N-Channel Compressor Ext. Detect question

Davids Question

Jun 27, 2017 00:00

Hello guys,

I have one project to use multi-channel signal and I need a limiter to limit the signal,

now the 4-channel signal input to the N-Channel compressor like the below picture:

my question is about the red dot which is use on the external detect signal and

which source should I choose (CH1 to CH4)?

One message from the wiki: https://wiki.analog.com/resources/tools-software/sigmastudio/toolbox/dynamicsprocessors/nchannel

"The block compares among the n channels and selects the one with the highest energy. In a 5-channel situation, the sound will be controlled and processed based on the loudest channel."

I am confused at "selects the one with the highest energy " it means if there are 4 channel input signal and one of them need to limit, the compressor will control it, but what if there are two channel signal need to limit? (one is the highest energy signal, and the other is over the limiter threshold but this signal energy is lower than the highest one, in this case I wish these two channel need to limit.)

Thanks for reading my question!

Regards,

Alvis.

Audio Signal Processors

n-channel

compressor

SigmaDSP Audio Processors

adau1701 sigmastudio

limiter

adau1701

0a07c29f619e4ce1724d149dd747ee56.png-494x323.png

Report

DaveThib Reply

Jun 27, 2017 00:00

Hello Alvis,

This is a very common way to do compression of multiple channels. For instance, with stereo signals, you get a loud sound on one channel but not the other. In that case you want both channels to reduce their gain by the same amount. This way the phantom center does not shift when one channel gets loud. So the loudest channels is the one that sets the amount of gain reduction but all the channels get reduced by the same amount.

So since this happens dynamically and quickly, you cannot switch from one to the other. What I suggest you do is to use a signal merger cell and add all four channels together. Actually, use a single control mixer so you can set the level of this summed signal and feed that into the sidechain input. Then the loudest one will be more dominate but the other channels will contribute a little to the level.

The other thing I noticed is that you have the delay line on the sidechain signal. This is usually not what you want to do. Usually you delay the audio and NOT the sidechain. This way the sidechain gets the signal first and integrates it and sends it to control the gain. This takes time and so a loud impulse will get through before it gets a chance to compress it. IF you delay the audio then the sidechain has time to lower the gain before the loud audio hits the compressor. So I suggest you delay each audio channel and not the sidechain.

Thanks,

Dave T

Report

Davids Reply

Jun 28, 2017 00:00

Hello Dave,

Thanks for your reply!

After I read your suggestion, I change the limiter and signal link way, like the below picture:

I take out the Delay buffer from the original link way(the red circle box), and I use the signal merger with a T connector

to the "Ext. Signal Detect", and the signal input to the limiter, the signal will input to the merger first then go into the splitter and split to the 4 channel limiter.

Anything should I notice, change or I misunderstand what you mean? Please give me some advice, very appreciated it!

Thanks!

Regards,

Alvis.