Stereo Compressor Linking: Summed L+R detection or master/slave?
I am making a stereo compressor using GEN and have a question regarding the input to the level detection section. What is more common:
1. Have the L+R channels summed together before going into the level detection / gain computing section (i.e. the gain reduction for the two independent channels is determined by the sum of the two channels)?
2. Have a single channel 'master' (probably the left) go into the level detection and compute the gain reduction, then have the second channel as a 'slave' (i.e. the gain reduction on the second channel is determined solely by the first channel)?
3. I looked at the reference document for omx.comp~ and the stereo linking is explained as follows: "In stereo mode, the gain control signal is derived from whichever channel is loudest"
In hardware compressor units I'm familiar with, the second is usually the case, but I was wondering if it is the same for the standard stereo compressor you'd find in a DAW.
p.s.
Not a huge fan of the terms master/slave but seemed appropriate to use the terms as they are common for describing compressors.
they way omx does it more often used than i once thought, but why do you care about what others do?
i find this mode strange, and while it might be a working solution in a bus compressor on a typical stereo mix it will stop making sense when you want to use the thing as 2-channel effect - or of you want to use only one channel (while the R key input is still connected to somewhere)
it shouldnt be too hard to implement the "maximum" mode in addition to "mono" (key inputs mixed) and stereo/2-channel.
p.s.:
as if slavery would end from using other words for it.
Thank you for the reply, you make a good point about being able to implement all the different keying options.