Handling of vocals in MPT. Anyone?

Started by Harbinger, April 12, 2008, 03:03:49

Previous topic - Next topic

Harbinger

For the first time in my long musical career, i attempted to mix spoken vocals in with a track, and was unable to bring out the syllables so that a listener could hear what was clearly being said. Now, i understand that sung vocals are different from the spoken word, because when we sing we tend to use nearly the same volume for phrases of syllables. So mixing them into musical tracks is not very complicated. But spoken vocals keep many of the oral nuances that music can easily drown out.

Anyone have the knowledge of handling both kinds of lyrics and spoken word when mixing with music? Do we need certain types of VSTs or preprocessing? Compressors? Limiters?

residentgrey

No two people are not on fire...AWWW!

[img]http://www.taralax.com/assets/gfx/tsbanner_xpmono.png[/img] Web and Graphic Design just for you!
I r GhostMech on there, forever scouting.

Saga Musix

I think we already have some nice threads here about that. Try the forum's search function. We don't have to say everything again and again... :lol:
» No support, bug reports, feature requests via private messages - they will not be answered. Use the forums and the issue tracker so that everyone can benefit from your post.

seventhson

That's where compression, limiting and equalization come into play.
Compression and or limiting has the effect of making the quieter sounds jump out more, so when done properly all the different words and syllables should be equally loud.
Use eq to get rid of unwanted frequencies so the vocal doesn't muddy up the mix and make it fit better.

LPChip

Quote from: "seventhson"That's where compression, limiting and equalization come into play.
Compression and or limiting has the effect of making the quieter sounds jump out more, so when done properly all the different words and syllables should be equally loud.
Use eq to get rid of unwanted frequencies so the vocal doesn't muddy up the mix and make it fit better.

Actually, thats not entirelly true.

If you compress sound, you'll destroy the dynamics even more. I just saw a video (see other post here on the forum) about compression and how it can destroy your vocals.

You're better of lowering all other sounds a bit in volume so you can normally raise the voice. Maybe a slight compression can be added, but don't overdo it, or it will only enhance the wrong frequencies making it muddy.
"Heh, maybe I should've joined the compo only because it would've meant I wouldn't have had to worry about a damn EQ or compressor for a change. " - Atlantis
"yes.. I think in this case it was wishful thinking: MPT is makng my life hard so it must be wrong" - Rewbs

älskling

I think the keywords are "when done properly" :P

Harbinger

That's very good info, LP. I'll hafta try it. Now for that video....

seventhson

Quote from: "LPChip"

Actually, thats not entirelly true.

If you compress sound, you'll destroy the dynamics even more. I just saw a video (see other post here on the forum) about compression and how it can destroy your vocals.

You're better of lowering all other sounds a bit in volume so you can normally raise the voice. Maybe a slight compression can be added, but don't overdo it, or it will only enhance the wrong frequencies making it muddy.

Ofcourse you'll "destroy" some of the dynamics, that's the whole point of compressing isn't it.
I said "when done properly", and that isn't quite the same as completely wrecking the sound (you're probably talking about that loudness war video) , but rather compress to bring out some of the softer bits so that you won't end up with a vocal that differs so much in volume throughout the sample that you can't make out what's being said.

LPChip

Quote from: "seventhson"
Quote from: "LPChip"

Actually, thats not entirelly true.

If you compress sound, you'll destroy the dynamics even more. I just saw a video (see other post here on the forum) about compression and how it can destroy your vocals.

You're better of lowering all other sounds a bit in volume so you can normally raise the voice. Maybe a slight compression can be added, but don't overdo it, or it will only enhance the wrong frequencies making it muddy.

Ofcourse you'll "destroy" some of the dynamics, that's the whole point of compressing isn't it.
I said "when done properly", and that isn't quite the same as completely wrecking the sound (you're probably talking about that loudness war video) , but rather compress to bring out some of the softer bits so that you won't end up with a vocal that differs so much in volume throughout the sample that you can't make out what's being said.

Then I've misunderstood you :) But I guess its never a bad idea to point out dangers like this.
"Heh, maybe I should've joined the compo only because it would've meant I wouldn't have had to worry about a damn EQ or compressor for a change. " - Atlantis
"yes.. I think in this case it was wishful thinking: MPT is makng my life hard so it must be wrong" - Rewbs

Sam_Zen

Spoken words indeed is different sound material than sung words.
Most of the time with a bigger range of dynamics.

But I think it's premature to immediately think of the stage of limiters, or compressors etc.
It's a recording of an analog source, so depending first on the mike-situation.
Recognizing the text depends first on proper pronunciation. Maybe the 's' sounds are quite weak in the take..
I don't think there's a common method for this, every recording should be judged on its own.

But I prefer to EQ (in a wide sense) the recorded material first, before thinking about enhancing volume etc.
An order of the routine could be :
1) Because it's an analog recording, first clean the noise from the 'silent' parts of the track.
2) Make some EQ correction. (This stage proves that starting with limiting or compression is inconvenient,
because some filter algoritm easily can lead to overloaded peaks of the file.)
3) If the sound is optimal, then it's time to adjust the overall volume. As residentgrey stated,
Normalizing is the obvious step. But because, especially with spoken vocals, there could be sharp peaks in
the track, this function is maybe possible from 92 to 99 %.
This is where a limiter can be useful. To lower 3 occasional peaks of 92 down to 62.
Peak Limiter (nagware) can do this.
4) Time for the volume maximalization. First use the Normalization of course, because it doesn't
disturb the volume relation between different parts of the track, just raising the value of each sample-byte.
5) Only now it's about a possible compression of the file. Or anyway necessary because of web-tranport limitations.

A tip from the production perspective : try a little bit of the standard 'chorus' effect on a text track.
0.618033988

Harbinger

I was hoping you would reply, Zen, because something told me you've been there done that. Very rich info. Thanks.

Quick quez:

2) EQ correction: what are we trying to achieve -- louder bass, or midrange?

Also, if i may add; PLACEMENT of a vocal track makes a difference as well. For my own composition (Once my Friend), where the vocal line did not hafta stick to the beat, i had a lotta leeway as to musical position. And actually, it took a couple of tries to put the syllable in their own "beat" where they could be heard (i even took out a couple of sounds, and i still didnt think it was audible!)...

Sam_Zen

The type of range depends on the EQ correction demands with each single track, so a common preset will be rough.

Introducing a binary aspect in this:
A 'louder bass' can be done by an increase of the volume of the lower area.
A 'louder bass' also can be achieved by a decrease of the volume of the higher area.

I agree about the placement. Very important.
It makes significant difference as impression, if the voice goes just before, on, or after the 'beat'.
0.618033988

LPChip

If you go for an EQ (which I would recommend) then its wise to use a parametric EQ. I can highly recommend Electri-Q for this. (if you're low on cash try their posifopit edition which is free).

Your goal would be to boost the frequencies in the voice that appeal most to you and fit in the mix best. On top of that, you would want to make the voice clear. In order to do that, I usually boost small peaks of a Q of 0.100 at the following hz: 1k, 2k, 4k, 6k, 10k. I change the volume of each according to how much is needed. This will make it sound crystal clear (works for everything actually).

If the voice already has alot of high frequencies, you can ADD one additional point with a big Q of say... 400 and make the volume -3db or something. That'll remove the high end, but due to the peaks you made before, it will rise out those frequencies and you'll get clarity in those frequencies. Ofcource, this is purelly theoretical. Play with the volume and see how things work out.

Although the same could apply on the lower frequencies, I usually don't apply this tecnique there because it doesn't need it. Too much clarity especially on those area's will make it sound artificial.

You can apply make-up, and add details to it, but if you apply too much, it will look fake. (thats the point I'm trying to make :D)
"Heh, maybe I should've joined the compo only because it would've meant I wouldn't have had to worry about a damn EQ or compressor for a change. " - Atlantis
"yes.. I think in this case it was wishful thinking: MPT is makng my life hard so it must be wrong" - Rewbs

Sam_Zen

0.618033988

Harbinger

Yes, and if you need to make a radio-transmitted voice, like i did for "Once, My Friend," decrease all freqs except the middle 3. (However, figure out which 3 according to the music behind it.) IIRC, i used a flat "midpass" instead of a sine shape. For a VERY old AM radio sound, decrease all but one of the midrange sliders...