3

I'm developing a class library, that allows to do some manipulations with streamed audio (it is working with buffered samples, retrieved from recording device - so basically I've got an bytes array) for researsh purposes.

The problem is that i need to add noise reduction - there is need to capture noise profile, and apply noise reduction to each sample before applying sample processing & analysis, and I was not able to find any algorithms/samples/libraries that could be applied.

Any help would be appreciated. Thanks in advance.

danyloid
  • 1,677
  • 3
  • 21
  • 47
  • 1
    its a fairly broad topic i think - what is your application of noise reduction? speech rec? – Aaron Anodide Oct 31 '11 at 13:35
  • @Gabriel music regognition stuff - at the current point it is finding a corresponding musical note & freequency shift. – danyloid Oct 31 '11 at 13:45
  • that's cool - i used to work on speech rec - one very simple approach would be to use an energy based silence detector and then basically subtract it from the rest, taking it as a constant background noise... is the pitch detection done with zero crossing or are you doing FTs? (apology in advance i was more of a programmer than a scientist on this stuff but i found it really fun to work on) – Aaron Anodide Oct 31 '11 at 13:50
  • @Gabriel thanks - I should probably try the energy based approach. seems it may be applicable in the case. // I've used the fft - i had a little experince with that, so it was more comfortable for me. – danyloid Nov 02 '11 at 12:51
  • You might want to check out [dsp.se], but in general list/"bikeshed"/shopping questions of this type are off topic on StackExchange. –  Jan 06 '12 at 14:47

0 Answers0