Using reverb on a separate bus? (software only)

[ghostwhistler]

What is the 'correct' (yes I know) way to use reverb?
Should I route desired tracks to a single dedicated reverb bus?
Should I then set the reverb on that track to 100% wet?

Having tried this I find that it sounds a bit much, but then if it's less than 100% you get the dry signals bleeding through which may be undesirable.

I could just put a reverb plug in on every track i wish to process, but that's a bit excessive in terms of system performance and possibly less efficient.

What is the accepted, if any, method (we are talking using Daw and software, not hardware at all)

[zerocrossing]

There’s no “correct” way to do things, as you already know, but the setup you describe is basically correct. You should have the reverb plugin set at full wet, can control the amount sent to it to control how much reverb each channel gets, and then control the level of the overall reverb by the level of the send.

[ghostwhistler]

There’s no “correct” way to do things, as you already know, but the setup you describe is basically correct. You should have the reverb plugin set at full wet, can control the amount sent to it to control how much reverb each channel gets, and then control the level of the overall reverb by the level of the send.

Thanks

[Benedict]

Should I route desired tracks to a single dedicated reverb bus?
Should I then set the reverb on that track to 100% wet?

Correct. The reverb has to be 100% wet. The send amount is then what sets how much of each part/track is sent to the "Room" that the song is happening in. Technically you can set every Send to 100% or Unity - 0dB - and then adjust overall reverb level at the Return but that is a bit untidy and limits options. Generally, the safe starting point is to default all Sends to -12dB and the Return to 0dB.

While you can put a reverb on every track there are several far more pressing problems than CPU. With a reverb on each track, you can have every sound in a different space esp if you have a different set of settings for each reverb unit. Even if every reverb unit has the same preset, there is no "gluing" which happens when you have all the separate sounds coming together in the Send reverb.



p.s. Take care to use the official Send and Return system in your DAW as some will add latency if you create another channel.

[ghostwhistler]

Thanks, that's super helpful

[imrae]

I agree with the above. Basically:

The long answer
There is no correct way, it all depends on your artistic goals, if it sounds good it is good, etc etc etc

The short answer
Put each reverb on a "send"

[Ah_Dziz]

The way you described it in your original post is the "usual" way of doing it. If things are sounding "a bit much" you can just turn down the entire aux track and scale back the reverb across the board. I almost always end up backing off the full return signal of reverb during the mix . You may also find that your reverb settings are "too much" and it's as good an opportunity as any to work on that. Don't forget to use other processing on your return. EQ can work wonders for uncluttering things while still have a very reverby sound.

The reason that the setup with tracks being sent to one (or 2 or 3 reverbs) by varying ratios and that reverb is set 100% wet is that it allows the maximum flexibility for multi track mixing to stereo. If you used reverb on each track with the same or similar settings, and decided you want to change something across the whole track, you'd then have to change a jillion reverbs one at a time. This does mostly stem from a time when people didn't have many reverbs or busses to put them on. So it utilized resources well while remaining flexible.

[prodsamper]

Making a reverb bus is less cpu intensive, but you are stuck with only 1 reverb setting for all of the tracks. I usually have a "ambience glue" reverb bus, and the "creative" reverb I just insert on the tracks.

[JerGoertz]

Don't forget to use other processing on your return. EQ can work wonders for uncluttering things while still have a very reverby sound.

Beat me to it, but I wanted to emphasize this. Processing the wet signal by itself can be useful, which you can't do if you have reverb as an insert.

[imrae]

A nice trick can be to insert a transient designer and reduce the level of transients entering the reverb. This can help smooth out "grainy" and early-reflection-heavy reverbs.

[swilow11]

I do wonder if sending multiple tracks to reverb or any such effect is worse for CPU than just assigning a reverb to each individual track. My understanding is that a send with multiple track routed into it gets processed on a single core which can result in a bottleneck of sorts. That said, I basically always use sends so no real point of comparison.

[_leras]

I do wonder if sending multiple tracks to reverb or any such effect is worse for CPU than just assigning a reverb to each individual track. My understanding is that a send with multiple track routed into it gets processed on a single core which can result in a bottleneck of sorts. That said, I basically always use sends so no real point of comparison.

I would say no. You'll have less instances processing audio.

Also if you have sends you can process the returning reverb. You can't do this on an insert. (Ok well technically if you have Bitwig you can, but that just means less obvious routing)

If you are running into CPU issues then perhaps adding per track, or as in Bitwig you can create a group of a track and an individual reverb, then bouncing these down would mean that you'd be able to turn off tracks more easily. My bottle necks when mixing have usually been more about group level processing where I'm using higher CPU FX.