Skip to content

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

Fix audio format (#2159)

Djraemona76f15c78527
diff --git a/runtime/python/onnxruntime/funasr_onnx/sensevoice_bin.py b/runtime/python/onnxruntime/funasr_onnx/sensevoice_bin.py
index 37c98a8d0..b3d8dcf6e 100644
--- a/runtime/python/onnxruntime/funasr_onnx/sensevoice_bin.py
+++ b/runtime/python/onnxruntime/funasr_onnx/sensevoice_bin.py
@@ -196,7 +196,24 @@ class SenseVoiceSmall:
return asr_res
def load_data(self, wav_content: Union[str, np.ndarray, List[str]], fs: int = None) -> List:
+
+ def convert_to_wav(input_path, output_path):
+ from pydub import AudioSegment
+ try:
+ audio = AudioSegment.from_mp3(input_path)
+ audio.export(output_path, format="wav")
+ print("音频文件为mp3格式,已转换为wav格式")
+
+ except Exception as e:
+ print(f"转换失败:{e}")
+
def load_wav(path: str) -> np.ndarray:
+ if not path.lower().endswith('.wav'):
+ import os
+ input_path = path
+ path = os.path.splitext(path)[0]+'.wav'
+ convert_to_wav(input_path,path) #将mp3格式转换成wav格式
+
waveform, _ = librosa.load(path, sr=fs)
return waveform
@@ -215,6 +232,10 @@ class SenseVoiceSmall:
feats, feats_len = [], []
for waveform in waveform_list:
speech, _ = self.frontend.fbank(waveform)
+
+ if speech is None or speech.size == 0:
+ print("detected speech size {speech.size}")
+ raise ValueError("Empty speech detected, skipping this waveform.")
feat, feat_len = self.frontend.lfr_cmvn(speech)
feats.append(feat)
feats_len.append(feat_len)