[FFmpeg-user] Detect if a wav file is silence, beep, or live human talking

Isaac Asimov iasimovsp at outlook.com
Sun Jul 17 18:14:38 EEST 2016


You can turn wav into raw audio pcm format like .s16 with ffmpeg, and with other software (like open source audio tool "sox") read the audio samples.
To know if the wav is silent, simply read all audio samples and check that each sample is under the threshold of volument you want (or check if the pcm file is filled all with zeros if you only want to check an absolute silence).

To detect beeps... you need to apply a FFT (like MDCT) to samples of audio files, in the search of the frequency of that beep.

To detect human voice, you can search the result of FFT to find human voice frequencies, or if you want something more accurate, you can search the frequencies of formants of the language that you want to detect, or use CMU Sphinx (a open source speech recognition software).

> From: petedao at gmail.com
> Date: Sat, 16 Jul 2016 11:46:24 +0800
> To: ffmpeg-user at ffmpeg.org
> Subject: [FFmpeg-user] Detect if a wav file is silence, beep,	or live human talking
> 
> Hi
> 
> Does anyone know how I can use ffmpeg to determine if a wav file consists
> of :
> 
> - silence
> - beep
> - live human talking
> 
> If that can be done, is it also possible to determine if the live human is
> a recording or live?
> 
> Thanks,
> Pete
> _______________________________________________
> ffmpeg-user mailing list
> ffmpeg-user at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-user
> 
> To unsubscribe, visit link above, or email
> ffmpeg-user-request at ffmpeg.org with subject "unsubscribe".
 		 	   		  


More information about the ffmpeg-user mailing list