Mid-side lop-side [solved, and how!]

Doesn’t seem like one for the main forum so hope one of you guys can help out.

For some reason I can’t fathom my output channel from my mid-side matrix is all lop-sided, as you can see from the screenshot (which isn’t as bad as it can get). I’ve used mid-side before, built a matrix before, but this is the first time I’ve done it from files recorded live with a Zoom H2N, taped to the singer’s mic stand, i.e. central to both him and the PA. (The thinking behind this was that the mid would pick up as much acoustic info as possible, leaving the side to pickup the PA and audience.) The Zoom gives you a stereo WAV with the mid/side as L/R and supplies a plugin to decode. The plugin is also lop-sided, so I split the files and built the matrix, with the same results.

I’m hoping that someone could look over my setup and point out if I’ve made a mistake anywhere.

Also if anyone’s had similar experiences, I’d like to hear about them too. Perhaps it’s just par for the course for a live setup and a consequence of the room shape. I’m just guessing here…

To centre things up I have to set the output channel panner to L10-R. Quite severe. And that leaves the applause between playing way over to the right.

All a bit undesirable, really. Hope someone can shed some light.


I guess it depends on what’s going on within the Zoom. A mid-side matrix (sorry if you already know this) should consist of three mono channels, not two stereo channels. However, if the Zoom is processing the signal, I would expect it to be output as just one stereo channel. Can you give me some more info about how the Zoom is outputting the audio?

For example, here’s how I do a mid-side matrix:

I did already know that but please don’t apologise, it’s better to explain too much than too little.

The mid-side side of the H2n has two built-in mics: a cardioid and figure-of-8, set up exactly as you would in the studio. The H2n stores everything in stereo files, so in MS-RAW mode mid/side -> L/R, which it then post-processes through a built-in internal matrix (which they release as a VST plugin too) so you can hear the file back in stereo and adjust the width through the line out. But the important point is that however you chose to monitor, you have an importable stereo file with separate mid and side channels available to with as you wish - build your own matrix or use the supplied VST. (For the sake of completeness I should add that what I’m describing is the H2n’s MS-RAW mode. You can store a ready-processed fixed-width stereo WAV but I don’t use it [edit: ha!] because you can’t do anything with it in the daw.)

Anyway, that’s how I understand it from the manual.

And it’s thanks to writing all that I thought I’d just check the manual and now I think I’ve sussed the problem: I’m a great big eejit! It seems that while ****ing about with the H2n the afternoon before the recording (because I haven’t used it that much) I accidentally set the damn thing to MS-STEREO mode. Never mind, at least it’s recorded but I’ve just learnt a hard lesson. Don’t **** about with stuff which was set just fine before you touched it. And never assume you did things the way you thought you did.

Madre mía! See sig…

I’m going to check this out now in Cubase but it explains so much…

Yep, that looks familiar, except I use an aux for the side-. Anyway, thanks but see previous post…

Are you absolutely certain you recorded in ms?

I would setup the recorder and do a test to see if I could reproduce it recording in ms and xy.

The only time I have lopsided results is when the mics are slightly out of alignment. That’s using 2 separate mics though. That shouldn’t be a problem for you. One thing you could try is offsetting the side a few samples at a time to see if you can get the image centered when summed.

Edit: just saw you recorded in a different format than thought. :sunglasses:

Yup. Must be getting more senile than I realised. At least it was recorded and in MS too, even thought it mixed itself down to stereo. Could’ve been a lot worse…

If I get noticeably better results running the matrix backwards, I’ll post back. It’ll provide an interesting postscript.

Voxengo msed is a good plug for decoding. :slight_smile:

Yes, thanks, I already have that and have used it to “listen in on” some classic tracks. Try it with Bohemian Rhapsody - very illuminating, never realised the piano was such a honky-tonk! So it was on the menu but nice to hear from someone else in the same vein. Do seem to remember it had quite a few options available. Pretty impressive for a freebie.

So I suppose the plan of attack would be to create MS versions of the stereo files through VoxMSED, print them then reimport and do the job as first intended.

Thanks for your help, bud,

This has turned out to be a better tip than I thought. I’d forgotten there was an Inline mode, which is making a beautiful job of reversing my earlier c-u.

For those of you who don’t know the plugin - it’s free and available in 64- and 32-bit versions - there are three modes.
Encode: l/r -> m/s
Decode: m/s -> l/r
Inline: l/r -> m/s -> l/r *

  • Essentially it allows you to rebalance the stereo width on the fly. Hello again, clarity!

Bit of an essential piece of equipment that I now understand the use of so much better than I did. Thanks, Woodcrest, you are the man!

A very :smiley: C