I leave it for the mix (with a 144dB dynamic range, it doesn't really make sense to do otherwise) unless the compressor itself is an integral part of the sound.

I'll gently compress things i KNOW i'm going to need to compress in the mix anyways (DI'd Bass, clean/acoustic guitars, vocals) to save myself having to run a bunch of extra plugs later on. My DAW is in dire-need of upgrading as-is so anything to save horsepower is a great thing in my books :)

If CPU power wasn't an issue, it might be different story but i've been more than happy compressing while tracking in the past so it'd take one hell of a killer soft-compressor to change my mind.

Each to their own, it really depends on the appropriate musical/production style...people mainly tracking acoustic jazz probably wouldn't want too much dynamic levelling early on in the process, for example.

I'm mainly involved with rock'n'roll of one sort or another and usually recording on 24 trk tape - for most jobs I'm slamming the drums onto tape so hard compression isn't necessary at tracking - but some drummers play unevenly and when tracking digitally I might find it necessary to limit or compress excessive peaks a little to get a really strong signal on kick or snare. Usually drum compression has to be tailored to each individual song depending on pace and dynamics, so I find it safer to compress at mixdown.

I don't often compress distorted/saturated guitar sounds at tracking as it's already compressed in a way (tape saturation works wonders here too), acoustic is a different matter depending on the part being played and the player.

I find that bass guitar usually needs a fair amount of compression at tracking even when it's going down on tape, & I'll usually compress bass at mixdown as well. Most of our clients like as even a bottom end as possible, and I'm sure some of you guys must have felt the dismay of seeing rickenbackers come out of cases, waiting for the thunderous "bonk" whenever A is played on the E-string...yuck. Different guitars will sound out certain notes louder than others, even using a D.I.

With vocals I'll almost always compress, for loud rock yelling more so than for a quiet, breathy styles.

Some synths (especially vintage analog units) need dynamic control if it involves sudden large changes in volume - synths are volatile and can suddenly generate huge amounts of bass power.

Since we're using tape most of the time (I know this isn't normal in most set-ups these days) it really helps to get a sound as close to what you finally want laid down at tracking, it makes an old-school "live" mixdown easier. I'm used to this, so to an extent I carry this approach over to the digital domain. It takes confidence and conservatism to get it right of course, but it means you put down what is already (subjectively!) a good sound and then make it a great sound rather than nervously recording a useable but garishly raw sound and then using lots of processing to get an acceptable one.

There's also the issue of noise when working with tape - if it definitely needs compressed it's better to do it when tracking since afterwards you'll be pushing tape hiss and rumble up in volume when you do compress. Usually only a concern when tracking on analog.

Incidentally I would always use hardware when possible, I like individuality.

I hasten to add that I use a minimum of compression when tracking anything, always of the "invisible" sort (no gain pumping or serious flattening of dynamics): too much means you can then never go back to a natural instrument sound - this would always reduce mix options and will probably annoy clients.

At the end of the day I want to put down a sound which is within a certain ballpark. That ballpark has already been discussed by all concerned and varies from project to project. I change the rules all the time depending on media, artist, song and instrument.

Waffle, waffle...etc.

For the most part I also agree with alimoniak. I most always use some kind of compression/limiting when tracking vocal, bass, room Mike's on drums. I don't much track anymore on analog and miss that saturation on certain instruments, like drums. Ditto on the guitars. When using tube amplifiers, the compression is an inherent part of the sound so no need to generally add any more. As with most anything, less is more and it's easier to add than to try to take away. Many of us have certain pieces of equipment that a combination helped to give this our characteristic sounds when used correctly. Something I think inherently we get much joy from doing. Since our equipment is our instruments, like anything else it takes years of practice and technique to achieve ones on characteristic sound.

Don't touch that dial!
Ms. Remy Ann David

pretty much the same here, unless it's a vital part of the sound i want to achieve or clipping problems due to too much dynamics (mainly singers) i leave it for the mix, it's easy to add and a pain to get rid of.

a.

Usually just the bass and perhaps a touch on the vocals as kind of a safety if the singer cant work the mic very well....I will compress the bejesus out of clean background guitar, but usually I've already made the decision where its going in the mix and what its purpose is going to be.