acoustic_comp: the frame number of data stream lf0 is not consistent with others

[PavanIISc]

I am getting this error. assert out_frame_number == frame_number AssertionError

(I am trying this system for Indian language)

Why am I getting this error. Can you please tell me what should I to change?

[RasmusD]

As the error in your output says: acoustic_comp: the frame number of data stream lf0 is not consistent with others: current 2783 others 1393

So you have a mistmatched number of frames on the file.

[dreamk73]

Did you use the WORLD vocoder to extract the lf0 values? How did you create the label_state_align files? The number of frames in each have to match. In WORLD lf0 values are computed every 5ms.

[ronanki]

Make sure the number of frames in each of lf0, bap and mgc are same. Use "x2x" in SPTK to find out the number of frames.

./tools/bin/SPTK-3.9/x2x +fa lf0/cmu_us_arctic_slt_text_01001.lf0 | wc -l ./tools/bin/SPTK-3.9/x2x +fa bap/cmu_us_arctic_slt_text_01001.bap | wc -l (assuming 16khz) ./tools/bin/SPTK-3.9/x2x +fa mgc/cmu_us_arctic_slt_text_01001.mgc | xargs -n60 | wc -l

[PavanIISc]

@dreamk73 I have used WORLD vocoder "extract_features_for_merlin.sh" script to extract lf0, bap and mgc.

[PavanIISc]

@ronanki I have checked the number of frames. They are mismatching. How to extract lf0, bap and mgc? which script should I use?

[ronanki]

This is the script you should use: https://github.com/CSTR-Edinburgh/merlin/blob/master/misc/scripts/vocoder/world/extract_features_for_merlin.sh

Also set sampling frequency to either 16000Hz or 48000Hz w.r.t the data you are using, as the default value is 16000Hz: https://github.com/CSTR-Edinburgh/merlin/blob/master/misc/scripts/vocoder/world/extract_features_for_merlin.sh#L31

I hope this solves your problem.

[abhigarg]

Also the wav files must be of single channel i.e. mono waves

[nasmsh]

@ronanki Hello there. I've already converted my files to mono waves and sample rate of 16kHz.

I've used this script as I'm using the world_v2 vocoder: merlin/misc/scripts/vocoder/world_v2/copy_synthesis.sh

But the frames are still different. The lf0 and mgc has no problem. Only the bap frame is not the same. Can I know how to correct this, please?

[anushaprakash90]

@nasmsh I am also facing the same problem of mismatched frames, despite using the correct scripts. How did you solve this?