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

On Tue, Dec 06, 2005 at 08:55:20AM -0500, Brett McCoy wrote: +1000. The best investment is a good mic. It doesn't have to be all that expensive if you do your research. I bought an MXL V67G on ebay for under $100 with shock mount and I've been pretty happy with the results. Google for that and "Harvey Gerst" and you'll know why I got it. Gerst's online writings are priceless for people doing recording on the cheap. Sure, I'd rather have a Neumann, but who has the money for that? Another important point about mics is that some mics are better suited to different voices or even different songs. Try what you have, there's no such thing as the one true mic. My preamps aren't anything special - I have a Mackie 1202 VLZ - it's fine for now, but a little harsh in the treble. I'll probably upgrade that before getting any more expensive mics. I often compress on the way in, using an FMR RNC. (But not always; sometimes I leave it until later and use SC4 mono). Vocals are pretty tolerant of large amounts of compression, and I find that having good solid levels in my headphones helps me get better pitch. So does having some reverb in the phones. I do in-computer reverb with hardware monitoring by setting up a bus for the reverb and setting its input to the hardware input, so it doesn't go to "tape". Then I listen to the results without reverb and sometimes decide it doesn't need any! But anything that helps an unconfident singer (i.e. me) get a better performance is worthwhile. I've tried all the reverbs... freeverb, gverb, TAP reverb, versatile plate... I don't have a favorite yet, they're all useful for vocals. A tip for using Freeverb: It may seem a bit plain and sometimes harsh for vocals by itself, but you can get very good sounds out of it with some tricks: First, set the "damping" really high, that makes it smoother. Now try adding some predelay - a simple delay plugin in front of the reverb, at 100% wet; try around 50 ms delay to start with - that makes the "room" a little bigger and more interesting. Lower the delay time to make it more subtle; bump it up to 100 ms to make it a more obvious effect. Finally, try post-filtering; follow the delay/reverb combo with a low pass filter (the simple one-pole filter from CMT will do fine); set the cutoff around 5000 Hz to start with and adjust to taste. Smooooth. Here's an example using the above freeverb recipe (please don't anybody make permanent links to this one as it's a rough mix of stuff that we're probably going to re-record due to some performance flubs and some too-hot levels you can hear a bit): http://slinkp.com/~paul/dont_get_sad_mix_20051206.mp3 I played guitar and sang backups (badly), my wife sang lead (beautifully). And no, we didn't write that, it's a cover of an Ida song. I wish I wrote that. Other technical notes: Lots of compression (SC1) on each vocal track. Maybe 6:1 ratio, attack around 30 ms, release around 120 ms. No EQ on Abby's voice. I forget if that was the MXL V67G or the MXL V57 (which I sometimes like better on female voices, it has less of an upper-mid presence peak). Some 10-band EQ on my voice due to poor mic choice during tracking (Sennheiser 421 which is a GREAT all-purpose mic but for this song I found it too present and sibilant in the upper mids for backing vocals). Which brings up another idea ... if you want a brighter reverb sound than I was going for in that track, like super-in-your- face radio pop, one problem is that voices can get really sibilant. TAP De-Esser can be really handy here. Put it on the vocals after compression, before EQ and before reverb. Then you can make really obnoxiously bright vocals without the ssssssibilancccccce tearing your head off. I don't have any recorded examples of this to share - yet :-) That's nice if the rest of your gear is already pretty good, but personally I don't worry about balancing line levels at this point. In my view it's a lot less important than having a good mic, good soundcard, and good preamp. -- Paul Winkler http://www.slinkp.com