The number of frames for the same audio differ between 8k system and 16 system.

Skip to first unread message

Jun 22, 2017, 2:52:55 AM6/22/17
to phnrec
Hi, all.

I have audios whose sample rate is 16k HZ, and feed into the EN system for phoneme recognition. I output the phoneme posteriors.
For CZ, HU and RU systems, I have to downsample the audios to 8k HZ. And also output the phoneme posteriors.  (Same frame numbers for the same audio among these three systems)

But, the frame numbers of the output for the same audio from the 8kHZ (CZ,HU and RU)  systems and the 16kHZ (EN) system differ.  Particularly, the former has 1 more frame than the latter.

What is wrong and could you please offer me some solutions?

Thanks in advance!

Petr Schwarz

Jun 22, 2017, 3:06:12 AM6/22/17



if you need the same number of frames I would just ignore the last frame for some of the systems. The last waveform frame is usually not full and I am not sure how it was implemented in phnrec. If it was zero padded or removed. The resampling algorithm may

have impact on the final number of frames.


Best regards,



You received this message because you are subscribed to the Google Groups "phnrec" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
For more options, visit

Jun 25, 2017, 8:50:38 PM6/25/17
to phnrec
Thanks, Petr.

在 2017年6月22日星期四 UTC+8下午3:06:12,Petr Schwarz写道:
Reply all
Reply to author
0 new messages