Hey this fabulous description (and it really is fabulous thanks for the effort to write it!) prompts a technical question from me for the developers:
I’m assuming this is the case, but when GP combines a stereo (L,R) pair to mono, is it truly doing a mono mixdown (this is where I’m assuming yes) ie
((L+R) + (L-R))
or is it “just” summing L+R?
…As I said I’m guessing the GP internal plugins employ correct mono mixing?
Thank you @kevinf, I’m glad it’s proving helpful! As for what happens when two wires are connected to one input, I did search this forum before I started my new layout, and I found these pre-GP4 posts. They seem to indicate that “summing” is the answer, not a true mono mixdown, but I’d love to hear an update from those with more intimate knowledge of GP4.
I decided to just go with the summing, using a Gain Control to compensate for the increased volume, and to my ears it sounds fine – what am I missing compared to a correct mono mix? Also my wiring varies - in some places it’s summing L and R signals into one mono input, but in other places it’s actually summing 2 or 3 Left outputs to one input and 2 or 3 Right outputs to another input. I didn’t want to clutter things up even further with instances of mono-maker plugins.
@dhj @npudar I wonder if you could clarify this for us. Does the summing approach (feeding two or more signals into an input) give the same effect as running those signals through a 4-channel mixer block at the same volume levels? When I sum two signals into a Gain Control, like in my global rackspace wiring diagram above, I set the gain to -6 dB to prevent overloading the input. But I don’t actually know if that works - maybe the -6 might only affects the output of the Gain Control? And then there’s @kevinf’s question about “correct mono mixing”.
Why don’t you just attach a meter widget to the output to make sure you’re not overloading, etc.
Also, all those plugins marked with (mono) should have their bus layout changed to be mono.
@dhj I do have those meters - see my global widgets diagram above - but my question is about whether the inputs could be getting overloaded even though I’ve attenuated the gain. In other words, does the gain control affect the input level or the output level?
Gain generally refers to output. Attenuation refers to input
Okay, that sounds right, so when I say I’m “attenuating” the Gain Control level, what I should be saying is that I’m decreasing the gain (on the output). So how would I know if the summed inputs to the Gain Control are overdriving it? The meter is only showing the output level.
Stick another gain control in between, set it to 0dB and attach a meter to it —
Great idea, thanks. So, as far as you know, is summing two feeds into one Gain Control input going to give me the same result as running those feeds into an Audio Mixer?
An audio mixer is just a multi channel version of a gain control. It was added in a newer version of GP
Very good to know! So let me rephrase my question: is summing two feeds into one Audio Mixer input going to give me the same result as running them into two separate inputs of the Mixer? (sorry for all the questions, I’m just trying to avoid audio degradation)
If you configure input to be mono and connect two wires, they are just summed.
Letting your ears decide what sounds right is probably the best approach anyway 
I did end up doing a listening test. I summed the L channel of a stereo loop into 9 daisy-chained Gain controls at +6 dB each, and from there into the left input of a Mixer, with the channel output gain at -49 dB, which made its resulting level match that of the loop’s R channel which went directly into the right input of the Mixer with 0 dB gain. I patched the Mixer’s outputs to the L and R audio outs, and listened to it in headphones. I was actually shocked to hear no difference whatsoever, between the two channels. I expected the L output to be completely distorted, since the level meters in most of the 9 Gain controls were pinned, but no, the audio was pristine. I have no idea how GP4 manages this!
If I may offer… GP only deals in digital not analog audio which cannot multiply distortion, only pass along any it receives.
It’s helpful to remember the nature of the beast… once you get past the A/D converter or the (kindasorta) Signal generator software (VST etc)… there’s no more audio. It’s just 1s and 0s. It isn’t even audio it’s just an incredible simulation.
Clipping is the only type of distortion inimically possible in digital audio because you get undefined values on analog overdrive (ie noise). “Distortion” on analog to digital conversion is utterly worthless unless you want that A/D converter as a noise/distortion sound.
On output you can have noise result occur at the D/A converter in the same way, but technically that’s actually analog noise in the sense that it’s unintended.
Having no specific knowledge, (perhaps devs can answer this?) I would not be surprised at all if Gig Performer uses 32 or even 64-bit audio processing internally which provides a LOT of mathematical “audio headroom” far more than you could ever run out of from 49dB of increase.
My original observation intended something about summing a stereo signal Vs creating a technically correct mono. All these issues arose many years ago with the introduction of stereo vinyl carving, and further with the introduction of FM stereo (and later the very technically weird AM stereo) and even later TV broadcasting (which was FM), creating radios and phono players that would be compatible with stereo source material, and the reality of consumer electronics manufacturing design component budgets.
Tl;dr theres a way to sum the two halves of an analog stereo signal to produce a mono signal that is guaranteed to never have any possible frequency (via phase) components doubled as as a result of that summing.
It’s almost trivial to do in purely analog electronics. It gets trickier when all you have to work with is 1s and 0s pretending to be analog electronics.
Mathematically the expression is:
==
Edit: it’s actually a little more complex.
(L+R) + (-(L-R))
Sum plus (inverted phase) difference
Sum = combine both channels
Difference = invert the phase of either channel then sum those
Note the L-R difference when summed with the L+R sum, must also have its phase inverted.
This is one of those things where once it’s done it’s done. Also you won’t really hear any difference at all if the L and R channels are completely different signals. You’ll only notice a difference when comparing the final mono compatible mix with a straight sum of the L and R — AND — the stereo source was a typical modern stereo mix. The straight sum typically sounds “heavier” or “bassy” compared to the mono compatible downmix because the signal components that differ from L and R are typically not mixed 100% L and 100% R as they were occasionally in early 1950s and 1960s recordings, and also they tend to be high frequency components.
If you then mix these sum and difference channels they produce a true mono signal. So what?
The mono signal will have all the benefits the vinyl and broadcast engineers of yore sought. In the old days this unbalanced signal buildup in a summed stereo signal could unexpectedly blow out transmitter tubes in a 100,000 W transmitter. See the effect of Kim Karnes “Bette Davis Eyes” on the entire FM broadcast business for the rather astonishing example of a song that suddenly affected the basis of broadcast chains all over the world.
True mono from ((L+R)+(L-R)) will be always guaranteed sonically balanced with regard to subjective perception of relative relationships of frequency components.
That’s important in a high RF amplification scenario like commercial broadcast. But it may be important in a typical stage, club or PA setup too, mainly because this “flatter” or “less punchy” (vs straight summed) signal may be easier to eq than a straight summed L+R signal. And to make it even more complicated in practice that’s going to vary with the stereo source material. So without a true mono you could dial in a satisfactory result with one sound (song, mix, patch… whatever) and you just change patches and all of a sudden you’re feeling like something went out of control again.
I guess it isn’t Gig Performer’s job to do any of this. For a stage setup it might make the most sense to locate a rack unit from the broadcast business that produces a true analog mono from stereo and just put that in front of your amplification after the final mix.
It’s not likely to be a light breaking through the clouds moment of change amping mono vs stereo sum, but I suppose that’s possible depending on the stereo mix. Most likely you’ll find that it makes eq contouring more predictable and expectable. EQ changes to amplification will seem to have a more linear response. And depending on stereo synth patches especially you may notice they sound more broad spectrum in general, especially if they have been designed specifically to translate well for mono compatibility.
Edit; that will probably translate to allowing you to control amplification headroom much more precisely just with a level change rather than needing both a level and eq change to prevent clipping as is more likely with a sum. Also this means during a set the tech is less likely to need to make changes per song or solo etc. true mono will tend to make the whole thing more controllable in an operation sense. Essentially you’ll get best performance from mono amplification by giving it a true mono (not sum) signal to amplify. But this actually opens even more cans of worms because it also goes to the nature of the stereo source material (designed for mono compatibility). Augggh! Dopamine hungry brains!…
Apologies for the brain dump I kind of opened a can of worms I suppose but I wanted to offer something back for your diligent interest in something kind of arcane and ‘boring’ that may actually have an important applicability for live performance.