Defining Ratio/Threshold Behavior in Compression

That's typically the way it is done. My preferred approach is slightly different to this but in the same kind of practice.

I'll get to nominal gain at peaks with the preamp output then into an LA2A which is set for gain at around 12 o'clock and peak reduction 3.5 to 5, so that it's just shaving 1 to 3 db off at the most and smooths everything out nicely, then I'll check the output to make sure I'm still at nominal levels, then into an 1176 for peak reduction. The 1176 is usually set with input at around 6 or 7, fast attack and medium release, 8:1 ratio for more dynamic singers and 4:1 for more controlled singers. Output until it's at the right nominal level. On Femvox I typically substitute a Tubetech CL1b for the LA2A. It just sounds prettier.

Most professional engineers I know, have learned from, have worked with, tutored, mentored etc are all in the same category of using more than one comp and/or limiter on a vocalist. It is a RARE thing to encounter a vocalist who only needs one or the other.

(the following isn't exact science, just common operating science) What's nominal? 0db in the analog world is -18db in the digital world. For those of us used to riding hot in the analog world, +3db analog is -15, +6db analog is -12db digital, and of course +9db is -9db digital. There's nothing over ZERO (or naught) in the digital world so -18db being the "zero" point, leaves 18 db of headroom above it for all the transient spikey goodness. So you choose your nominal point wisely and carefully and see how it works for you and your gear and gets you the optimal balance between noise floor diminishing and distortion onset.

I do mostly the same thing with an 1176 first and a LA2A in series on vocals. the 1176 is set for fast-ish attack and fast-ish release. 4:1 or 8:1 depending on how bad the transients are.  The LA2A is set for compression. Both hit about 3db on the gain reduction on normal vocals.

This seems to solve a lot of problems for me. The 1176 is fast enough to cut the 'plosives while not compressing so much that it's too noticable. the LA2A rounds off the slower volume changes and adds thickness.

This Saturday night I will be running my new Notre Dame this way:

ND > Ventura > Dragon > STA > Burl

Since it's pretty much all new to me, I will be experimenting with settings and gain staging on all pieces. Suggestions - particularly on how hard to push the STA, which is the noisiest piece in the chain - are always welcome and appreciated.

If I can get some good takes and my vocalist's permission, I will upload something to SC.

I don't track vocals with compression. Why you ask? Because if someone gives me a raw vocal with an appropriate vocal mic, I can sit it in a mix with a single compressor. When I get vocals that someone deicded to compress "a little" on the way in, I end up with a whole string of crap AND heavy automation to do the same thing.

Limiting can be fine. For a really dynamic singer, some really fast limiting can make mixing easier....but, I still don't recommend it for people recording themselves. I'd rather have the most natural vocal sound I can get to start. But, then I hate that sound that's so popular now--a 251 with all dynamics removed, where they belt it and it gets quieter than when they whisper...

Yes, gonna try this.

I never track through compression, but i do compress and add a touch of reverb on the mix going back to the singer's earphones for that very reason, johnkenn.

I need to just bite the bullet and buy a cheap reverb unit to send to the headphone cue...