as f:
startTime = datetime.datetime.now()
#1 frame of audio at 23.98 or 48048 sample rate
data_size = 2002
#position at which to examine
startOffset = int(round(
sampleRate * 1.001)) * 148 * byteWidth * numberOfChannels
startOffset -= (data_size * byteWidth * numberOfChannels * 2)
dataStartPos = dataStartPos + startOffset
f.seek(dataStartPos)
if dataStartPos > dataLength:
print "file is too short to analyze"
return
dataToReadChunkSize = data_size * byteWidth * numberOfChannels
dataRemaining = data_size * byteWidth * numberOfChannels * 6
while dataRemaining > 0:
data = f.read(dataToReadChunkSize)
print pcwave
dataRemaining -= dataToReadChunkSize
print "Next Sample"
for channel in range(0, numberOfChannels):
subdata = ""
for x in range(channel * byteWidth, len(data),
byteWidth * numberOfChannels):
# subdata.extend(data[x:x+3])
subdata += (data[x] + data[x + 1] + data[x + 2])
# print anaylyzeDataWithZeroCrossing(subdata, byteWidth, sampleRate)
result = ultimateFreqAndPowerForData(
subdata, byteWidth, sampleRate)
if math.isnan(result[0]) or result[1] < -65.0:
print '\tDisregard', round(result[0]), result[1]
elif result[0] < 1005 and result[0] > 995:
print '\tMatch', round(result[0]), result[1]
else:
print '\tOops', round(result[0]), result[1]
else:
print "The file", filename, "doesn't exist"
except IndexError as e:
print "No file was passed in as a variable"
def multichannel():
try:
filePath = sys.argv[1]
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
print "File is not a wave file"
return
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
timeCodeSampleRate = int(math.ceil(48000 * 1.001))
byteWidth = wavefile.numberOfBytesPerSample()
numberOfChannels = wavefile.numberOfChannels()
data_size = 48048 #1 second of audio
print timeCodeSampleRate
print pctimecode.convertUnsignedLongLongToTimeCode(
wavefile.timestamp, timeCodeSampleRate, 24)
second = 0
with open(filePath, 'rb') as f:
startTime = datetime.datetime.now()
f.seek(dataStartPos)
dataToReadChunkSize = sampleRate * byteWidth * numberOfChannels
dataRemaining = dataLength
while dataRemaining > 0:
if dataRemaining > dataToReadChunkSize:
data = f.read(dataToReadChunkSize)
else:
data = f.read(dataRemaining)
dataRemaining -= len(data)
print wavefile.timestamp
print pctimecode.convertUnsignedLongLongToTimeCode(
wavefile.timestamp + (second * data_size),
timeCodeSampleRate, 24)
for channel in range(0, numberOfChannels):
subdata = ""
for x in range(channel * byteWidth, len(data),
byteWidth * numberOfChannels):
subdata += (data[x] + data[x + 1] + data[x + 2])
print highestAndNextHighestFreqAndPowerForData(
subdata, byteWidth, sampleRate)
second += 1
else:
print "The file", filename, "doesn't exist"
except IndexError as e:
print "No file was passed in as a variable"
def findTwoPop():
try:
filePath = sys.argv[1]
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
print "File is not a wave file"
return
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
byteWidth = wavefile.numberOfBytesPerSample()
numberOfChannels = wavefile.numberOfChannels()
with open(filePath, 'rb') as f:
startTime = datetime.datetime.now()
#1 frame of audio at 23.98 or 48048 sample rate
data_size = 2002
#position at which to examine
startOffset = int(round(sampleRate * 1.001)) * 148 * byteWidth * numberOfChannels
startOffset -= (data_size * byteWidth * numberOfChannels * 2)
dataStartPos = dataStartPos + startOffset
f.seek(dataStartPos)
if dataStartPos > dataLength:
print "file is too short to analyze"
return
dataToReadChunkSize = data_size*byteWidth*numberOfChannels
dataRemaining = data_size*byteWidth*numberOfChannels*6
while dataRemaining > 0:
data = f.read(dataToReadChunkSize)
print pcwave
dataRemaining -= dataToReadChunkSize
print "Next Sample"
for channel in range(0, numberOfChannels):
subdata = ""
for x in range(channel*byteWidth,len(data),byteWidth*numberOfChannels):
# subdata.extend(data[x:x+3])
subdata += (data[x] + data[x+1] + data[x+2])
# print anaylyzeDataWithZeroCrossing(subdata, byteWidth, sampleRate)
result = ultimateFreqAndPowerForData(subdata, byteWidth, sampleRate)
if math.isnan(result[0]) or result[1] < -65.0:
print '\tDisregard', round(result[0]), result[1]
elif result[0] < 1005 and result[0] > 995:
print '\tMatch', round(result[0]), result[1]
else:
print '\tOops', round(result[0]), result[1]
else:
print "The file",filename,"doesn't exist"
except IndexError as e:
print "No file was passed in as a variable"
def smallSampleSize():
try:
filePath = sys.argv[1]
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
print "File is not a wave file"
return
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
timeCodeSampleRate = int(sampleRate * 1.001)
byteWidth = wavefile.numberOfBytesPerSample()
numberOfChannels = wavefile.numberOfChannels()
frameNumber = 0
print datetime.datetime.now()
with open(filePath, 'rb') as f:
startTime = datetime.datetime.now()
f.seek(dataStartPos)
data_size = int((sampleRate * 1.001) / 24) #1 frame of audio at 23.98 or 48048 sample rate
dataToReadChunkSize = data_size*byteWidth
#examine 160 seconds of data
dataRemaining = (160*timeCodeSampleRate)*byteWidth*numberOfChannels #dataLength
while dataRemaining > 0:
if dataRemaining > dataToReadChunkSize:
data = f.read(dataToReadChunkSize)
else:
data = f.read(dataRemaining)
dataRemaining -= len(data)
print pctimecode.convertUnsignedLongLongToTimeCode(wavefile.timestamp + (frameNumber * data_size), timeCodeSampleRate, 24)
frameNumber += 1
for channel in range(0, numberOfChannels):
subdata = ""
for x in range(channel*byteWidth,len(data),byteWidth*numberOfChannels):
# subdata.extend(data[x:x+3])
subdata += (data[x] + data[x+1] + data[x+2])
print ultimateFreqAndPowerForData(subdata, byteWidth, sampleRate)
print datetime.datetime.now()
else:
print "The file",filename,"doesn't exist"
except IndexError as e:
print "No file was passed in as a variable"
def multichannel():
try:
filePath = sys.argv[1]
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
print "File is not a wave file"
return
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
timeCodeSampleRate = int(math.ceil(48000 * 1.001))
byteWidth = wavefile.numberOfBytesPerSample()
numberOfChannels = wavefile.numberOfChannels()
data_size = 48048 #1 second of audio
print timeCodeSampleRate
print pctimecode.convertUnsignedLongLongToTimeCode(wavefile.timestamp, timeCodeSampleRate, 24)
second = 0
with open(filePath, 'rb') as f:
startTime = datetime.datetime.now()
f.seek(dataStartPos)
dataToReadChunkSize = sampleRate*byteWidth*numberOfChannels
dataRemaining = dataLength
while dataRemaining > 0:
if dataRemaining > dataToReadChunkSize:
data = f.read(dataToReadChunkSize)
else:
data = f.read(dataRemaining)
dataRemaining -= len(data)
print wavefile.timestamp
print pctimecode.convertUnsignedLongLongToTimeCode(wavefile.timestamp + (second * data_size), timeCodeSampleRate, 24)
for channel in range(0, numberOfChannels):
subdata = ""
for x in range(channel*byteWidth,len(data),byteWidth*numberOfChannels):
subdata += (data[x] + data[x+1] + data[x+2])
print highestAndNextHighestFreqAndPowerForData(subdata, byteWidth, sampleRate)
second += 1
else:
print "The file",filename,"doesn't exist"
except IndexError as e:
print "No file was passed in as a variable"
def multichannel():
data_size = 48000 #1 second of audio
frate = 48000.0
try:
filePath = sys.argv[1]
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
print "File is not a wave file"
return
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
byteWidth = wavefile.numberOfBytesPerSample()
numberOfChannels = wavefile.numberOfChannels()
print byteWidth, numberOfChannels
second = 0
with open(filePath, 'rb') as f:
startTime = datetime.datetime.now()
f.seek(dataStartPos)
dataToReadChunkSize = sampleRate*byteWidth*numberOfChannels
dataRemaining = dataLength
while dataRemaining > 0:
if dataRemaining > dataToReadChunkSize:
data = f.read(dataToReadChunkSize)
else:
data = f.read(dataRemaining)
dataRemaining -= len(data)
print "Next second", second
for channel in range(0, numberOfChannels):
subdata = ""
for x in range(channel*byteWidth,len(data),byteWidth*numberOfChannels):
# subdata.extend(data[x:x+3])
subdata += (data[x] + data[x+1] + data[x+2])
print anaylyzeData(subdata, byteWidth, sampleRate)
second += 1
else:
print "The file",filename,"doesn't exist"
except IndexError as e:
print "No file was passed in as a variable"
def listOfTimeCodeValuesForFile(fullpath): wavefile = PCWaveFile(fullpath) if wavefile.isValidWaveFile() == False: print 'File is not valid' return tcList = [] sampleRate = wavefile.fmtChunkInfoDict['nSamplesPerSec'] tcList.append(wavefile.timestamp) tcList.append(pctimecode.convertUnsignedLongLongToTimeCode(wavefile.timestamp, sampleRate*1.001, 24)) #23976 tcList.append(pctimecode.convertUnsignedLongLongToTimeCode(wavefile.timestamp, sampleRate, 24)) #24 tcList.append(pctimecode.convertUnsignedLongLongToTimeCode(wavefile.timestamp, sampleRate, 25)) #25 tcList.append(pctimecode.convertUnsignedLongLongToTimeCode(wavefile.timestamp, sampleRate*1.001, 30)) #2997 tcList.append(pctimecode.convertUnsignedLongLongToTimeCode(wavefile.timestamp, sampleRate, 30)) #30 return tcList
def formatInformationForFile(filePath):
wavefile = PCWaveFile(filePath)
if wavefile == None or wavefile.isValidWaveFile() == False:
badInfo = {}
badInfo["fileName"] = os.path.basename(filePath)
badInfo["fullPath"] = filePath
badInfo["bitDepth"] = 0
badInfo["sampleRate"] = 0
badInfo["bextChunk"] = False
badInfo["timeStamp"] = 0
badInfo["timeStampTC"] = "00:00:00:00"
badInfo["regnChunk"] = False
badInfo["channelCount"] = 0
badInfo["status"] = 'fail'
badInfo["errors"] = ["{} is not a wave file.".format(os.path.basename(filePath)),]
badInfo['fileLevels'] = {'status':'fail','layout':'unknown', "warningString":""}
badInfo['headTones'] = {'status':'fail', 'resultString':'Unable to analyze head tones on account of invalid format.', "warningString":""}
badInfo['pops'] = {'status':'fail', 'resultString':'Unable to analyze pops on account of invalid format.', "warningString":""}
return badInfo
info = {}
info['fileName'] = os.path.basename(filePath)
info['fullPath'] = filePath
info['bitDepth'] = wavefile.numberOfBitsPerSample()
info['sampleRate'] = wavefile.numberOfSamplesPerSecond()
info['bextChunk'] = True if wavefile.hasBextChunk() else False
info['timeStamp'] = wavefile.timestamp
info['timeStampTC'] = prettyTimeCode2398(wavefile.timestamp, info['sampleRate'])
info['regnChunk'] = True if wavefile.hasBextChunk() else False
if wavefile.hasRegnChunk() == True:
info['regnChunkUserTimeStamp'] = wavefile.regnChunkInfoDict['userTimeStamp']
info['regnChunkOriginalTimeStamp'] = wavefile.regnChunkInfoDict['originalTimeStamp']
info['channelCount'] = wavefile.numberOfChannels()
if info['bitDepth'] != 24 or info['sampleRate'] != 48000:
info['fileLevels'] = {'status':'fail','layout':'unknown', "warningString":""}
info['headTones'] = {'status':'fail', 'resultString':'Unable to analyze head tones on account of invalid format.', "warningString":""}
info['pops'] = {'status':'fail', 'resultString':'Unable to analyze pops on account of invalid format.', "warningString":""}
else:
info['fileLevels'] = analyzeLevels(filePath)
info['headTones'] = analyzeHeadAlt(filePath)
info['pops'] = analyzeTwoPopInEighths(filePath)
return addStatusToInformation(info)
def examine2pop():
data_size = 48000/24 #1 second of audio
frate = 48000.0
try:
filePath = sys.argv[1]
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
print "File is not a wave file"
return
print wavefile.getFmtChunkDict()
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
byteWidth = wavefile.numberOfBytesPerSample()
with open(filePath, 'rb') as f:
startTime = datetime.datetime.now()
f.seek(dataStartPos)
dataToReadChunkSize = data_size*byteWidth
dataRemaining = dataLength
while dataRemaining > 0:
if dataRemaining > dataToReadChunkSize:
data = f.read(dataToReadChunkSize)
else:
data = f.read(dataRemaining)
print anaylyzeData(data, byteWidth, sampleRate)
dataRemaining -= len(data)
else:
print "The file",filename,"doesn't exist"
except:
print "No file was passed in as a variable"
def analyzeTrackCorrelations(filePath):
wavefile = None
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
else:
return
if wavefile == None or wavefile.isValidWaveFile() == False:
return {"status":"fail", "resultString":"file is not a wavefile", "layout":"Unknown"}
sampleRate = wavefile.fmtChunkInfoDict['nSamplesPerSec']
channelByteWidth = wavefile.fmtChunkInfoDict['wBitsPerSample'] / 8
bitsPerSample = wavefile.fmtChunkInfoDict['wBitsPerSample']
numberOfChannels = wavefile.fmtChunkInfoDict['nChannels']
packetByteWidth = channelByteWidth * numberOfChannels
if numberOfChannels != 6:
if numberOfChannels == 2:
return {"status":"pass", "resultString":"File contains {} audio channels".format(numberOfChannels), "layout":"STEREO"}
else:
return {"status":"pass", "resultString":"File contains {} audio channels".format(numberOfChannels), "layout":"Unknown"}
isSigned = True
if channelByteWidth == 1:
isSigned = False
chunkPosition = wavefile.getDataChunk()['chunkPosition']
numberOfSamples = wavefile.getDataChunk()['chunkSize'] / packetByteWidth
#open the file, read through the packets and do a running mean of each track in the file
#what this does do is ASSUME that the input file is 24 bit signed data, otherwise 16 bit would be signed, but 8 bit wouldn't
#not that 8 bit is an issue
with open(filePath, 'rb') as f:
f.seek(chunkPosition)
blockWidthInSeconds = 1
samplesToReadBlockSize = (sampleRate * blockWidthInSeconds)/2
numberOfSamplesRemaining = numberOfSamples
bufferAmount = samplesToReadBlockSize if numberOfSamples > samplesToReadBlockSize else numberOfSamples
totalBlocksInFile = numberOfSamplesRemaining // samplesToReadBlockSize
shouldReadEntireFile = False
analysisSpacing = 0
numberOfAnalyses = 34
initialAnalyses = 0
if totalBlocksInFile < (2 * numberOfAnalyses):
shouldReadEntireFile = True
else:
analysisSpacing = totalBlocksInFile / numberOfAnalyses
startTime = datetime.datetime.now()
constantBitDepth16 = pow(2,15)
constantBitDepth24 = pow(2,23)
#totalsCounter = RunningTotalCounter(pow(2,(bitsPerSample-1)), samplesToReadBlockSize, numberOfChannels)
totalsCounter = RunningTotalCounter(constantBitDepth16, samplesToReadBlockSize, numberOfChannels)
# you need to constrain the size of the input as we will be calculating RMS values which require exponential math
while numberOfSamplesRemaining > 0:
#read a buffer's amount of samples
if numberOfSamplesRemaining > bufferAmount:
bytesToRead = packetByteWidth * bufferAmount
data = f.read(bytesToRead)
numberOfSamplesRemaining -= (len(data)/packetByteWidth)
if len(data) != bytesToRead:
logging.error("IO Error: bytes returned do not match bytes asked for")
#read a buffer's amount of samples
analysis = {}
for i in range(0,numberOfChannels):
analysis[str(i)] = []
# read the file or current analysis into memory and sum all of the samples and raise to the 2nd pow
if shouldReadEntireFile == True or (initialAnalyses % analysisSpacing == 0):
for i in range(0, bytesToRead, packetByteWidth):
subdata = data[i:(i+packetByteWidth)]
for chPos in range(0, numberOfChannels):
trackoffset = chPos * channelByteWidth
extendedValue = les24toles32Value(''.join(subdata[trackoffset:trackoffset+channelByteWidth]))
extendedValue = int(((1.0 * extendedValue) / constantBitDepth24) * constantBitDepth16)
analysis[str(chPos)].append(extendedValue)
maxValue = 0
track = 0
srcTrack = '0'
for i in [0,1,2,3,4,5]:
if int(srcTrack) == i:
continue
result = np.cov(analysis[srcTrack], analysis[str(i)])[0,1]
if result > maxValue:
maxValue = result
track = i
print srcTrack, "matches", track, "with", maxValue
initialAnalyses += 1
else:
data = f.read(packetByteWidth * numberOfSamplesRemaining)
numberOfSamplesRemaining = 0
return {"OK":"OK"}
def formatInformationForFile(filePath):
wavefile = PCWaveFile(filePath)
if wavefile == None or wavefile.isValidWaveFile() == False:
badInfo = {}
badInfo["fileName"] = os.path.basename(filePath)
badInfo["fullPath"] = filePath
badInfo["bitDepth"] = 0
badInfo["sampleRate"] = 0
badInfo["bextChunk"] = False
badInfo["timeStamp"] = 0
badInfo["timeStampTC"] = "00:00:00:00"
badInfo["regnChunk"] = False
badInfo["channelCount"] = 0
badInfo["status"] = 'fail'
badInfo["errors"] = [
"{} is not a wave file.".format(os.path.basename(filePath)),
]
badInfo['fileLevels'] = {
'status': 'fail',
'layout': 'unknown',
"warningString": ""
}
badInfo['headTones'] = {
'status': 'fail',
'resultString':
'Unable to analyze head tones on account of invalid format.',
"warningString": ""
}
badInfo['pops'] = {
'status': 'fail',
'resultString':
'Unable to analyze pops on account of invalid format.',
"warningString": ""
}
return badInfo
info = {}
info['fileName'] = os.path.basename(filePath)
info['fullPath'] = filePath
info['bitDepth'] = wavefile.numberOfBitsPerSample()
info['sampleRate'] = wavefile.numberOfSamplesPerSecond()
info['bextChunk'] = True if wavefile.hasBextChunk() else False
info['timeStamp'] = wavefile.timestamp
info['timeStampTC'] = prettyTimeCode2398(wavefile.timestamp,
info['sampleRate'])
info['regnChunk'] = True if wavefile.hasBextChunk() else False
if wavefile.hasRegnChunk() == True:
info['regnChunkUserTimeStamp'] = wavefile.regnChunkInfoDict[
'userTimeStamp']
info['regnChunkOriginalTimeStamp'] = wavefile.regnChunkInfoDict[
'originalTimeStamp']
info['channelCount'] = wavefile.numberOfChannels()
if info['bitDepth'] != 24 or info['sampleRate'] != 48000:
info['fileLevels'] = {
'status': 'fail',
'layout': 'unknown',
"warningString": ""
}
info['headTones'] = {
'status': 'fail',
'resultString':
'Unable to analyze head tones on account of invalid format.',
"warningString": ""
}
info['pops'] = {
'status': 'fail',
'resultString':
'Unable to analyze pops on account of invalid format.',
"warningString": ""
}
else:
info['fileLevels'] = analyzeLevels(filePath)
info['headTones'] = analyzeHeadAlt(filePath)
info['pops'] = analyzeTwoPopInEighths(filePath)
return addStatusToInformation(info)
def analyzeLevels(filePath):
wavefile = None
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
else:
return
if wavefile == None or wavefile.isValidWaveFile() == False:
return {"status":"fail", "resultString":"file is not a wavefile", "layout":"Unknown", "warningString":""}
sampleRate = wavefile.fmtChunkInfoDict['nSamplesPerSec']
channelByteWidth = wavefile.fmtChunkInfoDict['wBitsPerSample'] / 8
bitsPerSample = wavefile.fmtChunkInfoDict['wBitsPerSample']
numberOfChannels = wavefile.fmtChunkInfoDict['nChannels']
packetByteWidth = channelByteWidth * numberOfChannels
for title in ['MEHELP']:
if title.lower() in filePath.lower():
layout = 'MONO' if numberOfChannels == 1 else 'MULTI-MONO'
return {"status":"pass", "resultString":"The file contains {} in its file name and will not be analyzed.".format(title), "layout":layout, "warningString":""}
if numberOfChannels != 6:
if numberOfChannels == 2:
return {"status":"pass", "resultString":"File contains {} audio channels".format(numberOfChannels), "layout":"STEREO", "warningString":""}
else:
return {"status":"pass", "resultString":"File contains {} audio channels".format(numberOfChannels), "layout":"Unknown", "warningString":""}
isSigned = True
if channelByteWidth == 1:
isSigned = False
chunkPosition = wavefile.getDataChunk()['chunkPosition']
numberOfSamples = wavefile.getDataChunk()['chunkSize'] / packetByteWidth
#open the file, read through the packets and do a running mean of each track in the file
#what this does do is ASSUME that the input file is 24 bit signed data, otherwise 16 bit would be signed, but 8 bit wouldn't
#not that 8 bit is an issue
with open(filePath, 'rb') as f:
f.seek(chunkPosition)
blockWidthInSeconds = 4
samplesToReadBlockSize = (sampleRate * blockWidthInSeconds)
numberOfSamplesRemaining = numberOfSamples
bufferAmount = samplesToReadBlockSize if numberOfSamples > samplesToReadBlockSize else numberOfSamples
totalBlocksInFile = numberOfSamplesRemaining // samplesToReadBlockSize
shouldReadEntireFile = False
analysisSpacing = 0
numberOfAnalyses = 34
initialAnalyses = 0
if totalBlocksInFile < (2 * numberOfAnalyses):
shouldReadEntireFile = True
else:
analysisSpacing = totalBlocksInFile / numberOfAnalyses
startTime = datetime.datetime.now()
constantBitDepth16 = pow(2,15)
constantBitDepth24 = pow(2,23)
#totalsCounter = RunningTotalCounter(pow(2,(bitsPerSample-1)), samplesToReadBlockSize, numberOfChannels)
totalsCounter = RunningTotalCounter(constantBitDepth16, samplesToReadBlockSize, numberOfChannels)
# you need to constrain the size of the input as we will be calculating RMS values which require exponential math
while numberOfSamplesRemaining > 0:
#read a buffer's amount of samples
if numberOfSamplesRemaining > bufferAmount:
bytesToRead = packetByteWidth * bufferAmount
data = f.read(bytesToRead)
numberOfSamplesRemaining -= (len(data)/packetByteWidth)
if len(data) != bytesToRead:
logging.error("IO Error: bytes returned do not match bytes asked for")
#read a buffer's amount of samples
analysis = totalsCounter.emptyAnalysisDictionary()
# read the file or current analysis into memory and sum all of the samples and raise to the 2nd pow
if shouldReadEntireFile == True or (initialAnalyses % analysisSpacing == 0):
for i in range(0, bytesToRead, packetByteWidth):
subdata = data[i:(i+packetByteWidth)]
for chPos in range(0, numberOfChannels):
trackoffset = chPos * channelByteWidth
extendedValue = les24toles32Value(''.join(subdata[trackoffset:trackoffset+channelByteWidth]))
extendedValue = int(((1.0 * extendedValue) / constantBitDepth24) * constantBitDepth16)
analysis[str(chPos)] += (extendedValue**2)
totalsCounter.addAnalysis(analysis)
initialAnalyses += 1
else:
data = f.read(packetByteWidth * numberOfSamplesRemaining)
numberOfSamplesRemaining = 0
return totalsCounter.performAnalysis()
def analyzeTrackCorrelations(filePath):
wavefile = None
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
else:
return
if wavefile == None or wavefile.isValidWaveFile() == False:
return {
"status": "fail",
"resultString": "file is not a wavefile",
"layout": "Unknown"
}
sampleRate = wavefile.fmtChunkInfoDict['nSamplesPerSec']
channelByteWidth = wavefile.fmtChunkInfoDict['wBitsPerSample'] / 8
bitsPerSample = wavefile.fmtChunkInfoDict['wBitsPerSample']
numberOfChannels = wavefile.fmtChunkInfoDict['nChannels']
packetByteWidth = channelByteWidth * numberOfChannels
if numberOfChannels != 6:
if numberOfChannels == 2:
return {
"status":
"pass",
"resultString":
"File contains {} audio channels".format(numberOfChannels),
"layout":
"STEREO"
}
else:
return {
"status":
"pass",
"resultString":
"File contains {} audio channels".format(numberOfChannels),
"layout":
"Unknown"
}
isSigned = True
if channelByteWidth == 1:
isSigned = False
chunkPosition = wavefile.getDataChunk()['chunkPosition']
numberOfSamples = wavefile.getDataChunk()['chunkSize'] / packetByteWidth
#open the file, read through the packets and do a running mean of each track in the file
#what this does do is ASSUME that the input file is 24 bit signed data, otherwise 16 bit would be signed, but 8 bit wouldn't
#not that 8 bit is an issue
with open(filePath, 'rb') as f:
f.seek(chunkPosition)
blockWidthInSeconds = 1
samplesToReadBlockSize = (sampleRate * blockWidthInSeconds) / 2
numberOfSamplesRemaining = numberOfSamples
bufferAmount = samplesToReadBlockSize if numberOfSamples > samplesToReadBlockSize else numberOfSamples
totalBlocksInFile = numberOfSamplesRemaining // samplesToReadBlockSize
shouldReadEntireFile = False
analysisSpacing = 0
numberOfAnalyses = 34
initialAnalyses = 0
if totalBlocksInFile < (2 * numberOfAnalyses):
shouldReadEntireFile = True
else:
analysisSpacing = totalBlocksInFile / numberOfAnalyses
startTime = datetime.datetime.now()
constantBitDepth16 = pow(2, 15)
constantBitDepth24 = pow(2, 23)
#totalsCounter = RunningTotalCounter(pow(2,(bitsPerSample-1)), samplesToReadBlockSize, numberOfChannels)
totalsCounter = RunningTotalCounter(constantBitDepth16,
samplesToReadBlockSize,
numberOfChannels)
# you need to constrain the size of the input as we will be calculating RMS values which require exponential math
while numberOfSamplesRemaining > 0:
#read a buffer's amount of samples
if numberOfSamplesRemaining > bufferAmount:
bytesToRead = packetByteWidth * bufferAmount
data = f.read(bytesToRead)
numberOfSamplesRemaining -= (len(data) / packetByteWidth)
if len(data) != bytesToRead:
logging.error(
"IO Error: bytes returned do not match bytes asked for"
)
#read a buffer's amount of samples
analysis = {}
for i in range(0, numberOfChannels):
analysis[str(i)] = []
# read the file or current analysis into memory and sum all of the samples and raise to the 2nd pow
if shouldReadEntireFile == True or (initialAnalyses %
analysisSpacing == 0):
for i in range(0, bytesToRead, packetByteWidth):
subdata = data[i:(i + packetByteWidth)]
for chPos in range(0, numberOfChannels):
trackoffset = chPos * channelByteWidth
extendedValue = les24toles32Value(''.join(
subdata[trackoffset:trackoffset +
channelByteWidth]))
extendedValue = int(
((1.0 * extendedValue) / constantBitDepth24) *
constantBitDepth16)
analysis[str(chPos)].append(extendedValue)
maxValue = 0
track = 0
srcTrack = '0'
for i in [0, 1, 2, 3, 4, 5]:
if int(srcTrack) == i:
continue
result = np.cov(analysis[srcTrack],
analysis[str(i)])[0, 1]
if result > maxValue:
maxValue = result
track = i
print srcTrack, "matches", track, "with", maxValue
initialAnalyses += 1
else:
data = f.read(packetByteWidth * numberOfSamplesRemaining)
numberOfSamplesRemaining = 0
return {"OK": "OK"}
def analyzeTwoPopInEighths(filePath):
if not os.path.exists(filePath):
return {"status":"fail", "resultString":"file doesn't exist", "warningString":""}
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
return {"status":"fail", "resultString":"file is not a wavefile"}
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
timeCodeSampleRate = int(math.ceil(sampleRate * 1.001))
byteWidth = wavefile.numberOfBytesPerSample()
numberOfChannels = wavefile.numberOfChannels()
data_size = int(math.ceil(sampleRate * 1.001))/24 #1 second of audio
dataStartPosOffset = 0
normalizedTimeStamp = wavefile.timestamp #
currentPositionInFile = 0
dataLengthInSamples = dataLength / (numberOfChannels * byteWidth)
# print "dataStartPosOffset",dataStartPosOffset
toneStartTimeStamp = pctimecode.convertTimeCodeToUnsignedLongLong("00:57:30:00", timeCodeSampleRate, 24)
popSafeStartTimeStamp = pctimecode.convertTimeCodeToUnsignedLongLong("00:59:57:00", timeCodeSampleRate, 24)
hourOneTimeStamp = pctimecode.convertTimeCodeToUnsignedLongLong("01:00:00:00", timeCodeSampleRate, 24)
# print "normalizedTimeStamp",normalizedTimeStamp
# print "toneStartTimeStamp",toneStartTimeStamp
#assume file starts at 00:57:30:00, in the case that the file doesn't have a bext extension
#also gets into an issue of what to favor, bext over regn? I assume that the encoders user bext
#as you couldn't reliably expect a regn chunk
if normalizedTimeStamp == 0 or (normalizedTimeStamp % (timeCodeSampleRate * 60 * 60 * 24) == 0):
normalizedTimeStamp = toneStartTimeStamp
#we're going to read 3 seconds total
secondsInSamplesToRead = 3 * timeCodeSampleRate
# print "secondsInSamplesToRead",secondsInSamplesToRead
#make sure that the timestamp starts on the second, otherwise normalize the timestamp
normalizedTimestampOffset = wavefile.timestamp % timeCodeSampleRate
if normalizedTimestampOffset > 0:
dataStartPosOffset += (timeCodeSampleRate - normalizedTimestampOffset) * byteWidth * numberOfChannels
normalizedTimeStamp = wavefile.timestamp + (timeCodeSampleRate - normalizedTimestampOffset)
# print "(timeCodeSampleRate - normalizedTimestampOffset)", (timeCodeSampleRate - normalizedTimestampOffset)
# print "normalizedTimestampOffset", normalizedTimestampOffset
# print "dataStartPosOffset", dataStartPosOffset
# print "normalizedTimeStamp", normalizedTimeStamp
if normalizedTimeStamp > popSafeStartTimeStamp:
# print 'if wavefile.timestamp > toneStartTimeStamp:'
return {"status":"fail", "resultString":"timecode start exceeds pop location","warningString":""}
if dataLengthInSamples < (hourOneTimeStamp - toneStartTimeStamp):
return {"status":"fail", "resultString":"the file is too short", "warningString":""}
if normalizedTimeStamp < popSafeStartTimeStamp:
#seek to start time to analyze
sampleDifference = (popSafeStartTimeStamp - normalizedTimeStamp)
currentPositionInFile = sampleDifference + normalizedTimeStamp
dataStartPosOffset += (sampleDifference * byteWidth * numberOfChannels)
#lets only look at the 3 seconds of the file from 00:59:57:00 to 01:00:00:00
sizeToRead = secondsInSamplesToRead * byteWidth * numberOfChannels
sizeToRead = sizeToRead if dataLength > sizeToRead else dataLength
#this scans the head to make sure there is no signal in the head up to a sixteenth of frame before the pop
headResult = preScanHead(filePath, data_size, dataStartPos, dataStartPosOffset, byteWidth, numberOfChannels, sampleRate)
#this scans the tail to make sure there is no signal in the tail after an eighth of a frame after the pop
nextByteOffset = (data_size * 25 * byteWidth * numberOfChannels) #seek 25 frames from the initial start which would be 59:58:01
tailResult = preScanTail(filePath, data_size, dataStartPos, (dataStartPosOffset + nextByteOffset), byteWidth, numberOfChannels, sampleRate)
overallResult = {"status":"pass", "resultString":""}
suppressSuperfluousWarnings = False
if headResult['status'] == 'fail':
overallResult['status'] = 'fail'
overallResult["resultString"] += headResult["resultString"]
if tailResult['status'] == 'fail':
overallResult['status'] = 'fail'
overallResult["resultString"] += tailResult["resultString"]
if overallResult['status'] == 'pass':
overallResult["resultString"] = "Pop area contains no discernable problems."
twoPopCounter = TwoPopCounterAlt()
twoPopCounter.timeCodeSampleRate = timeCodeSampleRate
twoPopCounter.timeBase = 24
with open(filePath, 'rb') as f:
f.seek(dataStartPos + dataStartPosOffset)
dataToReadChunkSize = data_size*byteWidth*numberOfChannels
dataRemaining = sizeToRead
while dataRemaining > 0:
if dataRemaining > dataToReadChunkSize:
data = f.read(dataToReadChunkSize)
else:
data = f.read(dataRemaining)
dataRemaining -= len(data)
for channel in range(0, numberOfChannels):
subdata = ""
for x in range(channel*byteWidth,len(data),byteWidth*numberOfChannels):
subdata += (data[x] + data[x+1] + data[x+2])
analysis = ultimateFreqAndPowerForData(subdata, byteWidth, sampleRate)
twoPopCounter.addAnalysisForChannelAndTimeStamp(analysis, channel, currentPositionInFile)
currentPositionInFile += data_size
twoPopResult = twoPopCounter.result()
#likely that the pop is off as we are checking an area of up to a sixteenth of frame prior to the pop and area an eighth of a frame
#after the pop
if overallResult['status'] == 'fail' and twoPopResult['status'] == 'pass':
twoPopResult['status'] = 'fail'
twoPopResult['resultString'] = "Check your pop as it is likely off. " + overallResult['resultString']
if overallResult['status'] == 'fail' and twoPopResult['status'] == 'fail' and twoPopResult['foundPop'] == 'pass':
twoPopResult['status'] = 'fail'
twoPopResult['resultString'] = "Check your pop as it is likely off. " + overallResult['resultString']
popIntegrityResult = {"status":"pass", "resultString":""}
if overallResult['status'] == 'pass' and twoPopResult['status'] == 'pass' and twoPopResult['foundPop'] == 'pass':
#lets make sure the pop actually starts consists of a frame
nextByteOffset = (data_size * 23 * byteWidth * numberOfChannels) #seek 23 frames from the initial start which would be 59:57:23
popIntegrityResult = scanPop(filePath, data_size, dataStartPos, (dataStartPosOffset + nextByteOffset), byteWidth, numberOfChannels, sampleRate)
if popIntegrityResult['status'] == 'fail':
twoPopResult['status'] = 'fail'
twoPopResult['resultString'] = popIntegrityResult['resultString']
return twoPopResult
def analyzeTwoPopAlt(filePath):
if not os.path.exists(filePath):
return {"status":"fail", "resultString":"file doesn't exist", "warningString":""}
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
return {"status":"fail", "resultString":"file is not a wavefile"}
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
timeCodeSampleRate = int(math.ceil(sampleRate * 1.001))
byteWidth = wavefile.numberOfBytesPerSample()
numberOfChannels = wavefile.numberOfChannels()
data_size = int(math.ceil(sampleRate * 1.001))/24 #1 second of audio
dataStartPosOffset = 0
normalizedTimeStamp = wavefile.timestamp #
currentPositionInFile = 0
dataLengthInSamples = dataLength / (numberOfChannels * byteWidth)
# print "dataStartPosOffset",dataStartPosOffset
toneStartTimeStamp = pctimecode.convertTimeCodeToUnsignedLongLong("00:57:30:00", timeCodeSampleRate, 24)
popSafeStartTimeStamp = pctimecode.convertTimeCodeToUnsignedLongLong("00:59:57:00", timeCodeSampleRate, 24)
hourOneTimeStamp = pctimecode.convertTimeCodeToUnsignedLongLong("01:00:00:00", timeCodeSampleRate, 24)
# print "normalizedTimeStamp",normalizedTimeStamp
# print "toneStartTimeStamp",toneStartTimeStamp
#assume file starts at 00:57:30:00, in the case that the file doesn't have a bext extension
#also gets into an issue of what to favor, bext over regn? I assume that the encoders user bext
#as you couldn't reliably expect a regn chunk
if normalizedTimeStamp == 0:
normalizedTimeStamp = toneStartTimeStamp
#we're going to read 3 seconds total
secondsInSamplesToRead = 3 * timeCodeSampleRate
# print "secondsInSamplesToRead",secondsInSamplesToRead
#make sure that the timestamp starts on the second, otherwise normalize the timestamp
normalizedTimestampOffset = wavefile.timestamp % timeCodeSampleRate
if normalizedTimestampOffset > 0:
dataStartPosOffset += (timeCodeSampleRate - normalizedTimestampOffset) * byteWidth * numberOfChannels
normalizedTimeStamp = wavefile.timestamp + (timeCodeSampleRate - normalizedTimestampOffset)
# print "(timeCodeSampleRate - normalizedTimestampOffset)", (timeCodeSampleRate - normalizedTimestampOffset)
# print "normalizedTimestampOffset", normalizedTimestampOffset
# print "dataStartPosOffset", dataStartPosOffset
# print "normalizedTimeStamp", normalizedTimeStamp
if normalizedTimeStamp > popSafeStartTimeStamp:
# print 'if wavefile.timestamp > toneStartTimeStamp:'
return {"status":"fail", "resultString":"timecode start exceeds pop location","warningString":""}
if dataLengthInSamples < (hourOneTimeStamp - toneStartTimeStamp):
return {"status":"fail", "resultString":"the file is too short", "warningString":""}
if normalizedTimeStamp < popSafeStartTimeStamp:
# print 'if wavefile.timestamp < toneStartTimeStamp:'
sampleDifference = (popSafeStartTimeStamp - normalizedTimeStamp)
currentPositionInFile = sampleDifference + normalizedTimeStamp
dataStartPosOffset += (sampleDifference * byteWidth * numberOfChannels)
#lets only look at the 3 seconds of the file from 00:59:57:00 to 01:00:00:00
sizeToRead = secondsInSamplesToRead * byteWidth * numberOfChannels
sizeToRead = sizeToRead if dataLength > sizeToRead else dataLength
# print "sizeToRead",sizeToRead
with open(filePath, 'rb') as f:
f.seek(dataStartPos + dataStartPosOffset)
dataToReadChunkSize = data_size*byteWidth*numberOfChannels
dataRemaining = sizeToRead
twoPopCounter = TwoPopCounterAlt()
twoPopCounter.timeCodeSampleRate = timeCodeSampleRate
twoPopCounter.timeBase = 24
while dataRemaining > 0:
if dataRemaining > dataToReadChunkSize:
data = f.read(dataToReadChunkSize)
else:
data = f.read(dataRemaining)
dataRemaining -= len(data)
for channel in range(0, numberOfChannels):
subdata = ""
for x in range(channel*byteWidth,len(data),byteWidth*numberOfChannels):
subdata += (data[x] + data[x+1] + data[x+2])
analysis = ultimateFreqAndPowerForData(subdata, byteWidth, sampleRate)
twoPopCounter.addAnalysisForChannelAndTimeStamp(analysis, channel, currentPositionInFile)
currentPositionInFile += data_size
return twoPopCounter.result()
def analyzeHeadAlt(filePath):
if os.path.exists(filePath):
wavefile = PCWaveFile(filePath)
if wavefile.isValidWaveFile() != True:
return {"status":"fail", "resultString":"file is not a wavefile"}
dataStartPos = wavefile.getDataChunk()["chunkPosition"]
dataLength = wavefile.getDataChunk()["chunkSize"]
sampleRate = wavefile.numberOfSamplesPerSecond()
timeCodeSampleRate = int(math.ceil(sampleRate * 1.001))
byteWidth = wavefile.numberOfBytesPerSample()
numberOfChannels = wavefile.numberOfChannels()
data_size = int(math.ceil(sampleRate * 1.001)) #1 second of audio
dataLengthInSamples = dataLength / (numberOfChannels * byteWidth)
dataStartPosOffset = 0
normalizedTimeStamp = wavefile.timestamp #
# print "dataStartPosOffset",dataStartPosOffset
toneStartTimeStamp = pctimecode.convertTimeCodeToUnsignedLongLong("00:57:30:00", timeCodeSampleRate, 24)
# print "normalizedTimeStamp",normalizedTimeStamp
# print "toneStartTimeStamp",toneStartTimeStamp
#assume file starts at 00:57:30:00, in the case that the file doesn't have a bext extension
#also gets into an issue of what to favor, bext over regn? I assume that the encoders user bext
#as you couldn't reliably expect a regn chunk
if normalizedTimeStamp == 0 or (normalizedTimeStamp % (timeCodeSampleRate * 60 * 60 * 24) == 0) :
normalizedTimeStamp = toneStartTimeStamp
#the head length should be 150 seconds
secondsInSamplesToRead = 150 * timeCodeSampleRate
# print "secondsInSamplesToRead",secondsInSamplesToRead
if secondsInSamplesToRead > dataLengthInSamples:
return {"status":"fail", "resultString":"the file is too short", "warningString":""}
#make sure that the timestamp starts on the second, otherwise normalize the timestamp
normalizedTimestampOffset = wavefile.timestamp % timeCodeSampleRate
if normalizedTimestampOffset > 0:
dataStartPosOffset += (timeCodeSampleRate - normalizedTimestampOffset) * byteWidth * numberOfChannels
normalizedTimeStamp = wavefile.timestamp + (timeCodeSampleRate - normalizedTimestampOffset)
# print "(timeCodeSampleRate - normalizedTimestampOffset)", (timeCodeSampleRate - normalizedTimestampOffset)
# print "normalizedTimestampOffset", normalizedTimestampOffset
# print "dataStartPosOffset", dataStartPosOffset
# print "normalizedTimeStamp", normalizedTimeStamp
if normalizedTimeStamp > toneStartTimeStamp:
# print 'if wavefile.timestamp > toneStartTimeStamp:'
secondsInSamplesToRead -= (normalizedTimeStamp - toneStartTimeStamp)
if normalizedTimeStamp < toneStartTimeStamp:
# print 'if wavefile.timestamp < toneStartTimeStamp:'
secondsInSamplesToRead += (toneStartTimeStamp - normalizedTimeStamp)
#lets only look at the first 150 seconds of the file
sizeToRead = secondsInSamplesToRead * byteWidth * numberOfChannels
sizeToRead = sizeToRead if dataLength > sizeToRead else dataLength
second = 0
with open(filePath, 'rb') as f:
startTime = datetime.datetime.now()
dataStartPos += dataStartPosOffset
f.seek(dataStartPos)
dataToReadChunkSize = timeCodeSampleRate*byteWidth*numberOfChannels
dataRemaining = sizeToRead
counterDict = {}
for channel in range(0, numberOfChannels):
counterDict[str(channel)] = ToneCounterAlt()
while dataRemaining > 0:
if dataRemaining > dataToReadChunkSize:
data = f.read(dataToReadChunkSize)
else:
data = f.read(dataRemaining)
dataRemaining -= len(data)
currentPositionInFile = normalizedTimeStamp + (second * data_size)
# print currentPositionInFile, pctimecode.convertUnsignedLongLongToTimeCode(currentPositionInFile, 48048, 24)
for channel in range(0, numberOfChannels):
subdata = ""
for x in range(channel*byteWidth,len(data),byteWidth*numberOfChannels):
subdata += (data[x] + data[x+1] + data[x+2])
result = highestAndNextHighestFreqAndPowerForData(subdata, byteWidth, sampleRate)
toneCounter = counterDict[str(channel)]
toneCounter.addAnalysis({
"currentTimeStamp":currentPositionInFile,
"timeCodeSampleRate":timeCodeSampleRate,
"freqs":result,
"RMSForInputData":calculateRMSForData(subdata, byteWidth)})
second += 1
#Questions to answer 1) Does it have at least 60 secs of tone 2) Does it start at head 3) Does it have a Pop 4) Does it have other tones?
toneCounterResult = ToneCounterResultAlt(timeCodeSampleRate, 24)
return toneCounterResult.result(counterDict)
else:
return {"status":"fail", "resultString":"file doesn't exist","warningString":""}