AEC card input processing

scott180

Working on my first AEC card system.

Normally I put parametric EQ and compression as input processing on a mic. With the AEC card having auto gain, is compression really necessary?

What would recommended input processing be? If I put a paramentric on a mic off the dry side, I wouldn't need it on the AEC side correct? In the help file, it stated that it's better to have room/speaker EQ before the reference, and thus two sets of processing blocks are needed, one for the Far Side audio as it is mixed with any near side reenforcement.

Let me know if I'm making sense.

Dan Lynch

This is what I do. I assume that if a mic needs eq, limiting, etc. to make it sound ok in one location, it'll need the same thing to make it sound ok
in the other location too so I use stereo EQs and crossovers on the mic inputs to process both the dry and AEC inputs at the same time. The
crossovers give hpf, lpf, mute, and limiting all in one object.

You want the reference signal for the AEC to come from as close to the analog outputs as possible so that the reference will track any changes
you make to that signal. That's another case where stereo processing objects make it easy to process the local mix and the farside audio in the
same way without mixing them together first.



Here's the file if you want a closer look at anything. AECdemo.zip

Dan

scott180

So you're not using compressors as part of the input chain, you're using the auto gain instead? Or just some limiting in the the crossover block?

Thanks for the help. It's much more elegant than my setup.

Scott

Dan Lynch

I don't use compressors on speech mics unless a consultant or designer forces me to. It's some dude talking, not a rock and roll snare drum. Using a compressor on a speech mic is usually just the habitual use of the wrong tool for the job. If I want to prevent whacky peaks, I'll use a limiter. If I want to control the average volume, I'll use AGC followed by a limiter. Using AGC without a limiter is just begging for disaster. That's true of any processing which dynamically increases the gain of a signal.

Dan