Compressing vocals
Starting with automation before compression can help achieve a natural sound. Tape saturation before compression can help too.1
Use more than one compressor in series, starting off with ratios of 2:1.2
Warren sets a compressor to work on the loud parts of a vocal, doing about 3dB of gain reduction.3
Consider using a limiter (~20:1 ratio) to control rogue peaks.4
For general leveling, start off with 3:1. RMS compression and soft knee can be useful here.5
If the vocal feels too urgent, the release may be too fast. If it feels too squishy, the release may be too slow. Too fast of an attack can make it feels crushed.6
Use a faster setting with FET compression to catch transients or VCA compression for something a little cleaner.7 Marc Daniel Nelson actually uses MJUC with fast time constants doing just 1dB of reduction.8 Upward compression can be useful for bringing up the low level moments (just a little bit, like 3-6dB of compression).9 Use a multiband compressor on vocals to control different parts of the spectrum. Use Optical compression or Tube compression to do more leveling. Also see Mike Senior’s vocal balance workflow.
Parallel compression is very helpful. Michael Brauer uses up to five parallel compressors. Parallel limiting and clipping can be useful too.10