On percussion, I'd apply saturation first to clip or round off the high peaks, before compressing. But what about vocals?

My very limited experience in 1970s-80s recording studios is that vocals were recorded to tape "straight", with compression added at the mixdown stage. This implies that saturation comes before compression.

On the other hand, if you compress first then apply saturation (easy to do with plugins) you're bringing the vocals up to a more or less fixed level before "adding harmonics", which might sound more consistent.

Please discuss pros and cons of each!

 

Usually, saturation is meant as compression + distortion, so standard method as analog usually does:
Preamp (gain, saturation/distortion)
Eq (with saturation or clean digital)
Compressor (so it does compression either clean or with saturation (fet, opto..))
Saturation/color (as tape, for example)....

Even mastering on hardware, it means further improvements, further saturation....as you see, looks like saturation goes with gear. Or easier to say working with only gear - saturation and harmonics are everywhere, slight or not.... But compression->slight distortion means saturation... So, after.

Or no rules.
Just try, listen, and hear which of approaches is the best

 

I say try both, go with what you like. But probably sat first.

 

Logically thinking, might be so:
Create sine
Set threshold, ratio too much, fast attack..
Then you see that round peaks of sines become square-like, so square waveforms mean distortion, so saturation goes from compression.
Same with limiting, clipping. Saturation is due to it.
But you can just set color and distortion you need/want, then compress, or vice versa. It's up to you.

 

There is no right answer.

I would say "depends"

 

Having played around a little this evening with a vocal I recorded previously, I'm leaning in the direction of gentle saturation first to remove "spikes", followed by compression to help with leveling, followed possibly by more saturation as an effect.

I should say that I always use clean compression, rather than those plugins that try to emulate analogue gear. As an erstwhile audio designer, I don't see any point in replicating the undesirable side effects of poor circuits! And of course everything is much easier in digital; for example lookahead is straightforward, which it isn't in analogue electronics.

 

Don't want to throw sand in the machine, but what about the saturation added at the compression level?
I find that late '90s that type of saturation to be much more pronounced in pop productions against tape saturation,

logical deduction tells that would impact 70s-80s recordings big time

 

Sometimes you don't have to use either. Or only one if them.
Just as a reminder maybe.

 

You shouldn't do either, both are unnatural and were the product of technical shortcomings of analogue gear.

 

Just switch inserts around and listen…
This is the way!

 

Both in/early and out/late. Just like the input- and output transformers in great sounding consoles, preamps, compressors, etc. it's also subjective and up to taste. But personally i think the LA2A, real hardware as well as many plugin emulations, sounds fantastic! It saturates in a lovely way.

 

Last edited: Apr 4, 2023

How's that for a start?

 

Well, I like my compression to be as clean as possible, with a bit of lookahead to catch quick spikes before they happen. I don't see any point in emulating the unavoidable faults of analogue equipment. (I used to design analogue audio circuits for a living!) My ideal is a compressor that compresses cleanly and does nothing else; lately I'm using ToneBoosters Compressor v4 in its cleanest mode.

https://toneboosters.com/tb_compressor_v4.html

Actually any compressor generates unwanted frequencies by its very nature; essentially it's an amplitude modulator that modulates the amplitude of the vocal (or whatever) signal by a slow control waveform. As we know, amplitude modulation creates pairs of sidebands. The faster the control signal, the higher the sideband frequencies, and they may be audible. This is a theoretical issue and not one that can be got around by circuit or software design. And as the control signal is derived from the vocal signal, it inevitably contains some low frequencies from it. These can then modulate the vocal signal itself, causing harmonic distortion.

Having said that, the Kotelnikov compressor claims to work on a different principle, using a Hilbert transform for its RMS sensing, but I'm afraid the maths behind it is beyond me so I can't comment on it! I've played with this compressor and it certainly seems clean, but with its dual peak and RMS sections it seems too complex for everyday use. It's really intended for mastering.

https://docs.tokyodawn.net/kotelnikov-manual/

EDIT: Reading the manual, I now see that the ToneBoosters Compressor also uses Hilbert transformation in its clean mode!

 

Last edited: Apr 4, 2023

Seen you talking about this one before, just bought it. Holy crap. I needed this! What an amazing tool.

 

the post about no right answer is true for all tracks. But in a "general purpose" scenario, i will usually do daw eq->saturation->subtractive eq->compression - >additive/flavor eq/ whatever else like a reverb send or exciter, etc.

it's the order I always think about chaining them, even if i do not use every component. You have the benefit of not saturating or compressing unwanted noise/low volume frq content. Those two plugins get hit with only what they need. Otherwise you are compressing noise.

 

Last edited: Apr 5, 2023

I haven't read the whole thread but imho it's best to look at this not as a one arrangement or the other, but what will give you the sound you're looking for. You can use saturation to prep a compressor, but you might want to hit the compressor with all of the audio's transient's intact. It will trigger the threshold differently depending on how you do it. They're both good. A lot of compressors use saturation built into their algo, sometimes adding more is appropriate, sometimes no.

Old recordings definitely are NOT always "straight to tape", probably the opposite really. Typically the old school way is to get the exact sound you want onto the tape, including eq and compression. In this method mixing is more a matter of volume balancing as the song plays through.

 

Thanks for the links, no I haven't tried them (yet). They certainly seem to be aimed at clean compression -- but they don't use Hilbert transforms, FWIW!