Pre-dist and Post-dist EQ (Some SansAmp GT2 content)

[Vivek]

The MIMIC paper released by Fractal states that the building blocks of an Amp Model are

Pre-distortion EQ -> Distortion -> Post-distortion EQ


R.G. Keen said "PRE and POST distortion EQ: Filtering of the signal both before and after the distortion usually has at least as much if not more to do with how the distortion sounds than the actual method of distortion."


Teemu K analysed the SansAmp GT2 in his famous eBook

Here are the pre-distortion EQ curves from the GT2, as plotted by Sir Teemu



I can see some variants, by moving the basic shapes along the frequency axis. We can move the peak frequency of the Tube Screamer type curve

For the Tweed curve, we can imagine variants with different Notch frequencies and depths could lead to substantially different sounds



Question 1: Do amps use any other Pre-distortion EQ curves more than the two types shown ? Can you suggest other Pre-distortion EQ curves from other amps ?

Question 2: Is that Notch coming from the Pre-distortion Tone Stack since no other component in an Amp has that kind of frequency response ?

Then there is a stage that does frequency response shaping and distortion at the same time:



This looks suspiciously close to the frequency responses of Cab Sims

Question 3: Can you suggest some other post-distortion EQ curves that are used in real Amps ? Or do I just use other Cab Sim curves here if I want more types of Amp sounds ?


Question 4: Please help me build a short list of some famous Amps that have tone controls before the stage that creates most of the distortion, and other amps that have tone controls after the stage that creates most of the distortion. Then I could go to TSC ON THE WEB and see the corresponding tone stack response and test it at the correct location in the Amp Simulator.

Question 5: Are there graphs available of same Cabinet, same Mic but different Mic positions ? That would guide me on shape of filter for Cab Sim that has "Mic Placement" capabilities.


Thanks

[Andon]

To your last question, it of course depends on which mic you are using, and how close your placement is. Different mics have different frequency responses, and placement matters: center on tends to be brighter with more high ends but less bass, edge is smoother (more mids) but can also be muddy, and right up on the speaker will provide more low end response while backing off will lessen it. Your best bet is to hop into a DAW and mess around with a sine wage input as you move the mic around (or do it in real life if you can!)

When it comes to question 4, it would depend on your definition of EQ, and whether or not you mean all of the components that make up the tone of an amp (various high and low pass filters, etc.) along the way or just the directly controllable portions. Many/most amps and distortion pedals have a bass cut before the distortion circuit and a bass boost after it in order to maintain clarity of signal, but this is done without the input of the user; the EQ that you can control is almost always going to be after the distortion section (though there are examples of distortions where the controllable parameters for bass or mids are before the dirt circuit and the treble, etc. are afterward).

[Vivek]

Thanks Andon


A) I seek actual data on frequency response versus mic versus mic placement. I was hoping someone has published such data, maybe a manufacturer, maybe a PhD student.

Could you guide me to any published data if you know about it, please

B) by EQ, Fractal/Keen meant all components that shape tone, even those aspects that are not user adjustable but are chosen by the Amp designer.


Yes, bass control before distortion and treble after distortion is an excellent idea followed by many Amp designers. I seek more in-depth information like type of Bass control and Treble control used in this application.

Thanks.

[Vivek]

https://townsendlabs.com/why-mic-axis-matters/

I seek more information of this type.

[temol]

A) I seek actual data on frequency response versus mic versus mic placement.

You can generate a wav file with a sweep from 20Hz to 20kHz then apply IR in a DAW (Reaper for example), export to wav, then analyze frequency response.
There's many IR files with different mic positions.

Here you can see an example. 4 different mic positions, freq response generated in Audacity. Better than nothing....

[Vivek]

Here you can see an example. 4 different mic positions, freq response generated in Audacity. Better than nothing....

Thanks !!

Do you have information on :
Which speaker ?
Which mic position ?

for these graphs.

PS : I heard there are some Apps that can directly generate a frequency response curve from an IR. Example:

https://www.vguitarforums.com/smf/index.php?topic=21694.0

[Mark Hammer]

There's always this, which had 6-band EQ before and after clipping.

[Vivek]

That Akai D2G is such a fantastic idea !

D= distortion
2G = two graphic equalisers.

Last year, I was in touch with a person who owned one, and he was kind to post a YouTube video

https://youtu.be/xfLqxbsvolg

https://youtu.be/bdo8SAPhkXo


Then I used the proposed EQ curves of Akai D2G in a DSP project I was working on.

[ThermionicScott]

Question 4: Please help me build a short list of some famous Amps that have tone controls before the stage that creates most of the distortion, and other amps that have tone controls after the stage that creates most of the distortion. Then I could go to TSC ON THE WEB and see the corresponding tone stack response and test it at the correct location in the Amp Simulator.

All of the classic non-master-volume tube amps have the EQ well before the part that clips.  That's why their reputations for distorted tones vary so widely.

Don't take this the wrong way, but have you spent any time actually playing through tube amps?  It seems like you spend a lot of time trying to understand them through modelling.  Which would be like trying to understand sex by reading books about it.

[temol]

Do you have information on :
Which speaker ?
Which mic position ?

I do not have any information about speakers. Mesa 4x12 (V30 maybe?), SM57 1" from the grill cloth. Positions - cap, cone close to the dust cap, cone further away from the dust cap, edge. But I did not write down which is which on the example graphs.

https://wilkinsonaudio.com/products/gods-cab

[Mark Hammer]

A number of DOD's drive pedals differ by use of some pre-distortion resonant boosts in addition to post-distortion EQ-ing.  For instance, the FX66, shown below, employs this strategy.  Not QUITE as elaborate as the Akai, but working under the same general principles.  DOD used fixed resonant boosts, rather than variable, but the strategy is the same: accentuate some parts of the spectrum to make them distort more, and EQ to taste afterwards.

[teemuk]

Reminds me of Metal Zone. DOD, in general, owns a lot to Roland.

[Vivek]

A number of DOD's drive pedals differ by use of some pre-distortion resonant boosts in addition to post-distortion EQ-ing.  For instance, the FX66, shown below, employs this strategy.  Not QUITE as elaborate as the Akai, but working under the same general principles.  DOD used fixed resonant boosts, rather than variable, but the strategy is the same: accentuate some parts of the spectrum to make them distort more, and EQ to taste afterwards.

Thanks Mark !!!

What do Q5 and Q7 do ? They look unconventional and innovative !

Did you plot out the pre-distortion and Post Distortion EQ curves ?

Reminds me of Metal Zone. DOD, in general, owns a lot to Roland.

Dear Sir Teemu, thank you for publishing your eBook. I have gained a lot from it over the years.

Can you suggest some variants of the various EQ curves of the SansAmp, for example to add Vox, Dumble, Rockman and other variants to the available sounds.


Tom of ELECTRICDRUID has the best analysis of Metal Zone on the Interweb : https://electricdruid.net/boss-mt-2-metal-zone-pedal-analysis/

[idy]

Q5 and Q7 just gyrators, yes?

[Mark Hammer]

Yes.  I couldn't tell you exactly what frequency band they are boosting, but boosting is what they are doing.

[Vivek]

Q5 and Q7 just gyrators, yes?

I initially thought after a quick glance, that they somehow help to distort the signal, given that there is no DC bias to the base.

Have I totally misunderstood that part ?




Here's the famous gyrator calculator


http://www.muzique.com/lab/gyrator.htm

[Steben]

gyrators usually are used to mimick an inductor, resulting in RLC filters or in other words form bandpass or bandcut filters.

[Vivek]

gyrators usually are used to mimick an inductor, resulting in RLC filters or in other words form bandpass or bandcut filters.

Thanks Steven,

Do Transistor Gyrators have DC bias on their base ?

[PRR]

...there is no DC bias to the base....

That's gotta be a drawing error. Do we have another source?

[Vivek]

Around 2006, I was involved in development of firmware for one of the top 10 Digital MFX producers of that time.

I had done some research on fitting digital Biquad filters to some target pre-distortion and post-distortion curves that were given to me. I wrote a tool that would visually show the resultant filter created out of multiband Biquad filters.

It was then a manual process to tweak lots of sliders till the resultant curve matched the desired frequency response.

Here are some graphs from that project, along with some notes I wrote in 2006 :


Above are the Equaliser graphs of the 2101 Clean channel.  The pre-distortion equaliser is set at a low level of -10dB so that most of the frequencies pass undistorted through the non-linear stages. Notice however the hump in the pre- EQ around 7 KHz indicating that the highs are more distorted.

The Pre- EQ is set flat, with a low rolloff around 100 Hz. The resultant total equalisation is shown in bold red (but do remember that there is a nonlinear distortion stage in between the pre-distortion EQ and post-distortion EQ).

Below is a typical Nonlinear stage that would be placed in between the pre-distortion and post-distortion EQ




And here are some more Amp models

Above is the EQ charts for a Model trying to emulate the the high gain channel of the 2101.





and below is the EQ graph for one of the acoustic simulators



I am not sure about the accuracy of the EQ graphs given to me in 2006. My task was limited to fitting digital Biquads to the EQ curves I was given, and some similar work to generate about 20 different nonlinear transfer functions


I was now seeking more data on the pre and post distortion EQ curves of common amps.