Input Voltage/Gain Calibration for Guitar Plugins

[GrabtharsHammer]

Can someone recommend a cheap (but reliable) signal generator that I can use for the input calibration?
Or are there other ways to do this without having to buy a signal generator device? Perhaps a software that can use the audio interface's audio output to generate a 1VAC peak signal? Or an iPhone app?

[Ilya-v]

Hi GrabtharsHammer.

Yes you can definitely use your audio interface output as a signal generator, I can't believe it slipped my mind with all the VST tools at my disposal!
This 1kHz WAV test tone will do: http://www.mediacollege.com/audio/tone/download/
Even youtube 1kHz test tone will do.


Carefully adjust the output till a good DVM measures exactly 0.707VAC and loop it back into your High-Z Instrument/Guitar input with Gain at minimum, and boost in your DAW till you reach 0.0dbFS.
It's simpler than I though... I will edit the first post.

An iPhone or Android app will NOT do because the phone doesn't have an accurate output volume adjustment but rather limited amount of steps.

[GrabtharsHammer]

Some quick calculations for typical audio interface Instrument High-Z input values:

+4dBu headroom (1.736V Peak) with Gain trim at minimum, boost by 4.79db.
+5dBu headroom (1.948V Peak) with Gain trim at minimum, boost by 5.79db.
+6dBu headroom (2.185V Peak) with Gain trim at minimum, boost by 6.79db.
+7dBu headroom (2.452V Peak) with Gain trim at minimum, boost by 7.79db.
+8dBu (2.751V Peak) -> boost 8.79db
+9dBu (3.087V Peak) -> boost 9.79db.
+10dBu (3.464V Peak) -> boost 10.79db.

Here are my RME Babyface Pro's specs:
„- Maximum input level @+4 dBu, Gain 0 dB: +13 dBu
- Maximum input level @-10 dBV, Gain 9 dB: -5 dBu“

Did I understand it right that theoretically I have to boost by 13.79dB when the interface's input is set to +4 dBu? That seems very much...

[Ilya-v]

Yes, the higher the headroom in dBu (hence voltage headroom before clipping the analog preamp) the more boost you have to add in your DAW.
But most spec sheets are not 100% accurate so it would be much better to output a 1kHz sinewave 0.707vAC RMS (DVM measured) signal from one of the other outputs and loop it back to the input, then boost in your DAW to reach 0dbFS.

On the other side, the more you boost the more you hear the noise floor of your analog path.
But audio interfaces become better with the years and so is the analog input path noise floor.

[Alec246]

I find this thread very interesting, I have been searching for this information for ages!

But I'm a Noob. I got my Multimeter, Measured around VAC 0.700, and then I am trying to get my DAW to read 0.0 with that input, but with +12 on EQ, I'm still short of that, around -7.0, and Im getting tons of noise.

What am I doing wrong? I guess I need a more detailed tutorial :/ Maybe I'm even measuring frequency wrong.

Thanks

[Alec246]

Oh, BTW. I'm using an Audient iD14 Interface

1 x Discrete JFET Instrument Input ( Does that change anything?)

D.I (Channel 1):
D.I GAIN: 0 to 66 dB (inc. +10 dB software boost)
MAXIMUM INPUT LEVEL: +9 dBu (0.6% THD typical)
INPUT IMPEDANCE: >500kΩ Unbalanced
FREQUENCY RESPONSE: ±0.1 dB 20Hz to 22kHz
THD+N @ 0dBu (1kHz): <0.04% all musical 2nd and 3rd harmonics
SNR: 95 dB un-weighted, 98 dB A-weighted
1/4” JACK: TIP (Hot) & SLEEVE (Shield)

[aciddose]

1 x Discrete JFET Instrument Input ( Does that change anything?)

Typically a high-Z input will also have higher noise from EMI.

Any gain you add ITB will be subtracted from your noise ratio. So if you have -110 dB noise floor and you add 20 dB gain ITB, the resulting noise floor is then -90 dB.

That's why using the minimum input level may not be optimal. You need to find the maximum peak level (an outlier) generally approximately 3 dB above the usual peak level at the input.

That then depends upon the total circuit however. That level of gain may also not be optimal as far as the noise floor is concerned. It is impossible to get a full 24 bits without noise (144.5 dB) and ~21 is usually the best you can hope for from the highest end interface (-130 dB.) A minimal floor of -120 dB is a good aim if you have full control of all sources of noise (scientific equipment) but as far as guitars go you'll probably never come close.

That is why I described the ideal process as being one where you minimize the noise floor as a first step, then worry about bit depth and THD, then you can adjust the gain downward if needed ITB where you'll reduce the noise floor and maintain close to your ideal noise ratio.

[Alec246]

At the MICROPHONE PREAMPLIFIER Spec the iD14 has this:
MAXIMUM INPUT LEVEL: +12 dBu (0 dBFS digital Maximum)

At the DI Input
MAXIMUM INPUT LEVEL: +9 dBu (0.6% THD typical)

I get confused because the DI doesnt mention 0 dBFS, but THD instead, so I don´t know what value of headroom it has until 0 dBFS?

Also. Will a Notebook work as a Signal Generator if I output a 0.707VAC 1Khz Sine Wave from it Line Out into my Interface Guitar DI Input? I feel better using an independent device than looping a signal from the same Interface, I fear interference may occour.

Also, I would love a better detailed post to explain the process to Boost inside the DAW. Maybe using Reaper as an example? Since it´s the most accessible for everyone. Im trying Studio One but not for long.

Also, how can I check if I get a better SNR from raising the Gain Pot on my Interface vs Raising in the DAW?

Lots of questions I know, but you have opened a can of worm, and now we want to know everything about this because its something that no ones talks about!

Thanks

[aciddose]

I get confused because the DI doesnt mention 0 dBFS, but THD instead, so I don´t know what value of headroom it has until 0 dBFS?

All you need to do to find out is monitor the input and watch the meter in your DAW.

Will a Notebook work as a Signal Generator if I output a 0.707VAC 1Khz Sine Wave from it Line Out into my Interface Guitar DI Input? I feel better using an independent device than looping a signal from the same Interface, I fear interference may occour.

If interference occurred the device would be flawed and you should probably return it or contact support for a replacement/repair.

Also, I would love a better detailed post to explain the process to Boost inside the DAW. Maybe using Reaper as an example? Since it´s the most accessible for everyone. Im trying Studio One but not for long.

Some hosts like Cubase include an input gain knob. Hosts like Reaper do not. In Reaper you'll need to use an input plug-in like js:Volume Adjustment.

Also, how can I check if I get a better SNR from raising the Gain Pot on my Interface vs Raising in the DAW?

This requires the ability to accurately measure noise. There are various measurement types including weighted filters and special meters designed to measure noise levels. Peak levels won't give you a complete picture because it won't be obvious which spectral elements dominate the result. So a combination of peak, RMS and spectral analysis are needed all at once to give you part of the information you need to make a valid assessment of total noise level.

Once you can accurately measure the total noise level and understand the parts that make it up you can learn about their causes and adjust your input to minimize some of them. Most likely the majority of your noise is EMI when working with guitars and similar instruments (mics, etc) and minimizing this is definitely a whole can of worms because no matter what you to the instrument itself is essentially an antenna!

Minimizing issues with patch cables (low Z cables) and inline circuits (tone controls, volume pots, switches, sockets and contacts) can all play a part there as well as insulating and grounding the guitar cavity. Ultimately though you'll probably want to remove any active sources of EMI from your studio: the DAW itself is almost always a major source of EMI.

(If your DAW includes a wireless networking connection it's an active transmitter in the GHz range!)

[Peter - IK Multimedia]

Since AmpliTube was mentioned, the team has asked that I pass on clarification about why this does not apply to AmpliTube:

The fact that Spice runs on this convention is not applicable to a plug-in like AmpliTube. It could be 10Vpp = 0dBFS or 0.1Vpp = 0dBFS, it doesn’t mean a thing in this context.

In fact, this is a convention that applies when running the Spice model, and only there.
Analog simulation software like Spice and similar are only one step of all those implied in our modeling process to go from the schematics and real equipment/components to the finished DSP processor. Each model that is created has a specific input / output reference levels trimming that has nothing to do with the reference levels that are of course considered when doing the analog emulation of the circuit.

For AmpliTube models (and actually just about everything else we model) we also include the real equipment in the process (not just schematics) to be sure we know how to calibrate the I/O levels so that the user gets the most realistic experience from that piece of gear, real or modeled.

In AmpliTube, the setting of the input gain on the Hi-Z input of your interface (or on iRig HD 2 for example) has to be made right in the same way you would properly set the gain of any preamp in the world (mic or instrument) - exactly like the ABC rules of audio gain staging for digital audio: set the input gain so that the acquired signal on the digital recorder or DAW is as close as possible to 0dBFS without ever clipping. This ensures the best possible dynamic range and SNR.

That’s the basic rule, and it’s the best practice to set the input level of AmpliTube. We ensure doing this you get the most realistic experience from the models that are finely calibrated to exactly these conditions.

[Ilya-v]

But I'm a Noob. I got my Multimeter, Measured around VAC 0.700, and then I am trying to get my DAW to read 0.0 with that input, but with +12 on EQ, I'm still short of that, around -7.0, and Im getting tons of noise.

The noise floor will rise with the added gain in any case, whether you use the audio interface gain pot, boost in your DAW, or add gain in the amp simulation.
A very low noise floor analog input stage is a must for high gain amp sims, it is one of the most important parts that change practically in each new hardware revision of the Axe-Fx for example.

Yes you can use a laptop as a signal generator as long as the signal is clean without ground loop noises etc..
On the contrary, using the same device (audio interface) is much better than an external one for the reason above.

Yes it's +9dBu for the DI input with gain pot at minimum to 0dbFS.

Also, how can I check if I get a better SNR from raising the Gain Pot on my Interface vs Raising in the DAW?

Reach 0dbFS in your DAW with 1vAC Peak at input, once with the gain pot and once with a plugin boost and see what the meter says with nothing plugged at the input and amp simulation plugin bypassed.
The downside of boosting with the gain pot is you have less analog headroom for hot pickups like EMG.

[Ilya-v]

Since AmpliTube was mentioned, the team has asked that I pass on clarification about why this does not apply to AmpliTube:

The fact that Spice runs on this convention is not applicable to a plug-in like AmpliTube. It could be 10Vpp = 0dBFS or 0.1Vpp = 0dBFS, it doesn’t mean a thing in this context.

Each model that is created has a specific input / output reference levels trimming that has nothing to do with the reference levels

Hi Peter.

What are you saying? that Amplitube is completely random between models when it comes to input gain simulation? Is that what you call realistic?
If Amplitube is a schematic/component based simulation, a reference voltage HAS to be there for consistency between amp models and their amount of gain.
Surely if you use a component/schematic based simulation you often reuse the basic electronic components and don't re-engineer them for every new amp, unless you don't follow the original amp schematic at all.

Please ask your team to clarify this further, because most other respectable amp simulations don't hide this information and "Play hard but don't clip" doesn't cut it in this thread, this thread is about NUMBERS and the effort to achieve realistic gain from the simulated amps, read: NOT random.

[aciddose]

The output of your guitar is random.

It isn't really, but it depends upon many factors including how hard you play and where you set the tone pots or switches, which pickups are used all the way to which patch cables you use.

This is why everything you say is nonsense. The level your particular guitar will produce into various amp hardware has nothing to do with the number you get from a volt meter or relative to the reference level of your particular audio interface.

Why can't you understand that?

If you want to compare a real hardware amp vs. its "model" in software you need to compare them directly in A/B. If the hardware differs from the software you can tweak the input gain to either to match them up. This is an iterative process with multiple factors at play.

If you aren't doing that, why do you even care? The input gain pot on most amplifier hardware is so inaccurate (20% generally) that you'll have absolutely no ability to match up identical hardware units side-by-side without spending a huge amount of time tweaking the settings on both.

[Alec246]

Thanks Ilya-v for the reply!

I understand what he is trying to say. If what the Amplitube team say is correct, I could get my Low Output Single Coil, boost the Input Gain until it's like a TS808 is with Level all the way up, and play on a Fender Twin that will sound distorted on Gain 2.

That's not how a real Fender Twin would sound in real life, unless you have that TS808 Boosting the Input. So how do we get a value that means the sound I'm getting from the Amp Sim is the same as the one I would get from the Real Amp?

Also, when driven under such circunstances, I bet that the Modelling isn't done focusing on a Impossibly Hot Guitar Signal coming in, as a Real Amp would never be profiled that way, so if Amplitube does that, It's not reacting like a real Amp would. It's going outside the envelope of the Simulation, and the sound you get there may sound good, but it's Amplitube's Sound, not the Amp's

[aciddose]

Various amps respond differently to different input levels. This is why they all have input gain knobs. A guitar amp is nothing like a finely-tuned scientific piece of equipment like an oscilloscope or spectrum analyzer. It's a musical device and entirely focused on "trial and error", "tweaking" as an iterative process to get "the sound" you want.

This input gain knob is almost always raised significantly because guitars are for the most part going to produce extremely cool signal levels vs. line level or modular synthesizers (up to 10v peaks!)

Also: raising the gain ITB is very different than on the preamp for your interface. The preamp only increases the level of noise already present in the signal. The interface adds noise and so if it isn't driven with a proper level signal you'll always get worse S/N ratios.

The first step is to get the best possible signal/noise ratio at all stages; from the guitar, through to the instrument cable and any buffer/preamp all the way to the input gain on the interface itself.

You can only get the best S/N ratio if all those elements are perfect.

Then you will have a "maximized" signal level relative to 0 dBfs and you can decrease the level or if desired also increase it to get the desired effect ITB because adjusting the gain ITB will maintain the S/N ratio you already recorded from the input.