Skip to content

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

fix: support loading .pcm (16k 1c 16bit) audio files in load_utils.py (#2667) (#2668)

weilikaieb1574b813e2
diff --git a/funasr/utils/load_utils.py b/funasr/utils/load_utils.py
index c38f1d937..d208f7dc0 100644
--- a/funasr/utils/load_utils.py
+++ b/funasr/utils/load_utils.py
@@ -241,10 +241,19 @@ def _load_audio_ffmpeg(file: str, sr: int = 16000):
# This launches a subprocess to decode audio while down-mixing
# and resampling as necessary. Requires the ffmpeg CLI in PATH.
# fmt: off
+ pcm_params = []
+ if file.lower().endswith('.pcm'):
+ pcm_params = [
+ "-f", "s16le",
+ "-ar", str(sr),
+ "-ac", "1"
+ ]
+
cmd = [
"ffmpeg",
"-nostdin",
"-threads", "0",
+ *pcm_params, # PCM files need input format specified before -i since PCM is raw data without headers
"-i", file,
"-f", "s16le",
"-ac", "1",