zlei6/ComfyUI
1
0
Fork 0
mirror of https://github.com/Comfy-Org/ComfyUI.git synced 2026-08-24 10:03:55 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Partner Nodes] fix(ByteDance): encode stereo reference audio without doubling its duration (#15177)

Browse Source 
* [Partner Nodes] fix(ByteDance): encode stereo reference audio without doubling its duration

Signed-off-by: bigcat88 <bigcat88@icloud.com>
This commit is contained in:
Alexander Piskun
2026-07-31 21:57:45 +03:00
committed by GitHub
parent 0b73ae83a1
commit 6cedd34343
2 changed files with 89 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
comfy_api_nodes/util/conversions.py
View File

@@ -266,16 +266,14 @@ def audio_tensor_to_contiguous_ndarray(waveform: torch.Tensor) -> np.ndarray:
waveform: a tensor of shape (1, channels, samples) derived from a Comfy `AUDIO` type.
Returns:
Contiguous numpy array of the audio waveform. If the audio was batched,
the first item is taken.
Contiguous numpy array of the audio waveform.
Raises:
ValueError: If the waveform is not shaped (1, channels, samples).
"""
if waveform.ndim != 3 or waveform.shape[0] != 1:
raise ValueError("Expected waveform tensor shape (1, channels, samples)")
# If batch is > 1, take first item
if waveform.shape[0] > 1:
waveform = waveform[0]
# Prepare for av: remove batch dim, move to CPU, make contiguous, convert to numpy array
audio_data_np = waveform.squeeze(0).cpu().contiguous().numpy()
if audio_data_np.dtype != np.float32:
@@ -285,20 +283,21 @@ def audio_tensor_to_contiguous_ndarray(waveform: torch.Tensor) -> np.ndarray:
def audio_input_to_mp3(audio: Input.Audio) -> BytesIO:
waveform = audio["waveform"].cpu()
audio_data_np = audio_tensor_to_contiguous_ndarray(audio["waveform"])
sample_rate = int(audio["sample_rate"])
output_buffer = BytesIO()
output_container = av.open(output_buffer, mode="w", format="mp3")
out_stream = output_container.add_stream("libmp3lame", rate=audio["sample_rate"])
out_stream = output_container.add_stream("libmp3lame", rate=sample_rate)
out_stream.bit_rate = 320000
frame = av.AudioFrame.from_ndarray(
waveform.movedim(0, 1).reshape(1, -1).float().numpy(),
format="flt",
layout="mono" if waveform.shape[0] == 1 else "stereo",
audio_data_np,
format="fltp",
layout="stereo" if audio_data_np.shape[0] > 1 else "mono",
)
frame.sample_rate = audio["sample_rate"]
frame.sample_rate = sample_rate
frame.pts = 0
output_container.mux(out_stream.encode(frame))
output_container.mux(out_stream.encode(None))