Input Voltage/Gain Calibration for Guitar Plugins

[Ilya-v]

Most guitar players that use audio interface have absolutely no clue how to set their input gain for most realistic behavior from their VST amp simulations.
This has to stop, every guitarist should know that 1VAC Peak Analog = 0.0dbFS digital, for practically all amplifier simulations based on SPICE, that includes commercial software like Amplitube, ReValver, TH3, S-Gear, Mercuriall, Kuassa, etc.. and freeware like LePou, AcmeBarGig, Ignite, etc.

Here are few threads on other amp sim forums that mention this: Link 1, Link 2
EDIT: On page 4 of this thread there is proof from several developers that this is indeed the case.

AxeFX II is ever calibrated so a real world 1V is translated to 1V in the digital domain, that's why input level adjustment does NOT affect the amount of gain the simulated amplifier has and as a result sounds accurate with any guitar.
On the other hand, VST Plugins input level DO affect the amount of gain the simulated amp has.

The question is what is the correct input level???
The audio interface that I plug my guitar into has a Gain knob, the amp sim has an input volume level...
How do I know that my telecaster single-coil pickups don't behave like a Blackout/EMG in the digital domain?
The answer CALIBRATION.
Most Guitar VST plugins expect a 1VAC Peak (1kHz) in the real world to be full scale 0.0dbFS in the digital domain (in your DAW).
So we calibrate...

To calibrate I used a function generator to generate 1kHz 1VAC Peak (0.707V RMS) analog tone and feed that into the audio interface to calibrate the input gain on the Audio Track in the DAW.
It works like magic and all amp simulations sound like they were intended too, with proper amounts of gain that match their real world counterparts.

With 0.0dbFS input ReValver 4 will output 0.0 dbFS if the Pickup RMS is set to 0.707V, which actually means ReValver is at unity gain at 0.707.

1kHz 1VAC Peak Sinewave (0.707 AC RMS) at the input of the audio interface set to High-Z should peak at -0.1dbFS (just below clipping) in your DAW for amp simulations to sound close to the real amplifier, especially commercial ones like Amplitube, Revalver, S-Gear and TH3.
The reason for that because SPICE (including LTSpice and similar) treats the full digital scale as exactly 1VAC Peak.

How to calibrate:
1. Set you audio interface to High-Z and keep the gain pot on the audio interface at minimum, this maximizes headroom and minimizes noise floor.
2. Plug a function generator set to 1kHz Sinewave 1VAC Peak (0.707vAC RMS if you measure it with a DVM), into the audio interface.
3. Boost the digital input gain (NOT the interface Gain pot) or use an EQ plugin on your Audio Track till it reaches -0.1dbFS.

EDIT:
It completely slipped my mind, but you can use your audio interface output and just play a 1kHz sinus test tone WAV you can find easily on the web, or even a youtube 1kHz video, then loop it back with a cable to the Input.
Carefully adjust the output till you read exactly 0.707VAC with your DVM, loop it back to your Instrument Input then just boost in your DAW till you reach 0.0dbFS.

That's it, your calibrated!
Now load your favorite VST amp sim AFTER the EQ you used to boost the input signal.
The reason we don't use the input level inside the plugin because in your DAW you can actually see when the meter peak hits -0.1dbFS.

Remember by how much db you've digitally boosted the input, this is the default setting with your audio interface with the gain pot at minimum for 1VAC Peak unity gain.
I had to digitally boost around 8db to reach 0.0dbFS with a 1kHz 1VAC Peak at the input of my audio interface.
You can change guitars without worrying if you set the correct input gain in the amp simulator or on your audio interface, just remember the digital db boost value and zero on the audio interface gain pot on the High-Z input.
After calibration the amp simulations sounds fantastic and behaves accurately as intended.

Now the tech savvy guitarists among us with a function generator can calibrate their audio interfaces for proper input voltage.
A cheap function generator from ebay and a Digital Volt Meter (you should already have) is all you need to calibrate.

[Ilya-v]

Removed, Not important

[jacotzen]

Ciao Ilya thank you so much for keeping this topic 'alive' because it seems like even the developers/marketers are not too much interested into giving the right info on this topic, resulting in a huge amount of threads full of opinions totally biased by a bad gain structure. Since I'm pretty sure that if set properly anyone will hear no difference in an album context, I do believe that the major impact on a bad gain structure is on hi-gain distortion.
A couple of questions:
1. Is there another way to calibrate (avoiding the function generator)? Maybe by audio interface tech specs?
2. How come you suggest to keep the gain knob at minimum on your interface when generally is said to keep it close to the red light (and actually it is intended to work like this by the manufactures)?

[Ilya-v]

Hi!

Yeah the developers don't care about proper gain staging mainly because you have to use an external reference for calibration, yet on the other hand they pride themselves of having "hyper-realistic" amp simulation (Amplitube).
How in the world an amp simulation that's affected by input voltage can be realistic if every audio interface is different?
Without calibration a 50mV single-coil Telecaster can be as hot as a 3V EMG, that is 60 times the voltage!
If you adjust the gain pot on the audio interface, all guitars hit the red light making the equally hot.
Is a Telecaster equally hot as an EMG in the real world? No it's not, that's why we calibrate.

1. Unfortunately no, calibration is all about having an external reference, just like any other laboratory equipment calibration.

2. I keep the gain pot at minimum for maximum headroom of the audio interface for hotter pickups and I don't have to recalibrate each time I change the gain pot, it is mechanically stopped at minimum.
I only have to remember the amount of boost in the digital world for my audio interface.

[aciddose]

sqrt(1/2) = .707 = -3 dB.

So the trivial way to achieve this is simply play a note on your guitar and measure the typical level. Try finding the average and peak levels you produce.

Next, ensure you never clip the input on your interface by reducing the input gain to ensure your peak level is <0 dB.

Finally apply input gain on the channel in your DAW to get your peak (or average) level to sit at -3 dB.

For example you might have adjusted the input gain on your interface down -20 dB to ensure you avoid clipping the input playing heavy chords. The resulting peak gain might be -6 dB typically but the average of a single note might then end up near -12 dB. So you need +9 dB input gain in your DAW to make up for the -20 dB headroom you've used, resulting in your average single string sitting at -3 dB.

If your goal is to ensure the variations in inputs are reflected accurately you need to measure all the different devices you use and set your headroom and input gain based upon the loudest input device you use.

Setting up the interface to produce 0 dB for a 1v peak (RMS or peak? How does -3 dB enter into it?) signal is not a wise thing to do as you'll have no headroom left over and clip just about every device you input.

[jacotzen]

3. How do we cope with this issue when using other analog simulations like waves ssl?

P.S. My tascam us 144 has those specs:

Nominal input levels
MIC/LINE-GUITAR R in GUITAR position (1/4” jack, unbalanced) –51 dBu (TRIM=max) to –7 dBu (TRIM=min)

Maximum input levels
MIC/LINE-GUITAR R in GUITAR position (1/4” jack, unbalanced) +9 dBu (TRIM=min)

[Ilya-v]

Finally apply input gain on the channel in your DAW to get your peak (or average) level to sit at -3 dB.

Hi aciddose.
This defies the whole purpose of calibration and this thread.
The goal is to have a real world reference 1VAC without flying blind.
Measuring a guitar output with a DVM will not give anything remotely close to a real RMS value, because it is far from being a pure sinewave.

Setting up the interface to produce 0 dB for a 1v peak signal is not a wise thing to do as you'll have no headroom left over and clip just about every device you input.

On the contrary, you'll have the maximum headroom of your audio interface, and you are forgetting that DAWs work in floating point.
In your DAW you can boost 120db in the audio track then lower 120db on the master track while not clipping anything.

[aciddose]

Hi aciddose.
This defies the whole purpose of calibration and this thread.

No it does not.

The goal is to have a real world reference 1VAC without flying blind.

Are you talking about RMS or peak? The issue is during conversion and the amount of headroom you leave yourself. Setting up the interface to produce the maximum output with a 1v input is just plain stupid because the majority of input devices output far more than 1v peak levels. You'll end up with some negative amount of headroom and clip almost everything.

Measuring a guitar output with a DVM will not give anything remotely close to a real RMS value, because it is far from being a pure sinewave.

Actually it isn't anything related to being a pure sine and many DVMs worth more than $20 bucks have the ability to measure abs peak without trouble. Although I'd wonder why anyone would want to run backward or ass-first rather than setting up headroom and then adjusting the input gain ITB (in the box) once the signal has infinite headroom available?

Calibration is about accuracy without guess work.

Most importantly you need to be aware of what you're calibrating and why. If you're doing what I think you're doing, you have no idea why or how these devices work.

[Ilya-v]

Are you talking about RMS or peak? The issue is during conversion and the amount of headroom you leave yourself. Setting up the interface to produce the maximum output with a 1v input is just plain stupid because the majority of input devices output far more than 1v peak levels. You'll end up with some negative amount of headroom and clip almost everything.

I clearly mentioned 1VAC Peak in the first post several times.

You are absolutely wrong about the headroom, on the contrary, you'll have maximum headroom because the gain pot on the audio interface is at MINIMUM.
Boosting in the digital domain does NOT clip the signal because DAWs work in floating point.
In your DAW you can boost 120db on the audio track then lower 120db on the master track while not clipping anything.

This calibration is for GUITAR plugins which expect 1vAC Peak analog as 0.0dbFS digital.

[aciddose]

Yes, but the issue here is you're ignoring the far more important aspects of headroom and SNR and jumping ahead to plugging in a reference signal without knowing what each plug-in uses as a reference, nor what is optimal for the particular interface.

(I also doubt any plug-ins expect an average level of 0 dB as this would have been clipped in typical configurations. -6 dB to -20 dB would be far more practical.)

Adjusting the input should have nothing at all to do with a reference level but rather optimizing the conversion from analog to digital.

What you do with gain ITB is entirely separate and depends entirely upon what software you're using. Once you've set up the interface correctly for the levels of the various devices (guitars?) you'll be using you can then measure actual level vs. ITB level and compensate assuming you have a proper known reference level for the plug-in in question.

The easiest way to do this is probably as you say with a reference input with a known level as measuring peak levels ITB should be trivial. (In reality often not as trivial as one might hope.)

An alternative however is to measure the peak input level from any of the devices you want to use and compare it to the ITB recorded peak. This is far more likely to be easy to achieve for the majority of users as a function generator is not a common device while an abs-peak-metering DVM is quite common and affordable.

[Ilya-v]

@ aciddose
It seems you didn't bother to read the first post, but jumped straight into negating this method without bothering to understand it first.
Yes, I am talking about ITB (plugins) guitar amp simulations based on SPICE (most of them) which interpret the whole digital sale (dbFS) as 1vAC Peak, this is also mentioned in the first post.

If I would to calibrate for studio work: American Post: −20 dBFS = +4 dBu.
4 dBu is 1.736v AC Peak to -20dbFS digital.

But once again, we are talking about Guitar Plugins and calibrating to guitar plugins.
You are right about DVM is more obtainable, and that's all you need really, just measure 0.707VAC at the output of the audio interface and loop it back.

[aciddose]

No matter how many times you insist I did not read the post, it remains that I did read your post and I would not change a single thing about my posts no matter how many times I re-read it.

[Ilya-v]

3. How do we cope with this issue when using other analog simulations like waves ssl?

Well, this is a whole other story.
If I would calibrate for studio work: American Post: −20 dBFS = +4 dBu.
4 dBu is 1.736v AC Peak to -20dbFS digital.

In the dbFS wikipedia under the "Analog Levels" section you can see what most studios use.
https://en.wikipedia.org/wiki/DBFS

EDIT:
Actually, for Waves SSL you don't have to calibrate at all because maximum headroom with your audio interface is already 0.0dbFS.
Waves SSL doesn't expect any particular input voltage.

Maximum input levels
MIC/LINE-GUITAR R in GUITAR position (1/4” jack, unbalanced) +9 dBu (TRIM=min)

If the specs are correct, for guitar I would calculate as follows:
+9 dBu equals to 3.087VAC Peak, that will equal to 0.0dbFS with your audio interface.
We want 1vAC Peak to be 0.0dbFS so we boost in your daw by 9.8db.
If the specs of you audio interface are correct, boosting by 9.8db in your DAW should reach 0.0dbFS with a 1vAC Peak at the input.

My Focusrite Saffire has +8dBu headroom before clipping with minimum at the gain pot, equals to about 2.751VAC Peak, so I had to boost 8db in my DAW.
The math correlates to the calibration I did with the external hardware, so everything seems to be in order.

I used these calculators to help with conversion:
http://www.sengpielaudio.com/calculator-db-volt.htm
http://www.sengpielaudio.com/calculator-gainloss.htm

[aciddose]

Where do you get the idea about "SPICE" (this is almost complete nonsense but I'll give the benefit of the doubt) or the specifications for reference levels of specific plug-ins?

Do you have references you can cite? Do the manuals for the plug-ins provide this information?

[Ilya-v]

Do you have references you can cite? Do the manuals for the plug-ins provide this information?

Why yes.

In the first post (ahem) you will find links where the creator of much loved and used freeware guitar plugins like LePou, AcmeBarGig, Ignite Amps, clearly state that the simulation treats the full digital scale as 2VPP.
From previous knowledge and experience LTSpice/Spice also treats the digital scale in a WAV file (Input Voltage) as a 1VA Peak in the simulation.

Peavey ReValver is a component based simulation (also Amplitube) that's why they can create any amplifier/effect/electronics and sell it in the custom shop one by one.
From simple testing and deduction as can be seen in the revalver image in the first post, ReValver also treats 0.0dbFS as 1vAC Peak (0.707 VAC RMS) if you set the Pickup RMS dial to 0.707.

They all are basically a real-time Spice Interpreters.
Why Spice? Because it has fantastic tube simulation and used by many engineers for prototyping for many decades now.
No sane programmer wants to write a component based simulation from scratch for a mere guitar plugin, so they use.... you guessed it, Spice as the core to their simulation with a pretty GUI facelift.

Moreover, having played the real amps and watched endless youtube videos of the real amplifiers, I have a pretty good estimation of how much gain or headroom (clean amps) the simulated amp should have.