[question]Power amp resonance simulations and solutions

[VoivoD]

Hi.
Well as EQ based cabsim can be enough, what I think really isn't covered well by amp sims is the resonance (knob) of the power amp.

So I'm thinking about what solution could be taken to achieve this special effect (which one must understand from where it comes from).
So there are like EQ-based sims to achieve something similar, but unless they have some special magic into them, they are not gonna work.
Is there any other way to achieve like 90-95% of the real deal?
Maybe an IR would be ok? But loading IR's in stompboxes is kinda limited - notice the Mooer Radar very short impulse 1024 samples.
Could someone tap into their Amps fx loop, dial some combinations of knob values of resonance and maybe presence, capture it after a load box (or some kind of attenuator) and make some IR's out of it, that one could load to for example Mooer Radar?

[iainpunk]

i think that issue would probably solved if you were to use a load box on the amp that is not only resistive, but also inductive and even a tiny bit capacitive.
if you take an impulse response from that, it would incorporate the resonance in the amp sim you are using that IR for.

welcome to the forum,
cheers,
Iain

[Rob Strand]

I'm pretty sure you can replicate the effect with pure EQ ie. filtering.

It's a combination of:
- amp output impedance
- speaker impedance
- amount of gain (amplification not overdrive) / negative feedback in the amp
- the implementation of the resonance control.   I wouldn't call this a standard.
- where the control is set.

The thing to realize is it depends on the speaker.   Applying filtering effect measured from one speaker to another speaker can only ever be an approximation.  Probably OK provided your base-line response is representative of what you like.

If you picked a speaker then you can actually come-up with a reasonable approximation of the response.   The speaker is covered here,
http://www.aikenamps.com/index.php/designing-a-reactive-speaker-load-emulator

The old Marshall cab sim patent has similar info.


The output impedance of the ampiflier is a variable as well.   IIRC some commentary on that in the Marshall patent.

The last step is to combine that stuff with the resonance control itself.

[VoivoD]

@Rob Strand
No way you can do that with EQ. It basically makes amp out of control in these frequencies, making the speaker resonate with its own producing current (as far as I understand it, and used it, current me if I'm wrong). TBH it's more like a reverb than EQ. You can get the final frequency curve the same, but the sound/playthrough will not be the same, and the real thing is pretty magical.

@iainpunk
Something more or less like that.

[Rob Strand]

@Rob Strand
No way you can do that with EQ. It basically makes amp out of control in these frequencies, making the speaker resonate with its own producing current (as far as I understand it, and used it, current me if I'm wrong). TBH it's more like a reverb than EQ. You can get the final frequency curve the same, but the sound/playthrough will not be the same, and the real thing is pretty magical.

Maybe you are thinking of something different to me?

https://patents.google.com/patent/US5197102A/en
https://robrobinette.com/Voicing_an_Amp.htm#Resonance_Control
http://www.regiscoyne.com/tech/resonance/

[teemuk]

The main problem is that the (power amp's negative feedback -based) resonance control is not just a simple tone control but a low frequency "damping" control, and its effects are therefore dependent on several "non standard" variables.

We can easily estimate the "simple tone control" -effect, which is a shelving filter with range of difference between open and closed loop gains.

BUT...

What exactly is the open loop gain or closed loop gain of the amp (note that the impedance and therefore OLG vary at resonant f)?
We get a different answer with each amp.

What is the resonant f and at what Q?
We get a different answer with each speaker system.

Assuming we know all the above, what type of response we emphasize with the resulting gain function?
We get a different answer with each speaker system.

Assuming we can solve all that, how will the distortion profile vary between OLG versus CLG? How do we know at which point the non-linearity would start causing distortion?
We get different answers with each amp.

How much more speaker distortions, overshoots and "ringing" will there be when the cone is not damped vs. is damped?
Answer depends on speaker system and amp.

Etc.

So, we could interpolate between IRs of amp + speaker system with specific extremes of damping, but we would need infinite amount of such "profiles" to model the infinite amount of choices we have with different amps and speaker systems.

[VoivoD]

TBH peavey 5505 or 6505 + mesa v30 cab would be ok for me

@Rob

"The Resonance Control does for low frequencies what the Presence Control does for highs. The Resonance Control blocks the flow of low frequencies in the negative feedback circuit. Blocking low frequencies from the NFB circuit acts to boost those low frequencies at the speaker. Not only are the low freqs boosted but they become "harrier" with added harmonic distortion. The bass response also becomes audibly "looser" due to less speaker damping. Use the Resonance Control to subtly tweak your amp's low end."

Quite accurate description, not sure I understood yours.

[Rob Strand]

"The Resonance Control does for low frequencies what the Presence Control does for highs. The Resonance Control blocks the flow of low frequencies in the negative feedback circuit. Blocking low frequencies from the NFB circuit acts to boost those low frequencies at the speaker. Not only are the low freqs boosted but they become "harrier" with added harmonic distortion. The bass response also becomes audibly "looser" due to less speaker damping. Use the Resonance Control to subtly tweak your amp's low end."

Quite accurate description, not sure I understood yours.

That's more or less how I see it too.   The difference is he's using a descriptive language rather than scientific language.   That's why it might agree with the image in your mind.

"The Resonance Control does for low frequencies what the Presence Control does for highs."

"The Resonance Control blocks the flow of low frequencies in the negative feedback circuit. Blocking low frequencies from the NFB circuit acts to boost those low frequencies at the speaker."

On the face of it it's pretty straight forward: presence boost highs and resonance boost the lows.  Simple as that.  It's just an EQ.   

The first sentence carries a deeper equivalence than the second because both controls depend on the tube amplifier gain (open loop gain is the technical term).

However there's more to it. If you wrapped the presence and resonance control feedback network around an opamp, you would get a different response.   The reasons are:
- the opamp had a lot of gain so the response will follow the (simplified) frequency response based on just the caps and resistors in the feedback network.
- a tube amp has a low gain so that makes the actual response different to the simplified frequency response.   You need to allow for the reduced gain.    When you do this you will find as you increase the presence or resonance control the deviation between the opamp case and the tube-amp gets more and more.
- the response of the tube amp depends on the load.   The impedance of a speaker is not like a resistor it  is frequency dependent.
http://www.aikenamps.com/index.php/designing-a-reactive-speaker-load-emulator
  That makes the response deviate even further and the response depends on the speaker.
  To make things a litte more difficult to compare, the response in the opamp case will be relatively flat when the controls are backed off.
  In the tube amp case when the controls are backed off the response still isn't flat because of the amp/speaker interaction.
  One way to undo this effect is to compare difference (in dB) between the non-boosted response and boosted responses.
  Even with this modification the tube amp will be slightly different to the opamp.

The basic idea doesn't change between the opamp and the tube amp  The specifics is all about 'devils in the details" and the differences originate from the fact the tube amp has a low gain and the gain depends on the load (which is the speaker).

The short story is the opamp case is very simple to calculate.   The tube amp case is more difficult to calculate.   Both can be simulated.   Despite all the details the result is just a frequency response, an equalization curve.

" Not only are the low freqs boosted but they become "harrier" with added harmonic distortion. "
So that part is true also.  This is a non-linear effect which cannot be achieved with EQ.  It fits into the tube distortion aspect.
The question is if you don't emulate this effect how close is EQ base emulation of the resonance control to what you want to achieve.

"The bass response also becomes audibly "looser" due to less speaker damping"

This is where the descriptive language can really put you on the wrong track.    It does sound looser.    However, there's no magic to "damping" or "loosness".    The change in frequency response due to changes in damping is perfectly modelled with *just* a change in EQ.   

In these links, the extreme case of voltage drive, amplifiers with a low output impedance and high damping, are compared with the other extreme case of current drive, amplifiers with high output impedance and low damping.

http://education.lenardaudio.com/en/14_valve_amps_7.html

You can see from this figure the frequency response with high damping and no damping.   All you need to do is apply an EQ.   It goes further than that you can actually calculate the frequency responses and use maths to determine what that difference is.

http://copyright.lenardaudio.com/laidesign/images/a14/a14_vs_vdcd-response.gif

A tube amp is between the two cases of a low and high damping amplifier.    The addition of the resonance and presence controls does complicate things but the key point is: For any particular amp + speaker combination the effect can be modelled as an overall EQ.

[VoivoD]

Wow pretty big response, but in the end, I really disagree. Look, I fiddle with equalizers for 20 years, and yet not one time did it work like resonance in a power amp. You can't compare the frequency chart with how the freaking thing sounds. Maybe you can't hear it or whatever, yet I completely disagree with that simple EQ thing. Not sure from where you come from, from what gear, but that's completely wrong, and I will not waste time arguing about it. Take it or leave it.
Resonance is magical and is very important in the sound of an amp. Hey, I even heard solid-state amps that had very low damped lows, and the low end in them was all over the place. Cool effect, not really an EQ. The increase in the bass is from the speaker resonating from lack of control and because you dump less signal into the negative feedback. It's not the actual effect, it's just a side product, and if I would want that, I would have done the same in like few seconds with an EQ and don't even ask here.

[Rob Strand]

Wow pretty big response, but in the end, I really disagree. Look, I fiddle with equalizers for 20 years, and yet not one time did it work like resonance in a power amp. You can't compare the frequency chart with how the freaking thing sounds. Maybe you can't hear it or whatever, yet I completely disagree with that simple EQ thing. Not sure from where you come from, from what gear, but that's completely wrong, and I will not waste time arguing about it. Take it or leave it.
Resonance is magical and is very important in the sound of an amp. Hey, I even heard solid-state amps that had very low damped lows, and the low end in them was all over the place. Cool effect, not really an EQ. The increase in the bass is from the speaker resonating from lack of control and because you dump less signal into the negative feedback. It's not the actual effect, it's just a side product, and if I would want that, I would have done the same in like few seconds with an EQ and don't even ask here.

Maybe the thing you are looking for out of the resonance control isn't the EQ part - simple as that!    Tube amps are complex beasts.   There's lot going on.   

To me there's no denying the EQ aspect is there.   It's a well known thing and there's a reason why that guy wrote the page on voltage drive vs current drive.What I will say is if you tried to do a speaker sim by playing with an equalizer it's a tough job, a generic equalizer is perhaps not the tool for the job.  Yet we know a large number speaker sims are largely EQ.       I'm not saying you have got it wrong.   In the larger scope of things, there's more to it than just EQ.

There's also acoustical effects.   Rest your guitar up against a cranked amp and a whole freaky thing happens.   You don't have any expectation that's going to happen using a DSP setup and headphones.

As far as getting what you want.   Perhaps the things you want to emulate are the things that aren't EQ based.   That falls into the realm of modelling amps.   A whole world unto itself and probably outside of what is achievable with a simple pedal.