Lyra V2 – a better, faster, and more versatile speech codec

Re:Lyra V2 – a better, faster, and more versatile speech codec by orlp on Tuesday October 04, 2022 @12:12PM (#33082483)

You can bootleg your own fast lossless codec by doing delta-encoding on the raw PCM to get a lot of zeros and then feed it through an off-the-shelf fast compressor like snappy/lz4/zstandard/etc. It won't get remotely close to the dedicated audio algorithms, but I wouldn't be surprised if you cut your data size by a factor 2-4 and essentially no CPU cost compared to raw uncompressed audio.

[ Parent | on HN ]

Re:Lyra V2 – a better, faster, and more versatile speech codec by muziq on Wednesday October 05, 2022 @05:07PM (#33101255)

You’ve not done this before have you ?

[ Parent | on HN ]

Re:Lyra V2 – a better, faster, and more versatile speech codec by orlp on Thursday October 13, 2022 @10:15AM (#33191258)

I haven't, but now I have. I took https://opus-codec.org/static/examples/samples/music_orig.wa... from https://opus-codec.org/examples/. Then I wrote the following snippet of Python code:

    from scipy.io import wavfile
    import numpy as np
    import zstd

    sampling_rate, samples = wavfile.read(r'data/bootleg-compress/music_orig.wav')
    orig = samples.tobytes()

    naive_compressed = zstd.ZSTD_compress(orig)
    deltas = np.diff(samples, prepend=samples.dtype.type(0), axis=0) # Per-channel deltas.
    compressed_deltas = zstd.ZSTD_compress(deltas.ravel()) # Interleave channels and compress.

    decompressed_deltas = np.frombuffer(zstd.ZSTD_uncompress(compressed_deltas), dtype=samples.dtype)
    decompressed = np.cumsum(decompressed_deltas.reshape(deltas.shape), axis=0, dtype=samples.dtype)
    assert np.array_equal(samples, decompressed)

    print(len(orig))
    print(len(naive_compressed))
    print(len(compressed_deltas))

giving:

    17432876
    15518973
    12817602

Looks like my initial estimation of 2-4 was way off (when FLAC achieves ~2 this should've been a red flag), but you do get a ~1.36x reduction in space at basically memory read speed.

[ Parent | on HN ]