[Bugfix] Fix TorchCodec audio IO correctness (#58364)

Signed-off-by: Isotr0py <Isotr0py@outlook.com>
This commit is contained in:
Isotr0py
2026-09-29 11:34:48 -05:00
committed by GitHub
parent c3a4a36c20
commit 31876d8ea1
3 changed files with 67 additions and 36 deletions
+20 -3
View File
@@ -185,9 +185,15 @@ def test_load_audio_backend_matches_default(backend, dummy_audio_bytes):
ref_audio, ref_sr = load_audio(BytesIO(dummy_audio_bytes), sr=None)
audio, sr = load_audio(BytesIO(dummy_audio_bytes), sr=None, backend=backend)
assert sr == ref_sr
# Decoders disagree only on codec encoder-delay/padding, so torchcodec may
# emit a few extra trailing samples (e.g. ~192 for Ogg Vorbis). Compare the
# overlapping region, which must agree to float32 precision.
if backend == "torchcodec":
# Vorbis encoder padding is trimmed, so torchcodec must match the
# soundfile reference exactly, including the waveform length.
assert audio.shape == ref_audio.shape
np.testing.assert_allclose(ref_audio, audio, atol=1e-4)
return
# Decoders disagree only on codec encoder-delay/padding, so PyAV may
# emit a few extra trailing samples (e.g. ~192 for Ogg Vorbis). Compare
# the overlapping region, which must agree to float32 precision.
n = min(ref_audio.shape[-1], audio.shape[-1])
assert n > 0
np.testing.assert_allclose(ref_audio[:n], audio[:n], atol=1e-4)
@@ -201,6 +207,17 @@ def test_load_audio_default_preserves_vorbis_length(dummy_audio_bytes):
assert audio.shape == expected.shape
def test_load_audio_torchcodec_preserves_vorbis_length(dummy_audio_bytes):
"""Torchcodec must trim Vorbis trailing padding even on FFmpeg < 5.0,
matching the soundfile reference exactly."""
pytest.importorskip("torchcodec")
expected, expected_sr = load_audio_soundfile(BytesIO(dummy_audio_bytes), sr=None)
audio, sr = load_audio_torchcodec(BytesIO(dummy_audio_bytes), sr=None)
assert sr == expected_sr
assert audio.shape == expected.shape
np.testing.assert_allclose(audio, expected, atol=1e-4)
def test_load_audio_unknown_backend_rejected(dummy_audio_bytes):
"""An unknown backend must fail loudly instead of silently degrading."""
with pytest.raises(ValueError, match="Unknown audio backend"):