Modeler Time Domain Response Analysis

[James Freeman]

Cliffs comment on the Fractal forum on how he cannot share super secret information about time domain measurements got me interested in time domain measurements and what they can tell us about the modeler.

Helix Brit 2203;
In this example I have a few cycles of 82Hz sine at -10dB, the output is quite interesting.
First of all, the output was correctly (unless I'm wrong) phase flipped because the JCM800 has 3 inverting gain stages, but I have reversed it in this image below so we can see better what's going on.
What I notice immediately is some "pre-sag" before the amp starts to reproduce the tone, also, there is "post-ringing", both of these tell me that there are some elements (components?) that cause this sag and ringing, in a real amp it will be the power supply inductance and filtering capacitance.

I am very curious to measure other modelers and have a look at how they behave, I think time domain response has a lot to do with how the model feels to play.

82Hz & 400Hz.

Yes, it's another James Freeman thread, go ahead post your shitty memes and destroy beautiful science you banjo twanging luddites.

 

[PLX]

go ahead post your shitty memes and destroy beautiful science you banjo twanging luddites.

 

[norminal]

banjo twanging luddites.

 

[Sascha Franck]

Post-ringing yes, but pre-sag? Your pictures don't exactly reveal that, it's more looking like a (typically) "smeared" attack to me.

 

[James Freeman]

In the Fractal I'm going to turn things on and off, it should be much clearer what's doing what.

 

[TSJMajesty]

Cliffs comment on the Fractal forum on how he cannot share super secret information about time domain measurements got me interested in time domain measurements and what they can tell us about the modeler.

Helix Brit 2203;
In this example I have a few cycles of 82Hz sine at -10dB, the output is quite interesting.
First of all, the output was correctly (unless I'm wrong) phase flipped because the JCM800 has 3 inverting gain stages, but I have reversed it in this image below so we can see better what's going on.
What I notice immediately is some "pre-sag" before the amp starts to reproduce the tone, also, there is "post-ringing", both of these tell me that there are some elements (components?) that cause this sag and ringing, in a real amp it will be the power supply inductance and filtering capacitance.

I am very curious to measure other modelers and have a look at how they behave, I think time domain response has a lot to do with how the model feels to play.


View attachment 14233

82Hz & 400Hz.
View attachment 14234

Yes, it's another James Freeman thread, go ahead post your shitty memes and destroy beautiful science you banjo twanging luddites.

Those spikes remind me of this graph Cliff posted about plate suppressor diodes.
Is there any connection, or is it something entirely different?

 

[James Freeman]

What I see is a slight delay in the peaks but the DSP processing itself starts and stops when the original signal starts and stops, so this "pre-sag" as I called it might be something to do with the Native+DAW as I'm using Native for these tests.

 

[621]

i don't understand what it means maybe you can add some sound sample of those amp model

 

[James Freeman]

i don't understand what it means maybe you can add some sound sample of those amp model

Sure thing, here you go;

 

[Sascha Franck]

What I see is a slight delay in the peaks but the DSP processing itself starts and stops when the original signal starts and stops, so this "pre-sag" as I called it might be something to do with the Native+DAW as I'm using Native for these tests.

What if you dial in a clean sound?

 

[James Freeman]

What if you dial in a clean sound?

The last image was with the 2203 gain at 1, a clean sound, as you can see the processed waveform has almost no distortion.

 

[Sascha Franck]

The last image was with the 2203 gain at 1, a clean sound, as you can see the processed waveform has almost no distortion.

Oh, sure.
Yet, don't you think this could just be what guitar amplification is like (or at least some amps)?

 

[James Freeman]

Yet, don't you think this could just be what guitar amplification is like (or at least some amps)?

Yes, it might be the B+ sagging when the tubes draw current form the initial attack, that's why I assumed it is "pre-sag", but it might be not.
Why it starts below the center line though, could be something intentional so the amp model doesn't cut the initial attack so it has some small "run-in" towards the center line.

 

[the-trooper]

Sure thing, here you go;

Sounds flat and uninspiring. I bet pickups with it don't feel right under the fingers either.

 

[James Freeman]

It also occurred to me that the amp model is a complex polynomial equation that the DSP solves on the fly, so the Gibbs phenomenon and sinusoidal nature of the Fourier transform is what we see on the output.

 

[yeky83]

Cliffs comment on the Fractal forum on how he cannot share super secret information about time domain measurements got me interested in time domain measurements and what they can tell us about the modeler.

Helix Brit 2203;
In this example I have a few cycles of 82Hz sine at -10dB, the output is quite interesting.
First of all, the output was correctly (unless I'm wrong) phase flipped because the JCM800 has 3 inverting gain stages, but I have reversed it in this image below so we can see better what's going on.
What I notice immediately is some "pre-sag" before the amp starts to reproduce the tone, also, there is "post-ringing", both of these tell me that there are some elements (components?) that cause this sag and ringing, in a real amp it will be the power supply inductance and filtering capacitance.

I am very curious to measure other modelers and have a look at how they behave, I think time domain response has a lot to do with how the model feels to play.


View attachment 14233

82Hz & 400Hz.
View attachment 14234

Yes, it's another James Freeman thread, go ahead post your shitty memes and destroy beautiful science you banjo twanging luddites.

What you seem to be calling “pre-sag” just looks like group delay...? Beautiful science

 

[James Freeman]

What you seem to be calling “pre-sag” just looks like group delay...? Beautiful science

Yep, could very well be.

Now I'm curious what the AxeFx will output with its elaborate speaker emulation parameters.

 

[Sascha Franck]

just looks like group delay...?

What's that? Sorry, I've got no idea about DSP processing.

 

[Ed DeGenaro]

Yes, it might be the B+ sagging when the tubes draw current form the initial attack, that's why I assumed it is "pre-sag", but it might be not.
Why it starts below the center line though, could be something intentional so the amp model doesn't cut the initial attack so it has some small "run-in" towards the center line.

Im with you on this. Its like what the power supply does on the initial attack.
On the guytron it was actually after the initial attack. Until i changed the filtering.

 

[FractalAudio]

You're sort of on the right track but your wavelets are wrong. You can't use several cycles of a sine wave because they're discontinuous at the beginning and end. What you're seeing on the output is the response to those discontinuities.

You need a function that is not discontinuous otherwise you're simply exciting various resonances and that's what you're seeing.