Amp modeling vs. tape modeling

[inverseroom]

I've been playing through amp models for years now--first a SansAmp and now a Vox ToneLab--and I'm not alone in thinking these things now sound pretty great, or at least good enough so that on a recording they don't make people say, "That is SO not an amp." If you record in your house and can't mic a cab all day, Pods and the like are a perfectly acceptable substitute--and I happen to think the ToneLab is actually fantastic.

But a recent thread here sort of trashed the idea of audio tape saturation modeling, and I have to ask--why is this so much harder to model than amps? Is it, in fact, a greater technological challenge--and if so, why?...or are we merely more sensitive to the supposed sanctity of audio tape, and thus less willing to accept a digital substitute for "that sound"?

I say this as a tape agnostic, since the only tape I've worked with is 4-track cassette. Would love to hear what people think.

[rimbaud234]

There are several ways to do modeling, one way is to model the individual components of the device. So for a guitar amp or a compressor, they look at the schematic and try to emulate how all of the individual components work together using mathematical models of the components (tubes, resisters, etc).
The issue with tape machines and tape is that they're really complex and accurately modeling how all these things work together can be difficult and hard to make sound right.
(I know some folks who make some very highly rated emulations of classic gear and when asked about tape emulation, they've reiterated this to me.)

[xonlocust]

If you record in your house and can't mic a cab all day, Pods and the like are a perfectly acceptable substitute--and I happen to think the ToneLab is actually fantastic.

really? last time i heard a Pod i thought it was a total pile of shit. i'd still way rather mic whatever is there than use one (based on last time i heard one in use a couple years ago). seriously - are they better now, cuz i really thought it blew ass when i heard it...

[pscottm]

ahh but you are assuming that your amp models sound like those amps. i doubt they are any closer to what they emulate than tape emulation is to tape. there's no way a pod sounds exactly like a real ac30 cranked but it may sound damn good anyway.

here are some considerations..

w amp modeling, anything goes, there is no 'standard reference' of which all people are familiar, so it can sound totally alien and still sound like an amp out there somewhere amid the thousands and thousands of different amps.

tape is a singular reference, not 'anything goes'. when's the last time you calibrated an amp to factory specs?

amp modeling doesn't have to deal w transients or full freq range.

tape modeling can sound good, just not totally accurate, just like amp modelling.

modeling is modeling.

[KennyLusk]

I use both and I use it in different ways.

I use a Guitar Port (POD on a PC, basically) and it's a very useful tool to me. I have a tendency though to route the Guitar Port's audio to my amp and mic the cabinet though.

I also use RTR analog tape in my tracking and record building "processes".

I find a use for both worlds and love them both and don't have anything negative to say about any of the tools you're talking about.

I also have a tape saturation plug I use in Sonar and find it extremely useful to "warm up" a track that's "missing something" from time to time.

I'd say that all the tools available to us as musicians and engineers are excellent tools and just simply should be used if and when you need them.

Emulation is a good thing.

[inverseroom]

ahh but you are assuming that your amp models sound like those amps.

Oh, heck no. I don't care if my Twin model sounds just like a Twin. I do care if the guitar DOES sound like it's coming out of a mic'ed tube amp, though, which as everyone knows is a great sound, in its many forms. Besides, put two Twins side by side, and you have two different sounds. What I care about is "tube ampness," which I think the modelers more or less deliver...

[Knights Who Say Neve]

Re amp modelers: I've tried these (Pod Pro, v-amp, cyber-twin etc.) and though the sound was 60-70% there, the dynamics and playing feel were not. Useful in some situations, but rather uninspiring if you're used to tube amps. YMMV.

Re Tape modeling: You won't be able to model tape accurately until Digital recording itself becomes indistinguishable from analog. Until then it's still going to sound digital, because modeling is digital.

[inverseroom]

Well, I don't wanna turn this into an amp-modeling defense, though I do like my Tonelab a lot, and like it mainly because it excels at dynamics. The Bassman model is really particularly good--you get that sort of "sag" you get with a loud, clean amp.

That said, I'm sure Scott's right that the models aren't really accurate. I mean, if you want to play a particular amp, you gotta go do that. But, like KennyLusk, I find modelers to be useful tools that make a wide array of good sounds.

Anyway. What I'm curious about is, if the technological challenges of approximating the sound and feel of audio tape are too great to convincingly meet, then what are they exactly? Modeled "tube" distortion and compression, while imperfect, make a lot of people perfectly happy every day, whereas the tape models don't seem as widespread, even though the sounds of tape are very appealing and sought-after by musicians.

I just want one of the geekiest among you to say why you think the intricasies of the tape sound are so hard to reproduce in the digital arena, if you have any idea.

[Mr. Dipity]

I just want one of the geekiest among you to say why you think the intricasies of the tape sound are so hard to reproduce in the digital arena, if you have any idea.

It simple. Amp modeling is easier than tape modeling because when you are done, you have a product that you can sell to every bedroom guitarist in the world.

Tape modeling only sells to bedroom tapeopers. There are fewer of them.

[rimbaud234]

Not to throw fuel on a fire (and disregarding people's like or dislike of U2 which is often discussed here), but the Edge is a guy whose playing and tone I greatly respect. He can also buy and sell just about everyone of us here... and our gear. He can also afford a wall of AC-30s mic'd by the best in the business.

http://www.line6.com/textArticle.html?i ... ory=artist

Discuss amongst yourselves.

[shedshrine]

"I wanted to hate it. I can't .... This thing is amazing."-Tape Op Forum.
Focusrite Liquid Channel ad.

"Liquid technogy is a unique hybrid design that goes way beyond modeling since it can dynamically recreate sonic behavior ...with Dynamic Convolution to ...impulse response etc..."

So Line 6 licenses the technogy, markets it to home recordists using " warmth" and "fatness" not to mention "phatness" and kidney shaped and pro racked version boxes dubbed the Line 6 Endless Oxide or something, and one dial has the tape widths and types, theres a 7.5, 15, and 30 ips control, and a third tape deck type dial where you bring up your Revox, Studer etc....

[Knights Who Say Neve]

Anyway. What I'm curious about is, if the technological challenges of approximating the sound and feel of audio tape are too great to convincingly meet, then what are they exactly? Modeled "tube" distortion and compression, while imperfect, make a lot of people perfectly happy every day, whereas the tape models don't seem as widespread, even though the sounds of tape are very appealing and sought-after by musicians.

I just want one of the geekiest among you to say why you think the intricasies of the tape sound are so hard to reproduce in the digital arena, if you have any idea.

I guess it's like trying to model your favorite cotton shirt, using polyester...

[vvv]

Not to throw fuel on a fire (and disregarding people's like or dislike of U2 which is often discussed here), but the Edge is a guy whose playing and tone I greatly respect. He can also buy and sell just about everyone of us here... and our gear. He can also afford a wall of AC-30s mic'd by the best in the business.

http://www.line6.com/textArticle.html?i ... ory=artist

Discuss amongst yourselves.

I found this interesting from the above link:
"The main guitar sound throughout the new U2 single "VERTIGO" played by The Edge, was the result of the combination of the Line 6 DM4 stompbox, a vintage echo unit, a vintage '60s Telecaster and a vintage Fender Amplifier. This guitarist likes to work with cutting edge technology as well as vintage pieces.

-Dallas Schoo " (Italics mine.)

[jmiller]

I'm not against modeling, but a few years ago I sold my amp and got an amp modeler, figuring it would give me better tone options and save space, not to mention not having to mic my amp at three in the morning. A couple years ago, I plugged into my wife's amp (a fender 112 or something) and it was like hearing God's voice through the speaker. I bought a 300 dollar Danelectro DM25 and my modeler sits in the closet now. That amp is killer.

It seems to me that tape modeling usually just sounds like a plugin, at least the ones i've used. There are so many variables with tape- formula, speed, NR type, and alignment. I think alignment is particularly interesting, since it's so idiosyncratic.

I guess to reiterate what others have said, it seems that tape is so much more complex to model properly.

[xonlocust]

I just want one of the geekiest among you to say why you think the intricasies of the tape sound are so hard to reproduce in the digital arena, if you have any idea.

there are about 8 pages to sift through on the topic here (by those far geekier than i):

http://recforums.prosoundweb.com/index. ... 1570/0/0/0

and a more succinct post from jakob erland here:

http://recforums.prosoundweb.com/index. ... 04b24f8461

The beloved "tape sound" is frightfully hard to emulate thruthfully in math-based systems (plugins, effects)

Why?

Because the phenomens you hear as a specific tape "sound" are mostly beyond the reach of simple algorithmic description. This because there are quite many parameters that influence each other (and the "sound") in non-linear manners.

Non-linear means that a given parameter will influence another parameter (or "sound") in a way that can be correlated one way for a certain range, and then gradually changing into correlation the opposite way for another range.

Such parameters, that are not expressable with simple linear functions are VERY hard to implement as math. And a whole range of odd parameters like these interacting is much worse.

Convolution techniques (factor analysis) won't help you here either, as they can only do a static (and thus linear) description of the system under investigation.

So yes, the only way of getting the 456/15IPS sound would be to use just that.

On the other hand - you would still need to know what specific tape machine - and at what settings - to compare to. Magnetic systems as well as calibration vary a lot - and a Studer A800 will sound very, very different from a Fostex A8, even with both running 456/15IPS

Jakob E.