m/s encoder decoder recommendations

[jmh]

I want to play with m/s stereo I intend to record side on one channel and mid on the other of a stereo pair. It's easy enough to split into 3 tracks in studio, but does anyone have a favored plugin to do this?

[andy cross]

The free Voxengo MSED works for me.

[ivanoff]

+1

[TotalSonic]

If you need an automateable SAW native plugin for this then the Sonoris MS Codec is excellent and very affordable - https://www.sonorissoftware.com/product/ms-codec/
Otherwise, as mentioned above, the freeware MSED VST plugin from Voxengo works very well in SAWStudio as well - https://www.voxengo.com/product/msed/

Best regards,
Steve Berson

[John Ludlow]

I've heard of mid-side processing for many years, but I've never actually tried it (and I wasn't quite sure what it was or how it worked). Now, with a more clear idea of its potential and use, I want to try to use it to expand the soundstage of some ambient music I created that has never 'felt' quite expansive enough to me. I've downloaded and installed Voxengo MSED successfully.

I presume that one sure way of doing it would be to engage one copy of MSED to encode the stereo pair, make dynamic or frequency changes to mid/side, and save the result to a new 'stereo' file. Then, open the new file and engage MSED again on it - this time in decode mode, and write the changes to a new L/R stereo file.

As I envision it, this should work - but it wouldn't be ideal from a work-flow perspective. I might have to re-do the 'encode' process multiple times in order to get the right dynamic or frequency changes (which would happen post-encode but pre-decode). I might not know I didn't like something about it until it was 'decoded' back into actual stereo. I feel pretty sure that this should be able to be done in a single pass. What's the best way to do it?

I could open the 'encode' copy 'pre-FX', and the decode copy 'post-FX'. I could arrange the order of effects (if there even were any others) such that 'encode' happened before them and 'decode' happened after them. I at least think this would work. That way I could hear the post-decode result while I was adjusting dynamics, etc, in realtime, pre-decode.

Or - maybe there's a better way using channel routing or the effects bus? I haven't done much of that, but I'm open to learning if there's a reason that it's superior. If you've done it, what do you think is the best way, and why?

ALSO - one thing I still don't understand about M/S processing is: what happens to left/right differentiation after the 'encode' process, which renders all but the mid sounds as a single track? Since the 'S' track now contains all the sound that was not in common, how can the 'decode' process re-render what is exclusively far-left and far-right with 'side' input that is only one mono track? Are the left and right sounds encoded somehow other than just being 'not-mid'? Thanks!

[jmh]

My understanding of m/s is pointing a directional mic at the source (mid) and a figure 8 pattern perpendicular (side) to it it you are recording on mono tracks on playback you use the mid track panned in the center and the side you play back on a second track panned to one side then you playback the side on a third channel but you hit the phase button so it is the opposite polarity and pan it to the other side. If you lift the faders of tracks 2&3 by the same amount you should introduce more stereo.

If you record both mics as L&R of a stereo pair they are 'encoded' but do not playback as stereo. You can decode them easily in saw by just selecting R or L and as described above. The use of a plugin can tidy up the process within a single track - and I suspect (not having used the plugin yet) do some additional things.

You now have me wondering what kind of results you might get applying effects to the mid or side before decoding?

[John Ludlow]

Somehow the 'side' is rendered as a single, mono, channel when 'encoded'. Or, at least, everything I've read during the past few days resolves the not-mid portion to a single, mono, track - and the mid portion to a second track. But everything I've read starts with stereo. The portion of the signal that is the same on both left and right tracks becomes 'mid'. Everything else is 'side'.

By the way, I've tried it now, using Voxengo's plug-in, employing the second method I suggested might work. MSED has buttons you can push to turn off either or both of the mid and side. And, when I listen to them separately, the mid channel is the middle of the soundstage, and the side channel is the sides - as expected. But, I had also presumed that there would be a way to indicate how far to the left and right of center the 'middle' extended. I haven't found it yet, at least, in MSED. That's a little disappointing.

Oh - but, when the m and s are decoded back into stereo again - somehow there are distinct right and left sounds. This, in spite of having been stored together on the s track. Go figure. MSED has three modes of operation: encode, decode, and in-line. In-line seems to be 'pass through'. So, I tried 'encoding' on the pre-fader fx copy, and 'inline' on the post-fader fx copy to see what the encoded side signal sounds like. I had expected something that was not intended to be played aloud - but it wasn't. So, I don't know how they do it.

Also, my original idea of changing dynamics and/or tone of either or both of just mid and just side is thwarted in my setup - since there is no immediate way to modify only one channel of a stereo pair. So, some console wizardry (that has not occurred to me yet...) is going to be necessary to do the encoding - editing - and subsequent decoding all inline. Any ideas about this would be much appreciated!

MSED also gives you the option of raising the relative volume of just mid or just side. So, I was able to increase the apparent width of the soundstage just by turning a gain dial on the 'side' channel of the encoding instantiation of MSED.

Anyway, but - yeah, in theory you can do things like compress just one channel or the other. If you consider this a mastering tool, then maybe the vocalist, in the center, periodically gets too far from the mic (and it's too late to fix it in the mix or re-record it) so you compress just the mid channel and the voice develops clarity and consistency - without the guitar and horns becoming overbearing.

But, what I imagine I want to use it for is to widen the apparent soundstage itself by chopping out the sides, and operating on them without touching the mid. Ideally, at the very extreme edges there would be blurry high-end reverberative content that the ear wouldn't concentrate on, but would subliminally make everything feel wider. I've heard it in other people's music, but I wasn't sure how they managed it. I wouldn't want for it to compete with the middle - just add to the feeling of bigness. But - that's why I want a dial to define where the middle ends and the sides begin.

And - in the real world, no stereo image is only exactly in the middle or else only at one extreme side or the other. There's always variation across the stage. I think Voxengo must arbitrarily make that decision as to what goes on M and what goes on S. In my imagination - the perfect MS functionality would allow me to decide where the middle ends and the 'side' begins across the soundstage. MSED gets me part way there but I wonder whether other plug-ins have more control over that.

I can also imagine recording a single instrument in stereo, encoding it into m&s, and doing different reverb things to the side than I do to the mid so that the room can feel really 'live' without blurring the detail of the performance in the mid. All just ideas though.

[John Ludlow]

Oh - I see that there is a very clear explanation, in Steve Berson's Sonoris link (above), for how encoding and decoding is done in MS. I think I'm just out of luck for hoping for a 'side dial'.

[jmh]

Oh - I see that there is a very clear explanation, in Steve Berson's Sonoris link (above), for how encoding and decoding is done in MS. I think I'm just out of luck for hoping for a 'side dial'.

I was wondering why you wanted to encode then decode. There has got to be many compressors that you can patch after the encoder an un-link the L & R channels to make them operate mono/mono. I'll probably fire up my rig tomorrow and I'll see if any of the plugins I have operate in that mode (not that I have many).

[John Ludlow]

I was wondering why you wanted to encode then decode. There has got to be many compressors that you can patch after the encoder an un-link the L & R channels to make them operate mono/mono. I'll probably fire up my rig tomorrow and I'll see if any of the plugins I have operate in that mode (not that I have many).

I was imagining that if I widened the mid to an extreme, then all that would be left in the side channel was the extreme left and right. But, I would never get there the way it is done. There would be gradations of everything that wasn't precisely in the middle. So, if an instrument is panned somewhat right, the mid channel would contain as much gain as was common for both left and right, and the remaining gain would be assigned to the side. The result would be that the instrument sounded right of center - but not at extreme right. What I really want is two sets of sounds that are each unique themselves to one side. Whereas I don't want to disturb anything in the middle. I'm hoping to do it within the context of the existing piece rather than just slapping something new on the side - so it will feel as if it belongs. In the event that I could isolate them using MSED, then I could modify them by pulling a lot of the frequency spectrum out of them, running it through a reverb, and mixing it back in - resulting in filling the sound stage completely from l to r. I'll have to find a different way.

I will say, though, that just increasing the gain on the side channel has made it feel wider.