[linux-audio-user] How to get good vocal sound

[Dave Phillips]

Bill Allen wrote:

> I've been struggling with a technique problem. How do you get your 
> vocals to sound good? I know that you can't make a silk purse from a 
> sow's ear, but you can make the best of what you've got.

Hi Bill:

You also have to remember that you don't hear yourself the way others 
hear you. I'm not especially fond of my voice, but I know how to hit the 
right notes and other people seem to like it.

The *sound* of a voice can be effective enough as a carrier of 
personality. Bob Dylan is often used as an example of a not-so-beautiful 
voice that's perfectly suited to his material. And it's worth mentioning 
that Dylan does know how to sing (check out Girl From The North Country).

> What combination of plugins and settings do you use to get the best 
> sound from vocals particularly in Ardour?

As you say, there are many to choose from. I like Tim Goetze's Versatile 
Plate Reverb, it's easy to control and has some sweet spots that 
complement my voice nicely. I also often use the SC4 compressor 
(pre-fade) and the TAP Scaling Limiter (post-fade) on the master track.

As general advice: Pick an effect, preferably one with only a few 
controls, and learn what each control does to the sound. In the case of 
reverbs, think about the ambience you want to create and make your 
initial settings accordingly. Yes, you'll have to experiment, but IMO 
that's the best way to learn about these things.

> I know that this is one of those subjective questions for which the 
> best answer is try it out yourself and find what sounds best, but 
> there are so many plugins (an embarassment of riches), each of which 
> has many settings, that a brute force search of all the combinations 
> would take forever - not to mention that after a while, my poor ears 
> become exhausted with the effort and refuse to hear differences 
> anymore. So what I'm really looking for is good starting points to 
> work from.

It's easy to be overwhelmed: LADSPA, DSSI, VST/VSTi, all provide that 
embarassment of riches. But I figure that the pros don't have the time 
to learn everything about everything, so I don't expect it from myself. 
I think you should plan carefully: think about what sort of sound you 
want to project, think about its characteristics, then start working 
towards it with a minimum number of effects. A little 'verb and some 
judicious compression go a long way towards making a good basic sound, I 
find that for my purposes I use little else. Maybe someday I'll find a 
use for chorus and delays, but they'll probably get used on instruments, 
not on my voice.

A few random tips: Stand when you sing. Use a pop filter (I should 
follow my own advice). Breathe deeply, you can always erase the noise. 
Don't raise your chin when you try to reach notes on the high end of 
your range, it tightens the vocal mechanism and works against your 
attempt. Relax, you won't sing (or play) well with excess tension. 
Consider your input chain, i.e., decent microphones, preamp if 
necessary, good hot signal, etc. Don't record with effects in, you'll 
fool yourself and it will be harder to fix (I like that Ardour forces 
that behaviour). Key your music to complement your available range.

WRT planning: Do you want to recreate the sound of a small group playing 
in a bar ? Or a big band performing in a concert hall ? A rock band 
playing an auditorium ? A jazz quartet in your living room ? Each of 
these scenarios has some salient characteristics that you should try to 
describe and understand in relation to your selected effects. Try 
analyzing the vocal sound on some of your favorite recordings in the 
same manner.

Yes, you have to study up on some terminology, but you don't have to 
become a DSP engineer. That's what we have Steve Harris here for anyway. 
;-)  And don't worry about understanding everything right away. It's 
enough to know what a compressor does, the rest you'll learn by fiddling 
with settings and *listening* to the results. For instance, I use the 
SC4 compressor but I'm still sort of blank about exactly what the knee 
radius does. Always more to learn. :)

> One combination that I like is GVerb to get depth and L/C/R Delay to 
> get width. Even with those two getting the settings right takes time. 
> For comunication, I've included a jack rack with some settings I've 
> found that work OK. I would love your critiques and suggestions for 
> other setups.

GVerb is more complex than I'd advise for a start, but it is a fine 
effect and definitely worth learning. Have you tried Freeverb or any of 
the other 'verbs from the LADSPA collection ? The TAP reverb is also 
superfine (with presets!), but again  I'd suggest something even simpler.

If you really have problems with intonation and pitch acuity you can 
always try fixing it in the mix with a pitch shifter plugin and Ardour's 
automation control, but that's getting elaborate. You can also try that 
infamous vocal "fixer" from Antares, but I don't know if it will run 
under any of the current support systems for VST under Linux. It costs 
$$ though, and frankly at that point I'd suggest singing lessons.

You can also use Ardour's excellent editing tools to correct small (and 
not so small?) errors in timing. A bit tedious, but worth the effort.

I listen a lot to the music made by other members of this list. I love 
the sound of Pete Bessman's F4, I'd like to know how he got it. I'm also 
fond of The Girls, their material is great and their singing is perfect 
for it. I like the overall sound they achieve, though it's quite 
different from my aims. And I'd commit heinous and unconscionable acts 
just to be able to play and sound like Steve Doonan. Lots to be learned 
from the folks using the same software (more or less). So many different 
genres are represented at http://lam.fugal.net, it's a terrific 
resource: if you like a particular piece, you can just write to the 
composer and ask how s/he did it.

Well, there's my contribution to the thread. HTH. :)

Best,

dp