How to mix super-loud source tracks?
May 29, 2011 9:57 PM
How best to assemble bastard-pop/mash-up style mixes from source material with heavily compressed dynamics?
I've been playing with mash-up style mixes in Reaper but working with modern, brickwalled source material (e.g., anything by Gaga) is a challenge—there isn't any headroom to lay anything else down on top of it. Just playing with the levels on the source tracks isn't enough to get clean output on the final mix—I either get clipping or I get mush.
Are there specific production techniques that'll help me get a reasonable result when mixing really loud source tracks?
I've been playing with mash-up style mixes in Reaper but working with modern, brickwalled source material (e.g., anything by Gaga) is a challenge—there isn't any headroom to lay anything else down on top of it. Just playing with the levels on the source tracks isn't enough to get clean output on the final mix—I either get clipping or I get mush.
Are there specific production techniques that'll help me get a reasonable result when mixing really loud source tracks?
I've never used Reaper, but if you are having a problem with clipping, then its not that the source material is brickwall-limited that's your problem, it's that you are doing something wrong in Reaper. Think about it: If the source material is bouncing off digital zero, and you reduce its level, boom, there's your new headroom. Inside the computer, its all just numbers, you can renormalize to whatever level you want. That doesn't mean it will sound good, but it won't clip. Or, thinking about it a different way, DJs mix these songs, right? Even though the amps in clubs only have X headroom, they still pull them into a mixer with other stuff and don't clip the amp.
Now, beyond the issue of straightup clipping, I'm going to guess that p3t3 is on the right track: there is probably just too much spectral information in those tracks. 'Mush' is often a symptom of too much music stepping on each other, rather than music just being too loud. When you have those wide synth stabs, pads everywhere, and huge kicks, you use up all the audio spectrum, so yeah, probably the first thing to try is EQ. DJ mixers have been equipped with kill switches since ancient times for exactly this reason. Here are some things to try:
-Carve out space for your material with EQ.
- Expander (sorry if this is too obvious but hey, just in case)
- Run the source mix through a compressor, but feed the comp's sidechain with a submix of your material, so that it ducks the original mix on the loud transients of your stuff. Obviously if you do this too extremely it will sound all pumpy, but a lot of released stuff sounds like this anyway these days.
- If you want to get fancy, combine these tricks: using a multiband compressor (apparently Reaper ships with one called 'ReaXComp'), you can tame sub-parts of the spectrum, and using EQ'd program to feed the sidechain, you can control the taming with sub-parts of the spectrum.
posted by jeb at 6:32 AM on May 30, 2011
Now, beyond the issue of straightup clipping, I'm going to guess that p3t3 is on the right track: there is probably just too much spectral information in those tracks. 'Mush' is often a symptom of too much music stepping on each other, rather than music just being too loud. When you have those wide synth stabs, pads everywhere, and huge kicks, you use up all the audio spectrum, so yeah, probably the first thing to try is EQ. DJ mixers have been equipped with kill switches since ancient times for exactly this reason. Here are some things to try:
-Carve out space for your material with EQ.
- Expander (sorry if this is too obvious but hey, just in case)
- Run the source mix through a compressor, but feed the comp's sidechain with a submix of your material, so that it ducks the original mix on the loud transients of your stuff. Obviously if you do this too extremely it will sound all pumpy, but a lot of released stuff sounds like this anyway these days.
- If you want to get fancy, combine these tricks: using a multiband compressor (apparently Reaper ships with one called 'ReaXComp'), you can tame sub-parts of the spectrum, and using EQ'd program to feed the sidechain, you can control the taming with sub-parts of the spectrum.
posted by jeb at 6:32 AM on May 30, 2011
You can combine an acappella with parts of the whole track as well if you can find one.
posted by Not Supplied at 7:24 AM on May 30, 2011
posted by Not Supplied at 7:24 AM on May 30, 2011
(I agree with p3t3)
In Reaper's list of JS plugins, look for the spectrum analyzer plugin. You can use this to look for frequency ranges that are particularly "hot", and EQ them out from one or more of the source tracks. This will help you cut down on a large part of the mush without too much effort.
posted by vanar sena at 3:13 AM on June 21, 2011
In Reaper's list of JS plugins, look for the spectrum analyzer plugin. You can use this to look for frequency ranges that are particularly "hot", and EQ them out from one or more of the source tracks. This will help you cut down on a large part of the mush without too much effort.
posted by vanar sena at 3:13 AM on June 21, 2011
This thread is closed to new comments.
If the stuff you're using is too compressed for that approach, you could try plugins for frequency isolation or noise removal (where you can define the "noise" via audio sample to use it as a more specific kind of EQ isolation). But those methods sometimes strip too much out.
posted by p3t3 at 2:58 AM on May 30, 2011