The Big Guitar Amp Sim Roundup + Review

[Anderton]

I have been hearing for a while that Guitar Rig is older tech and behind the times compared to other products on the market.

Not compared to a Fender Twin Reverb

I was under the impression that emulations from that era weren't exactly spot on.
From what I hear from people who are more knowledgeable, emulations still aren't there yet.
You obviously know way more than I ever will, but your statement above sounds like empty rhetoric to me. That Fender may be a rather simple machine, emulating every little thing that happens when you turn it on in software isn't.
(Just my 2 cents, please correct me if I'm wrong)

Actually, that was my point A Fender Twin Reverb is "older tech," and "behind the times" compared to newer amps that have newer features or whatever. But, nothing does what a Fender Twin Reverb does, so I don't care how old it is or what tech it used.

Similarly, nothing does what Guitar Rig does. Whether someone likes what it does or not is a separate issue, but just as I keep some old tube amps around because nothing sounds quite like them, some of the older software has sounds I can't get any other way.

[Anderton]

But on the flip side, there are things that amp sims can do that you simply can't do with a tube amp.

I agree there, I'm just saying that while preamp modeling may be said to be mature and refined, there is still much to be learned and accomplished in modeling tube amp output dynamics. It really hasn't even begun in any detailed sense. What is there works really well for most music, but it falls apart when you try for a blown out tweed sound or anything outside the "normal" operating conditions — so amp simulation is not finished, there is still room for much more innovation.

So here's what I don't understand...so much effort is going into creating models of things that already exist. I understand the main motive is to have the sound of expensive/difficult-to-lug-around gear in something that works in your computer, and that's great. But when people designed the Fender Twin or Marshall, did they think "okay, that's it, guitar amps can never get better, might as well all go home now"?

To me, what's exciting about amp sims is that they give the freedom to do with effects and amps what CGI does with film. We can make sounds that don't exist in the real world with amp sims. It's encouraging that software companies are venturing beyond our shared cultural confines, whether it's Scuffham not imitating any specific amp, or Line 6 including more non-traditional effects.

I have no real interest in something that captures every nuance of a tube amp, because I already have a tube amp! I want something that will let me come up with sounds I haven't heard before. That's what I love about amp sims.

[guitarzan]

But on the flip side, there are things that amp sims can do that you simply can't do with a tube amp.

I agree there, I'm just saying that while preamp modeling may be said to be mature and refined, there is still much to be learned and accomplished in modeling tube amp output dynamics. It really hasn't even begun in any detailed sense. What is there works really well for most music, but it falls apart when you try for a blown out tweed sound or anything outside the "normal" operating conditions — so amp simulation is not finished, there is still room for much more innovation.

So here's what I don't understand...so much effort is going into creating models of things that already exist. I understand the main motive is to have the sound of expensive/difficult-to-lug-around gear in something that works in your computer, and that's great. But when people designed the Fender Twin or Marshall, did they think "okay, that's it, guitar amps can never get better, might as well all go home now"?

To me, what's exciting about amp sims is that they give the freedom to do with effects and amps what CGI does with film. We can make sounds that don't exist in the real world with amp sims. It's encouraging that software companies are venturing beyond our shared cultural confines, whether it's Scuffham not imitating any specific amp, or Line 6 including more non-traditional effects.

I have no real interest in something that captures every nuance of a tube amp, because I already have a tube amp! I want something that will let me come up with sounds I haven't heard before. That's what I love about amp sims.

It's not that they aren't capturing every nuance — it's that they crap out totally
at a certain point rather than going into the huge mega overdrive of an out-of-spec tweed (for example). Ignoring output dynamics leaves modeling incomplete and unusable for some genres.

[telecode]

I have been hearing for a while that Guitar Rig is older tech and behind the times compared to other products on the market.

Not compared to a Fender Twin Reverb

I was under the impression that emulations from that era weren't exactly spot on.
From what I hear from people who are more knowledgeable, emulations still aren't there yet.
You obviously know way more than I ever will, but your statement above sounds like empty rhetoric to me. That Fender may be a rather simple machine, emulating every little thing that happens when you turn it on in software isn't.
(Just my 2 cents, please correct me if I'm wrong)

I don't think amp sims even come close to modeling the nonlinear stuff tube amps do when they are pushed well beyond their design limits — power amp sag and recovery, output transformer saturation, speaker motional impedance, cab & speaker compression/limiting/distortion, etc. Some amp simulations attempt to do sag or whatever, but it's usually pretty weak. Most stuff always sounded great on even the first generation modelers, but if you try to push into Neil Young or garage band territory the sims still don't even attempt to get you there IMHO. Those kind of dynamics are just not represented in any useful way yet.

That statement only applies to specific genres and type.of guitar sounds. I think I'd you were playing live and wanted that feedback heavy Neil and Crazy Horse sound,.you would probably be usually a real amp.

I feel that for recording, where you are not making a guitar player track (i.e. the electric guitar is in the mix as a background instrument) the amp sims work well.

[telecode]

But on the flip side, there are things that amp sims can do that you simply can't do with a tube amp.

I agree there, I'm just saying that while preamp modeling may be said to be mature and refined, there is still much to be learned and accomplished in modeling tube amp output dynamics. It really hasn't even begun in any detailed sense. What is there works really well for most music, but it falls apart when you try for a blown out tweed sound or anything outside the "normal" operating conditions — so amp simulation is not finished, there is still room for much more innovation.

So here's what I don't understand...so much effort is going into creating models of things that already exist. I understand the main motive is to have the sound of expensive/difficult-to-lug-around gear in something that works in your computer, and that's great. But when people designed the Fender Twin or Marshall, did they think "okay, that's it, guitar amps can never get better, might as well all go home now"?

To me, what's exciting about amp sims is that they give the freedom to do with effects and amps what CGI does with film. We can make sounds that don't exist in the real world with amp sims. It's encouraging that software companies are venturing beyond our shared cultural confines, whether it's Scuffham not imitating any specific amp, or Line 6 including more non-traditional effects.

I have no real interest in something that captures every nuance of a tube amp, because I already have a tube amp! I want something that will let me come up with sounds I haven't heard before. That's what I love about amp sims.

I agree with this. I have done it a couple of times. Basically,used the tool to make sounds that you might not be able to make with a regular amp and pedal setup. Thrown stuff together and just see what happens but it can only be used in certain situations. I find I am usually tweaking and adjusting to mimic particular guitar sounds that
are consideredd appropriate for the style.

[guitarzan]

… I think I'd you were playing live and wanted that feedback heavy Neil and Crazy Horse sound, you would probably be usually a real amp.

I feel that for recording, where you are not making a guitar player track (i.e. the electric guitar is in the mix as a background instrument) the amp sims work well.

Yes, for most stuff amp sims sound great, but the fact that that they'll go full flat square wave instead of into gooey power amp sag overdrive shows their limitation and failure to model the output dynamics. I am not talking about feedback — I have always been able to get feedback through monitors if I want. It is the dynamics, the slippery push-pull of power amp, output transformer, and speaker cab that is missing completely.

It just seems a shame to me that amp sims are only capable of tighter controlled tones that could easily be produced and recorded at low volume with most any amp and the right pedals, where they fall apart at the types of tones that require select amps played full out to the point of near failure — the very place a sim would be most useful to me.

I was hopeful 20 years ago that modeling would get there soon. I have been pushing for output dynamics modeling on this site for 14 years, but there has been little progress. If some developer is looking for an area where they can make a real contribution to guitar amp simulation — this is it. There are a few work-arounds in place that suffice most of the time, but until someone can model a dying tweed, I remain unconvinced. Output Dynamics!

[telecode]

It slowly dawned on me, that all these variations and abilities to change mic positions, or grid configs inside any tube (Peavey revalver) is an obfuscation,

Sorry, can you explain what does this mean? Is Amplitube 4 a "better" sim modeler than Revalver?

Wow, Revalver 4 is $100 compared to Amplitubes $150.
https://peavey.com/products/revalver/

[Tj Shredder]

There are a few work-arounds in place that suffice most of the time, but until someone can model a dying tweed, I remain unconvinced. Output Dynamics!

This argument is a bit like the complaints about virtual pianos. There is always a physical and spacial component which is impossible to model, as its related to how you physically play it. You can touch a real piano in a way no keyboard controller will be able to be touched. If I pluck a piano string, its a different way to playing than hitting a key and getting a sampled plucked piano string. If you need the physical connection you need the real thing. You cannot model slamming the guitar into your amp, which had been a fashion some time ago...
That is why I was so positively surprised about Trash 2, you could get extreme with it, far away from normal operation of a sim. If you fiddle around with some extreme wave shaping curves, you might even be able to find something close to a dying tweed, but more likely you find something better... I would always go for the better, its easier to find and is better...

[telecode]

There are a few work-arounds in place that suffice most of the time, but until someone can model a dying tweed, I remain unconvinced. Output Dynamics!

This argument is a bit like the complaints about virtual pianos. There is always a physical and spacial component which is impossible to model, as its related to how you physically play it. You can touch a real piano in a way no keyboard controller will be able to be touched. If I pluck a piano string, its a different way to playing than hitting a key and getting a sampled plucked piano string. If you need the physical connection you need the real thing. You cannot model slamming the guitar into your amp, which had been a fashion some time ago...
That is why I was so positively surprised about Trash 2, you could get extreme with it, far away from normal operation of a sim. If you fiddle around with some extreme wave shaping curves, you might even be able to find something close to a dying tweed, but more likely you find something better... I would always go for the better, its easier to find and is better...

I personally don't the electric guitar & VST vs physical piano & VST is a good comparison. Maybe a better one could be acoustic guitar & VST or samples. not the cheap Chinese Korean acoustic shit guitars, but the high end US made Martin or Gibson. *some* of then play and sound like no other and can't be captured digitally IMO.

[guitarzan]

This argument is a bit like the complaints about virtual pianos. There is always a physical and spacial component which is impossible to model, as its related to how you physically play it. You can touch a real piano in a way no keyboard controller will be able to be touched. If I pluck a piano string, its a different way to playing than hitting a key and getting a sampled plucked piano string. If you need the physical connection you need the real thing. You cannot model slamming the guitar into your amp, which had been a fashion some time ago...
That is why I was so positively surprised about Trash 2, you could get extreme with it, far away from normal operation of a sim. If you fiddle around with some extreme wave shaping curves, you might even be able to find something close to a dying tweed, but more likely you find something better... I would always go for the better, its easier to find and is better...

Amp modeling development has spent all of these years perfecting the pre-amp while the rest of the amp is still just roughly sketched out. Maybe a bit of generic compression and some filtering for the power amp, a static IR for the speaker and cab. In a real tube amp there are loads of complex and evolving dynamics going on in the output section, especially in the vintage ones with little or no negative feedback, that are not represented at all in many amp sims.

Come on — it's been 20 years — model the other 2/3 of the amp. At least try...make an advancement, do something innovative, put the kind of elastic dynamics into the power section that belong there. Better would be great, but I'd be happy for now if they'd just start by puting something there other than the relatively static status quo.

[Anderton]

I think there is an inherent difference between an amp and a model. An amp derives its sound from physical elements, whereas the amp depends on electronic elements. It's sort of like trying to model a rock with a plant

In theory, it would be possible to model every nuance of a physical device. But that would require an immense amount of CPU and incredibly complex algorithms. How do you model the glue drying in a cabinet over a period of years, or changes in humidity that affect the speaker cone?

When technology allows that level of detail in the creation of a virtual device so that is has the "look and fell" of a physical device, I believe industry will be more interested in creating hyper-realistic sex robots, not hyper-realistic Fender Twins. Follow the money

[guitarzan]

I think there is an inherent difference between an amp and a model. An amp derives its sound from physical elements, whereas the amp depends on electronic elements. It's sort of like trying to model a rock with a plant

In theory, it would be possible to model every nuance of a physical device. But that would require an immense amount of CPU and incredibly complex algorithms. How do you model the glue drying in a cabinet over a period of years, or changes in humidity that affect the speaker cone?

When technology allows that level of detail in the creation of a virtual device so that is has the "look and fell" of a physical device, I believe industry will be more interested in creating hyper-realistic sex robots, not hyper-realistic Fender Twins. Follow the money

I'm not talking about glue or wood, I'm talking about the way the very real and electro-magnetic dynamics evolve in a tube amp. I'm no tech, but it seems to me the signal doesn't go flat square wave against the rails in a vintage amp, the transformer saturates, the rails "give", the power amp sags, now the transformer is no longer saturated, the signal again heads toward clippping but again overwhelming the rectifier...etc [EDIT: shortly after I posted this the term "cascading brownout" entered my head. I don't know if that is technically what I am describing, but seems like it must be related somehow]. It's just not there at all in modelling that I'm aware of. If it's too much for current CPU's, then perhaps a hybrid approach like the Trueno synth...digitally controlled hardware. If it can't be done at all, then amp modelling has already past it's peak and will remain secondary. Maybe that's just fine.

EDIT: It could probably all be done with a combination of compression, limiting, and clipping — just in a complex evolving and elastic way.

[audiojunkie]

This argument is a bit like the complaints about virtual pianos. There is always a physical and spacial component which is impossible to model, as its related to how you physically play it. You can touch a real piano in a way no keyboard controller will be able to be touched. If I pluck a piano string, its a different way to playing than hitting a key and getting a sampled plucked piano string. If you need the physical connection you need the real thing. You cannot model slamming the guitar into your amp, which had been a fashion some time ago...
That is why I was so positively surprised about Trash 2, you could get extreme with it, far away from normal operation of a sim. If you fiddle around with some extreme wave shaping curves, you might even be able to find something close to a dying tweed, but more likely you find something better... I would always go for the better, its easier to find and is better...

Amp modeling development has spent all of these years perfecting the pre-amp while the rest of the amp is still just roughly sketched out. Maybe a bit of generic compression and some filtering for the power amp, a static IR for the speaker and cab. In a real tube amp there are loads of complex and evolving dynamics going on in the output section, especially in the vintage ones with little or no negative feedback, that are not represented at all in many amp sims.

Come on — it's been 20 years — model the other 2/3 of the amp. At least try...make an advancement, do something innovative, put the kind of elastic dynamics into the power section that belong there. Better would be great, but I'd be happy for now if they'd just start by puting something there other than the relatively static status quo.

I agree with this! The focus is always on the amp. The cab (for example), is just as important, but almost always an after thought--very little innovation there.

[telecode]

I think there is an inherent difference between an amp and a model. An amp derives its sound from physical elements, whereas the amp depends on electronic elements. It's sort of like trying to model a rock with a plant

In theory, it would be possible to model every nuance of a physical device. But that would require an immense amount of CPU and incredibly complex algorithms. How do you model the glue drying in a cabinet over a period of years, or changes in humidity that affect the speaker cone?

When technology allows that level of detail in the creation of a virtual device so that is has the "look and fell" of a physical device, I believe industry will be more interested in creating hyper-realistic sex robots, not hyper-realistic Fender Twins. Follow the money

I would rather jack off to a hyper realistic 60s amp than a hyper realistic Katy Perry.

(Unless they made a hyper realistic Raquel Welch circa 70s sex bot. Now that's a whole different ball game. )

[guitarzan]

I agree with this! The focus is always on the amp. The cab (for example), is just as important, but almost always an after thought--very little innovation there.

Yep, they definitely need to do something dynamic for the speaker and cab — but they need to get the output section right first, then figure out what would be happening on the other side of that output transformer [EDIT: also the speaker driver and output section need to be quite interactive]. And it's not linear, they need to consider motional impedance (failure to represent the effects of motional impedance was Ted Weber's biggest beef with amp modeling) as well as many other factors including mechanical compression, limiting, and distortion from the cab.