Unverified50% confidenceFactExact time
原版 Whisper 的单次处理窗口仅为 30 秒，需要通过滑动窗口拼接处理长音频
1
Sources
50%
Confidence
Long-term
Relevance
7/11/2026
First Seen
Sources
MOSS-Transcribe-Diarize：0.9B端到端多说话人语音转写模型
twitterlmsysorg7/9/2026
Related Claims
Unverified在 M1 芯片上运行 Whisper base 模型的转录速度可达音频时长的 10-15 倍，即 1 分钟音频仅需 4-6 秒完成转录77% similarUnverifiedWhisper small模型在现代笔记本电脑的CPU上处理一段10分钟音频通常只需1-3分钟76% similarUnverifiedWhisper的语音识别模型只能处理16kHz采样率的单声道PCM浮点数据格式的原始音频波形68% similarUnverifiedThe human ear's discrimination threshold for inter-room audio synchronization delay differences is approximately 30 milliseconds.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/486966API
curl https://kongchang.com/api/v1/knowledge/claims/486966MCP
get_claim(id=486966)