search

xanguera / BeamformIt

BeamformIt acoustic beamforming software
353 stars 111 forks source link

.del files format explained #28

Closed ChrisSpraaklab closed 1 year ago

ChrisSpraaklab commented 1 year ago

Hi @xanguera,

Thanks for BeamformIt, it works really well for my application. I do however have one question: could you explain the format of the .del files? What do the values mean? I am trying to use that data as a feature for diarization. Example of one row: 0 -> 0 1.000000 0 0.126987 4 0.125331 0 0.139378 0 0.175643 0 1.000000 -2 0.111881 -4 0.139887 0 1.000000 0 0.163803 0 0.157673 0 0.124941 2 0.124990 0 0.106618 0 0.120836 0 0.124682 There are 16 channels in my audio, and I see that reprented in this .del file. I get that the first number indicates a window of 250 ms. What I don't understand however are the values [0, 2, 4, -4] in front of every decimal value. I am guessing the decimal value is the delay info?

Thanks in advance.

xanguera commented 1 year ago

Hi, you are welcome. As per the documentation: "It contains a delay for each scroll segment and their GCC-PHAT values". This means that the integer value is the optimum time delay applied to each channel and the floating point value is the correlation strength with the reference channel (i.e. how similar that piece of audio is to the channel taken as reference).

yours,

Xavier

On Fri, 3 Nov 2023 at 11:39, ChrisSpraaklab @.***> wrote:

Hi @xanguera https://github.com/xanguera,

Thanks for BeamformIt, it works really well for my application. I do however have one question: could you explain the format of the .del files? What do the values mean. I am trying to use that data as a feature for diarization. Example of one row: 0 -> 0 1.000000 0 0.126987 4 0.125331 0 0.139378 0 0.175643 0 1.000000 -2 0.111881 -4 0.139887 0 1.000000 0 0.163803 0 0.157673 0 0.124941 2 0.124990 0 0.106618 0 0.120836 0 0.124682 There are 16 channels in my audio, and I see that reprented in this .del file. I get that the first number indicates a window of 250 ms. What I don't understand however are the values [0, 2, 4, -4] in front of every decimal value. I am guessing the decimal value is the delay info?

Thanks in advance.

— Reply to this email directly, view it on GitHub https://github.com/xanguera/BeamformIt/issues/28, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACITV7EO7QSIIEFCI5RNSDTYCTJX7AVCNFSM6AAAAAA64JP5TWVHI2DSMVQWIX3LMV43ASLTON2WKOZRHE3TMMBQGAZDGNY . You are receiving this because you were mentioned.Message ID: @.***>