Dynamic response cab sim?

[Fancy Lime]

Hi there.

I was wondering if there is a design for a cab sim out there that not only mimics the frequency response of a guitar cab but also the dynamic response  of a closed back cab at high volume. The biggest contribution here would be what the cab people call "compression" and what is in our terms soft clipping. Wouldn't be too hard to do, would it? Couple of 2N7000's or BS170's wired back to back to ground (gate and drain tied together, the soft-clipping diode way), push that with the right amount of voltage swing and we're done. But is that even worth the trouble? I mean, we usually feed the cab sim with an already clipped signal. So would we even hear the difference of yet another mildly clipping stage?

EDIT: We may want to add a treble bypass cap to the MOSFET clipper, so that only the low frequencies are "compressed" as the would be in a real closed-back cab.

Cheers,
Andy

[ElectricDruid]

I don't know, but I doubt it.

Reducing a cabinet's response to a single frequency response is a big approximation to begin with, since I doubt the frequency response is identical at all power levels, so it should really be some kind of family of curves. Plus, as you say, you can start to get non-linear behaviour too.

My guess would be that when people get into that level of detail for a cab sim, they're working in DSP land.

[bool]

Oh, compression.
In a daw, a tri- or even dual-band comp will get you there fast; otherwise you can play with the idea of a muff-style tone control with a soft-clip pair across the lpf.
Even a baxandall tone control can be "enhanced" this way.
https://patents.google.com/patent/US5509080

[Fancy Lime]

I don't know, but I doubt it.

Reducing a cabinet's response to a single frequency response is a big approximation to begin with, since I doubt the frequency response is identical at all power levels, so it should really be some kind of family of curves. Plus, as you say, you can start to get non-linear behaviour too.

My guess would be that when people get into that level of detail for a cab sim, they're working in DSP land.

If we wanted to get an accurate representation of all the behavior of, say, a Mesa 4x12 with Celestion vintage 30's, then yes, DSP is pretty much the only reasonable way to do it. But for a simple analog cab sim, what I would be trying to do is something that "makes the guitar signal sound better in roughly the same way that a typical (whatever that means) guitar cab does". I'm just wondering if for that sort of thing we would even want to do anything other than frequency shaping. Or in other words, if we could design a "perfect" real cab, would we want it to compress? We certainly want it to have a non-linear frequency response, as anyone who has plugged their Marshall head into a full-range PA cab can attest.

Andy

[ElectricDruid]

But for a simple analog cab sim, what I would be trying to do is something that "makes the guitar signal sound better in roughly the same way that a typical (whatever that means) guitar cab does".

Yeah, fair enough. That's a realistic goal. And probably doesn't need anything much more than tweaking the frequency response. Like you say, for an ideal cabinet, we'd like to leave the cabinet distortion out anyway.

[Elektrojänis]

The biggest contribution here would be what the cab people call "compression" and what is in our terms soft clipping.

The compression in speakers is not all clipping. When driven hard thermal compression is usually significant. The voice coil heats up and it's impedance changes. This can be quite significant as the voice coil is usually quite small for the power it has to handle (and most of that power will be converted to heat).

[dschwartz]

I personally find speaker artifacts unpleasant. Most of the dynamic effects happen when the speaker is actually being damaged or it is already damaged.. and even the fanciest IR are linear and no one complains of lack of "speaker dynamics".
Interestingly, heavy filtering gives the impression of compression, many users of my Omnicabsim pedal had asked me if it has a compressor inside..
I wouldn't bother to add dynamic effects to it, and would rather imitate power amp like clipping.

[Fancy Lime]

The biggest contribution here would be what the cab people call "compression" and what is in our terms soft clipping.

The compression in speakers is not all clipping. When driven hard thermal compression is usually significant. The voice coil heats up and it's impedance changes. This can be quite significant as the voice coil is usually quite small for the power it has to handle (and most of that power will be converted to heat).

Hi Petri,
Yes, sure these effects exist as well. But would we want to emulate them if we could? I mean, we *can*.  Digitally that would be very easy and even possible with a very slow analog compressor. But I'm guessing there is a reason why 15W through an 8x12 full stack sounds better to many people than 150W through a 2x12. So emulating speakers at their thermal or mechanical limit may not be all that good for the sound. So why emulate it, right?

Hyvää,
Andy

[Fancy Lime]

I personally find speaker artifacts unpleasant. Most of the dynamic effects happen when the speaker is actually being damaged or it is already damaged.. and even the fanciest IR are linear and no one complains of lack of "speaker dynamics".
Interestingly, heavy filtering gives the impression of compression, many users of my Omnicabsim pedal had asked me if it has a compressor inside..
I wouldn't bother to add dynamic effects to it, and would rather imitate power amp like clipping.

Me thoughts exactly. I am still a bit fuzzy on what frequency response curve I would want for a cabsim. Many of the available cabsims seem to be sort of similar but a few have very different curves. I'm gonna have to SPICE the $hi+ out of this, once I get the damned program running.

Cheers,
Andy

[dschwartz]

I personally find speaker artifacts unpleasant. Most of the dynamic effects happen when the speaker is actually being damaged or it is already damaged.. and even the fanciest IR are linear and no one complains of lack of "speaker dynamics".
Interestingly, heavy filtering gives the impression of compression, many users of my Omnicabsim pedal had asked me if it has a compressor inside..
I wouldn't bother to add dynamic effects to it, and would rather imitate power amp like clipping.

Me thoughts exactly. I am still a bit fuzzy on what frequency response curve I would want for a cabsim. Many of the available cabsims seem to be sort of similar but a few have very different curves. I'm gonna have to SPICE the $hi+ out of this, once I get the damned program running.

Cheers,
Andy

There are 3 key factors i consider when designing a cabsim. One is the actual speaker freq response ( which are available in datasheets) , then, the physical cabinet resonance (open, closed) which has an effect on the bass response, and the power amp Damping effect (which peaks the low end and boosts the high end) ..all three considered, make a pretty damn good cabinet sound, and really good for live sounds..if you add the microphone response (additional lpf and peak at the high mids) it gets better for recording.

[Rob Strand]

Digitally that would be very easy and even possible with a very slow analog compressor. But I'm guessing there is a reason why 15W through an 8x12 full stack sounds better to many people than 150W through a 2x12. So emulating speakers at their thermal or mechanical limit may not be all that good for the sound. So why emulate it, right?

When you listen to multiple speakers you are listening to the sum of multiple speakers.   The distance from each speaker to your ear is different which creates a time delay.  The listening angle (to the axis of the driver) is different so basically you are listening to the sum of many slightly different speakers which are delayed from each other.   There might also be some near field and coupling effects thrown in.  And then the effect of the floor and the room in general.    So there's all those complications which do make an impact.     They are all linear effects, no need to consider compression or distortion.   (To me) standing near a large stack of speakers does sound better than a 1x12 even at low volumes.

The nonlinearities come from the magnetic field changing at different excursions.  There's also non-linearity from the suspension.     People have come-up with all sort of models for this.   Look-up non-linear Thiele-Small.    The effect largely occurs at low frequencies.   At mid and high frequencies the speaker doesn't actually move much

When you have a low frequency *and* a high frequency signal at the same time the low frequencies will affect the high frequencies as you get a modulation effect.   This isn't unlike a diode clipper.   There's also FM effects where the physical movement of the speaker modulates the high frequencies by the Doppler effect; plenty of papers on this.

The thermal effect is slow and causes a drift.   If you put a resistance in series with the speaker it achieves the same effect as a hot coil speaker..   If you do this it doesn't make enormous tonal changes, it just changes the bass a bit; how much effect depends on the amp damping factor.

IMHO, the secondary issues aren't where the problems are on a cab sim.    Like dschwartz mentioned the IR responses sound pretty good and they are totally linear.     The IR's can also capture some of the effects of a multiple speaker array.

EDIT: Some links on non-linearity

http://audiomobile.net/wp-content/uploads/2014/08/Klippel_Nonlinearity_Poster.pdf

http://www.klippel.de/fileadmin/klippel/Files/Know_How/Literature/Papers/Prediction_of_speaker_performance_at_high_amplitudes_01.pdf

http://www.klippel.de/fileadmin/klippel/Files/Know_How/Application_Notes/AN_21_Bl_Shift.pdf

[bool]

First and the all-important Q: do you really want/need all these nonlinearities?

I recall sound engineers were putting miked/recorded guitars thru ampfarm plugins to make them "sit still" in a radio-oriented pop-rock mix context in late 90's already! Iow to even out the unwanted amp/mike nonlinearities and to "add" some that people want to hear but don't come straight out of the amp...

I guess this approach was just the opposite of what you're up to...

[ElectricDruid]

When you listen to multiple speakers you are listening to the sum of multiple speakers.   The distance from each speaker to your ear is different where creates a time delay.  The listening angle (to the axis of the driver) is different so basically you are listening to the sum of many slightly different speakers which are delayed from each other.   There might also be some near field and coupling effects thrown in.  And then the effect of the floor and the room in general.    So there's all those complications which do make an impact.     They are all linear effects, no need to consider compression or distortion.   (To me) standing near a large stack of speakers does sound better than a 1x12 even at low volumes.

Those tiny time delays are going to put some notches and peaks into the frequency response, especially at higher frequencies (where the wavelengths are shorter). Subtle movement of your head would move those peaks and notches, giving a more "alive" sound - basically a mild phaser gently applied to your signal. By the time you're boogieing about in front of the speakers, the phase shifts will be pretty significant!
Since this effect can only occur when you've got multiple sound sources producing the same signal, it's no surprise that a 4 x12 sounds better (more interesting?) than a 1 x12.

[Rob Strand]

Those tiny time delays are going to put some notches and peaks into the frequency response, especially at higher frequencies (where the wavelengths are shorter). Subtle movement of your head would move those peaks and notches, giving a more "alive" sound - basically a mild phaser gently applied to your signal. By the time you're boogieing about in front of the speakers, the phase shifts will be pretty significant!
Since this effect can only occur when you've got multiple sound sources producing the same signal, it's no surprise that a 4 x12 sounds better (more interesting?) than a 1 x12.

It has a fairly strong effect on the response.

If you just consider these things you get some pretty extreme looking frequency responses:
- the phase-shift (time delay) due the different distances
- the slightly different levels due the different distances
- the reduction in high frequencies due to the off-axis response of speaker

There's some nice pics of off-axis responses here,
http://rutcho.com/speaker_drivers/supravox_t215_rtf_64/supravox_t215_rtf_64.html

This is for a 21cm speaker.  Image shifting the curves down in frequency by a factor of (21/30=0.7) for a 12" driver.   If you are standing close to the speakers the angle from the ears to the driver axis can be fairly significant, especially to the driver down low on the ground.   You can see from those curves that the high frequencies are strongly attenuated off-axis.  For an array no many speakers are contributing to the high frequencies.   When you are far from the speaker the angle becomes less and you get more highs.

One caveat:  off-axis mic'ing will give a much different response to these distant off axis plots.  The mic is in really close, in the near-field, and the response there is quite different to the far-field response.

Of course in a real room the situation is more complex.   You also have the response of your ear to consider. You ears will enhance the high frequencies for sounds approaching from the sides, which undoes the speaker roll-off a little bit.

I looked at this stuff about 20 years ago.  I've got a model which sums up the response based on the items listed above.  It include a model for the off-axis response as well.   You basically specify the positions of the drivers and the listening position and it sums up the pressure responses.   It assumes a flat response driver so you have to multiply the result by the response of the driver.    I could also add a reflection from the floor.  The result was a very unintuitive looking response and *much* larger changes to the response than you would expect.   If I get time I'll pull out this stuff post a plot.


This is a set-up with all drivers spaced 350mm between centers.    The listening position is at 3m away from the speakers, the ear around head height and a fraction off the centre in the left/right sense.   The speakers are on the ground (first driver center about 160mm from base) and build up higher and wider as more drivers are added.


What's not included: no floor or wall reflections,  no modelling of diffraction loss from the enclosure, no ear response mods.

There is a hack in the off-axis response to prevent the attenuation going too low.

The responses are normalized around 800Hz.

Just to emphasize the driver has a flat response.   The response of the guitar speaker would add (in dB) to what is shown.  So the rise in the response you see on a guitar speaker at high frequencies would raise the level of the highs.

[ElectricDruid]

Very interesting, Rob. Thanks.

[bool]

The off-axis stuff is the one that has some relevance; that's what you usually do when miking; and in that french speaker curves you can see what it does to the - say 2-4k freq range. Guit. speaker curves will be somewhat different, and the mikes used as well; but the same-ish "eq signature" will be there...

Not that any of this matters for the die-hard "want to emulate tube amp and cabs" camp.

[Fancy Lime]

Hmm, this turned out interesting! I never thought about the whole "interference between cones" thing. Explains why a single direct recorded guitar often sounds more "realistic" or more "live" through earphones or monitors when you add a tiny bit of chorus or very subtle flanging. More of a trick for mixing that for a cab sim, though. I mean, that is exactly the sort of thing we normally want to minimize by close miking with a (super)cardiod mic. But for a "virtual room mic", adding a bit of chorus and reverb is probably a good idea, if done right. I don't see me using that on anything other than a single clean or lightly overdriven guitar. Having two distorted guitars and then adding that "virtual room mic" would probably go muddy real fast.

Andy

[Rob Strand]

To me the response has like a pink noise filter shape it  (ie slope 3dB/oct or 10dB/dec) but at some point tapering off at low frequencies.

One thing which stands out  is the bass boost.   That comes about because the bass pretty much always adds, whereas the highs are subject reduction due to off-axis  response and due to cancellations.   

The baffle diffraction loss will cut the bass a bit.   The larger stacks will have a larger frontal area and the diffraction loss will be pushed down to lower frequencies compared a smaller enclosure.  So that further exaggerates the bass on a larger stack.
https://www.trueaudio.com/st_diff1.htm

When Marshall started using these stacks was it the bass boost of the enclosure which inspired them to put the presence boost and low-cut voicings in their amps?   That creates a scenario with bass cut before distortion and bass boost after (from the speakers).   That was the first thing that crossed my mind when I saw these curves.

More of a trick for mixing that for a cab sim, though. I mean, that is exactly the sort of thing we normally want to minimize by close miking with a (super)cardiod mic.

It makes you wonder what aspects do we want to keep and what aspects do we want to reduce on a cab sim!

[Fancy Lime]

...
It makes you wonder what aspects do we want to keep and what aspects do we want to reduce on a cab sim!

Exactly. A big part of frequency shaping in amplifiers is to iron out the unwanted frequency shaping elsewhere in the chain of sound, including of course the cabs. That is why a guitar amp through a flat response PA cab sounds awful but guitar through a flat response PA amp + flat response PA cab sounds... well, still not good but a lot better than the same cab with a guitar amp.
When building a complete direct recording setup (which is what I a up to at the moment, hence the question about cabsim in the first place), consisting of a studio preamp plus cabsim, it makes sense to me to make them work together for a good result rather than work against each other. We are not limited by the physical limitations of real speakers here.  The preamp I am designing at the moment actually has a relatively flat response compared to "real" amps. So my inkling is that the corresponding cabsim will have a response that will only partially resemble that of a real cab. The goal is, after all, to make the combination sound good. And "good" in this context is pretty much defined by what we are used to hearing on our favorite records, therefore: Amp+Cab+Mic (very well and neatly defined then, isn't it?).

Cheers,
Andy

[Rob Strand]

When building a complete direct recording setup (which is what I a up to at the moment, hence the question about cabsim in the first place), consisting of a studio preamp plus cabsim, it makes sense to me to make them work together for a good result rather than work against each other.

Agreed.  I've been saying stuff along those lines for a long time.

The preamp I am designing at the moment actually has a relatively flat response compared to "real" amps. So my inkling is that the corresponding cabsim will have a response that will only partially resemble that of a real cab.

Yes, when people plug their pedals into a cab sim, or pedals which have built-in cab sims you get the same issue.  The whole amp tone shaping thing is discarded.   

Amps aren't flat at all, even with all the controls at 12:00.  It's like you need an amp voicing switch you can pull in.  That way the amp part and speaker part can be separated.   

On a tube amp the signal at output of the power amp has more bass boost and more treble boost than the signal at input of the power amp.  So the amp/speaker interaction is yet another variable.   IIRC some of the Red Box cab sim models tried to capture this difference depending on which sockets you used.