Guitar Amp Sim for Amp Players?

DSP, Plugin and Host development discussion.
RELATED
PRODUCTS

Post

Can this be done? (in your opinions):

What I am always looking for is a guitar amp simulator for amp players.

It's not a myth — there exists amps that virtually never hit a wall, no matter what you throw at them. I think it is a balance of the perfectly inadequate output transformer in conjunction with the perfectly matched struggling power supply (also no negative feedback). They seem to buck the signal off the rails — the harder you push, the harder they buck. Doesn't ever turn to pure mush, just increasingly bizarre behaviors that truly rock.

The clearest and most exagerated example I can think of is Neil Young's Live Rust album. You can hear Neil heat up the amp by listening to Powderfinger, then hear the full sagging Deluxe glory in the following tune, Cortez the Killer. Pretty much all of that is coming from an early Fender Deluxe with an Echoplex into the front acting as a clean boost (there's a lot more to his rig, but this is the meat of it).

These types of amps are increasingly rare birds, and unruly loud when operating in this mode (can be obnoxious while tinkering — think 115db bagpipe/banjo), and I really believe many people would very much enjoy a more manageable version of the experience if it could be even 80% emulated.

The answer is probably not in the spice models, because we are talking way off spec. (basically a sustained system-wide failure — rolling cascading brownout?)

It seems like it should be within the realm of possibility... some combo of brickwall limiting, wave shaping, envelope generator / VCA, parallel gain stages... somehow.

A friend and I are going to tinker with the idea, but neither of us are amp techs nor coders, so we will be limited to synthedit and open source, just hoping for more of a fun proof-of-concept outcome.

It's an untapped niche as far as I can tell — someone should do it right.
:pray:

Please — no corporate enforcer goons insisting it's already been done. This is no way directed at or about you in any way, shape, or form. What you got just ain't it.
Last edited by guitarzan on Sat Sep 07, 2019 6:44 am, edited 1 time in total.

Post

The only way to get this done is, you need to do it yourself. Those corporate enforcer goons would not do it anyway, as they are fine with what they have.
They would only listen to you if you are a multimillion guitar hero yourself.
“The new amp sim designed by GuiTarzan, if you wondered how to get his sound, buy this breathtaking new sim NOW!” People will rave about the Gui and the Tarzanian sound....

Post

GuiTarZaNic Flux Inducer: ENGAGED!
:lol:
No DSP knowledge at all, unfortunately. No retained programming skills of any kind, really.

Post

I think everything can be done, if you don't taking into account of what it's going to cost in the end. If there ever was a face for "law of diminishing returns" in guitar amp sims, specific performance (I have those one or two beefs and gripes with the amp sim industry too) it is this one. Why no one is doing it is because of developing costs, and the end result that maybe just will please you and me, and maybe Neil Young. To refine and better something, or copying an aspect of real power amp to 100 % - say better i by 5 % - is not going to end up costing 5 % more. It's going to cost 500 % more, and be laid upon the asking price.

I don't think anyone would cough up 1 million dollar for buying an app that will do this down to a tee, and a couple of other side effects, and "fringe" sounds that has been left out before. I have no doubt that it will be possible to do this, it's just a matter what it costs in pure cash in the end. Sunken costs, white elephant, not "in-house", a world of pandoras boxes is opened.

I e in the end: It's not just worth it.

But so it is with real world boutique amps such as Dumble, Trainwreck et al. Bettering a small early nuisance or annoyment from a select few players, by 5 % doesn't make the amp cost 5 % more. It costs 250 % more in the end. For a difference that is so subtle that it is neglible.

Post

Maybe so for a corporation, but it's not necessarily going to cost anybody anything but time. I don't care one bit about "modeling" a specific amp. I would just like to see someone simulate amp behavior that most people seem totally unaware of because of the rarity (not all tweed era amps will do this) and impracticability (115 db bagpipe-banjo).

Personally, the only reason I even have guitars is to light up an amp in this way. In my world, the amp is the instrument and guitar is the controller.

People would love it if they could get just a taste of it, I really believe that. It is not subtle in operation — not at all. It has always struck me as being raw analog synth-like at it's extremes... wave shaping, amplitude envelopes...

Most importantly, it's the most fun you can have with a guitar. I think the fun would still be there even in a crude emulation if it was well executed.

Post

guitarzan wrote: Fri Sep 06, 2019 8:46 pm The answer is probably not in the spice models, because we are talking way off spec. (basically a sustained system-wide failure — rolling cascading brownout?)

It seems like it should be within the realm of possibility... some combo of brickwall limiting, wave shaping, envelope generator / VCA, parallel gain stages... somehow.
I have limited experience with guitar amps and tube circuits in general, but since I kinda know a thing or two about Spice-like circuit simulation, I'm going to claim that you almost certainly got it backwards and Spice-like simulation is almost certainly the only thing that actually has some hope of working properly.

While it depends a bit on how realistic device models one is willing to use, in many cases Spice can handle somewhat realistically anything that doesn't involve magic smoke. The particular stuff that you describe might be a "failure" from the point of view of amplifier design, but from the point of circuit simulation the circuit is working just fine.

The downside of this approach is that if you want to model something like power supply sag, you need to put the whole circuit including it's power supply into a single simulation and the computational expense tends to grows pretty fast when the simulation gets larger (ie. doubling the circuit size, you might theoretically end up with 8 times the CPU cost in the worst case, although practical circuits are rarely quite that bad). There's a hard cap of how complex the circuit can grow before you simply can't run it in real-time anymore. The fact that you also need to simulate at high sampling rates to deal with aliasing doesn't exactly help either.

Either way, lots of intelligent hard working people have tried to "fudge" such behaviour from simple modules for something like 20 years now. That's basically what every older (and probably a few newer ones) amp model does. As far as I can tell, some of the newer ones (in particular those people complain are CPU hogs) actually might do some proper simulation too, but if they don't put the whole amplifier in a single simulation that's almost certainly a performance trade-off.

Post

All I know is every amp modeler or sim I have ever tried stops just short of the kind of behavior I am referring to.

For one thing, the sag as usually represented is heard, but not really felt. It should have the feel under the fingers of a triggered envelope.

Also, sag is just the beginning. Listen to the wave shaping going on in Neil Young's wildest stuff. I think sag contributes, but there is more going on. I don't even remember where this is coming from (I've talked to everybody about this stuff) but I also want to blame the power amp slew rate going AFU because of voltage fluctuations.

Maybe at a certain point the voltage sag causes the power amp to behave more like a Voltage Controlled Amp (VCA), and then the voltage fluctuations would act as an Envelope Generator (sag), while also causing the slew rate to shift into AFU wave shaping mode.

In the end, I don't really care about circuit behavior, all I care about is guitar signal behavior. I think it can be fudged — in fact, I think that may the best bet in a sometime-during-my-life way.

Run it through a good speaker cab sim / IR and it'll be absolutely close enough for rock'n'roll IMHO.

I'd be happy whatever way it took to get there. Maybe hybrid modeling / fudging. My goals have never really lined up with amp "modeling" — some of the old sims just seemed funer and more r&r to me — aliasing aside. Definitely far easier on the CPU.

I guess I'm suggesting something like a boutique niche plugin, similar to old school cool boutique guitar pedals, where it isn't expected to appeal to everyone or to perfectly model anything specifically, just be cool and fun.
Last edited by guitarzan on Sat Sep 07, 2019 7:20 pm, edited 1 time in total.

Post

What about a hybrid of algorithm and Vectorial Volterra Kernels?
Currently trying to turn noise into music. :neutral: Is boutique the new old?

Post

The Noodlist wrote: Sat Sep 07, 2019 6:53 pm What about a hybrid of algorithm and Vectorial Volterra Kernels?
Is that like what Nebula does? I always thought speaker cab modeling would go that route someday. Maybe... it was always too CPU intensive for me to get much use out of it back when it first came out. I haven't kept up with it's progress at all, though.

Post

It's important to bear in mind that when you hear Neil Young that you're not just hearing the 6L6 modded Deluxe. Pre-deluxe is an old Fender reverb with a 12AX7 and a pair of 6V6 valves and, apparently, the type of 6V6 is critical to the sound. Signal passes to a Magnatone and an old trannie amp (can't remember offhand) and the former is usually mixed higher live and studio (although the Deluxe and Baldwin are what Young hears most of on stage). All of this stuff is well-covered in various forums.

I don't know if you'd want to model all of these (never mind other things in the signal path). Of course, you'd have to model the interaction of the amp setup with the incredibly microphonic Firebird pickup too.

Post

Yeah...I posted a whole thing about his rig somewhere here on KVR a while back. The SS amp is a big ass Baldwin Exterminator — according to what I've read, only Neil hears that one, it doesn't make it into the mix — theory is that one is just to kick in the feedback. I think I read that was replaced by a Mesa Boogie more recently. The Magnatone 280 has that killer stereo vibrato — when the pitch raises in one channel it falls in the other — that is about 20% of the mix according to an interview with his tech (you can hear it as the subtle modulation under everything). Plus the outboard Fender Reverb and the tube Echoplex...few other pedals for specific tunes. The whole thing is basically a hot rodded surf rig — crazy man.

It would be cool if it were all modeled, but that's beyond what I'm suggesting now. I'm really not suggesting "modelling" anything right now, but it would be fun to have. I would call the whole modeled rig the "Lazy Horse". :)

EDIT: I do have a pair of vintage spec Firebird pickups — I'd like to put them in my SG.
Last edited by guitarzan on Sat Sep 07, 2019 10:51 pm, edited 2 times in total.

Post

I know what you mean about "playing the amp". As a bass guitarist if I have a good/great amp I can get a usable tone regardless of the bass...but a deluxe bass into a sh!tty amp just sounds strained and crappy.

Have you tried the Airwindows plugins, particularly Power Sag? There are some Gearslutz threads on other Airwindows plugs discussing how to use them as modular bits that, in series, emulate some pretty crazy analog behavior.

Post

I absolutely do want to try some of those Airwindows plugs. Besides PowerSag, I saw one called Slew, and maybe one called TubeSag or something. Yeah, seems like top notch and unusual stuff. I just emailed someone a couple days ago about those. I'll check out that link ASAP. Thanks.

Post

Zen-Drive Pedal Boutique in Acustica Nebula form. Amps are another area.
Ultimate Guitar Mega Pack Collection .
We only have to look at the advancements in visual technology, like CGI and animations. Hopefully, with more powerful computers and advancements in software, more realistic amp sims will become available.
Currently trying to turn noise into music. :neutral: Is boutique the new old?

Post

guitarzan wrote: Sat Sep 07, 2019 3:28 pm ... In my world, the amp is the instrument and guitar is the controller...
Most definitely. Very few people thinks like this. I've always thought that the amp together with the guitar forms ONE instrument, and the other can't be excluded. The cable from the guitar to the amp is the 7th string so to speak. I've always wanted that the amp sims should include the warts and all from the real world amps. The thing with electric guitar in my world, is that - provided you have the right real amp and real settings - is that while you're playing you should be able to sound both NICE and clean and "linear" and with the split second movement of your pick attack you can sound UGLY, if you'd like. And of course, mostly in between. You should be able to get hideous out there sounds with the pick attack, and go totally mayhem...


https://www.youtube.com/watch?v=sqHz7cUw4Ls

Post Reply

Return to “DSP and Plugin Development”