Beamforming in CUDA

Hi all, I am working on a wildlife conservation project where we want to be able to collect acoustic noises (animals mostly but human speech would be nice too) from a remotely piloted aircraft/glider (small uav with no noise when gliding) using a Jetson TK1. I have a separate post under embedded systems where we discussed creating a microphone array but I thought I would open up a thread for my next task and that is processing the audio on the TK1 using beamforming. By the way, my original post can be found here:

https://devtalk.nvidia.com/default/topic/805524/embedded-systems/jetson-tk1-microphone-array/

Would somebody be able to give me a hand in writing the application for processing the audio that I collect from my aircraft? Again this is work related and we are allowed to perform this function over the park system where this project is primarily taking place. I don’t have the first idea where to start as i’m not much of a programmer but other people told me that CUDA is what I want to use. What I want to be able to do is filter out the background noise that’s collected from the array. Any help would be appreciated! Thanks.

In my experience beamforming is an extremely good match for CUDA/GPUs (at least the beamforing algos I’ve worked on). So I think you’re on the right track here.

First time I hear about this “beam forming” ;). What is the noise like ? Perhaps provide a little wave file so we can listen to it ? Anyway if the noise is pretty much the same, you should give goldwave a try. It has a noise filtering option. Select a small part of the noise, copy it to the clipboard then apply noise filtering with clipboard option. See if that helps. (Goldwave probably uses fourrier transform series to convert sound into cosinus/sinus waves and then probably filters out the noise waves… works very well !