Problem in decimation

[ashley-b]

Hi, let me start of by saying am not a DSP expert.
But I have noticed scipy.signal.decimate gives poor results with large decimation values

But when I change

PyOctaveBand/PyOctaveBand.py

Line 54 in d9b36e6

sd = signal.decimate(x, factor[idx])

to

sd = signal.resample(x, (int)(len(x) / factor[idx]))

I see this improvement

I believe it is related to the note about not using decimate with values high than 13
https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.decimate.html#scipy.signal.decimate

[jmrplens]

Hi @ashley-b,

You are right, I had not seen that limitation.

Thank you very much for analyzing the results and coming up with the solution!
If you want, you can make a Pull Request or else I'll edit it directly mentioning you.

[ashley-b]

No problems, I created a PR #4