Problems with stereo versus mono

[Solomon]

I’m afraid there may be no good solution to this problem, but I figure I’ll toss it out there and see what everyone else’s experience has been.

I have a stereo-mono switch in my global rackspace which, when set to mono, replaces both the L and R outputs with a mono sum of both (plus a -4dB attenuation to compensate for the resulting loudness gain). Already, the -4dB is a compromise because the exact attenuation needed will depend on how wide the original stereo image is and on specific phase relationships.

However, speaking of phase relationships, now we get to the real issue: summing to mono causes a lot of phase cancellation, and some of my instruments (think acoustic pianos and Hammond) end up sounding very thin and hollow. Can anything be done about this?

One solution I’ve thought of would be to go through all my instruments and all my effects, and come up with both a stereo and a mono version of each. For example, for acoustic piano, maybe I would select a completely different plugin that uses monaural samples, and then select which one to use based on the state of my stereo-mono switch. For stereo effects such as tremolo, I could have two instances of the plugin, one with settings optimised for stereo and a second one with settings optimised for mono. However, this sounds terribly complicated in practice and would also require extra time for troubleshooting every time I add a new instrument or effect to make sure that the gain of both versions is approximately the same. Furthermore, if I’m recording the show, I’m going to want to record in stereo even when I’m sending mono output to the FOH, which means I’ll need both the stereo and mono versions of everything running simultaneously, which might double my CPU usage. I think this is the only possible solution that would produce the ideal result.

Another idea I’ve had would be to find some plugin that does automatic phase correction across the entire spectrum. For transients and noise I don’t think it matters, but for tonal signals, this plugin would need to identify all the sinusoidal content in the signal and rotate the phase of every individual sinusoid in so as to be in phase between the L and R signals. (Note that some sounds that have intentional stereo phase differences would presumably be harmed by this.) So far I haven’t found a plugin that does this. I’m tempted to try MAutoAlign (from Melda), which has a featured called “Spectral phase compensation”:

MAutoAlign performs something previously unimaginable - it actually fixes the phase for each frequency in all tracks, if you ask it to. During the analysis it now not only estimates the ideal delays and polarity, but also the ideal phase rotation to minimize cancellations. This means it can compensate for phase differences caused by 3D nature of the instruments being recorded, by the microphones and other recording equipment, and even processing you perform before MAutoAlign.

However, from what I can tell, MAutoAlign requires an “analysis” action before it actually starts doing the work, which suggests to me that it’s probably only meant for production, not live use. And I’m not sure what its CPU use and latency might be.

There’s an old thread in the forum that recommends bx_solo (from Brainwork), but from reading its description, all it does is basic mid-side processing, so I don’t think it does anything with phase alignment of individual frequencies.

A third solution might be just to discard the R channel and use only the L, which might work fine for some sounds but would definitely be problematic for others.

As I stated above, I’m not feeling optimistic about this, but I’d love to know what others have done. The main performance venue where I’m gigging these days does have a stereo PA, but I’m assuming that there will be lots of situtions where the PA is mono, and I want to be ready.

[pianopaul]

Are you sure that phase issues come from stereo to mono changing?
What happens when you do that conversion on your Audio Interface directly?

[LeeHarvey]

I understand what you mean. I also tried around and finally came to the conclusion that you wrote, among other things: I only use the right channel for mono.
I make my sound selection in this mode. My experience is that if a sound sounds good in mono, stereo is also good. Of course, you have to do without special effects that only work in stereo.

Unfortunately, I can’t contribute any more and I’m curious to hear how others solve the problem.

[themaartian]

As you suspected, MAutoAlign’s “spectral phase compensation” is not suitable for live use. I also wouldn’t expect MAutoAlign to be super helpful when fed a full mixdown stereo signal. It works well for bass, when blending a DI signal with a wet amp signal. It’s also really good for individual mic signals from drum kits (except, of course, the overheads).

You may have more success if you create your stereo mix taking certain issues into consideration. This article from Sonible provides some useful information.

Avoiding the Collapse: From Stereo to Mono (Compatibility)

[jlgrimes]

Mutility is a more direct way of doing it.

i dont really have any issues that way (or nothing i never noticed other than mono being mono)

[TheWWT]

I saw a video on YouTube this morning that discusses the kinds of problems that can occur when you go from stereo to mono. While it was discussed in the context of dance club music, the issue and the solutions extend to a lot of live performance scenarios where it’s extremely common to output in mono.

There is no free lunch here. You have to do the work to adjust each of your sounds independently to work in mono. There are tradeoffs that you cannot avoid making.

You might need to do extra work to create two different versions of a sound, one each for mono and stereo. If you are trying to make one sound work with both, you might need to give up some of the wideness you create with stereo effects—because certain effects are incompatible with mono—in favor of effects that work in both stereo and mono but that aren’t exactly what you want in either. Some tradeoffs will be easier to make than others, such as changing unison detuning from an even to an odd number of voices. Some may be more difficult, such as getting a reverb, chorus, or phaser that sounds OK in both stereo and mono.

The video I linked above has a lot of good guidance. I just found it this morning, like I said, and now have been listening to a lot of my sounds in mono and realizing that the root of some of my live performance issues is directly related to me using sounds that just don’t sound good in the mono PA setups I end up playing through.

Like the saying goes, you have to deal with the situation you have, not the one you want to have. Good luck to you!

[Solomon]

Thanks, that’s a great video which directly addresses all my questions (but doesn’t give any easy answers, as expected).

It gives a lot of food for thought though. I’m concluding that the best strategy is to design and test each sound for mono compatibility as I’m building my rackspaces rather than seeking to find some magic plugin that can rescue mono-incompatible sounds after the fact.

Another idea would be to have a mono-switch widget in the global rackspace which the local rackspaces can then use to make certain alternate choices (such as bypassing or replacing problematic stereo-widening plugins, replacing problematic wide instrument patches with a mono version, etc.)

It raises a question: does anyone have a favourite mono or mono-compatible acoustic piano or acoustic strings plugin? It seems to me that of a piano or a string section is sampled with a coincident pair of mics (e.g. X-Y or mid-side) then it should be mono-compatible and still give a pleasing (if not fully immersive) stereo image. So maybe a piano or strings plugin that gives alternate mic choices would be the way to go.

[pianopaul]

With Pianoteq you can build a mono Piano just using 1 microphone.