Modeling / Analog / Digital / Solid-State Community Posting

Discussion in 'Modeling Amps, Plugins and Apps' started by jaytee32, Aug 10, 2015.

  1. codamedia

    codamedia Friend of Leo's

    Posts:
    4,896
    Joined:
    Apr 4, 2009
    Location:
    Western Canada
    Fair enough... to be honest I didn't realize they had completely moved away from tubes on those amps ;)

    How a person understands it may be accurate, but how to relay it to others can be very difficult. Just look at our Vypyr exchange :D
     
  2. jaytee32

    jaytee32 Tele-Meister

    Posts:
    406
    Joined:
    May 30, 2011
    Location:
    Netherlands
    Indeed, if the only thing important to you is the tone, you can skip the whole thing and just listen to the amp and say "I like it" or "I don't like it". Or maybe some more colorful phrases.
     
    davidchagrin likes this.
  3. jaytee32

    jaytee32 Tele-Meister

    Posts:
    406
    Joined:
    May 30, 2011
    Location:
    Netherlands
    Haha! I'll change it from a Vypyr to a Fender Mustang ... none of those have had tubes yet, we can at least avoid that confusion. Thanks again for pointing it out.
     
  4. Jim Dep

    Jim Dep Friend of Leo's

    Posts:
    3,208
    Joined:
    May 23, 2010
    Location:
    New York
    A few years back I researched the logistics of what / how a Fender Mustang attempts to model their featured amps, and they digitally do model the actual circuits of the prototype amp, unless Fender was not being truthful in the info they disclosed.
     
  5. Allthesound

    Allthesound Tele-Afflicted

    Posts:
    1,658
    Joined:
    Oct 21, 2011
    Location:
    Connecticut
    @jaytee32 I'm not going to go through all your posts point by point to to spotlight your inaccuracies it would take up too much of my time, so i will sum up here in one post:

    What is commonly accepted in the music world as modeling in 2018 is what is known as DSP/CPU based algorithmic modeling. This is where a discreet analog circuit is modeled component for component ie resistors . capacitors , diodes ect ect as complex math's and then complied into algorithms to make it possible for them to process the signal in real time (with minimal latency). This has nothing to do with solidstate or analog it exists entirely in the digital domain.

    The first commercially available modelers were analog modeling synthesizers that came out in in the mid to late 90's like the Nord Lead . Yamaha AN1X , Roland JP8000 ect ect.
    Because they were all the rage modeling quickly became the buzz word , Digitech was one of the first to market a guitar amp as a modeling amp with the Johnson Millennium and later the Digitech 2112/2120. But in truth it was not a modeler it was a digitally controlled solid state /tube premp section with DSP based effects. The analog guitar signal was passed through the DSP and fancy EQ curves were applied to crudely recreate the tone stacks of different amps.
    Not at all component level modeling but more acurately signal processing.

    The line 6 POD's that followed used proper DSP based component level modeling technology and amps like the Mustang & Katana do as well.

    All DSP modelers need either a tube or solid state amp after the D/A stage to amplify the signal but the amp stage has nothing to do with whats going on in the modeling process.

    So to qoute the last point of your first post:

    "Finally we call a modeler any amp that can be configured to "sound like" other amps, or that allows you to construct a very wide range of specific tones, by choosing different combinations of settings. Given this definition of modeling, it could be achieved using many different techniques, for example:
    using an emulator principle to simulate various amps, effects, etc as described above
    using digital (DSP) techniques to transform an input signal into the same output signal an amp or effect would have produced. Technically speaking, this would be emulating the end-to-end signal transfer function, rather than the electronic circuit itself.
    using a profiling technique to copy how another amp responds to the guitar
    having an extremely flexible tone stack, power stage, or switchable internal circuitry that allows a wide range of amp sounds to be created."



    Modern day DSP amp modeling in fact exactly models (or as closely as possible) the electronic circuit component for component. You have that point completely backwards. (Which is why i said you don't seem to understand the underlying principles of modeling)

    Using your definition of modeling things like the Marshall JMP1 , Egnater Tweaker, Mesa boogie Triaxis , Black Widow MGP-1, just about every amp in a pedal stomp box , should be included in the modeling sub forum.
    That would only muddy the waters, they belong in their respective Amp/ Pedal forums.

    So again in 2018 we define amp modeling as something that only exists in the digital domain. That is all.

    If you want to better understand first hand how competent level amp models work before they are complied into algorithms watch this video on how to take a DI guitar track run it through the model and output it to a wave file , all you need the project file in the video description and download LT Spice.
    Try playing around with the values of the modeled resistors and caps to see how it effects the tone. After you have a wave file generated you need to play it through a IR (Impulse responce) loader so it sounds like its coming through a cab.

     
    Last edited: Feb 10, 2018
    beagle and luckett like this.
  6. jaytee32

    jaytee32 Tele-Meister

    Posts:
    406
    Joined:
    May 30, 2011
    Location:
    Netherlands
    Hi @Allthesound, thanks for your answer and also for the youtube video at the end, if I get time I will try it. Wasn't aware that such sims were available for the hobbyist, although I did already know about Spice, I could have thought about it myself. No need to explain the compiled code, DSPs, or algorithms to me, that's the stuff I work with every day, I even got asked once (professionally) to review a book on computer algorithms.

    About what is "commonly accepted", if you read the post you'll see I made some references to that, but we didn't allow ourselves to be hindered by that - after all, there are lot of commonly accepted things that are just plain wrong.

    What you say about a lot of what we wrote being equally applicable to pedals is true. We didn't worry about that as there is a different forum for pedals, this posting is in an amplifier forum so we didn't feel the need to talk about anything else. I don't know if it's still on their website, but when we wrote the post, Tech 21 talked about their "analog modeling" Sansamp pedal and also how that same circuit was at the heart of their Trademark series amps, which would make the Trademark the analog modeling amp brother of their analog modeling pedal. Note: analog. I didn't make that up.

    What you write about modeling the components is more or less what Peavey did for Transtube, except that's also an analog technology. See

    https://peavey.com/support/technotes/hartley/Chapter_3.pdf

    You wrote:

    This has nothing to do with solidstate or analog it exists entirely in the digital domain.​

    It's true that digital modeling exists entirely in the digital domain, but it is not true of modeling in general, as the Tech 21 and Peavey products illustrate.

    The point you make about how digital modeling is done is not complete. The Kemper profiler does not work that way - you can take a Kemper profiler and create a new profile of an amp using just the Kemper, the other amp, and a microphone. You don't have to tell the Kemper what the internal circuitry is in order to do this, so it doesn't know the internal circuit. As we pointed out in the post, what the Kemper is doing is modeling not the individual components, but the end-to-end transfer function. Technically, if you did know the individual transfer functions and the circuit, and you convolved all these functions mathematically to compute the end-to-end transfer function of the circuit, you would get the same answer as the Kemper's transfer function.

    As far as I know, Kempers are also based on DSP chips.

    I know that the Blackstar ID series amps do things exactly the way you describe in your last post, they are modeling the individual circuits, they've said as much in their product documentation, so this is as far as well can tell, a fact. The Kemper models the end-to-end transfer function. This must be a fact since you can do it yourself without knowing what the to-be-modeled amp's circuit is.

    Both are the Blackstar and Kemper are digital. The Transtube also models (to some extent) the individual components, but is analog. I've never seen an authoritative description of how the Mustang, Vyvpr, Valvetronix, CODE, Katana do it. If you want to build cheap, you want to use a variant of the Kemper technique, as there you only need to compute the results of the end-to-end transfer function, whereas in the other approach you have to compute the effect of all the components, in real time as you pointed out. Quoting from

    https://www.modernmojoguitars.com/digital-modeling-of-amps-kemper-vs-fractal/

    Fractal doesn’t have to do this at all — they model components in a schematic and use tweaked parameters to make it sound better. This is a much more involved simulation, and thus why they need to use higher-powered Analog Devices SHARC DSPs for the simulations. Kemper don’t need this as they have analyzed the distortion characteristics from the real profiled sounds and generate the sounds at a higher modeling level — which is why he only needs to use a Freescale DSP running at a lower clock speed (400 MHz). It is a much more elegant solution which is not only more efficient, but allows modeling of almost any distortion device.​

    A lot of this has already been discussed before in the threads I quoted, which is why I asked to read those threads first before responding to something you disagreed with. You can say

    "So again in 2018 we define amp modeling as something that only exists in the digital domain. That is all."​

    Except that's not all, as I've shown in the post. Despite your definition, Tech 21 defines their stuff as modeling and is not in the digital domain, and I've seen many define the Boss Katana NOT to be a modeling amp. Which IMO is looney, given that the guts of the Katana appear to come straight out of the GT-100 pedal, which Boss itself says "Advanced COSM amps that model vintage classics and deliver highly expressive tones".

    We tried to bring a bit of uniformity in the terminology in our community post, given all the disagreement and hype surrounding the term "modeling".
     
    LooseJack and davidchagrin like this.
  7. luckett

    luckett Banned

    Posts:
    3,614
    Joined:
    Jun 14, 2011
    Location:
    .
    That's not a bad thing to accomplish, but given your clear lack of understanding of digital technology, you are not the one who should be doing this.
     
  8. Junkyard Dog

    Junkyard Dog Tele-Afflicted

    Age:
    46
    Posts:
    1,611
    Joined:
    Mar 7, 2016
    Location:
    USA
    Just to clear up some confusion, the letter D in Class-D amplifiers does NOT stand for digital. Class-D amplification is based upon analog principles, and there is no digitization of the audio. The letter D actually stands for...nothing really. Basically before Class-D you had Class-A, Class-B, Class-AB, and Class-C...so next came Class-D.
     
    jaytee32 likes this.
  9. jaytee32

    jaytee32 Tele-Meister

    Posts:
    406
    Joined:
    May 30, 2011
    Location:
    Netherlands
    The only thing I can see in our exchange that might lead you to make such a statement, is that I didn't know that the definition of "digital" had been expanded to contain signals where the information was not in the bits themselves, but in the time domain. This is not a lack of understanding, I understand perfectly well how PWM works, I had just never heard anyone call it digital before. PWM and Class-D amplifiers are by the way not part of the original post. I've yet to see anybody point out an inaccuracy (besides older Vypyrs having tubes) in that post.

    I don't mind being shown wrong and we don't mind editing the post if it's inaccurate. Please be specific.
     
    Last edited: Feb 11, 2018
    davidchagrin likes this.
  10. Allthesound

    Allthesound Tele-Afflicted

    Posts:
    1,658
    Joined:
    Oct 21, 2011
    Location:
    Connecticut

    At least Kemper knows the difference between "modeling" and what they do "profiling" . Nowhere will you find Kemper in any of their descriptions or marketing literature referring to the Kemper as a modeler. Kemper uses an audio file recorded from another amp that is processed through proprietary & very sophisticated convolution algorithms to generate a profile of the amp. Its a completely different process. But you should have known that being the expert on computer algorithms you claim to be.

    Though different its given a home in the Modeling forums because it has nowhere else to go , there is nothing else on the market that does what it does in the way that it does it.

    Using your loose definition of modeling one could call the Marshall JTM45 a modeler because it was modeled of the Fender Bassman circuit or the early Fender Tweeds could be called modelers because the were modeled off the old RCA Receiver Tube Manual schematics.

    Obviously your mind is not going to be changed and you will buy into some manufactures marketing gimmicks as proof positive you are right , but at the end of the day if a modeler is not rooted in the digital domain then it's really doesn't belong in the modeling sub-forum.

    I can only guess whoever named the subforum didn't anticipate that someone would come along split hairs and challenge the intent, if they had they probably would have called the subforum Digital Modeling Amps, Plugins & Apps. Leave off the first part and you have Plugins & Apps , both digital, it stands to reason what the intent of the Modeling part was.
     
    beagle likes this.
  11. luckett

    luckett Banned

    Posts:
    3,614
    Joined:
    Jun 14, 2011
    Location:
    .
    That is incorrect. Class D utilizes digital signals. It is not based solely upon analog principles.
     
  12. Junkyard Dog

    Junkyard Dog Tele-Afflicted

    Age:
    46
    Posts:
    1,611
    Joined:
    Mar 7, 2016
    Location:
    USA
    No, that is a common misunderstanding. D was just the next letter in the alphabet and has nothing to do with "digital". Probably was named by engineers and NOT marketing experts ;)
     
  13. luckett

    luckett Banned

    Posts:
    3,614
    Joined:
    Jun 14, 2011
    Location:
    .
    You clearly don't understand the technology being used. The output of Class D amps are controlled with a digital signal.
     
  14. jaytee32

    jaytee32 Tele-Meister

    Posts:
    406
    Joined:
    May 30, 2011
    Location:
    Netherlands
    Part of the process (modeling a transfer function using DSP technology) is exactly the same, part of it is very different (modeling the individual component transfer functions and transporting the signal through them, vs modeling the end-to-end entire-amplifier transfer function). We made the distinction of a profiler in the post, but in the end, the intent of the thing is the same.

    Nope ... read the post. We called this a clone, because the intent of a JTM45 is not to be able to sound like a whole lot of different amps. The point is to sound like one specific amp. Like I said, in the threads I asked to read before commenting, we discussed exactly this point, along with most of the other objections you raised. :)

    Actually it's the other way around - we didn't want to buy into the marketing hype. That's why some of our definitions don't agree with what the manufacturer itself calls something ... anyway, what you say about obviously your mind is not going to be changed is wrong because

    This is a good point, as the subtitle of the forum actually says "it's a digital world" ... so you are correct, in this forum, we should focus on digital modeling. Whether you and I agree about whether analog modeling exists is to that extent irrelevant.

    Unless any of the other post authors object -- they've been real quiet, I wonder if they even still read this forum -- I'll make a new version and while there will still be a bit of the reference to other types of modeling, I'll move that into an appendix or something. Thanks for the discussion.
     
  15. codamedia

    codamedia Friend of Leo's

    Posts:
    4,896
    Joined:
    Apr 4, 2009
    Location:
    Western Canada
    Actually, "The output of Class D amps can be controlled with a digital signal", but they don't have to be.
    Class D integrates nicely with digital... but on it's own, it is not digital.
     
    Junkyard Dog and pondcaster like this.
  16. Junkyard Dog

    Junkyard Dog Tele-Afflicted

    Age:
    46
    Posts:
    1,611
    Joined:
    Mar 7, 2016
    Location:
    USA
    That "D" sure confuses a lot of people. That and probably the fact that class-D amplifiers are integrated into "digital" consumer audio products, e.g. MP3 players, phones, etc. The class-D amplifier may be controlled by digital signal, but the output is most certainly a continuous time analog signal. It's basically just a switching amplifier, but rather than confuse you further with a technical discussion, I'll reference the following:

    The "D" in "class D" often gives rise to the misnomer "digital amplifier" when referring to a class D circuit. In actuality, although the input may be encoded in a two level signal, it does not contain any digital information.

    M. Bloechl, M. Bataineh and D. Harrell, "Class D Switching Power Amplifiers: Theory, Design, and Performance," IEEE SoutheastCon, 2004. Proceedings., Greensboro, North Carolina, USA, 2004, pp. 123-146.
     
  17. luckett

    luckett Banned

    Posts:
    3,614
    Joined:
    Jun 14, 2011
    Location:
    .
    Provide a source for your quote.

    It's obvious that they can be controlled with a digital signal, because they are. That quote doesn't say there are Class D amps controlled with something other that a digital signal. To make this an easy exercise for everyone who thinks Class D is not digital, show me a production Class D amp that does not utilize a digital signal to control the output transistors. There are obviously lots of people who don't understand Class D that make inaccurate comments about the technology, especially consumer oriented or non-technical sites. A quick quote grab that doesn't provide context is a bit dubious when there are plenty of electronics industry sources that explain how Class D utilizes digital signals.
     
  18. Junkyard Dog

    Junkyard Dog Tele-Afflicted

    Age:
    46
    Posts:
    1,611
    Joined:
    Mar 7, 2016
    Location:
    USA
    Would you accept an example from one of the largest semiconductor manufacturers in the world?

    "ST’s analog class D amplifiers enable a higher efficiency, thus reducing the power dissipation and the need for larger heatsinks."

    http://www.st.com/en/audio-ics/class-d-audio-power-amplifiers.html
     
    telemnemonics likes this.
  19. luckett

    luckett Banned

    Posts:
    3,614
    Joined:
    Jun 14, 2011
    Location:
    .
    I never said anything about the output of a Class D amp being digital. But that's a cute straw man argument. It's very obvious that the output is analog.

    Now explain your previous comment about how there is no digitization of the audio.

    We've already covered this. A binary PWM signal with discrete on/off states is digital. A signal that switches between on/off most certainly contains digital information.
     
  20. luckett

    luckett Banned

    Posts:
    3,614
    Joined:
    Jun 14, 2011
    Location:
    .
    Back up your quote grab and entertain me by explaining to me how this chip works and how it doesn't utilize digital signals.
     
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.


  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.