Calibrating Input Level for Plugins

[laxu]

So what you need to do is:

- put a signal generator on track 1, route it to output 3. (probably want to turn your monitors off just to be safe), set the level of the sine wave so it is at 0dBFS.

- run a line level cable from output 3 to your Lehle, as you are doing.

- rather than plugging your Lehle to the amp, plug this cable straight to your audient instrument input.

- follow my steps above

This is to determine the level hitting the amp, it’s the one that matters. The level coming from the loadbox isn’t really important for capturing an amp, if you’re doing a pedal capture and want to retain accurate output level modelling then you’d consider it.

Once you know the level hitting your amp, you can determine how much you (or anyone else) needs to adjust to reach unity.

For the level coming from the loadbox, just avoid clipping

I tried this, but skipped the Lehle as it doesn't really matter here.

Output 3 -> Instrument Input 1.

With the signal generator (Logic Test Oscillator plugin) set to output at 0 dB through Output 3, the input I see on the Instrument Input 1 is -38 dB in Logic. To get this track to unity, I need to turn up the preamp by 38 dB. Logic reports it a bit weirdly as a scale of -8 (audio interface gain on zero) to 50 dB. Specs say the gain range is 58 dB.

With the sine generators default -12 dB level, Input 1 sees predictably -50 dB.

 

[bikescene]

I tried this, but skipped the Lehle as it doesn't really matter here.

Output 3 -> Instrument Input 1.

With the signal generator (Logic Test Oscillator plugin) set to output at 0 dB through Output 3, the input I see on the Instrument Input 1 is -38 dB in Logic. To get this track to unity, I need to turn up the preamp by 38 dB. Logic reports it a bit weirdly as a scale of -8 (audio interface gain on zero) to 50 dB. Specs say the gain range is 58 dB.

With the sine generators default -12 dB level, Input 1 sees predictably -50 dB.

Is your Volume 2 on the interface maxed out? The online manual seems to indicate that Volume 2 is linked to Output 3+4 and/or the 2nd headphone out.

 

[MirrorProfiles]

I tried this, but skipped the Lehle as it doesn't really matter here.

Output 3 -> Instrument Input 1.

With the signal generator (Logic Test Oscillator plugin) set to output at 0 dB through Output 3, the input I see on the Instrument Input 1 is -38 dB in Logic. To get this track to unity, I need to turn up the preamp by 38 dB. Logic reports it a bit weirdly as a scale of -8 (audio interface gain on zero) to 50 dB. Specs say the gain range is 58 dB.

With the sine generators default -12 dB level, Input 1 sees predictably -50 dB.

Is your Volume 2 on the interface maxed out? The online manual seems to indicate that Volume 2 is linked to Output 3+4 and/or the 2nd headphone out.

Yeah, check you HW output levels are fully maxed out. Sometimes they are linked to speaker output controls, or have software controlled levels. Also check if they have a switch to go between -10dbV and +4dBu etc.

I’d test with the Lehle and correct cables just to factor it all in.

 

[laxu]

Is your Volume 2 on the interface maxed out? The online manual seems to indicate that Volume 2 is linked to Output 3+4 and/or the 2nd headphone out.

I had it left where I usually set my headphones level...

Thanks! With the volume maxed, Input 1 sees -11.4 dB.

 

[MirrorProfiles]

I had it left where I usually set my headphones level...

Thanks! With the volume maxed, Input 1 sees -11.4 dB.

Looks like the headphone outputs are a lot weaker than the speaker outs.

If you use the speaker outs (1+2) you should get about 11dBu out of the Lehle (you can verify by running a sine wave back into input 1 and there should be a 1dB difference).

Making captures at 11dBu is pretty ideal and should be easy (aside from having to unplug your monitors!).

 

[laxu]

Looks like the headphone outputs are a lot weaker than the speaker outs.

If you use the speaker outs (1+2) you should get about 11dBu out of the Lehle (you can verify by running a sine wave back into input 1 and there should be a 1dB difference).

Making captures at 11dBu is pretty ideal and should be easy (aside from having to unplug your monitors!).

There is no difference between the line outs.

Basically the interface works like this:

• If headphones are connected to headphone out 1, then line outs 1/2 on the back are disconnected and audio is routed to hp out 1.
• If headphones are connected to headphone out 2, then line outs 3/4 on the back are disconnected and audio is routed to hp out 2.
• If instrument input in front is connected, mic/line input 1 on the back is disconnected.

I tried swapping it around so that it goes Output 1 -> Input 1, and the result is the same. With a 1 KHz sine generator at 0 dB out, I get -11.4 dB on Instrument input 1.

If I connect Output 1 -> Input 4 (mic/line, these are all identical), with Input 4 gain on zero, then I see -15.1 dB on that input.

 

[MirrorProfiles]

There is no difference between the line outs.

Basically the interface works like this:

• If headphones are connected to headphone out 1, then line outs 1/2 on the back are disconnected and audio is routed to hp out 1.
• If headphones are connected to headphone out 2, then line outs 3/4 on the back are disconnected and audio is routed to hp out 2.
• If instrument input in front is connected, mic/line input 1 on the back is disconnected.

I tried swapping it around so that it goes Output 1 -> Input 1, and the result is the same. With a 1 KHz sine generator at 0 dB out, I get -11.4 dB on Instrument input 1.

If I connect Output 1 -> Input 4 (mic/line, these are all identical), with Input 4 gain on zero, then I see -15.1 dB on that input.

Ah got it. It still seems like some level is getting lost somewhere, if you were sending -12dBFS out and getting approx -11dBFS back in it would make sense. the Lehle stuff tends to be 1:1 on the transformer ratio so it SHOULD be 11dBu on your output (i.e. 1dB louder than the maximum the DI input can take).

 

[laxu]

I tried the whole chain again with the Gigrig Wetter Box, and now the gains are pretty much identical. I'll see if I can get a capture made.

The key was maxing Volume 2, and maxing output 3 volume from the EVO Control app.

 

[laxu]

Thanks for all your help! I got a capture made and comparing it with the Wetter Box going to audio interface vs direct to amp, the tones are very close.

 

[MirrorProfiles]

"our method for calibration is just to pretend SNR can be optimized beyond the laws of physics and its better to avoid accuracy all together"

 

[2dor]

View attachment 32507

"our method for calibration is just to pretend SNR can be optimized beyond the laws of physics and its better to avoid accuracy all together"

Oh boy, oh boy ... some people are really butt-hurt aren't they?

 

[MirrorProfiles]

Oh boy, oh boy ... some people are really butt-hurt aren't they?

It’s concerning when a company who develops plugins and prides themselves on accuracy is just repeating tropes that don’t make any sense.

With so much choice available these days, it’s easy for customers to vote with their wallets on this kind of stuff. There are several topics that seem to prove as a good litmus test for how serious a company is with their work.

 

[2dor]

There are several topics that seem to prove as a good litmus test for how serious a company is with their work.

For sure

 

[Sedaxel]

"our method for calibration is just to pretend SNR can be optimized beyond the laws of physics and its better to avoid accuracy all together"

When a fool walks a path, the path finally ends, but the fool keeps on walking.

My god.

 

[DLC86]

I guess some companies have fear that users start noticing that their accurate amp sims are not that accurate after all

 

[2dor]

I guess some companies have fear that users start noticing that their accurate amp sims are not that accurate after all

There's always a reason behind this type of statement (firm, dismissive) and, in this case, I think your observation is very spot-on.

 

[Sedaxel]

There's always a reason behind this type of statement (firm, dismissive) and, in this case, I think your observation is very spot-on.

I'd not discard simple and pure ignorance either.

 

[2dor]

I'd not discard simple and pure ignorance either.

True, true. Sad to see this coming from a dev though.

 

[Deadpan]

IMO this is a bigger problem than we know.

Are people jumping on the train for attention?

Conspiracy theory city?

 

[[Nathan]]

View attachment 32507

This is so wild... I'm no expert but this is my caveman summary of the small evolution of info

1) Amp sims rise in popularity.. people start to notice their trials or paid software isn't as good as other people demoing it. A lot of people would have ditched the ampsim if it was bad. For the rest of people who are happy to tinker they either messed with the input trim / gain on the amp sim / smashed plugins before the amp sim to sweeten it... any number of random things to try and get it in the ballpark, all guesswork.

2) Ed / other people start championing that hey these amp sim things have specific levels they were calibrated at. If you use a multimeter you can set your interface to match the amp sim so you're hitting it at the right level. You can still boost or attenuate to taste but at least you can hit it at what they were hitting it at (HUGE IMPROVEMENT OF INFORMATION RIGHT HERE)

3) People don't want to use a multimeter (fair enough). Ok a decent compromise is that most interfaces hover around +12/+13dBu of input headroom on instrument inputs (which ironically for super hot outputs is close enough to where a guitar signal can peak... maybe closer to +10dBu but not insanely far off the mark)... so for the vast majority of caveman guitarists using a scarlett / uad / rme / audient / ssl / motu you can just leave it at 0 and then adjust the plugin, how handy.

To me the conversation was kind of over at this point. If you wanted to ignore the spec, cool... if you wanted to use it, cool.... its all there for you to take or leave. You can tell a lot of people still resented this topic for some reason but ok.

4) Old mates video comes out all of a sudden making a big deal about SnR and how using 0 gain on interfaces is wrong.... (read above how we even got there in the first place... its mostly right and if your interface sucks go back to #2 like it was originally discussed with a multimeter... set your behringer interface to its least bad setting and measure it)

For DiBiQuadro to champion "use your ears bro" is baffling, why do we want to aspire to Step1, it was bad. I'd have much more respect for a company who can actually come out with a nice streamlined solution for all of this instead of tripling down against it.