Some questions

[Johndirr]

What a nice project! I'm was looking for a tool that can diff audio files and display their differences.

Since you wrote in the project description that the parameters for feature extraction are like this:

Parameters:

n_mels: 128 (frequency bins)
hop_length: 512 (frame shift)
win_length: 2048 (frame size)
Sample rate: 16kHz

I wonder why the code has other values?

    def __init__(self, file1_path, file2_path):
        self.file1_path = Path(file1_path)
        self.file2_path = Path(file2_path)
        
        # Mel-spectrogram parameters
        self.sr = 16000
        self.n_mels = 128
        self.n_fft = 2048
        self.win_length = 400
        self.hop_length = 160 # 512
        self.similarity_threshold = 1.0 # 1.5 
        #self.distance_threshold = 0.3
        # Store audio data
        self.sr1 = None
        self.sr2 = None
        self.segments = []

I also wonder how the whole process scales? It is possible to diff two audio files of the same movie (uncut and cut edition) in a reasonable time?

Thank you