need help understanding some mid/side compression questions

[sweetjesus]

Originally Posted by sweetjesus

Hi there,

I have an API2500 and ran into this tidbit of advice on the API tips and tricks thread:


"another trick i learned about (from mcnair, i think) was running it in an M/S matrix, but with the stereo set to 100%.

at that point, it should be the same as just running it stereo, but man, sounds TOTALLY GREAT."

I'm trying to figure out exactly how to do this using Logic and a M/S encode/decoder tool. Am I meant to use the M/S tool to send the mid signal to the compressor then bring it back to the computer and add it back to the Sides?

If anyone has the time to break it down into simpleton speak in a step by step guide, I'd be very appreciative.

So I had a bit of a think about what this may entail.

I'm guessing that by linking the L/R at 100% its to avoid any stereo action happening, so therefore the user must be talking about affecting the mids and using the 100% L/R function to avoid adding any additional side information.

So basically compress the mid signal, leave sides alone.. Am I right?

[mml]

By doing this you would be triggering the compressor with the mid information but affecting everything, so no weirdness with the stereo field. Split to m/s in daw, send out to compressor where mid is left, right is side, send back in and decode to monitor the processed track. Give yourself plenty of headroom going to the compressor because the mid channel will be pretty hot. Both elements get sent to the compressor, not just the mid.

[sweetjesus]

Thanks heaps! Now that makes loads of sense.. I had no idea that a M/S signal was just two mono signals..

You have raised my curiosity however, how the heck does a mono signal for the sides deal with the difference in level between something that's panned a little further to one side or the other..



thanks for explaining how to use the m/s compression !

Originally Posted by mml

By doing this you would be triggering the compressor with the mid information but affecting everything, so no weirdness with the stereo field. Split to m/s in daw, send out to compressor where mid is left, right is side, send back in and decode to monitor the processed track. Give yourself plenty of headroom going to the compressor because the mid channel will be pretty hot. Both elements get sent to the compressor, not just the mid.

[sweetjesus]

okay so i tried this trick and it really didn't do anything for my stereo image.

i got a lot more enhancement by just bumping the sides up half a db.

[mml]

If you wanted stereo enhancement, you want to just compress the middle, which will have the peaks anyway. In software, you pick a dual mono plug, unlink the sides, and compress only the left. Then change the output gain separately to taste. The method you described earlier will just keep your image stable but is designed to NOT alter the spread.