diff --git a/server.py b/server.py index 89d78f8..6c09515 100644 --- a/server.py +++ b/server.py @@ -82,18 +82,10 @@ def process_audio(audio_bytes: bytes) -> tuple: 处理音频文件 返回: (audio_array, sample_rate) """ - # 使用 soundfile 读取音频 audio_io = io.BytesIO(audio_bytes) - audio, sr = sf.read(audio_io) - # 转换为单声道 - if len(audio.shape) > 1: - audio = audio.mean(axis=1) - - # 重采样到 16kHz - if sr != SAMPLE_RATE: - import resampy - audio = resampy.resample(audio, sr, SAMPLE_RATE) + # 使用 librosa 读取音频(支持更多格式:WAV, WebM, MP3, FLAC 等) + audio, sr = librosa.load(audio_io, sr=SAMPLE_RATE, mono=True) return audio, SAMPLE_RATE