VCA compressor w dual detector and crest control

Hi all,

I've been thinking about a multiband comp w/ dual detector and crest control for a long time and have done some partial work on it.

It will once be used to complete my In-ear finalizer design somewhere else on this forum. The idea is tpo make a 3-band stereo comp that will sound as good as possible with as little controls as possible. Its time constants are chosen for use on stereo mixes, not for individual channels.

The VCA comp thread made me look at it again.

I'll throw it in as it is, far from complete or correct, so other people can look at the principles.

All reactions are very welcome!

Schematic:

http://users.pandora.be/Rogy/IEMFinalizer/Multiband%20comp/Comp%20w%20dual%20detect%20&%20crest.pdf

Some explanation and some maths (might be far from correct- check it out!)

http://users.pandora.be/Rogy/IEMFinalizer/Multiband%20comp/Multiband%20stereo%20compressor.doc

Best,


Rogy

Hi Rogy,

It's an interesting idea to work on.

If I understand it you are equating the fast rms value as your peak value and varying the thresholds of the two as your peak/rms ratio?

What do you think of the idea of doing it with just one detector and maybe stacking overlapping nlr circuits and varying their levels into the Ct pin?

Hi Crusty,


you understand it correct.

Your way of doing it should also be possible. But since I need a stereo compressor, I will use 2 4301's so I have two VCA's, six opamps and two detectors. That's why I used the second detector and an absolute value circuit.

To get the crest control working, I vary the DC voltage that feeds the Threshold pots. I need three dual linear threshold pots (low, mid and hi threshold) for the fastRMS detector and changing the DC voltage to them with an LM317/337 dual supply is the simplest and most accurate solution to having a Crest factor control which is equal for all bands.


Greetz,


Rogy

what do the crest control? that`s new for me.

[quote author="Rogy"]It will once be used to complete my In-ear finalizer design somewhere else on this forum.
.... not for individual channels.

All reactions are very welcome![/quote]
you have my attention as I too am in IEM mode

why no Peak detection at all ?
will you let the transmitters take care of that ?

I think some absolute peak clamping is worth while and to do it during the 3 band stage may make it more palatable.

I currently use the RNC in SN mode and I can see how a 3 band SN mode would be better.

Front panel indication of the 3 band GR might be necessary for the operator during a live show ...
UNLESS the whole thing is designed to be specific to IEM and only kick in during higher and very high levels ... not just as a effect.


Has anyone used the Junger D series
I saw one being used in an IEM system
http://www.junger-audio.com/


Has anyone use a Rane
PERSONAL MONITOR PROCESSOR ?
http://www.rane.com/mm42.html

Hi 12afael,

The crest factor of audio is the difference between the average and the peak level of the signal.

With the Crest control of the compressor, the difference in dB between the threshold of the peak detector and the threshold of the RMS detector can be set.
In this case: the peak detector's threshold can be fom 2 to 12 dB above the RMS detector's threshold.

At +2dB, the compressor will mainly be controlled by the peak detector, at +12dB the compressor will mainly be controlled by the RMS detector (of course all dependent of the type of signal hitting the compressor).

Kev,

The time constants of the FastRMS detector are very fast, and because the timing cap is so small the logging (or averaging) of the FastRMS detector is defeated. So although the component is still an RMS detector, it operates in Peak mode in this particular application.

About the metering, I completely agree with you: we need gain reduction per band, and I would even integrate a LED indicating whether the fastRMS or slowRMS triggers the sidechain at a given point.

Greetz,


Rogy

yeah ... coool

if you have a list of interest participants ... stick me on it

thanx Rogy.