Compress music harder with quiet vocals

[zydeceltico]

Hi All,

I thought I read a long time ago about someway to use parameter modulation and ReaComp to compress a band buss harder when the Lead vocal gets quieter but I can't seem to find that info anymore.

Thoughts and ideas?

Thx!

Tim

[Coachz]

I don't know how to do that but why not just automate the volume of the vocal and ride the fader to keep it right in the pocket where you want it ?

https://www.youtube.com/watch?v=Un719m_7pNc

[ashcat_lt]

Use the vocal signal to modulate the threshold of ratio or whatever you want. What exactly is the question?

[Coachz]

Yeah I think he would want the lead vocal to modulate the inverse of the compression so that as the vocal goes down the compression would go up if that's what he's trying to achieve.

[ashcat_lt]

Threshold would want to go the same direction as the vocal. Ratio would want to be inverted. I don't think either is a great idea. I'm also not sure the PM audio detector is going to give you a usable envelope. Worth a try. Maybe with a negative delay for some lookahead on the modulation source.

[tls11823]

Originally Posted by Coachz

...why not just automate the volume of the vocal and ride the fader to keep it right in the pocket where you want it ?

Agreed. Turning down everything else to get out of the way of the vocal would create a really unnatural sound.

Back when I listened to FM radio and its associated blather of advertisements, I recall hearing that sort of thing as they were talking about a great sale on mattresses or an upcoming monster truck rally. It would start out with loud music that would go up and down in volume as the announcer did his thing. Every time he paused for even a bit, the music would come blaring back for a second and then go back down.

Even before I knew anything about audio processing, I knew that this was a bad idea, as it distracted from the message.

So yeah, if the vocal's too quiet in places, bring it up rather than bringing everything else down to match it.

Now, if the music is actually too loud, then I'd recommend putting it in a folder or create a buss to control everything but the vocal, and then automate that.

[vdubreeze]

It would be easy to have a compressor respond in either direction to a vocal signal's presence in general. I have doubts that the actual level difference between two vocal sections (that 's a real triggerable difference and not mostly softer as in lack of punchy delivery and little in meter difference) could trigger this difference usefully. Much better to simply manually automate the parameters as needed.

[zydeceltico]

Originally Posted by tls11823

Agreed. Turning down everything else to get out of the way of the vocal would create a really unnatural sound.

Back when I listened to FM radio and its associated blather of advertisements, I recall hearing that sort of thing as they were talking about a great sale on mattresses or an upcoming monster truck rally. It would start out with loud music that would go up and down in volume as the announcer did his thing. Every time he paused for even a bit, the music would come blaring back for a second and then go back down.

Even before I knew anything about audio processing, I knew that this was a bad idea, as it distracted from the message.

So yeah, if the vocal's too quiet in places, bring it up rather than bringing everything else down to match it.

Now, if the music is actually too loud, then I'd recommend putting it in a folder or create a buss to control everything but the vocal, and then automate that.

Hey Everybody - It's just a single very dynamic lead vocal on top of two finger picked acoustic guitars and three tracks of Session Strings Pro. The guitars and strings are really rich and in one spot the vocal dives from rather big to almost a whisper for effect.

I have Waves Vocal Rider - and KINDA like the effect of lifting and lowering the vocal level - much the same as automating - however I'm not crazy about the dynamic so I'm not using it this go-round.

What finally felt right and good for me was side-chaining ReaEQ on a single band at 1kHz. ReaEQ on a folder containing all the strings as they are more prominent than the guitars at the particular spot in question. Put a new send from Lead Vocal to ReaEQ. Then I used parameter modulation to drive down that 1kHz band everytime the Vox hit. I like the way it breathes more. Doesn't sound as flat to me like automating vox levels - which I do sometimes - just not this time.

[zydeceltico]

Originally Posted by vdubreeze

It would be easy to have a compressor respond in either direction to a vocal signal's presence in general. I have doubts that the actual level difference between two vocal sections (that 's a real triggerable difference and not mostly softer as in lack of punchy delivery and little in meter difference) could trigger this difference usefully. Much better to simply manually automate the parameters as needed.

I would still like to know if anybody knows how to make ReaComp work "backwards" so that the louder the vocal signal the less it compresses and the quieter the vocal signal the more it compresses.

Ideas?

[zydeceltico]

Originally Posted by Coachz

Yeah I think he would want the lead vocal to modulate the inverse of the compression so that as the vocal goes down the compression would go up if that's what he's trying to achieve.

yep. exactly.

[Tod]

Originally Posted by zydeceltico

I would still like to know if anybody knows how to make ReaComp work "backwards" so that the louder the vocal signal the less it compresses and the quieter the vocal signal the more it compresses.

Ideas?

Hi zydeceltico, can you explain why you would want to do that?

[zydeceltico]

Originally Posted by Tod

Hi zydeceltico, can you explain why you would want to do that?

Just because? LOL - - seriously though - dynamics. I do a lot of acoustic-driven singer/songwriter stuff. Acoustic guitars and vocals. Lots of level-dynamics in the vocals.

Occasionally (for years) I have come across conditions where it would just be more elegant to have a "smart" compressor that knows the vocal is "loud" so it doesn't need to compress "the band" as much during those sections as when I - vocally - get to quieter parts.

I have done this with automation, multing, Vocal Rider-type plugins, etc. ...and have been happy with the results 95% of the time - - - but every once in awhile I wish I could figure out how to do it "the other way."

And I KNOW I have read a Reaper article somewhere years ago of a way to do this with parameter modulation but now I can't seem to find that reference anymore.

It's strictly for a side-chaining condition.

[ashcat_lt]

Originally Posted by zydeceltico

I would still like to know if anybody knows how to make ReaComp work "backwards" so that the louder the vocal signal the less it compresses and the quieter the vocal signal the more it compresses.

Ideas?

I answered that question. Modulate the threshold based on the vocal. Do you have questions about PM in general or routing or what?

[zydeceltico]

Originally Posted by ashcat_lt

I answered that question. Modulate the threshold based on the vocal. Do you have questions about PM in general or routing or what?

Hey Ash -

I must've missed your response earlier.

Cool - yeah - that makes sense. I'm guessing I would want to set "direction" to negative? I'm not at the console right now so this is all from memory.

But yeah - will setting direction to negative get the threshold to move in the opposite way? Re-reading your original post. Can you explain "negative delay?"

And how to "invert the ratio?"

Thanks

Tim

[ashcat_lt]

Originally Posted by zydeceltico

But yeah - will setting direction to negative get the threshold to move in the opposite way?

If you're automating threshold, then no. You want that to go down when the vocal level goes down so that it digs in to the music program more as the vocal gets quieter.

Can you explain "negative delay?"

JS Time Adjustment set to a negative value so that the detector sees the change in vocal level coming, starts its attack routine, and hopefully gets where it needs to be by the time vocal actually comes in.

And how to "invert the ratio?"

Set direction to negative here so that ratio gets bigger with smaller vocal inputs. IF you choose to modulate the ratio.

BUT are there actually vocals pretty much constantly through the piece? What do you suppose happens during the pauses between lines, or any instrumental breaks in the song? The vocals are REALLY quiet then.

[vdubreeze]

Having tried a bunch of versions of this kind of thing, my experience is that you can only make the music change very slightly before it sounds "desky". Splitting the movement into groups gives a little more before it does, but I've never found it to work by itself it's the vocal that is inconsistent. Always had to make up the difference getting to the desired balance by adjusting the vocal itself as well, and then decided what's the point of the music verging on sounding like it's slightly going up and down if I have to hit the voc track anyway. Lots of success getting guitars or whatever to duck a db or two, at one frequency or full, when the vocal was present, but not to compensate for the vocal, just to drop the guitars, or whatever, a hair to be less close to the vocal. If there's a problem track one would generally need to have *that* track affected by another track acting as key. Not sure if having the problem track as key to other tracks will yield enough fix.

Originally Posted by zydeceltico

I would still like to know if anybody knows how to make ReaComp work "backwards" so that the louder the vocal signal the less it compresses and the quieter the vocal signal the more it compresses.

Ideas?

Even though any parameter can be triggered negatively, I find it much easier to PM the fx from full dry to full wet than to make the threshold or ratio fine tuned to where it's perfect during PM changing from the main setting and back. I'll make it go from dry to wet and then adjust the compressor manually to get the setting to do what it's supposed to when wet. . And instead of having it go from the original setting to a second setting, have the original, normal normal compressor setting handled by another instance of a compressor that doesn't get automated. But that's just me : )