Reverb - Insert or Send - When and Why

[GuitarRuss]

I've been amazed at the difference between the sound of reverbs sent as a send or an insert. I don't deal with many tracks and had always thought that the sends were largely about saving Cpu power by sharing the plug between tracks. I run a quad opteron so I had more than enough power to keep adding inserts. Last week, I decided to try using the reverb on a send instead after watching "Mix it Like a Record" by Charles Dye, because he talks about how it help to glue your mix by sharing the same reverbs between tracks.

Of course, I understand that inserts totally color the sound and sends you get the original track plus the effects on top which will sound different.

My question is this. What are your guidelines for when you choose one over the other? In some cases one does sound a lot better. What are your preferences for send or insert for different types of reverb (convolution, plate, hall etc.) This subject really got me curious. Of course you can also use the sends/effects-channel to do other stuff, like EQ the reverb itself, or cut out the low end, or pan. etc. What are your ideas about this as well?

Edit: Much of the difference I was hearing had to do with the fact that I hadn't allowed the proper wet/dry balance between the channels and hadn't bypassed the other effects when I was testing. So ignore the stuff about reverb being different on sends and returns. You can still read my post through the other people quoting them, however this way you skip that stuff and get to the more interesting subject of treatment of your effects channel. My bad...

[tINY]

I don't use reverbs on inserts.

The inserts get used for things like delay, chorus, harmonizers, etc - anything you only want on one source.



-tINY

[lowfreq33]

Put it on a send, make sure that the mix control on the verb is 100% wet. Use your sends to control how much of what goes to the verb. Inserts are only for compressers and eq type effects. Time based effects should always be on a separate track.

[doorknocker]

Originally Posted by lowfreq33

Time based effects should always be on a separate track.

Not necessarily. Much like you would record say a delay-effected guitar with one mic on one channel, you could also insert say a delay on a track and get a similar effect. Just do what sounds best in the same way that there are no rules whether a comp should follow an EQ, etc

[Unclenny]

Originally Posted by doorknocker

Not necessarily.

I agree.

Sometimes I might have a guit part, for instance, that needs to be in 'another space'. A verb set up like a phone booth or something else equally foolish can be a good effect.

Generally, though...verbs on sends.

[Wayne]

Originally Posted by GuitarRuss

I've been amazed at the difference between the sound of reverbs sent as a send or an insert.

ITB, what is the difference you're seeing? Other routing-chain differences aside, would there be a difference doing the wet/dry blend in one spot vs the other?

[warhead]

I really only insert EQ's and compressors, things that I'm using to change the entire timbre and dynamic range of a track. The only exceptions for me would be when I want to heavily affect a track and soak it in a flange, chorus, delay etc. Even then, I find I can make things wet enough when needed and I might rather set up an fx send so I can maybe dabble a few other tracks in at times.

War

[ambiguity]

it's not about saving resources, it's about putting the elements in the same space.

Sure, you could insert the same verb on a bunch of tracks and fiddle with the dry/wet ration on the verb itself, but what would be the purpose of that?

[phaqu]

I send, then return to a channel, not to the return input (old school, on a board).

[GuitarRuss]

Originally Posted by dragan

it's not about saving resources, it's about putting the elements in the same space.

It's a great thing that it actually does both. I was inspired to try it in the first place after Charles Dye talked about how it can help unify elements of your mix. I would say after trying it, that that is certainly true...

However, there still are situations where both options should be considered for different sonic results.

[GuitarRuss]

Originally Posted by phaqu

I send, then return to a channel, not to the return input (old school, on a board).

I would guess that this means that you add the character of your board's preamps twice (once on input and again on send) which might be great if your board's preamps are great. However, if the board's preamp's are less than spectacular it would mean a double dose of the negative sonic qualities of your preamp as well. Interesting.

I'm ITB because of budget reasons but sometimes save enough money to spend on studio time OTB. Is my guess right, or is there some other logic behind that choice?

[Djembe]

Another thing to bear in mind - with a program like Cubase, if you insert a stereo effect on a mono track, you're only going to get a mono effect.

[GuitarRuss]

Originally Posted by Djembe

Another thing to bear in mind - with a program like Cubase, if you insert a stereo effect on a mono track, you're only going to get a mono effect.

Yes, that's definitely true. I guess there could be workarounds like splitting to a stereo track which would be worthwhile in certain situations like a stereo convolution that changes the signal alot. I recently also discovered the dual stereo panners in Nuendo. They are great. You can pan both sides of your stereo track independently which is great if you want to change sides or bring them towards the center. I've been recording onto stereo tracks, instead of dual mono, for the specific purpose of being able to get the most out of some of those IR's. This stereo dual panner is a great little bonus to Nuendo but I'm sure the other programs have it too.

[PheelTheMusic]

I'm still really not sure I see the insert/send difference.

If you use a reverb as an insert set at 50% dry/wet are you not hearing exactly half of the unaltered original signal and exactly half of the fully wet signal? If you use a send and have the original signal and the send volume both at 100% is this not effectively the same thing? That is 50% of each dry and wet as it reaches your ears in both cases.

I'm shocked to read everyone on here saying otherwise. Can someone plase clarify??? How does using an insert "change" the original signal. I believe it's just basically a 2 channel mixer within the reverb plug and the level relationship between the two channels (one 100% dry and one 100%wet)are changed by the dry/wet knob.

[MessianicDreams]

Originally Posted by PheelTheMusic

I'm still really not sure I see the insert/send difference.

here's the main clincher most people haven't mentioned yet:


INSERTS are (mainly) PRE-fader
(time-based effect) SENDS are (mainly) POST-fader.


So if you have it as an insert (pre) then you can't control how much of the signal is sent to the reverb. If you have it as a send, you can control this.

Also think about what happens when you move your volume fader. If you've got your 'verb as an insert, when you pull the fader down you're pulling down the sum of the wet and dry, therefore not changing how much you're "exciting" the reverberant space (this makes more sense if you think off your reverb as a echo chamber with a physical loud speaker and a physical mic). When you pull down your fader with a post-send, you're exciting the room less.

[PheelTheMusic]

GuitarRuss: It seems to me that your comparing apples to oranges. In your recommendation, the result would be a 100% wet signal on the insert track but only a %50/%50 dry/wet ratio on the send. If you bring the volume fader of the original track down and still haver the prefader send going to the bus, and the plugin set to 100% wet, the sound (I would imagine) will be EXACTLY the same as the insert at 100%. Are you proposing this is not the case?

Originally Posted by GuitarRuss

This is a common confusion. To understand it the difference sonically try this. Download SIR (free convolution) and get a convolution of a telephone speaker. When you put it as an insert the whole signal will get crushed down until it sounds like it's being squeezed through an old telephone handset speaker. There will be no low end, only thin nasally mid-range...

Now take the same convolution and put it as a send. You will still have your original signal plus a thin little telephone sound. Now you can take that telephone sound and pan it left or right and your original signal is still there bass, treble and all.

Seriously, go try this, it will make it very, very clear what the difference is.

A mixer might create an effects channel and dial up a distortion and a flanger. Then he uses a send from the drum bus and activate it for the bridge of the song to give it an effect. If you put these two effects as inserts your bass drum and snare will lose a lot of power so you won't be able to use much effect. If you use a send, your bass drum is still thumping just fine and you've got this crazy effect that you can play with now as it's own seperate entity. You can roll of the low end off this channel (before or after the flanger), if you don't like the flanged bass drum. etc.

Hope this helps.

[VictorQ]

I can't help but to just point out, that ITB with any linear time-invariant process, there will be effectively no difference in sound between a 50/50 dry/wet mix from the inserted plugins internal 'mix' knob and one obtained with a send at 0 dB (plus a 6 dB attenuation on the master fader). Any difference, that would turn up in a null-test should be due to differences in the internal resolution of the plug vs. the host mix-buss.

Now the difference in functionality will only manifest itself when non-linear or time-varying stuff enter into the equation. Automation envelopes will cause some difference if the automation is different (ie. with a send you can put the same envelope on both the track and the return to get the same result as an insert).

Linear time-invariant processes include eq, changing the gain (by a fixed amount), mixing signals (in fixed proportion), most delays (and delay-derived effects such as chorus, flanger, comb filter etc.), most reverbs.

Non-linear stuff is: harmonic distortion and saturation, dynamics processing (gate, comp, lim).

Being "linear" is the same as "preserving linear combination". That meens something like if you effect a mix of signals you get a mix of the signals processed individually, and if you change the gain and then process you get the same as if you process and then change the gain.

Some reverbs er non-linear (fx. the TC NonLin) and would not have this property. I don't have a powercore anymore, so I can't test it. I'd love to hear somebody describe the difference in this case!


Anyway with a linear verb the difference is only in functionality, not in sound...

[phaqu]

Originally Posted by GuitarRuss

I would guess that this means that you add the character of your board's preamps twice (once on input and again on send) which might be great if your board's preamps are great. However, if the board's preamp's are less than spectacular it would mean a double dose of the negative sonic qualities of your preamp as well. Interesting.

To somewhat of a degree, yes. The reason I did it that way was to have control over the verb at my fingertips on a fader instead of reaching over to an outboard piece to adjust levels and I still had my original dry signal. That and to be able to eq again if I needed to.

I'm ITB because of budget reasons but sometimes save enough money to spend on studio time OTB. Is my guess right, or is there some other logic behind that choice?

What you can do is set up your rough mix and tracks and groups so when you go to mix at the big place,either through a console or a summing box, your ready to just start patching. OTB ,to my ears is worlds better than ITB.

[Benmrx]

Originally Posted by GuitarRuss

My question is this. What are your guidelines for when you choose one over the other?

Maybe this has already been mentioned, but personally, I like to automate levels going to my time-based effects: Reverbs, Delays, etc. and setting them up on an aux makes it pretty damn easy, especially in a DAW.

That would be one reason.

[Kiwiburger]

I use Cubase4 and I frequently test reverbs as an Insert effect over a VSTi track (which is stereo and therefore works).

What I find is that often the amount of reverb I need is about 1% or 2% before it sounds excessive. If the plugin mix can only select to the nearest percent, then that's a very coarse adjustment. Only as issue with certain plugs.

If I set the same reverb up as a send, I can set the Send at around -12dB, and the return at around -12dB, and I have plenty of control over the mix.

Apart from that - I don't hear a lot of difference in the sound quality, and if I just want a specific reverb effect on a single track, I see no reason to avoid using an insert effect.

I totally agree with the comments about using certain convolution impulses as a "pipeline" insert effect. Guitar cabs and speaker effects in particular. This definately requires use as an insert effect.

[Wayne]

Originally Posted by GuitarRuss

... You can do this test with a onboard pitch shift too. Insert = the whole pitch will be different at 100%. Send at 100% equals original pitch plus the new harmony coming from the send. These will absolutely not null each other out!! Try that in a null test and get back to me... The same is true (although much less dramatically) of some acoustic convolutions.

But what you are describing is the difference between 100% wet and a percent wet/dry mix, not a difference of where the effect is done.
Mute the original (dry signal) in the 'send scenario and you will be back to one 'pitched tone. thumbsup

[GuitarRuss]

Is anyone doing anything interesting, treatment wise, to their reverb effects channels, or is pretty much just reverb and occasionally panning? As a relative newcomer, I thought maybe there are some really cool tricks to get more out of reverb on effects channels.

[Benmrx]

Originally Posted by GuitarRuss

Is anyone doing anything interesting, treatment wise, to their reverb effects channels, or is pretty much just reverb and occasionally panning? As a relative newcomer, I thought maybe there are some really cool tricks to get more out of reverb on effects channels.

I'm compressing/eq'ing all the time on effects sends and/or returns. I love squashing the shit out of a reverb return, followed by some HP/LP filters.

[Kiwiburger]

GuitarRuss - maybe I missed it somewhere, but what DAW and convolution verb are you running?

I'm guessing that what you are talking about is simply the difference between mixing a 100% wet effect over the top of a dry track (Send), and processing the whole track through a reverb and balancing the wet/dry.

As somebody has pointed out - if you Send is pre-fader, you can do the 100% wet thing, so there is actually zero difference in where you place the effect. As long as you achieve the same balance, they sound about the same, if not sample accurate identical.

The end result is your wet and dry are mixed together in whatever ratio you choose. Whether than happens within the reverb algorithmn, or in the DAW mix bus is the only difference.

But I agree that if your DAW sends are defaulting to post fader, then you can't achieve 100% wet with Send. Impulses that are designed to be 'Pipline' impulses require 100% wet, with no dry mixed in at all.

If you are layering a reverb with send, and mixing in some dry sound within the reverb - and IF your DAW lacks plugin delay compensation (it's not Protools is it?) then you are going to get phase issues, which might explain the difference in eq you are hearing ...

[GuitarRuss]

Originally Posted by Kiwiburger

GuitarRuss - maybe I missed it somewhere, but what DAW and convolution verb are you running?

I'm guessing that what you are talking about is simply the difference between mixing a 100% wet effect over the top of a dry track (Send), and processing the whole track through a reverb and balancing the wet/dry.

As somebody has pointed out - if you Send is pre-fader, you can do the 100% wet thing, so there is actually zero difference in where you place the effect. As long as you achieve the same balance, they sound about the same, if not sample accurate identical.

The end result is your wet and dry are mixed together in whatever ratio you choose. Whether than happens within the reverb algorithmn, or in the DAW mix bus is the only difference.

But I agree that if your DAW sends are defaulting to post fader, then you can't achieve 100% wet with Send. Impulses that are designed to be 'Pipline' impulses require 100% wet, with no dry mixed in at all.

If you are layering a reverb with send, and mixing in some dry sound within the reverb - and IF your DAW lacks plugin delay compensation (it's not Protools is it?) then you are going to get phase issues, which might explain the difference in eq you are hearing ...

Thanks to you and other posters for helping me straighten that out. I rechecked and the differences had to do with not muting the other effects on the channel and not having my wet at 100% on send and 50% on insert. I deleted some of my previous posts to save people from wasting time. I also put a note on my original post.

However, I'm still interested in the original idea which was how to get the most out of using reverbs on send channels by combining with other techniques. Also, there is that handy little pre/post switch on the send channel which I think can have more than a few uses in the right situations. What are the right situations for you to choose post fader over pre-fader on an effects send.

[GuitarRuss]

Originally Posted by Benmrx

I'm compressing/eq'ing all the time on effects sends and/or returns. I love squashing the shit out of a reverb return, followed by some HP/LP filters.

I'll have to try that. What styles do you find that works for? The HP/LP filters and EQ are great because you can get so much more control that any of the built in controls that you will get on the reverbs. I never thought of compressing reverb...

[GuitarRuss]

Originally Posted by Wayne

But what you are describing is the difference between 100% wet and a percent wet/dry mix, not a difference of where the effect is done.
Mute the original (dry signal) in the 'send scenario and you will be back to one 'pitched tone. thumbsup

Thanks for the clarification. So basically just the insert when you want all of the signal processed for stuff like pitch/time fx and powerful convolutions like the telephone speaker where a 50/50 wet/dry isn't enough. This certainly simplifies things.

[santibanks]

Originally Posted by Kiwiburger

I use Cubase4 and I frequently test reverbs as an Insert effect over a VSTi track (which is stereo and therefore works).

What I find is that often the amount of reverb I need is about 1% or 2% before it sounds excessive. If the plugin mix can only select to the nearest percent, then that's a very coarse adjustment. Only as issue with certain plugs.

I dont know if the levels are different with inserts and the S/R signal path.

but the thing is, s/r works a bit like a mixer. You can just use one setting for a reverb and apply it over more then 1 track. Just mix it in with the send knob to get the right dry/wet signal you are after.

You can only do this with inserts if the processor unit has a dry/wet facility. This sounds strange as you might think they all have but thats not the case. My Boss SE50 units don't have it as a knob. Have to tweak the presettings to achieve the effect (though I use them more for live then for studiowork)

[Benmrx]

Originally Posted by GuitarRuss

I'll have to try that. What styles do you find that works for? The HP/LP filters and EQ are great because you can get so much more control that any of the built in controls that you will get on the reverbs. I never thought of compressing reverb...

Pretty much everything I do could fall under the umbrella of "indie rock".