top of page

Progress Report 2

The compressor has been coming along nicely. We have been doing tests in Matlab and have successfully coded a non-realtime compressor with user set attack and release times, ratio, and threshold settings. One problem we have run into is doing stereo processing. In Matlab, we have only been able to get a mono compressor to work correctly, but doing stereo processing appears to be streamlined in the Juce library. This library has also allowed us to create prototype Audio Units that can be used in a DAW for realtime processing, so we are currently working on adapting our Matlab code to the C++ Juce AU strucure.

 

On the left are some sample signals run through our prototype non-realtime compressor with varied attack, release, ratio, and threshold settings.

 

In building the Audio Unit, to do real time processing we are doing buffer processing, i.e. dealing with small windows of the signal at a time. This directly relates to what we learned in class for the FFT. We are also planning on taking an impulse response of an analog compressor in the Audio Studio and using the FFT of this multiplied with the FFT of the incoming singal to result in a sonic coloration that is characteristic of the analog compressor. This implements both use of the FFT and convolution theorems into our project (and will hopefully make the compressor sound better).

 

So far, the coolest thing we have seen is the ability to actually make an AU plug-in show up in Abelton Live (a DAW) and process audio with it. Also, making nonfunctional compressors that distorted the signal in really interesting ways was pretty cool.

 

bottom of page