DIGITAL FEEDBACK TO IMPROVE THE SOUND REPRODUCTION OF AN ELE

[lilltroll]

Version: 0.2
Status: Under development
License: Custom Licence
Download: https://github.com/lilltroll77/DCF

This project aims to improve the sound quality/reproduction from an loudspeaker by giving the power amplifier an output impedance which appears substantially greater than zero for low frequencies. The output resistance can appear to have both positive or negative (Like in -3 Ohm !!!), as well as the output reactance can appear to have both positive and negative (Like in -j*4 Ohm) that together would sum to - (3 + j*4) Ohm, or maybe more easy to understand |5| Ohm with the phase angle of -143 degress between Voltage and Current for a frequncy.
For further read:

Patent:

http://www.wipo.org/pctdb/en/wo.jsp?wo=2007073341
Or read it in european www.espacenet.com
Images:

[lilltroll]

The first block that is needed is a stable digital BP-filter in the region 15-100 Hz. That is used to add or remove virtual mass, compliance and mechanical damping to the loudspekaer.

[lilltroll]

So I have blocks for the feedback loop - now its time to work with the blocks for the forward section. I will need to resample the data here.

[lilltroll]

Working with a special softfloat format that can give a parametric filter with a SNR>= 100 dB in the feedback loop.

[lilltroll]

I have ordered 2 more "USB Audio" cards, and will send them to 2 guinea pigs in sweden to find out if they are able to reduce distortion as well.

[shawn]

Think this has potential for these guys. Protect your IP, they charge use for there's, to a point.
@100DB dynamic SNR you stand a great chance of success. I definitely think Audio tools,
help defrenchiate XCORES from allelse(ALICE). There is more information in Audio than Video.
So when i started seeing people apply xcore to audio I started paying serious attention. Everyone
wants big fast video, butt for the most part that postprocess. I am going to use a xcore in my leslie,
so that makes me a looper. Tie that into my TC2290, you see what I mean.

[lilltroll]

We tested to include the feedback code into the USB Audio 2.0 XMOS Code ver. 1.45.
I didn't believe that it would be possible to just add a 8:th thread and run in at full load @192 kHz without getting timing problems - but it went perfect.
Disabling S/PDIFgives another free thread. So it's actually some free amount of MIPS to add DSP code on the L1 chip.
Even I could follow and understand the ver. 1.45 of the code without the need of understanding USB2.0. I broke up the signal path at the I2S and added new channels and added the DSP stuff.

[lilltroll]

With the new USB API from TheSycon I desided to go Open Source, and push it to GIT.

First is a MATLAB GUI, where I try to use the API in the 1.22 driver

[lilltroll]

The GUI is working, the acess to theSycon API is working, but the enumeration of USB devices seen from the API doesn't work.
Have never done that part earlier.