deepgeek/gradio-webrtc
2
0
Fork 0
mirror of https://github.com/HumanAIGC-Engineering/gradio-webrtc.git synced 2026-02-05 18:09:23 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix audio type conversion (#259)

Browse Source 
* Fix conversion between audio dtypes

* Run Pytest in CI

* Add pytest tests path in pyproject.toml

* Fix usages

* Use other PR's test format (more or less)

* Support legacy arguments

* Fix pyproject.toml and test location

* Omit `test` arg in CI, given by pyproject.toml

---------

Co-authored-by: Freddy Boulton <alfonsoboulton@gmail.com>
This commit is contained in:
Václav Volhejn
2025-04-09 16:00:23 +02:00
committed by GitHub
parent fdf6bea1c6
commit 58bccddd93
9 changed files with 128 additions and 43 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
backend/fastrtc/websocket.py
View File

@@ -12,7 +12,14 @@ from fastapi import WebSocket
from fastapi.websockets import WebSocketDisconnect, WebSocketState
from .tracks import AsyncStreamHandler, StreamHandlerImpl
from .utils import AdditionalOutputs, CloseStream, DataChannel, split_output
from .utils import (
AdditionalOutputs,
CloseStream,
DataChannel,
audio_to_float32,
audio_to_int16,
split_output,
)
class WebSocketDataChannel(DataChannel):
@@ -31,14 +38,12 @@ def convert_to_mulaw(
audio_data: np.ndarray, original_rate: int, target_rate: int
) -> bytes:
"""Convert audio data to 8kHz mu-law format"""
if audio_data.dtype != np.float32:
audio_data = audio_data.astype(np.float32) / 32768.0
audio_data = audio_to_float32(audio_data)
if original_rate != target_rate:
audio_data = librosa.resample(audio_data, orig_sr=original_rate, target_sr=8000)
audio_data = (audio_data * 32768).astype(np.int16)
audio_data = audio_to_int16(audio_data)
return audioop.lin2ulaw(audio_data, 2) # type: ignore
@@ -122,14 +127,13 @@ class WebSocketHandler:
)
if self.stream_handler.input_sample_rate != 8000:
audio_array = audio_array.astype(np.float32) / 32768.0
audio_array = audio_to_float32(audio_array)
audio_array = librosa.resample(
audio_array,
orig_sr=8000,
target_sr=self.stream_handler.input_sample_rate,
)
audio_array = (audio_array * 32768).astype(np.int16)
audio_array = audio_to_int16(audio_array)
try:
if isinstance(self.stream_handler, AsyncStreamHandler):
await self.stream_handler.receive(