ok, now it doesn't make too much sense to have a submix compressor shaping only the snare. if the snare needs shaping, you do it on it's own track. so back to square one: you feed the submix comp a tonal-balanced drum mix, what triggers it? short answer is the same: whatever signal passes the threshold at any given moment (kick, snare, percussion, maybe a sum of two elements hitting at the same time). The release greatly affects what happens after, that's very important to hear.

now an useful answer: why do you need a submix compressor? there are plenty of top40 number ones that are mixed without submixes (very track goes to the master channel). and here comes the TASTE (and by extension, your -lack of- vision): if you know what you want to achieve, do everything required to achieve that. But just because you read about submixes, top/down mixing, parallel whatever, doesn't mean it applies your case. Doing for your music what others do for their own, will not make your music better. If you take scheps, cla, ghenea, hanes, whatever worldwide famous engineer, they don't even work the same at that level. Each one has it's own workflow and methods. That's why is so so vital to experiment alone.

 

Spoiler: Here are a few values and clues:

Drums

fast attack times
Live bass drum (bad drummer)
Attack Release Ratio Threshold
50ms 300ms 8:1 -15dB

Live Bassdrum (more assertiveness)
Attack Release Ratio Threshold
100ms 300ms 8:1 -20dB
Live bass drum (for distortion = limiter use)
Attack Release Ratio Threshold
1-10ms 0.5-10ms infinite:1 -5 - 0dB

Bassdrum
Attack Release Ratio Gainreduction
50ms 300ms 5:1 -4 - -10dB for loud beats
10-30ms 200ms 4:1 - 8:1
5ms 50ms 4:1 3-6dB

Due to closemiking (microphone close to the sound source), the natural sustain is lost due to initial reflections. The "Release Time" must not be so long, so that the next beat is not influenced. The "Release Time" must also not be shorter than the "Fade Time" of the noise gate. Otherwise the beat will become louder towards the end. The larger the "Ratio", the more linear the sustain curve.
For typical mid tempo pop and light rock, the "Attack Time" should be 2ms or less. A medium "Ratio" should be selected. Then the bass drum will sound like a drum machine.

For fast pop, R'n'B, and dance, choose an "Attack" of 5-10 ms, so the kick sound is not compressed. The "Ratio" should be set high. The bass drum can be added to the mix at a low level and still assert itself.

For hard rock and heavy metal a short "Attack" time is recommended. The "Release" time should also be short (30ms or less). The "Ratio" should be very high.

Snare
Attack Release Ratio Threshold
5-20ms 150ms 4:1 - 6:1
20-30ms 200ms 3:1 - 6:1

For a good drummer, high compression threshold, only on loud beats the compressor should regulate. Insecure drummer: low threshold, soft-knee characteristic curve

For pop music, a quiet, soft, smooth snare is required. This can be achieved with the following compressor settings: "Attack" below 2 ms, "Release" 1 sec. or more, "Ratio" to medium setting. The attack phase (transient) is also compressed with these settings. The snare is then easy to control in the mix.

A snappy, lively snare is needed for dance and R'n'B. The following settings are necessary for this: "Attack" must be long (5-15 ms), "Release" on the other hand very short (less than 200 ms), "Ratio" set to medium. The attack phase (transient) is thus not compressed: The snare "pops."
Finally, for hard rock and heavy metal you need a fat snare with long sustain (decay time). This is achieved with the following settings: "Attack" is short (under 2 ms), Release is long (2 sec.) and "Ratio" is high (8:1). A very strongly compressed snare can be achieved this way.

Toms
Attack Release Ratio Threshold
5-10ms 0.5-1s 5:1
5-20ms 150ms 4:1 - 6:1

 

It depends on you that what do you expected from the comp? Is it subtle or aggressive? glue? pumping or squashing?Color?
If it's about 1 db gain reduction then it's not pumping anymore, it's kissing

 

just want to clarrify something

when people say "practice" - they actually mean , you need to train your ears - by doing the work (practice compression in this case)

 

Not at all.

All sounds that cross the threshold.

The same as for other compressors, depending on the source sound and/or level.

Not necessarily. This depends on the attack time.

Err, no because in this case the compressor changes the sound, the development of the dynamics.

With a compressor (on a bus, formerly known as subgroups) you can even out level differences. At a given peak you can lower the RMS (longer attack) or raise it (short to no attack time). Or the other way around, at a given RMS you can raise the transients/peaks (longer attack time) or lower them (short to no attack time).

But no matter how you set it, it changes the sound (especially) of percussive sources (attack, decay). If you use a compressor on vocals to even out the level differences, the lower parts become more audible. In this case you do it more for the level (even if this is also a change in sound).

 

A compressor is a dynamics management device. It is not (primarily) the supplier of an effect. Even if transient shaping, otherwise "pumping", can sometimes come as a bonus from its application to (portions of) a mix, the perceived "livening" still depends on shallowing a given signal's dynamic envelope, reducing its initial "contour". True effects either derive a secondary signal from a primary source, allowing the two to be combined, or they process, modulate, the primary entirely. A compressor does neither, but controls amplification of a source signal (over threshold) against attack and release settings. The compressor is an aid to the balance engineer. However, while in most instances "technically ideal" compressor-against-unigain introduction will remain undetected by audience, "overdoing" their functionality can, in some cases, manipulate a signal in support of musical expression. While the latter can be desirable as a program enhancement, the compressor remains a sound engineering tool, and is still not an effect per se.

 

I usually stack 2 or 3 compressors on buses. The first should have a fast attack and release, hard knee, and a high enough threshold to tame those peaks which can be problematic for the next compressor which has a slow attack, release to taste, relatively soft knee and low threshold to shape and glue the bus.

 

everything is can be an "effect" if you use them creatively, as simple as using EQ for widening tool (EQ L/R separately, M/S mode). Compressors, even the mixing console has been used as "distortion machine" since the 60s. Ever heard of "just letting the signal pass through" ?

 

rule of thumb with compressors. It makes loud sounds softer and soft sounds louder. So adding a compressor on a bus you could do that to bring out more audible room noise, instead of adding db's to the room track making your head room less... as an example

 

dont know if this was posted but OP should watch this