def processVideo(inputFile, outputFile, tempDir):
global frameRate
command = "ffmpeg -i '" + inputFile + "' -qscale:v " + str(
FRAME_QUALITY) + " " + tempDir + "/frame%06d.jpg -hide_banner"
subprocess.call(command, shell=True)
command = "ffmpeg -i '" + inputFile + "' -ab 160k -ac 2 -ar " + str(
SAMPLE_RATE) + " -vn " + tempDir + "/audio.wav"
subprocess.call(command, shell=True)
sampleRate, audioData = wavfile.read(tempDir + "/audio.wav")
audioSampleCount = audioData.shape[0]
maxAudioVolume = getMaxVolume(audioData)
if frameRate is None:
frameRate = getFrameRate(inputFile)
samplesPerFrame = sampleRate / frameRate
audioFrameCount = int(math.ceil(audioSampleCount / samplesPerFrame))
hasLoudAudio = np.zeros((audioFrameCount))
for i in range(audioFrameCount):
start = int(i * samplesPerFrame)
end = min(int((i + 1) * samplesPerFrame), audioSampleCount)
audiochunks = audioData[start:end]
maxchunksVolume = float(getMaxVolume(audiochunks)) / max(
maxAudioVolume, 1e-10)
if maxchunksVolume >= SILENT_THRESHOLD:
hasLoudAudio[i] = 1
chunks = [[0, 0, 0]]
shouldIncludeFrame = np.zeros((audioFrameCount))
for i in range(audioFrameCount):
start = int(max(0, i - FRAME_SPREADAGE))
end = int(min(audioFrameCount, i + 1 + FRAME_SPREADAGE))
shouldIncludeFrame[i] = np.max(hasLoudAudio[start:end])
if (i >= 1 and shouldIncludeFrame[i] !=
shouldIncludeFrame[i - 1]): # Did we flip?
chunks.append([chunks[-1][1], i, shouldIncludeFrame[i - 1]])
chunks.append([chunks[-1][1], audioFrameCount, shouldIncludeFrame[i - 1]])
chunks = chunks[1:]
outputAudioData = []
outputPointer = 0
mask = [
x / AUDIO_FADE_ENVELOPE_SIZE for x in range(AUDIO_FADE_ENVELOPE_SIZE)
] # Create audio envelope mask
lastExistingFrame = None
for chunk in chunks:
audioChunk = audioData[int(chunk[0] *
samplesPerFrame):int(chunk[1] *
samplesPerFrame)]
sFile = tempDir + "/tempStart.wav"
eFile = tempDir + "/tempEnd.wav"
wavfile.write(sFile, SAMPLE_RATE, audioChunk)
with WavReader(sFile) as reader:
with WavWriter(eFile, reader.channels,
reader.samplerate) as writer:
tsm = audio_stretch_algorithm(reader.channels,
speed=NEW_SPEED[int(chunk[2])])
tsm.run(reader, writer)
_, alteredAudioData = wavfile.read(eFile)
leng = alteredAudioData.shape[0]
endPointer = outputPointer + leng
outputAudioData.extend((alteredAudioData / maxAudioVolume).tolist())
# Smoothing the audio
if leng < AUDIO_FADE_ENVELOPE_SIZE:
for i in range(outputPointer, endPointer):
outputAudioData[i] = 0
else:
for i in range(outputPointer,
outputPointer + AUDIO_FADE_ENVELOPE_SIZE):
outputAudioData[i][0] *= mask[i - outputPointer]
outputAudioData[i][1] *= mask[i - outputPointer]
for i in range(endPointer - AUDIO_FADE_ENVELOPE_SIZE, endPointer):
outputAudioData[i][0] *= (
1 - mask[i - endPointer + AUDIO_FADE_ENVELOPE_SIZE])
outputAudioData[i][1] *= (
1 - mask[i - endPointer + AUDIO_FADE_ENVELOPE_SIZE])
startOutputFrame = int(math.ceil(outputPointer / samplesPerFrame))
endOutputFrame = int(math.ceil(endPointer / samplesPerFrame))
for outputFrame in range(startOutputFrame, endOutputFrame):
inputFrame = int(chunk[0] + NEW_SPEED[int(chunk[2])] *
(outputFrame - startOutputFrame))
didItWork = copyFrame(inputFrame, outputFrame, tempDir)
if didItWork:
lastExistingFrame = inputFrame
else:
copyFrame(lastExistingFrame, outputFrame, tempDir)
outputPointer = endPointer
outputAudioData = np.asarray(outputAudioData)
wavfile.write(tempDir + "/audioNew.wav", SAMPLE_RATE, outputAudioData)
command = f"ffmpeg -framerate {frameRate} -i {tempDir}/newFrame%06d.jpg -i {tempDir}/audioNew.wav -strict -2 -c:v libx264 -preset {H264_PRESET} -crf {H264_CRF} -pix_fmt yuvj420p '{outputFile}'"
subprocess.call(command, shell=True)
for chunk in chunks:
if EDL:
if (chunk[2] == True):
edlFrameNumber += 1
writeELD(chunk[0], chunk[1], edlFrameNumber)
continue
audioChunk = audioData[int(chunk[0] *
samplesPerFrame):int(chunk[1] *
samplesPerFrame)]
sFile = TEMP_FOLDER.name + "/tempStart.wav"
eFile = TEMP_FOLDER.name + "/tempEnd.wav"
wavfile.write(sFile, SAMPLE_RATE, audioChunk)
with WavReader(sFile) as reader:
with WavWriter(eFile, reader.channels, reader.samplerate) as writer:
tsm = audio_stretch_algorithm(reader.channels,
speed=NEW_SPEED[int(chunk[2])])
tsm.run(reader, writer)
_, alteredAudioData = wavfile.read(eFile)
leng = alteredAudioData.shape[0]
endPointer = outputPointer + leng
outputAudioData.extend((alteredAudioData / maxAudioVolume).tolist())
# Smoothing the audio
if leng < AUDIO_FADE_ENVELOPE_SIZE:
for i in range(outputPointer, endPointer):
outputAudioData[i] = 0
else:
for i in range(outputPointer,
outputPointer + AUDIO_FADE_ENVELOPE_SIZE):
outputAudioData[i][0] *= mask[i - outputPointer]
outputAudioData[i][1] *= mask[i - outputPointer]