The case for delay after reverb for studio recording. Please add your experiences.

[Ed DeGenaro]

Parallel or delay/verb for me.
But I try to avoid verb whenever I can

 

[DLC86]

This thread is rife with bullshit statements that are basically just saying:

"it doesn't really make a difference, except for the times when it does"


Which is just a free pass to get away with any kind of meaningless statement.

Took me 2 seconds to bounce out a clip - delay with modulation going before a reverb vs after. When the reverb is before, you hear the reverb tail shift around in a way that sounds absolutely shite (like what happens if you add chorus to an IR). When you have the reverb after the delay, the reverb tail isn't being thrown around in the same way.

Delay + Reverb Test

From Ed S

samply.app

Why would changing the order of processing cause an issue anyway? just do whatever needs to be done, its not like a limitation exists to begin with.

The only bullshit statement here is saying that it makes a difference using an extreme case to show you're right, that's the literal definition of cherry picking.
Saying in which cases it's exactly the same, in which cases it's so close that it doesn't matter and in which cases it does matter is just the representation of the reality we can all observe, and that's exactly what I did in my post and in the video... if you call this "bullshit statement" you have serious comprehension issues.

It's normal for the difference to be obvious in your sample cuz moving the delay time by half a second is closer to pitch shifting than to modulation, and it's one of the cases I already mentioned anyway: "extreme modulation".
Try with a guitar signal and a typical modulation of a few ms and let's hear if it's still so obvious, i.e. take a typical scenario without cherry picking, cuz no one uses a half second modulation on a snare.

I didn't watch it because it wasn't in English. But are you saying your null tests reveal that they don't null??

There are (non-auto)translated english subtitles in all my videos... anyway, here's the spoiler: expectedly they perfectly null when there are no non-linearities, they don't when there are instead.
But the point is: are you able to hear the difference when they don't? I bet in most non-cherry-picked cases you won't be able.

And by the way, with modulation, even two identical samples recorded with the same delay/reverb order and identical settings on both won't null (the "why" is explained in the video), does that mean they'll sound different?

 

[DLC86]

even on non-modulating/saturating delays/verbs

No, in that case there's no difference, the null test in my video proves it.

 

[DLC86]

I'm not saying that sometimes it won't matter which order they're in, but trying to push a narrative that it doesn't matter is 100% wrong

No one ever said that it doesn't matter in any case, re-read the entire thread if necessary.

 

[MirrorProfiles]

The only bullshit statement here is saying that it makes a difference using an extreme case to show you're right, that's the literal definition of cherry picking.

It’s no different to you cherry picking an outlier case where it backs you up. I’m not making claims about it being indistinguishable, and you’re saying that it’s only possible to make it so under very precise conditions. I am saying that requiring such conditions to prove the point is bullshit.

It's normal for the difference to be obvious in your sample cuz moving the delay time by half a second is closer to pitch shifting than to modulation, and it's one of the cases I already mentioned anyway: "extreme modulation".
Try with a guitar signal and a typical modulation of a few ms and let's hear if it's still so obvious, i.e. take a typical scenario without cherry picking, cuz no one uses a half second modulation on a snare.

This is broken watch territory. “try using a space echo/binson/memory man/2290/DM-2 m/PCM42 but don’t use any of the features or appeals of the device that might reveal it” They all have unstable pitch, or non linearities that will matter. And if they don’t, they offer the ability to take advantage of them.

I agree that sometimes it doesn’t matter which order they’re in, but most of the time it’s probably best to assume that it might make a difference. Then you can use the gear as it was intended.

 

[MirrorProfiles]

and it’s not just pitch modulation that’ll easily expose it. Look at the modulation options on a 2290 (panning, volume, delay time) - you’ll hear if that’s happening before a reverb or after. Think of multi tap delays each with their own modulations and panning. It matters.

 

[MirrorProfiles]

strymon/eventide/meris pedals

LOL at anyone using these purely for basic/subtle delays. we’ve had literally decades worth of R&D to make delays that do so much more than just a straight delay.

Companding, modulation, LFO’s, envelopes, panning, pitch shifting, chorusing/flanging, freely messing with delay time (a la EP-3 or tape machines with a variac), bit crushing. And that’s before we even get to different approaches to reverb which can be its own can of worms. Think how few reverbs are true stereo. So many classic ones like 224/PCM70 sum the inputs to mono, and unless they are true stereo, the stereo field placement is going to be affected. So again, it just massively limits the situation where it won’t matter.

Quite often complex delays will give you options on where (or even which delay taps) get processed with reverb. I guess these cases are also all cherry picking.

 

[DLC86]

It’s no different to you cherry picking an outlier case where it backs you up. I’m not making claims about it being indistinguishable, and you’re saying that it’s only possible to make it so under very precise conditions. I am saying that requiring such conditions to prove the point is bullshit.

Where did I cherry pick? This discussion started with you arguing that it's an oversimplification to say that with subtle modulation/distortion the difference is negligible. And to prove it you made a sample with extreme modulation applied on a snare. Ok...

This is broken watch territory. “try using a space echo/binson/memory man/2290/DM-2 m/PCM42 but don’t use any of the features or appeals of the device that might reveal it” They all have unstable pitch, or non linearities that will matter. And if they don’t, they offer the ability to take advantage of them.

If taking as an example a typical scenario (which means the default settings of 99% of delay pedals and algorithms or the typical behaviour of a real tape delay) is a "broken watch", then your sample is a clock with no hands at all, won't be right not even once a day.

 

[DLC86]

and it’s not just pitch modulation that’ll easily expose it. Look at the modulation options on a 2290 (panning, volume, delay time) - you’ll hear if that’s happening before a reverb or after. Think of multi tap delays each with their own modulations and panning. It matters.

If the delay and reverb are true stereo you won't hear the difference on those at all, panning and volume are LTI, thus commutative. About modulation, already answered several times.

 

[MirrorProfiles]

Where did I cherry pick? This discussion started with you arguing that it's an oversimplification to say that with subtle modulation/distortion the difference is negligible. And to prove it you made a sample with extreme modulation applied on a snare. Ok...


If taking as an example a typical scenario (which means the default settings of 99% of delay pedals and algorithms or the typical behaviour of a real tape delay) is a "broken watch", then your sample is a clock with no hands at all, won't be right not even once a day.

I’d just say that saying the order of delay and reverb processing making a negligible difference is only true in specific circumstances, and on the whole wouldn’t really benefit anyone to follow that assumption. The fact that saying the exact opposite is equally true just shows how little use it is:

“the order of delay and reverb processing WILL make a difference, except in some situations (depending on the gear and how it’s used).”

IMO it’s better to consider the order of processing than not.

 

[MirrorProfiles]

If the delay and reverb are true stereo you won't hear the difference on those at all, panning and volume are LTI, thus commutative. About modulation, already answered several times.

Exactly. So it assumes that these are a given. and often they are not, either by creative use by the user (which is why they’re editable), or as part of the design of the processor in use.

 

[DLC86]

I’d just say that saying the order of delay and reverb processing making a negligible difference is only true in specific circumstances, and on the whole wouldn’t really benefit anyone to follow that assumption.

What you call "specific circumstances" is actually the grand majority of cases in typical guitar-related scenarios. Or do you think that varying the delay time by half a second with an LFO is used as frequently as the subtle wow and flutter of a tape delay?

IMO it’s better to consider the order of processing than not.

Didn't ever say the contrary... and I explained the whole reason I brought this up in one of my first posts:

the misconception is that most people think those differences are caused by the fact that delaying a reverb and reverberating a delay are two different things, when they're actually the same exact thing.
It's like saying that 2x4 is different than 4x2.

In case it's not clear, in that bit, with the words "delay" and "reverb" I refer to the "raw" delays and reverbs, not to all the non-linearities you can add to them.

Exactly. So it assumes that these are a given. and often they are not, either by creative use by the user (which is why they’re editable), or as part of the design of the processor in use.

I didn't assume anything! Stop putting words in my mouth and thoughts in my head!
You made an absolute statement ("you’ll hear if that’s happening before a reverb or after") and I basically just said "it depends".

 

[MirrorProfiles]

What you call "specific circumstances" is actually the grand majority of cases in typical guitar-related scenarios. Or do you think that varying the delay time by half a second with an LFO is used as frequently as the subtle wow and flutter of a tape delay?

A Roland Space Echo at any settings is going to have pretty noticeable pitch artefacts. Same with an Echorec. Most people using an EP-3 are setting the feedback high and mangling the delay time to do weird fx. And I’d imagine anyone who has bought a 2290 has done so specifically so they can modulate the panning or volume or alter the envelope, or modulate the delay time enough to cause artefacts. Someone with an H3000 is probably altering the pitch and modulating all kinds of parameters. A memory man barely has any knobs and even still it’s very easy to cause obvious pitch changes.

Anyone using a stereo delay without using a true stereo reverb is going to hear a difference. As I said, reverbs like 224’s or PCM70’s or RMX16’s etc sum the inputs to mono.

I don’t think any of those are uncommon pieces of gear, nor used for them.

In case it's not clear, in that bit, with the words "delay" and "reverb" I refer to the "raw" delays and reverbs, not to all the non-linearities you can add to them.

This is essentially “delaying the signal” vs using an actual delay.

I get exactly what you’re saying but in practice the number of exceptions mean it’ll always depend. And musicians LIKE to use gear that is non linear and modulates in all kinds of ways.

 

[KingsXJJ]

I’m beginning to be sorry that I asked.

 

[DLC86]

A Roland Space Echo at any settings is going to have pretty noticeable pitch artefacts. Same with an Echorec. Most people using an EP-3 are setting the feedback high and mangling the delay time to do weird fx. And I’d imagine anyone who has bought a 2290 has done so specifically so they can modulate the panning or volume or alter the envelope, or modulate the delay time enough to cause artefacts. Someone with an H3000 is probably altering the pitch and modulating all kinds of parameters. A memory man barely has any knobs and even still it’s very easy to cause obvious pitch changes.

Anyone using a stereo delay without using a true stereo reverb is going to hear a difference. As I said, reverbs like 224’s or PCM70’s or RMX16’s etc sum the inputs to mono.

I don’t think any of those are uncommon pieces of gear, nor used for them.

This is essentially “delaying the signal” vs using an actual delay.

I get exactly what you’re saying but in practice the number of exceptions mean it’ll always depend. And musicians LIKE to use gear that is non linear and modulates in all kinds of ways.

Ok, here's a blind test, for each sample number (1,2,3,4) tell me which one is delay>reverb or reverb>delay between A and B.

Delay Reverb order test – Google Drive

drive.google.com

Should be pretty easy since I used a bit of drive, compander and modulation on the delay and modulation on the reverb too, right?
The modulation was not even so subtle since depth was set to 100% in the delay, an amount you won't find on any well-maintained tape delay (for sure not on the binson that I've played in the past).
The reverb model was the Gold Plate at default settings.

For anyone with a fractal, after the results I can also share the preset in case you wanna make a comparison while playing, just by switching the multiplexer channel on scene 1.

 

[Orvillain]

There are (non-auto)translated english subtitles in all my videos... anyway, here's the spoiler: expectedly they perfectly null when there are no non-linearities, they don't when there are instead.
But the point is: are you able to hear the difference when they don't? I bet in most non-cherry-picked cases you won't be able.

I don't read translations for anything except J-pr0n. Sorry.

Which algorithms did you manage to get perfect nulls with??

 

[DLC86]

Which algorithms did you manage to get perfect nulls with??

All of them if you turn off drive, modulation, compander and pitch shifting

 

[DLC86]

I don't read translations for anything except J-pr0n. Sorry.

Learn italian then you lazy ass

 

[Orvillain]

@DLC86 please articulate your central assertion.

 

[Orvillain]

Learn italian then you lazy ass

No. I don't like the Italian language. It sounds like gobbling dicks.