Calibrating Input Level for Plugins

[James Freeman]

I think a lot of the perceived “hotness” of an EMG is from the preamp clipping the loud notes, and the softer ones being closer in volume for a more even sound.

In EMG's defense, they still have a good amount of headroom, just the very peaks are clipping.
The waveforms above are me digging hard in on all open strings with the intention to clip everything, which is not 'playing'.
Still, passives are undoubtedly better for anything edge of breakup or dynamic open chord crunchy sounds ala AC/DC or any form of Blues where digging in and retaining dynamics is important.

 

[MirrorProfiles]

In EMG's defense, they still have a good amount of headroom, just the very peaks are clipping.
The waveforms above are me digging hard in on all open strings with the intention to clip everything, which is not 'playing'.
Still, passives are undoubtedly better for anything edge of breakup or dynamic open chord crunchy sounds ala AC/DC or any form of Blues where digging in and retaining dynamics is important.

Yep, and anything high gain is likely going through several (carefully balanced) stages of clipping anyway. Even lower gain sounds are going to end up squared off once they’ve been through a full chain of gear. Still, we’re talking about digital, and preserving that information is important.

 

[zolko60]

I took a look at UAD Apollo 8 manual. The max. HiZ input level is +12dBu.

RME FF400 HiZ max. input is +15dBu (12dB pad engaged). No pad - way too hot.

What is the sensitivity of Tonex Pedal input if people say you have set it -15dB from default value of 0dB?

What is the sensitivity of Fractal at 0dB position?

What about Headrush?

 

[James Freeman]

498T in a Les Paul clips the input

Can you try running through a BOSS pedal?
Preferably with a fresh battery or a power supply.
All boss pedals are buffered and will probably clip the signal before Helix does.

Also, there will probably be a difference in clipping 'sound' when clipping the BOSS transistor buffer compared to the Helix voltage rail clipping.
We should test this.

The SD SH-4 is not as hot as the 498T but this is what I got.

Going through a BOSS pedal will still have more headroom than an EMG 81.

 

[MirrorProfiles]

Can you try running through a BOSS pedal?
Preferably with a fresh battery or a power supply.
All boss pedals are buffered and will probably clip the signal before Helix does.

Yep, clips the input buffer and produces a level below what would clip Helix.

Tested with an NS-2 as my other Boss pedals are either lower impedance than 1 MOhm (DS-1) or have that annoying bleed through (old SD-1). Also tested an MXR Sugar Drive (600kOhm) with the buffer enabled. Seemed to produce a slightly louder signal than the NS-2 but not sure if it has more headroom or is just adjusting gain.

 

[James Freeman]

Another thing, when going through a unity gain transistor buffer we add another stage of random noise.
I'm seeing the expected 3dB increase in noise floor, nothing to do about it, physics is inescapable.

 

[James Freeman]

Oooh, another good find.
The -5.5dB pad also lowers the noise floor by 3.5dB, so boosting back 5.5dB only increases noise floor by 2dB.
So using a buffered pedal is useless, just enable the PAD and boost by 5.5dB at the input to retain the same gain, 2dB higher noise floor is nothing for the already super quiet -127dBFS RMS noise floor and you gain 5.5dB of real headroom.

It would be nice if the Helix would boost 5.5dB automatically when the pad is engaged.

 

[MirrorProfiles]

just enable the PAD and boost by 5.5dB at the input to retain the same gain, 2dB higher noise floor is nothing for the already super quiet -127dBFS RMS noise floor and you gain 5.5dB of real headroom.

Is the input impedance the same when using the pad? and is this the case for all Helix models?

 

[James Freeman]

Is the input impedance the same when using the pad? and is this the case for all Helix models?

Measured input impedance is about 830k with or without the pad, that's with the 1M setting in Helix.

If you want to measure yourself, solve for R3.
R1 and R2 are two halves of a 500k volume pot in a box.
Run a test signal with pot at full volume, turn pot down until you read -6db in DAW (half voltage), measure both halves of pot, solve for R3.

I actually used a writing utensil... you can tell by my perfect hand writing how long I haven't used one.

PS.
The switchable input impedance values are real.

 

[DLC86]

If you want to measure yourself, solve for R3.
R1 and R2 are two halves of a 500k volume pot in a box.
Run a test signal with pot at full volume, turn pot down until you read -6db in DAW (half voltage), measure both halves of pot, solve for R3.

What's the difference between this method and (the far easier) just measuring the resistance between tip and sleeve?

 

[James Freeman]

You can't measure resistance from outside if there is a series coupling capacitor or a more complex RC network at the input, you have to use AC signal at say 1kHz and calculate the equivalent impedance that causes a known voltage drop.

Typically people use a series pot and calculate the input impedance as part of a voltage divider, but I already had a 500k volume box so there is another parallel resistor (R2) to the input impedance.
Also, I already see the dBFS values form the Helix in my DAW and know that -6dB is half the voltage so I don't need to measure voltage drop, just measure the two halves of the pot in the box and solve for R3.

BOSS TU-3 measured input impedance is 2Meg @ 1kHz.

 

[Orvillain]

Sooooo... I finally got around to doing this. My setup is a little unusual, because I use one interface for my outputs, and one interface for my DI signal input.

Presonus Quantum - Line Output 2: A -17dB 1kHz sine-tone results in 500mV AC on multi-meter
Neve DI Box going into Antelope Audio Discrete 8 SC, input 8 set to mic mode, with the gain at 0dB results in: -18.9dB signal on the Reaper mixer meter
So to get my original -17.0dB signal, the preamp on the Antelope needs to be +2dB when I'm using the Neve DI box

to get the Helix gate to open (With the @Freeman -15.3dB threshold) then I need +4dB on the preamp, or inside Helix.

So effectively, Helix Native needs a +4dB boost, according to the method laid out in this thread.

 

[Orvillain]

Next thing.... took the same sine-tone source.... sent it to my reamper.... sent the signal from my reamper into the now calibrated and understood DI level... and I now found the output level knob position on the reamper that gives me the same -17dB signal, and it's about 3 o'clock on the knob.

 

[James Freeman]

That's the way little grasshopper, follow the way of the volta... I mean chi.

 

[Orvillain]

That's the way little grasshopper, follow the way of the volta... I mean chi.

lolololol.

Of course with these levels, ToneX now complains that my input level is too low... so .... gee... thanks IK.... ???

 

[James Freeman]

ToneX is ... special.
Can you 'Next' without adjusting the level?

 

[Orvillain]

ToneX is ... special.
Can you 'Next' without adjusting the level?

Yeah. I'm trying it out now. But I suspect (chatting with @MirrorProfiles on this now) that ToneX will have sent it's signals at a louder level than expected, so the capture will probably have more gain than the real amp, compared to my guitar going into the amp.

 

[MirrorProfiles]

Yeah. I'm trying it out now. But I suspect (chatting with @MirrorProfiles on this now) that ToneX will have sent it's signals at a louder level than expected, so the capture will probably have more gain than the real amp, compared to my guitar going into the amp.

It should play back fine, but the training files will probably be hitting the amp with more gain than you intend to use it. And then you're hitting the model a bit softer which evens it back out.

Jason Sadites has this exact problem - he's running the test signal into the amp REALLY hot and than having to back down the input level by 15dB to resolve it.

You either pick IK's calibration, which is quite hot (I just use 12dBu and its on the lower end of what IK wants to see, but still tickles the orange LED's). Or you ignore their calibration and use your own, which is fine for your set up but others will need to adjust their levels to get the same response. Which is basically what causes threads like this to exist in the first place.

Would be great if IK had a preference in their settings where you could choose different internal calibrations.

 

[JT]

lolololol.

Of course with these levels, ToneX now complains that my input level is too low... so .... gee... thanks IK.... ???

The level meters for Tone-X really just a weird crap shoot for me. Like the meter onscreen vs. level meter on individual interface inputs was just I don't even know. The screen says too red, the interface is barely hitting green. Who the F knows what you are supposed to be focusing on as the right answer. No discernable successful results when using this or that as the baseline. Just up in the air, cross your fingers and hope for the best type of sh!t.

 

[James Freeman]

The ToneX Pedal has a reference input dBu and acts as an audio interface, there should be ZERO guessing when capturing with it.