Ableton Forum

GREAT tip AJ, many thanks!

AdamJay wrote:

djadonis206 wrote:what do you mean make the result mono?

you will notice the "L" in the 2nd Utility plugin is highlighted.
this is taking just the LEFT of the signal and summing it mono.
so i've really already made the result mono in the 2nd Utility plug by default.

Holla back!

I threw the resulting loop into Recycle and and converted just the left channel to mono

Word - Peace ADAM!

How would you get JUST the bass sound to be left? Could it be done easily enough in an editor like Cooledit and just invert the stereo track???

Patch wrote:How would you get JUST the bass sound to be left? Could it be done easily enough in an editor like Cooledit and just invert the stereo track???

Patch, it doesnt work that way.
If the bass sound is in the mono field, it will be removed. It can not be isolated like the rest of the sounds. And typically speaking, bass is usually mono.

play with it though, you can get different results than whats in the tutorial .als Utility is the most simple effect in Live.

That's the utility trick?

That's one of the oldest remix tricks in the book... usually used to remove vocals. Take the left track and add the right track with the phase inverted. Take the right track and add the left track with the phase inverted. This has actually been a preset in sound forge's channel converter since I first owned it (back with version 4.5) And it's a common hack way to make karaoke CDs.

Now show me a way to easily extract only the stuff down the middle and then I'll be impressed.

brecht wrote:That's the utility trick?

That's one of the oldest remix tricks in the book... usually used to remove vocals. Take the left track and add the right track with the phase inverted. Take the right track and add the left track with the phase inverted. This has actually been a preset in sound forge's channel converter since I first owned it (back with version 4.5) And it's a common hack way to make karaoke CDs.

Now show me a way to easily extract only the stuff down the middle and then I'll be impressed.

I know your not out to be mean or a smart ass - I appreciate the knowledge Adam dropped

not everyone has a PC full of presets or are 'that' into Karaoke - !?!? Come on dawg, chill

Now show me a way to easily extract only the stuff down the middle and
then I'll be impressed.

There is a way to do it with utility.... Someone posted ages ago....

Believe it has something to do with 2 tracks....

Not my thing so can't be bothered working it out....

-Ben

you can actually get the middle on its own reasonably easily if you have somekind of modular DSP environment (MAX, reaktor , whatever)
[edit] yes you can use 2 Live tracks too [/edit]

Do the same thing that the utility trick does (or use the utility trick) to get the side audio - left from right, right from left, then mono it . To get the middle on it's own >> take the source audio and mono that - put it out of phase and add it to the utility audio.

so effectively you have

middle and side M&S (unprocessed)
and side S

M&S - S = M

I use minus here because adding an out of phase version is effectivley subtracting.

The sad thing is that you dont get 'just the vocals' in the Middle channel, you get a ton of stuff such as snare components, guitars .. etc. Not a usable acapella output.

I made a plugin for myself to do this (with a bit of in-train EQ) but it's pretty useless in most cases.

Having just spent several fruitless hours trying to extract the center, I've come to the conclusion that it can't be done, not by addition and subtraction anyway. So, to save you the bother...

Imagine a track with three sounds panned left, right and center (L,R and C).

The left channel contains L + C/2 and the right channel contains R + C/2

Adding these togther gives L+R+C.
Subtracting one from the other gives either L-R or R-L (The 200% width setting on Utility gives L-R on one channel and R-L on the other)

No matter how you combine these factors you can never get C on it's own
(L-R)+(R-L)=0
(L-R)-(R-L)=2(L-R)
(L+R+C)+(R-L)=2R+C
(L+R+C)-(R-L)=2L+C
etc.

all you ever come out with is the sum of the 2 channels, the difference between them, or one channel or the other on its own. or a multiple of one of those.

re the original tip, you don't need the second utility to create a mono signal, just invert one of the outputs on the first utility.

To extract L or R or C ONLY you'd need a third signal (equation) that is not a linear combination of the oter two.

you can extract the center info as I stated in my post .. just dont expect it to sound like some vocals on their own. It will just sound like a hell of lot of stuff!

Use adams method or similar to get just the 'side' signals

then subtract that from the original source .. you are left with the center. 'subtraction' in Live is done by reversing the phase of the entire track. read my post again

track 'm&s' - track's' = track 'm'

we already have track 's' via the utility trick, so invert that STEREO s output (nt mono-ified) and invert both sides and add it to the original.

it's not hard, but it doesn't sound good - even with EQ.

Sorry to dissagree with you Angstrom, but in your equation

M=L+C+R - ie everything added tigether to make mono.
S=L-R (or R-L) - subtracting left from right cancels the C signal which is equal in both (this is what the 200% width setting does)

so M-S is 2R+C (or 2L+C) ie. the left (or right) stereo channel * 2.


I think theAnimal has it right, you need some nonlinear element, but I'm buggered if I can work out how to do it.

no, M stands for middle , as in Center

side means L&R ie , things that are present in either left or right but not the center. Terminology from miking (middle and side)

I am correct, I can either link to an asd or a vsti which I made to prove as much.

think about it : in your terminology
we have an original track of L+R+C -

Aj's utility trick produces:
L+R - C or more plainly, the product "L+R "

so L+R+C (original track) - L+R (product from above actions) = ?

But Aj's trick gives L-R not L+R. One channel (let's say the right) is inverted. so subtracting the side signal from the summed mono can cancel out L but doubles R.

This isn't just theoretical musings, I have tried it in practive and this is the result I got.

Anyway I don't want to get into a flame war. I'd love it if you could prove me wrong, but I'm not convinced so far.

Middle & side miking is something different. There you're calculating two signals from two other signals. Here you're trying to calculate three signals from two other signals, i.e. to solve a system with 3 variables, but you only have 2 equations:

1L + 1/2C + 0R = left channel
0L + 1/2C + 1R = right channel

It's mathematically impossible to completely solve this. You can eliminate 1 variable but not more. Just my 0,02€, but I'm learning a lot from this thread.