KVR Audio

Ilya-v · Post by **Ilya-v** » Mon Apr 10, 2017 8:47 am

Hi izonin.
If you have Apogee or Mytek can you please check the RMS noise floor on the High-Z Inputs with gain at minimum using the freeware SPAN analyzer VST?
It's no fun spending 2000$ on an Apogee if I can build a 5$ treble booster with a super quiet opamp and de-emphasis it in DAW for the same result.

izonin · Post by **izonin** » Mon Apr 10, 2017 8:54 am

A studio I work with has the Apogee. I can check the floor next time I have a recording session.

The method you describe... wouldn't this boosting and cutting (non linear phase) introduce harmonic distortion? I've tried this with a high quality Yamaha preamp before the interface and the result wasn't pretty.

Ilya-v · Post by **Ilya-v** » Mon Apr 10, 2017 8:56 am

Yes probably, but for electric guitar 0.1% vs 0.0001% THD isn't much of a difference.
The Yamaha preamp has to be quieter even when boosted thn the interface input noise floor for this to work.
Meaning, you should not see any difference in noise floor analysis of your audio interface with or without the boosted signal before it.

izonin · Post by **izonin** » Mon Apr 10, 2017 9:00 am

It's exactly electric guitar I've tried it with. This introduced too much distortion on the highest frets. Not good at all for clean jazzy stuff. But for heavily overdriven guitar sounds it might work.

Ilya-v · Post by **Ilya-v** » Mon Apr 10, 2017 9:17 am

There are many factor you have to consider when doing pre-emphasis especially with another external unit.
Important things to watch for:
1. The treble boosted signal should not be distort before the next stage at all frequencies.
2. The treble boosted signal should not add to the noise floor of the next stage.

If the distortion was audible it wasn't in the 0.1% range but much higher, the Yamaha preamp was simply distorted even before the audio interface.
The GE-7 I use only has 4.5v headroom because it runs on 9V power supply, so when I boost the 6.4kHz shelve to max (+9db) I'll probably distort the highs with 1.5v Peak from the guitar, if of course I reach than.
But 4.5V Peak equals to +12dBu and my interface maximum input is +8dBu so the interface clips before the GE-7, I should see it immediately.

This whole Pre/De-Emphasis process should be as clean as possible with big headrooms.

jacotzen · Post by **jacotzen** » Mon Apr 10, 2017 1:43 pm

I don't know about you but my 'problem' is that I have no intention to spend other money on products that I don't know if they work as they claim to. Personally I don't give a f*k if the tone is identical to the real deal, if it's a bit fizzy or what, first I care if the product is usable: latency and gain staging play a big rule in this and that's why I'm here for trying to understand if this gain staging issue must be relate to the audio interface or to the software or both. Everyone knows that the more you spend the more you pretend but at the same time I know for sure that sometimes small changes can produce large improvements. At the end is all about knowledge (that I don't have and I'm trying to learn and understand).

jacotzen · Post by **jacotzen** » Mon Apr 10, 2017 4:35 pm

For Ilya: "Your input signal picks up some of the output signal, maybe through ground, but that's another topic."

Actually I experimented a bit and realized that the metronome is recorded when monitoring a track loaded with an amp sim even if there is any instrument plugged into the audio interface... if I disconnect the headphones it's ok (that's why when I tried to output the signal to the built in output the problem didn't occur), I researched more and I found that this issue is related to the 'cross talk bleeding' (I learned only NOW that there are a lots of topic on the subject).

Now I don't want to drive you mad but it seems like your audio interface if affected by this... a lot of audio interfaces actually are :/

Think about: you use an amp sim, your gain seems always low comparing to the real world, you end up adding gain or distortion to have more signal especially when soloing and you end up summing two signals!

EDIT: for your information this problem occurs even on a so called professional audio interface apogee duet (730 euros

) https://discussions.apple.com/thread/16 ... 5&tstart=0

Ilya-v · Post by **Ilya-v** » Mon Apr 10, 2017 5:05 pm

Yes I am well aware of the crosstalk between the outputs.
To minimize if not eliminate this crosstalk I turned the volume pot (Monitor) on the front of the audio interface to maximum and control the volume with the mix-control software (driver).
With the volume at maximum there is no signal bled to ground through the resulting voltage divider hence minimum crosstalk between outputs.
So when I listen with headphones there is zero bleed to the monitors, I just mute the output to the monitor via software.

I tried to record the high-z input shorted with a metronome clicking at 0.0dbFS, then I boosted the track by 70db to hear the noise floor.
I can hear some faint click of the metronome but it is BELOW the noise floor so I can see it on the frequency analyzer, the reason I can hear it because noise acts like dithering thus allowing me to hear sound quieter than the loud noise.
Anyway, the bleed from the output to input on the Saffire is way below the noise floor, it's completely irrelevant.
But that topic is for another thread.

jacotzen · Post by **jacotzen** » Mon Apr 10, 2017 5:52 pm

Ilya, do you have any suggestion on how to manage the gain chain if I plug this preamp into the interface just for testing another signal path? Obviously I will use the clean channel so that I can put an amp sim on the insert

here the specs:
http://hughes-and-kettner.com/wp-conten ... 5_korr.pdf

Ilya-v · Post by **Ilya-v** » Mon Apr 10, 2017 6:06 pm

I would not recommend adding anything before the audio interface especially if it has transistor/tube gain stages, UNLESS it is quieter the the noise floor of the audio interface, which is highly unlikely because analog-to-digital-converters are designed to have very low noise floor.

Voxengo SPAN spectrum analyzer is freeware and a great tool. http://www.voxengo.com/product/span/
Please download it and report your noise floor RMS value as in the images I've posted earlier.

jacotzen · Post by **jacotzen** » Mon Apr 10, 2017 6:27 pm

Ilya could you be so gentle to provide basic instructions on what to do step by step please? I'm a newbie and I want to avoid to alterate the measuring due to my lack of experience/knowledge?

Ilya-v · Post by **Ilya-v** » Mon Apr 10, 2017 6:32 pm

On the upper left corner of this website there is an Envelope symbol next to your user name, this is called a private message (PM).
We can continue the out of topic conversation there.

Ilya-v · Post by **Ilya-v** » Tue Apr 11, 2017 11:28 am

Direct proof by a well known amp simulation company Kuassa.

Hi Kuassa.

What dbFS your amplifier simulator treats as 1VAC Peak (2VPP)?
Most VST Amp Simulators tread 0.0dbFS as 1VAC Peak, is it the same with your amp simulations?
The reason for that is I calibrate my audio interface with a function generator with a 1VAC at input and boost in DAW till it reaches 0.0dbFS, the result is the simulation behave like a real amp.

Thank you,
Ilya.V

Hi Ilya,

Wow you noticed that.
About the 1VAC yes definitely. All of our amps are able to have input more than that actually for a more overdriven tone. But make sure to adjust it via input volume knob on top left of the plugin and leave the input gain on the soundcard at 1VAC like you mentioned.

Enjoy!
Best,

Kuassa Teknika
Kuassa Support Team

Yeah, not only I noticed but it is important for a company to share that information for people who actually want a realistic simulation.

Still waiting for more replies from other companies.
Meanwhile I know for sure that LePou, Ignite, Kuassa, ReValver treat 0.0dbFS as 1VAC.
This is no surprise, they are all indeed SPICE based.

Ilya-v · Post by **Ilya-v** » Tue Apr 11, 2017 11:59 am

You can include Mercuriall in that;
https://mercuriall.com/cms/node/9

They specifically state "The modeling technology includes State-Spice simulation of electric circuits.." and have clear LTSpice images in the slide show.

http://atoragon.blogspot.co.il/2016/02/ ... audio.html
In this link search "LTSpice", the developer himself speaks about how he created these simulations.

They are no joke either, Mercuriall is one of the best amp simulations around.

Ilya-v · Post by **Ilya-v** » Tue Apr 11, 2017 2:17 pm

Got a reply from the man himself.

Hi Mercuriall.

What dbFS in DAW the amplifier simulator treats as 1VAC Peak (2VPP)?
Most VST Amp Simulators treat 0.0dbFS as 1VAC Peak in the digital world at
the grid of the first tube stage, is it the same with your amp simulations?
The reason for that is I calibrate my audio interface with a function
generator with a 1VAC at input and boost in DAW till it reaches 0.0dbFS, the
result is the simulation behave like the real amp, yet I still have the full
headroom of my audio interface.

Thank you,
Ilya-V

Hi Ilya,

Yes, this is true when the Input Volume is 1. Other Input Volume values give a multiplication factor equal to the value of this controller.

Kind regards,
Mercuriall Audio Software.

Mercuriall, 0.0dbFS = 1VAC Peak in the simulation.

I hope people starting to understand the importance of input calibration if realistic amp simulation is desired.

Also, some participants of this thread that called my theory "nonsense", "naive", "stupid" and "flawed in rational" can take their words back.

Input Voltage/Gain Calibration for Guitar Plugins