The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Question about gain staging in plug-in chain
Old 4th March 2019
  #1
Gear Head
 

Question about gain staging in plug-in chain

There seem to be different schools. Some say that if a plug-in (one that is not emulating hardware) clips, it's fine due to floating point, and others don't want any plug-in to go red even if it doesn't on the channel.

Currently I'm stuck somewhere in between gain staging as I am producing, which sometimes cripples the flow I'm in. Other times I'm actually exporting the multitracks from Ableton and am even contemplating getting Mixbus to do the mixing in there. I look at VU on the individual channels and LUFS on the master but I don't really know why.

I guess there's no right and wrong, but how do you view gain staging? Do you let your plug-ins clip? If so, why?

Thanks
Old 4th March 2019
  #2
Lives for gear
Quote:
Originally Posted by Sonny Casanovas View Post

I guess there's no right and wrong, but how do you view gain staging? Do you let your plug-ins clip? If so, why?
I view gain staging as using EQ. Why not use it? Is there something to lose? I don't think so.

As you can typically always add gain when you've gain staged, in a 32/64-bit float world there is seldom something you lose. Its harder to prevent clipping or non-optimal levels in the processing ("signal") chain after the fact. Simply because even if you can lower the master fader, if there's a problem earlier in the processing, you need to go back and fix it. If you set the levels when you start out and maintain them except for when intending not to, this is seldom needed. This is especially true for when you record to 24bit audio, in Live or when exporting. If anything clips at this point, you've committed that. If you gain stage you have already prevented this.

In addition, when you maintain those levels between processing stages you won't be lured by a simple loudness increase that your treatment is beneficial.
Old 4th March 2019
  #3
Gear Head
 

Quote:
Originally Posted by Mikael B View Post
I view gain staging as using EQ. Why not use it? Is there something to lose? I don't think so.

As you can typically always add gain when you've gain staged, in a 32/64-bit float world there is seldom something you lose. Its harder to prevent clipping or non-optimal levels in the processing ("signal") chain after the fact. Simply because even if you can lower the master fader, if there's a problem earlier in the processing, you need to go back and fix it. If you set the levels when you start out and maintain them except for when intending not to, this is seldom needed. This is especially true for when you record to 24bit audio, in Live or when exporting. If anything clips at this point, you've committed that. If you gain stage you have already prevented this.

In addition, when you maintain those levels between processing stages you won't be lured by a simple loudness increase that your treatment is beneficial.
Thanks for answering Mikael. This brought me to a side note. How do you do it? Are you keeping an eye on the fader or do you use some other unit? Because as I mentioned I look at the VU and depending on the source, 0 VU differs enormously. Sometimes I'm almost clipping early in the signal chain so I adjust to land on a negative number that is close to where I should be. If I then add gain with the utility tool I don't get the channel fader and the VU meter to coincide. Should I be looking at lufs instead? Or ditch nthe VU altogether and strictly have my eyes on the dB scale in Live?
Old 4th March 2019
  #4
Gear Guru
 
Brent Hahn's Avatar
 

As a sidebar to this, someone recently suggested a comparison between the saturation behavior of a hardware tape simulator vs that of a tape-sim plugin, with both being hit with a +9 input level. I don't know the first thing about software design, but is +9 into a plugin even possible?
Old 4th March 2019
  #5
Lives for gear
Depends on how the metring is set up. Is -20 in the DAW the same as 0 in the analog world?
Old 4th March 2019
  #6
Gear Guru
 
Brent Hahn's Avatar
 

Quote:
Originally Posted by elegentdrum View Post
Depends on how the metring is set up. Is -20 in the DAW the same as 0 in the analog world?
I don't know - I took it to mean 9 dB past "in the red" in the plugin, but maybe that's not what it means.
Old 4th March 2019
  #7
Lives for gear
 
stinkyfingers's Avatar
 

Quote:
Originally Posted by Brent Hahn View Post
I don't know the first thing about software design, but is +9 into a plugin even possible?
Floating point allows for hundreds (or thousands, for double) of dB over 0 dBFS.
.
Old 4th March 2019
  #8
Lives for gear
 

Side benefit of Unity Gain Staging ....... no level change when bypassing the plugin. Allows more accurate evaluation of what that plugin is doing.
Old 4th March 2019
  #9
Gear Guru
 
Brent Hahn's Avatar
 

Quote:
Originally Posted by stinkyfingers View Post
Floating point allows for hundreds (or thousands, for double) of dB over 0 dBFS.
So can I take that to mean that "floating point" takes an incoming signal that's hotter than 0 and treats it as if it's not? I admit to not having a clue about the whole floating point thing; I've always worked in PT where you simply don't push plugins into the red or things get ugly.
Old 4th March 2019
  #10
Lives for gear
 

Quote:
Originally Posted by Brent Hahn View Post
So can I take that to mean that "floating point" takes an incoming signal that's hotter than 0 and treats it as if it's not? I admit to not having a clue about the whole floating point thing; I've always worked in PT where you simply don't push plugins into the red or things get ugly.
In PT starting with the new audio engine which I believe was v11 it too got floating point processing as opposed to the older TDM system which was fixed.

With float the signal sort of fits within 24-bits but those bits can then scale up and down, if that makes sense. I honestly don't know of a better way to describe it, but "hotter than 0" sort of is and isn't the right way of looking at it, at the same time.

So looking at a channel meter other than one for an output "0dB" I think will often mean that if you send that signal to a physical output you're pushing the conversion from float to fixed and then from fixed to analog to the maximum value possible. If that channel meter goes above 0dB then you will indeed be clipping the output.

However, because floating point processing "scales", or "floats", you can push the signal beyond your 0dB on oyour channel meter and then scale it back by lowering it later on the master output. That's why (ignoring plugins) you can have your channel fader push signals to say "+10" and if you then lower the master output fader to -15 your final signal will sit at -5dBFS and will not clip, neither will it have clipped before that output fader.

Quote:
Originally Posted by Brent Hahn View Post
I don't know - I took it to mean 9 dB past "in the red" in the plugin, but maybe that's not what it means.
I did a test once using my hardware Neve 33609 compressor and the UAD-2 version, and what I did was to level match the two and then push them equally hard. So the above could indeed mean past "in the red", but it could also mean +9dBVU. If it's +9dBVU and the plugin is built to emulate the hardware then it's possible that 0VU on the plugin equals something like -18dBFS, which in turn means that when you read +9dBVU on the plugin the actual value in the channel is -9dBFS, so you still have 9dB headroom before clipping.
Old 4th March 2019
  #11
Lives for gear
 

Quote:
Originally Posted by Sonny Casanovas View Post
Thanks for answering Mikael. This brought me to a side note. How do you do it? Are you keeping an eye on the fader or do you use some other unit? Because as I mentioned I look at the VU and depending on the source, 0 VU differs enormously. Sometimes I'm almost clipping early in the signal chain so I adjust to land on a negative number that is close to where I should be. If I then add gain with the utility tool I don't get the channel fader and the VU meter to coincide. Should I be looking at lufs instead? Or ditch nthe VU altogether and strictly have my eyes on the dB scale in Live?
If you're worried only about clipping then you should just stay away from 0dBFS throughout the signal chain. You're not losing anything by doing that.

The way to get it done without using too many braincells (a problem for me since I've got not good brain) is to just start off the mix process by playing back all channels and lower all faders while looking at the master output channel meter. Is it close to zero? Pull down the faders. Then when you're mixing you can simply adjust levels as you go along; if you compress an instrument you'll lower its level so use makeup gain to bring it back to where it sounds as loud as without compression. If you start boosting 4k on that electric guitar then in the EQ plugin window bring down the output so it again sounds equally loud, just with a different tonal balance.

This procedure is generally good, because you can then easily toggle the plugin on/off to hear what the actually desired effect does to the signal. If you also hear the level drop or increase you're more likely to think that "Oh, that 4k boost sounds nice" where in fact you just liked that it got louder...

And so by going through the process this way you actually end up staying pretty close to the same levels throughout the mix process and staying away from clipping.

With a decently set up monitoring environment (i.e. interface and speakers etc) you should be getting loud enough playback.

If your plugins actually do emulate some sort of analog non-linear behavior like saturation then check the plugin manufacturers notes on when that happens. If the plugin has a virtual VU meter for example then it's probably going to start saturating the signal noticeably more above 0VU. Check to see where that plugin's 0VU is on the absolute digital scale. Chances are it will be around -18dBFS average. So this means that a sinewave that is -18dBFS should read 0VU in that plugin. Push the sinewave beyond that level and you should then get more non-linear behavior.

PS: I don't use Ableton, but I would imagine the same should apply.
Old 4th March 2019
  #12
Gear Head
 

Quote:
Originally Posted by mattiasnyc View Post
If you're worried only about clipping then you should just stay away from 0dBFS throughout the signal chain. You're not losing anything by doing that.

The way to get it done without using too many braincells (a problem for me since I've got not good brain) is to just start off the mix process by playing back all channels and lower all faders while looking at the master output channel meter. Is it close to zero? Pull down the faders. Then when you're mixing you can simply adjust levels as you go along; if you compress an instrument you'll lower its level so use makeup gain to bring it back to where it sounds as loud as without compression. If you start boosting 4k on that electric guitar then in the EQ plugin window bring down the output so it again sounds equally loud, just with a different tonal balance.

This procedure is generally good, because you can then easily toggle the plugin on/off to hear what the actually desired effect does to the signal. If you also hear the level drop or increase you're more likely to think that "Oh, that 4k boost sounds nice" where in fact you just liked that it got louder...

And so by going through the process this way you actually end up staying pretty close to the same levels throughout the mix process and staying away from clipping.

With a decently set up monitoring environment (i.e. interface and speakers etc) you should be getting loud enough playback.

If your plugins actually do emulate some sort of analog non-linear behavior like saturation then check the plugin manufacturers notes on when that happens. If the plugin has a virtual VU meter for example then it's probably going to start saturating the signal noticeably more above 0VU. Check to see where that plugin's 0VU is on the absolute digital scale. Chances are it will be around -18dBFS average. So this means that a sinewave that is -18dBFS should read 0VU in that plugin. Push the sinewave beyond that level and you should then get more non-linear behavior.

PS: I don't use Ableton, but I would imagine the same should apply.

Great, thanks! One other issue aside from just clipping is that I notice the high end in my mixes often get lost if I just look at the vu and dB meters. It seems that a hat doesn't have the same impact on the vu scales as a bassline for example. I guess I'm asking if there's some
"rule of thumb" to follow except for mixing with my ears rather than my eyes.
Old 4th March 2019
  #13
Lives for gear
 

Well I think it's a mistake to balance the mix using meters. That's just way, way the wrong way. Like you said, use your ears.

And 'yes', different types of content will feel and look differently on the same meter. "Gain staging" is really just a way to get to a point where your signals don't distort or do something 'funny' without you wanting them to do that. So that's why I said you can often pretty much just drop the entire mix while looking at the master output and stay away from zero, and take it from there. Of course that's assuming you've already done a sort of very rough balance of all instruments. It really doesn't have to be perfect, but just a rough balance.

But to reiterate: It's really two separate things - to set your levels to avoid clipping and allow some plugins to work properly / to set your levels so that instruments are balanced against each other in a way that's aesthetically pleasing to you (i.e. "mixing").

And I suppose I should have pointed this out:

If you lower faders so that the master doesn't clip that obviously won't help levels going into your plugins, since your plugins are pre-fader normally. So, the other way of going about it is to simply lower clip/event-gain (or whatever it's called in Ableton) or use a "trim" function at the beginning of the signal chain before inserts (could be built-in like in Cubase, or could be a plugin).

So there are really two main areas to look at when setting just some basic gain, and that's before or after your inserts.

But "mixing is mixing", and you shouldn't worry about what a hihat reads on a VU meter relative to a bass if the balance is right and you have no issues with clipping or plugins.
Old 6th March 2019
  #14
Lives for gear
Quote:
Originally Posted by Sonny Casanovas View Post
Thanks for answering Mikael. This brought me to a side note. How do you do it? Are you keeping an eye on the fader or do you use some other unit? Because as I mentioned I look at the VU and depending on the source, 0 VU differs enormously. Sometimes I'm almost clipping early in the signal chain so I adjust to land on a negative number that is close to where I should be. If I then add gain with the utility tool I don't get the channel fader and the VU meter to coincide. Should I be looking at lufs instead? Or ditch nthe VU altogether and strictly have my eyes on the dB scale in Live?
I used to just set up the levels with a meter & fader plug-in (trim) first thing and set this or the audio source to hit my decided RMS target, then maintain this level going in and out processing, unless I intend to change it. I ignore peaks totally as I'm controlling that later.

Since a few weeks I have Hornet's TheNormalizer inactivated on every track in my start template. The instances in the different tracks are set to the same group per instrument type, so all of the same type will analyze when I hit one of them and automatically set the pre-determined target level. This way all my faders can be on about the same position when starting a mixdown session. The effort with all this is minimal, so little money well spent.

Last in each track I have that meter plug-in that I glance at to see that I'm close when processing. I don't really trust Live's meters, but I need to give the new 10 meters a chance.

Last edited by Mikael B; 7th March 2019 at 01:07 AM..
Old 6th March 2019
  #15
Lives for gear
Quote:
Originally Posted by mattiasnyc View Post
Well I think it's a mistake to balance the mix using meters. That's just way, way the wrong way. Like you said, use your ears.
There's certainly truth in that, but sometimes you need to see it in order to "get it". And some issues that can become problems later you may not hear necessarily. It's however indeed a mistake to "mix with your eyes" in majority of situations, so I'm with you on this.

Last edited by Mikael B; 7th March 2019 at 01:08 AM..
Old 9th March 2019
  #16
Gear Maniac
 

Don't let things clip. No, it doesn't matter in floating point but eventually you're going to have to turn it down because once it goes out to your sound card it will not sound good and summing clipped signals results in a clipped signal. You can at this point just turn down the master fader but since it's super easy to just not let any channel clip even between plugins (really, this requires minimal effort at most, especially if you're doing level matching before/after the plugin which you should be) why not do it there?

Avoiding clipping does not require any specific levels and certainly not as low as the levels commonly advocated by those in favor of analog style gain staging. I just stay an arbitrary number of db below zero. I don't even know what the number is most of the time (usually when I check it's somewhere between -1 and -3 but can be lower.) The meters tell me when I'm clipping and I back off.

An important thing to remember about clipping not happening in floating point is that this only applies in the box. If you record an external signal that is clipping, that signal is clipped. You may be recording to 32-bit but your audio input is going to be 24-bit integer. It's only ok to clip after you're in the floating point realm. But as said already, rather don't clip anywhere.

Beyond that - does a plugin manual say the plugin works best with a particular input level?

Yes - do what it says.
No - don't worry about it. Choose an input level that sounds good to you.

No one has ever presented a technical explanation of why an analog emulation would need gain staging by default. Usually there's some mumbling about how the analog version needs gain staging and if the model is accurate then so does the emulation. But I don't buy that. Gain staging in analog is needed due to the fact that when driven too hard the circuits misbehave. That is, they stop working in the expected way (the way that would be modeled) and start overloading in unpleasant ways. This is not the same as behaving non-linearly.

Saturation, which is a type of overload behavior and one intentionally modeled, is not the same thing. If it were then no one would use saturation in analog. It would be as important to avoid as overloads due to incorrect gain staging. Saturation is something that happens while circuits are still within their tolerance but have been pushed hard enough to generate pleasant distortions.

Now I could be wrong here. And if so then I eagerly await someone explaining why, using technical details of exactly what's happening in the analog device and why the modeled device would behave the same.

Volume matching between a plugin enabled and disabled is almost always important. I'm not fully sold on having to volume match all EQ adjustments. If you're doing a subtle boost in a particular frequency range I believe the argument of "louder sounds better even when it's not" doesn't apply, or at least is overstated. You've adjusted the frequency balance and that's the most important indicator of whether it sounds better. If doing large amounts of boost or over wide ranges then volume matching matters. So I don't worry about it usually when doing more subtle EQ.

But volume matching when using distortion, compression, saturation, this is very important.
Old 9th March 2019
  #17
Lives for gear
 

Quote:
Originally Posted by strange loop View Post
Gain staging in analog is needed due to the fact that when driven too hard the circuits misbehave. That is, they stop working in the expected way (the way that would be modeled) and start overloading in unpleasant ways. This is not the same as behaving non-linearly.

Saturation, which is a type of overload behavior and one intentionally modeled, is not the same thing. If it were then no one would use saturation in analog. It would be as important to avoid as overloads due to incorrect gain staging. Saturation is something that happens while circuits are still within their tolerance but have been pushed hard enough to generate pleasant distortions.

Now I could be wrong here. And if so then I eagerly await someone explaining why, using technical details of exactly what's happening in the analog device
You can pick up a manual most likely and it'll tell how a particular device functions and what the limitations are. Typically historically a large amount of gear was made to be clean. The 'linear' behavior of an amp was you push a signal into it, add gain to it, and the signal comes out essentially the same but louder. The non-linear behavior happens when it's driven to the point of outputting something that is no longer the same but louder and instead has added junk in it. You can call that saturation if you want, or "misbehaving", or "overloading", it doesn't matter; from the standpoint of input =/= output it's "non-linear" and "distorting".

It's really just a semantics game in this context. Take your pick.

Quote:
Originally Posted by strange loop View Post
and why the modeled device would behave the same.
Because we don't model analog devices to simply input a clean signal and output the exact same thing without the "junk". The whole point of modeling analog is to capture everything that isn't accurate and clean. If there's an amp in real life that's essentially linear through 0VU and up +10dB or whatever, but then starts to generate harmonics, then why would we model that amp without those harmonics? If all we wanted was clean amplification we'd just pull up the fader in our DAW 10dB. Done.

No, what we want is everything that the analog device added, and if we liked the distortion characteristics (or call it "saturation" or whatever) then that's what we model. But unless the device was specifically one designed to saturate, then we really still are often looking at a range where it performs nominally and with minimal distortion and above that distortion begins.

Of course you could argue that something like a tape-emulation plugin always distorts the signal in the sense that input never equals output. And that makes sense. Of course it should add something at moderate levels, if it didn't why use it? But a maker of such a plugin would probably not program it to "distort linearly", or "consistently". Instead I'm betting they'd program it so that there's a hint of tape character imparted on the sound at all levels, but when pushing the signal beyond a certain point distortion increases non-linearly, because that's how the analog device worked. So going from -30dBFS to -20dBFS adds far less distortion to the signal than going from say -13dBFS to -3dBFS, even though the amount of gain is the same.

You can look at the UAD-2 software manual for the Studer tape deck plugin for example and it'll tell you how it's "calibrated".

But I have a feeling you likely won't accept anyone's explanation for why a plugin would be coded to act one way through a virtual 0VU at -18dBFS (for example) and then distort more and non-linearly beyond that, so..
Old 9th March 2019
  #18
Gear Maniac
 

I have an analog modeled plugin where the manual says to use it with particular input levels for optimum behavior and I follow those instructions. I have others which do not.

This is not about non-linear distortion, something my first post made it clear I am aware of and why we would want it. This is about the claim that if you don't use 0dBVU input (or whatever arbitrary level) to all analog modeled non-linear processing plugins, even when the designer doesn't say you need to, you are damaging your sound.

Your explanation of the difference in distortion when applying the same internal gain to different input levels is good and fits my observations of the difference in effect when sending in different levels and applying the same level of boost inside the plugin.

So to move this forward I'd like to ask - if I go in at -18dBFS and apply 16dB of internal gain why is this better** than going in at -8dBFS and applying 6dB of internal gain? There may be differences due to the higher input level. But are we sure about that? Ultimately the same level of -2dBFS is being pushed through the circuit. Perhaps you could expand on your explanation and give some detail on why this would be significantly different.

** Note that when I say "better" I am again specifically referring to the common gain staging claim that doing the latter inherently damages your sound in undesirable ways as opposed to the desirable effects of saturation. Also note that this is a different question to "why would this be significantly different?" Different is not necessarily better/worse.

Even assuming that the non-linear behavior means more distortion at the same overall gain level for a higher input signal with lower internal gain, I should be able to back off on the internal gain to say 2dB or 3dB (just throwing out random numbers here, not saying those would be the appropriate levels) and achieve a similar level of distortion.

I'm also curious as to how this applies to analog modeled compressors, which I know do have some non-linear distortion but not at even close to the levels of a saturation/distortion effect.
Old 9th March 2019
  #19
Lives for gear
 

I use many 'analog' based plugins. -18db is common. [I've even gone to -23 LUF for some projects recently].

Bottom line I'd say. There is other aspect to Gain Staging [or rather UNITY Gain Staging], than just driving at proper [nominal] levels.

When A plugin is bypassed, we don't want any volume level difference [the ole 'louder sounds better' is a deception from what is actually happening].

Second would be consistent workflow.

If your Source Track averages -18dB, it will hit your first plugin at nominal level [0dB].

Unity Gain Stagings says IN=OUT. So if the Plug ADDs or SUBtracts Gain ... adjust OUTPUT back to UNITY.
You are ready for the next plugin.

Even if you change the ORDER in the Chain ... that is not a problem. Bypassing is not a problem.

note: Even the 'Digital' plugins can be used at that SAME Level.

If you want to DRIVE the Plugin ... do it with the Plugins INPUT section ... then rebalance the Plugs OUTPUT to UNITY.

Your signal path will be balance/ organized and flexible.

It'll help keep SANITY during the Session.
Old 9th March 2019
  #20
Lives for gear
 

Quote:
Originally Posted by strange loop View Post
This is not about non-linear distortion, something my first post made it clear I am aware of and why we would want it. This is about the claim that if you don't use 0dBVU input (or whatever arbitrary level) to all analog modeled non-linear processing plugins, even when the designer doesn't say you need to, you are damaging your sound.
And who here has made that claim?

Nobody.

Quote:
Originally Posted by strange loop View Post
if I go in at -18dBFS and apply 16dB of internal gain why is this better** than going in at -8dBFS and applying 6dB of internal gain? There may be differences due to the higher input level. But are we sure about that? Ultimately the same level of -2dBFS is being pushed through the circuit. Perhaps you could expand on your explanation and give some detail on why this would be significantly different.
Right, but you have to consider what is being modeled and where gain is applied.

If you're looking at a tape saturation plugin that doesn't have some sort of variable "saturation" knob then the amount of saturation distortion is most likely going to be related to the input level. So if you're asking if it makes a difference between the two above you have to look at where the gain is applied if it's not before the plugin. If it's after the modeled tape heads then 'no' it won't be the same thing.

If you're looking at an amp model then it's possible, or likely, that what's modeled is the amplification circuit. In that case it would seem to me that it is the gain in the plugin that gives you saturation, not necessarily input level. Of course, that doesn't mean that there aren't other circuits modeled, like an EQ section (in the case of a 1073 for example) and an output adjustment.

But regardless of the above; IF a plugin is coded to add distortion that we like, crunch or harmonics or whatever you want to call it, then I argue that as long as it's not a dedicated distortion plugin (like a modeled guitar pedal) then it's going to behave like the analog device it's modeled after, and likely that device is going to have a nominal operating level, and the question then is what that corresponds to in the digital domain. It seems many 'calibrate' to about -18dBFS (higher in some cases).

Quote:
Originally Posted by strange loop View Post
** Note that when I say "better" I am again specifically referring to the common gain staging claim that doing the latter inherently damages your sound in undesirable ways as opposed to the desirable effects of saturation.
How is that a common claim? I see you keep saying this, yet I don't recall seeing anyone saying it..

Quote:
Originally Posted by strange loop View Post
Even assuming that the non-linear behavior means more distortion at the same overall gain level for a higher input signal with lower internal gain, I should be able to back off on the internal gain to say 2dB or 3dB (just throwing out random numbers here, not saying those would be the appropriate levels) and achieve a similar level of distortion.
That depends on where that "internal gain" is. If it's after the modeled circuit that distorts the signal then "no", you won't "undo" the distortion.

Quote:
Originally Posted by strange loop View Post
I'm also curious as to how this applies to analog modeled compressors, which I know do have some non-linear distortion but not at even close to the levels of a saturation/distortion effect.
It's probably entirely dependent on the individual plugins. "Read the manual" is key here. Fortunately it doesn't take long.
Old 9th March 2019
  #21
Gear Maniac
 

It's true that no one made that claim in this thread. I thought it reasonable that we did not need to stick to only the statements made in this thread.

If it helps, forget about where the claim that "You must always hit an analog modeled plugin at 0dBVU" comes from and just address the statement itself (which I think you have even if not intending to.)

The context of my post should have made it clear the internal gain I am referencing is not after the circuit since I stated that it is affecting the gain pushed through (i.e. into) the circuit. Possibly wasn't as clear as intended.

I also referenced the fact that of all my plugins which do some kind of saturation or distortion only one manual mentions a need for a specific input gain (and gives a meter and gain knob in the plugin to control this.) I hope you are now convinced that I do in fact pay attention to the contents of manuals.

Never seen a compressor manual which says it's required.

Quote:
Originally Posted by mattiasnyc View Post
If you're looking at a tape saturation plugin that doesn't have some sort of variable "saturation" knob then the amount of saturation distortion is most likely going to be related to the input level.
This is fair. I don't have any saturation plugins, including my tape saturation plugin, that don't have a gain knob before the saturation circuit or a saturation drive knob but I expect there are some, perhaps many, that don't. Mine isn't just for tape saturation, it does tube and others as well, so dedicated plugins are perhaps different.

Quote:
If you're looking at an amp model then it's possible, or likely, that what's modeled is the amplification circuit. In that case it would seem to me that it is the gain in the plugin that gives you saturation, not necessarily input level. Of course, that doesn't mean that there aren't other circuits modeled, like an EQ section (in the case of a 1073 for example) and an output adjustment.
That matches my expectation and I believe answers my question. It is not necessary by default to hit the plugin inputs at -18dBFS or whatever matches 0dBVU. What's important is the overall level you drive through the circuit. And this is, in my experience, controllable once you're inside the plugin. I haven't seen a saturation or distortion device that does not contain a gain/drive knob which controls how hard you push the saturation/distortion.

If a plugin manual states that it is calibrated to a lower level I'll use that lower level. Else

Quote:
It seems many 'calibrate' to about -18dBFS (higher in some cases).
I will feel safe in assuming that since they did not say they've calibrated to a particular value that their plugin works at all reasonable levels (and as I said in my first post I am a firm believer in staying under 0dBFS.)
Old 13th March 2019
  #22
Lives for gear
Quote:
Originally Posted by strange loop View Post

No one has ever presented a technical explanation of why an analog emulation would need gain staging by default. Usually there's some mumbling about how the analog version needs gain staging and if the model is accurate then so does the emulation. But I don't buy that. Gain staging in analog is needed due to the fact that when driven too hard the circuits misbehave. That is, they stop working in the expected way (the way that would be modeled) and start overloading in unpleasant ways. This is not the same as behaving non-linearly.

Saturation, which is a type of overload behavior and one intentionally modeled, is not the same thing. If it were then no one would use saturation in analog. It would be as important to avoid as overloads due to incorrect gain staging. Saturation is something that happens while circuits are still within their tolerance but have been pushed hard enough to generate pleasant distortions.

Now I could be wrong here. And if so then I eagerly await someone explaining why, using technical details of exactly what's happening in the analog device and why the modeled device would behave the same.
The modeled device doesn't necessarily behave the same due to aliasing, which the analog modeled device simply does not have. This is ridiculously true with saturation and distortion. Well behaved devices will make this a non issue with oversampling. Others will not. Input levels may very well work out very differently with at least some plug-ins. Most if not all dynamic effects that behave differently with different entry levels will not null with a level adjusted copied signal chain (So the compared "signals" have the same levels after processing).

Gain staging is just something I do. I think this simple thing changed digital audio for me.
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
Saturations / Newbie Audio Engineering + Production Question Zone
4
d_q8r / Newbie Audio Engineering + Production Question Zone
2
smokingrobot / Electronic Music Instruments and Electronic Music Production
0
redgrovesound / So Much Gear, So Little Time
5

Forum Jump
Forum Jump