Why know about Mid Side Processing?
And this is a powerful concept.
If you’re interested in getting a vinyl version of your masterpiece out to the world, it will help you to understand why certain conditions have to be met for your mix.
Another one of the benefits of Mid Side Mastering is that it has the potential for remix opportunities, saving you time, cost, frustrations. (Also: ever try to open an old project and get told by your faithful DAW you can’t open this anymore?)
Before you get all excited and send me mixes you can’t open anymore, you need to know some limitations. This page shows what I can do and what I can’t do with this setup.
For any two channel stereo file:
To get the MID portion, we pan the left and right channels to the center. This is also called ‘summing to mono’.
To get the SIDE portion, we invert the polarity of one (and only one) of the channels first and then pan everything to the center. This is also called the ‘difference’ signal.
Now we effectively have two different ‘mono’ signals.
We also have two approaches with these two signals we can use to do mid side processing.
We take the MID portion we created above and route or pan it LEFT. We take the SIDE portion we created and route or pan it RIGHT. This step is ENCODING.
If you’re following along, you may sense that we have pretty convoluted and lopsided mix going on right now. Well - you’re right! But we’re not done...
Recall we now have the MID on the left, and the SIDE on the right. Like before, pan both to the center. This will result in the original LEFT channel and we route or pan this to the left (where else?!).
To get the original RIGHT channel, invert the polarity of one of the signals, then pan to the center. This will result in the original RIGHT channel and we route/pan it RIGHT. These last two steps are DECODING. We have our original mix back! (Note: there may be additional fixed level adjustments to be made to deal with various panning laws).
Confusing? Yeah! This routing can get pretty complicated in any DAW. Fortunately Voxengo’s free plugin simplifies this....
Above: Voxengo's MSED is set to ENCODE for mid side processing.
Below: another instance is set to DECODE.
...as does WaveLab’s.
Above WaveLab's STEREO TOOLS is set to ENCODE while,
below, another instance is set to DECODE.
Before I explain the whole point of this routing and panning madness, let's go over....
Review the SETUP: We’ve created a MID and a SIDE portion to give us two different files.
We take the MID portion and leave it as is (centered). We place that in the first track lane of a (new empty) DAW project.
Next we take the SIDE portion and invert the polarity of one channel. We place that in the second track lane of the DAW - lined up immediately and exactly below the first to be in perfect sync. This is important. It’ll be fun too - believe me!
You may ask won’t inverting the polarity on the SIDE just give me back the original file? No. Here’s why: Remember we panned to the center summing to mono. So we are inverting the polarity on a composite signal. The result will give us that out-of-phase ‘wash’.
When these two files are played back together - in perfect sync - we get the original stereo file back. Again, plugins like Voxengo make this easier.
This approach is pretty simple too. Aside from the ‘SETUP’, there’s no more I need to do. All complicated encoding and decoding is bypassed. It gives me much more control. I like to call this approach: MID SIDE PROCESSING IN PLACE.
It’s the approach I prefer.
In the First Approach I talked about encoding and decoding a file. But it’s what happens in between the encoding and decoding where the magic can happen. This is the mid side processing part. By placing a processor like EQ, or Compression, with independent left/right control in between, I can have new control over a mix not available with conventional processing techniques.
Below - signal flows from top to bottom.
The Waves Renaissance EQ is set for independent left and right control
which will separately control
the MID (phantom center)
and the SIDE signals separately.
For example, by placing an EQ and adjusting a band up in just the “left” channel, I will actually bring up the frequencies in the MID portion - the phantom center. It's possible to raise vocals post-mix by raising the frequencies of the vocals. But it will also bring up frequencies and overtones of any other instruments in that range. So you see the possibilities and limitations now...
My favored Second Approach offers more control. Even if a processor has no left/right controls, I can still process the mid and side separately. I can also do something like this:
Let’s say you send me a mix where the vocals, bass, and drums are nicely centered, but the mix still favors one channel. You wonder if I can balance things out a little more without upsetting the center. I can! With this ‘In-Place’ setup, I can pan the stereo element of your mix (the lower track). This evens out the left/right balance of the mix without disrupting the placement of centred instruments. This kind of control isn’t possible with the First Approach.
Even this can be accomplished with a plugin:
Now, I can’t pan individual instruments, or add independent delays, for example, but you get the idea.
For the creative thinkers reading this, this offers lots of “what if’s”.
There’s also lots of opportunities for destruction - really total destruction. Some engineers will totally avoid mid side processing. Some engineers understand the challenges, take steps to deal with the situation, and know what to listen for.
Knowledge - and Experience - is Power.
Proud to be affiliated with these organizations: