Why do two single unit compressors need to have matched pairs of attenuators for stereo?

[canidoit]

Just curious as to why stereo linked units need to have matched pair optical attenuators or even tubes for vari-mu, etc?

Doesn't one of the units take control of the behaviour of the other channel when it comes to compressing?

Thank you.

 

[David Kulka]

Most vari-mu compressors don't have a stereo link option.

But even in the case of a model that does, having matched components is important. Yes, one unit "takes control' of the other in stereo mode but that is only the side chain control signal and the problem is, what happens downstream? If the units are mismatched there still can be differences in threshold, ratio, attack / release times, and even frequency response, that may cause something panned to center to wander left or right when compression is occurring.

 

[JohnRoberts]

So the stereo image does not shift because L channel and R channel compression isn't the same.

JR

 

[Newmarket]

There is a single control signal(s) but if the two gain cells are not identical then channels are not compressed the same.

 

[canidoit]

So theoretically, the two compressors are independent. Just the knobs are what is linked.

Why don't they just make it so that one compressor tells the other compressor how to behave for stereo, wouldn't that make it much easier and remove the stereo matching issue or is it much difficult to do or not worth it?

 

[ruffrecords]

So theoretically, the two compressors are independent. Just the knobs are what is linked.

Why don't they just make it so that one compressor tells the other compressor how to behave for stereo, wouldn't that make it much easier and remove the stereo matching issue or is it much difficult to do or not worth it?

That is effectively what the stereo link function does - it basically ensures that both gain control elements get sent the same control voltage. But if they are not matched then one channel may reduce the gain more than the other which may result in a stereo shift.

Cheers

Ian

 

[gyraf]

..the "kink" in gridvoltage-vs-gain in remote-cutoff (vari-mu) tubes is non linear and varies wildly from tube to tube, making stereo tracking near impossible without careful selection.

This the reason why so few stereo varimu comps exist, and the reason for the lateral/vertical switch on the 670 (and the G22) which distributes channel differences to M/S in stead of L/R

 

[Newmarket]

So theoretically, the two compressors are independent. Just the knobs are what is linked.

Why don't they just make it so that one compressor tells the other compressor how to behave for stereo, wouldn't that make it much easier and remove the stereo matching issue or is it much difficult to do or not worth it?

That is how it is done. But the gain cells themselves are independent. The way to make them have the same response would require monitoring the output of the slave channel and correcting that wrt the master channel. Ie have the channel in a feedback loop. Very very challenging in analogue. Now if you go all Digital/Plugin then you can bypass all the uncertainties / intolerances and have the ideal response.

 

[Matt Syson]

There are many non subtle issues to be sorted when trying to compress or limit in stereo as the attenuating elements must match and track each other very well otherwise image shifting as mentioned. Also the sidechain DC should be the summation of BOTH AC parts of the sidechain after the rectification and threshold and ratio controls otherwise out of phase signals in the aC domain can cancel if you just sum the AC part before rectification and the ratio circuitry. Using it on M/S signals is a workaround but may introduce other issues.