Speaker Diarization Output

[Rasipuram]

Hi

When we run this command for speaker diarization, it generates a png file. Does it generate any segment file with time frame and speaker id? If not, how do we generate it? pythonw audioAnalysis.py speakerDiarization -i data/pyAudioAnalysis/sampleAudioWav.wav

[andreaschandra]

any update on this?

[bb-work]

I would also like the answer to this question. Ideally, I'd like to split the input file into separate files containing each snippet of speech.

[manad-mishra]

Any update?

[scut-salmon]

Any update?

[tyiannak]

Yes u can do this by:

• using the sequence of flags (variable cls) to generate speaker segments (use flags2seg() function)
• then use scipy.io.wavfile.write() to write each segment returned by flags2seg to a separate wav file

[zhengqun]

I USE 2400 seconds dataset to train ,and the result is speech[98.2...] [24.168.....] WHY?

[SushmitaDey]

I tried flag2seg and generated segments ; but the code is assigning wrong speakers to the segments.Help me find out a path to solve the problem in speaker Diarization to identify correct speakers as fast as possible.

[khushi810]

What would be the value of the window in the function flags2segs(flags, window), while getting segments with its labels?

Thank you in advance :)

[vikram-ray]

I think flags2seg is changed to labels_to_segments.