DarwinNE/BK2034_GPIB

Increase the resolution of the 1/3 octave analysis for low frequency.

Closed this issue · 1 comments

Currently, the 1/3 of octave analysis for very low frequency done by octave.c is affected by a huge error, since only one spectrum is acquired between 0 Hz and 6.3 kHz. It would be better to acquire the spectrum twice, one below 800 Hz to have a good representation for low frequencies, and one up to 25kHz, to have information on all the audio band.
The acquisition should be done automatically and data gathered during the two passes should be blended together.

Done! Maybe we should still see wether a correction has to be applied when the number of FFT bins in a computed interval is very small.