How do modeller manufacturers ensure that their models sound like the real thing?

Discussion in 'Digital & Modeling Gear' started by NelsonP, Dec 13, 2019.

  1. NelsonP

    NelsonP Member

    Messages:
    370
    Joined:
    Jan 28, 2017
    As a spin off from a conversation in another thread, I realised that have never really considered how a modeller manufacturer determines whether a model is a true and fair reflection of the amp, cab and mic combination they are attempting to model.

    So , without giving away any trade secrets, how is it done?

    Do you look at:
    - the final result - i.e. does the Fender Deluxe model sound the same as a mic'd Fender Deluxe?
    - each element individually - ie, does the preamp produce the same output based on a certain input, same for the power amp, the cab, the mic etc
    - both of these
    - something else?
     
  2. Briandress

    Briandress Member

    Messages:
    429
    Joined:
    Nov 11, 2016
    ha they dont. because they dont
     
  3. kdm1218

    kdm1218 Supporting Member

    Messages:
    5,155
    Joined:
    May 9, 2016
    Location:
    TX
    When you have nothing better to do than come and troll in a thread you obviously don’t care about...

    To the OP, I don’t know the answers but would definitely be interesting to read on. I know I’ve read from Line 6 they only models amps they can actually purchase, work with, upkeep, etc, which is why no Dumble model in their line.
     
    Steirnen, DeadLazy, Calaban and 6 others like this.
  4. MojoRisin

    MojoRisin Member

    Messages:
    1,105
    Joined:
    Jan 29, 2019
    Ditto. I'd assume it's a combination of lots of oscilloscopes and listening while metaphorically turning dials and pushing levers until things line up.
     
    Eric Rowland likes this.
  5. phil_m

    phil_m Supporting Member

    Messages:
    10,728
    Joined:
    Jun 29, 2012
    Location:
    East of the Sun, West of the Moon
    Extensive A/B testing.
     
    RicardoDiazHimself likes this.
  6. mbenigni

    mbenigni Member

    Messages:
    7,110
    Joined:
    Jul 6, 2015
    It depends on the product. In some cases, there really isn't much effort to ensure a one-to-one match with a real world amplifier. This can be true with low end products that just aren't all that convincing (and are priced accordingly) or high end products that are meant to get you in the ballpark with an excellent, but more or less non-specific tone (Roland often goes in this direction.)

    Products that aim to mimic specific, real world amplifiers fall under two paradigms: modeling and profiling. Of the latter, there's really only one: Kemper. Here, functions are selected and functional coefficients are modified to ensure that the output is as similar as possible to that of the real amplifier for any given input. (Your "final result" option above.)

    Effective modeling requires not only that amplifier subsystems (preamp, power amp, cab, mics, etc.) convincingly mimic their real world counterparts; things have to be even more granular to ensure that individual controls respond as expected. To these ends, DSP coders model individual electronic components (diodes, capacitors, etc.) and their interaction, test these units, build virtual amp circuits comprised of systems of these units, and ultimately test the combined systems vs. real amps built from corresponding physical components. If all goes according to theory... the resulting amp model through a capable transducer (i.e. power amp and speaker*) should pass a blind test vs. the real amp, and the controls should function the same way.

    *This gets sketchy where amp and cab modeling is concerned: you can't get an electrical signal into a person's ears without some kind of amplification and speaker system. And that amplification and speaker system will have its own sonic signature and limitations. Which will always complicate the question as to whether you really nailed the sound of the original amp/ cab.
     
    Last edited: Dec 13, 2019
  7. -Hawk-

    -Hawk- Member

    Messages:
    524
    Joined:
    Jul 20, 2018
    Triple-blind YouTube videos with extra goatee.
     
  8. mbenigni

    mbenigni Member

    Messages:
    7,110
    Joined:
    Jul 6, 2015
    :rotflmao

    Yeah, that's what I was trying to say.
     
  9. Briandress

    Briandress Member

    Messages:
    429
    Joined:
    Nov 11, 2016

    To wit they are really not comparing the amp models to the amps they are intended to model rather the schematically correct circuit for that amplifier. Keeping in mind that most oft and heatedly debated models are those where their analog counterparts are quite far from one for one tonal accuracy it makes sense that there would not be a whole lot of comparisons going on. I dont believe that these companies are testing their plexi models against five or six different plexi amps that would all sound different in their own right.
     
  10. Digital Igloo

    Digital Igloo Member

    Messages:
    3,864
    Joined:
    Oct 4, 2010
    Location:
    Woodland Hills, CA
    Everyone drink!

    First of all, the notion of nailing the sound and feel of all three core elements of an amp—preamp, power amp, and cab—completely goes out the window when the user is responsible for the biggest sonic factor of the three. If you have a real Revv Generator 120 and cab but you're playing a model of the Revv Generator 120 preamp and power amp through anything but the Revv Generator cab you're used to, it's all but impossible to convince you you have a real Revv Generator 120 and cab in the room. It's a lesson in futility and people like our friend Briandress here like to blindly slag modeling without the context that's 100% required for these kind of conversations. He might as well insist that vinyl sounds better than WAV files—but not because they do in any empirical sense, but via some anecdotal black box where his uncle's $100,000 turntable-based home stereo system sounds better than his cousin's Microsoft Zune through earbuds. But I digress...

    Because modeling companies have no control over the myriad of (excellent to obscenely horrific) playback systems their modelers will be plugged into, they generally focus their efforts on nailing the preamp and power amp elements. Because both have dozens or hundreds of components that can be measured by extremely expensive and sensitive test equipment, if you know what you're doing, and you understand how the various components interact with one another, given enough sophisticated DSP tools, you can meticulously recreate all of those interactions in the digital realm. In the end, if the test equipment can't tell the difference (along with the occasional null test), you know you're really close.

    Then you perform your own double-blind listening tests. Then you bring in more people for double-blind listening tests. If necessary, you bring in golden-ear recording and mix engineers for even more double-blind listening tests. And if something's not right, you go back and tweak until it is.

    When guests tour YGG, @benadrian sets up a listening station in our studio. People can try the listening tests for themselves (real amp vs. modeled amp with an identical playback system—usually the real cab in another room). In the vast majority of cases, no one can reliably pick which is which. If they do, it's generally because something in the real amp has changed over the years since it was first modeled (tubes, biasing, etc.).

    So quite literally, whenever you read someone in the YouTube comment section whining about how "tº0Bz RuL3, d!gi‡@L suX0Rz!", they're inadvertently commenting on the inadequacy of playback systems, over which professional-level modeling/profiling companies like Atomic, Fractal, Kemper, Line 6, and Strymon have zero control.

    TL; DR: The preamp model is all but dead-on, the power amp model is all but dead-on, but if you plug it all into a cheap plastic OEM PA speaker (deceitfully marketed as "FRFR"), we can't be held responsible for the cab portion.
     
  11. Briandress

    Briandress Member

    Messages:
    429
    Joined:
    Nov 11, 2016
    @Digital Igloo to be fair i did not say that digital is no good. I actually exclusively play digitally produced sounds, Helix Native included. My conjecture was based on the premise that there would be no way to compare your plexi model to a plexi amp because there is not one the just makes one specific sound. They all sound different right? So if there is variance in sound from one amp to the next likely there would be similar variance from a model to a "real" amp.
     
    uab9253 and zuel69 like this.
  12. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,652
    Joined:
    Oct 13, 2007
    This is a no-brainer. I specifically know of two modeler manufacturers who use physical amps played through physical cabs as their reference. To evaluate the amp model, they will play through a neutral power amp and use the same physical cab to do an A/B comparison with the physical amp. One of these two manufacturers has an extensive collection of amps and cabs for this purpose. The other might, but I don't have that information.
    The procedure for evaluating cab sims should be identical. That is the technique I use. Whether modeler manufacturers do that A/B is a question I cannot answer. Maybe their reps here will chime in?
     
  13. Digital Igloo

    Digital Igloo Member

    Messages:
    3,864
    Joined:
    Oct 4, 2010
    Location:
    Woodland Hills, CA
    Correct. The model is, ideally, a picture-perfect recreation of the sound and feel of one specific serial number of one specific amp's preamp and power amp (including what condition it might be in).
    Yep. Also—and this is really important—the amp needs to be loaded by its matched cab, because it changes the amp's behavior. You can't just use some generic load box.
     
    stilwel, asilker, aizenx and 4 others like this.
  14. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,652
    Joined:
    Oct 13, 2007
    That is an integral part of the procedure I described: play the amp through its cab, then play the modeler through a neutral power amp and the same cab.
    Agreed.
     
    RonnieF, stilwel, aizenx and 3 others like this.
  15. Jerrod

    Jerrod Silver Supporting Member

    Messages:
    12,180
    Joined:
    Jan 7, 2002
    Why would amp-to-amp variances preclude the manufacturer from making comparisons?
     
    macdolfan904 likes this.
  16. Elric

    Elric Supporting Member

    Messages:
    3,709
    Joined:
    Sep 15, 2004
    Developers typically throw their algorithms into the “Tone Wars” Octagon and there they battle to the death. This way, over time, only the strongest algorithms survive and earn inclusion into the products used to bust out old Dokken riffs in your makeshift living room studio that you and your crew gather in on Saturday nights for Beer jams.

    Alternatively, some developers use “science” but most people do not put much stock in that in the age of feelings.
     
  17. NeedmoreCrunch

    NeedmoreCrunch Member

    Messages:
    1,159
    Joined:
    Jan 18, 2012
    Location:
    Pacific Northwest
    Tone is subjective, no one can ensure. Designers do their best to assimilate.
     
    JaiRamana and Barquentine like this.
  18. Briandress

    Briandress Member

    Messages:
    429
    Joined:
    Nov 11, 2016
    but do they all do a blind test with every single amp they have a model of? I guess they do?
     
  19. dspellman

    dspellman Member

    Messages:
    8,058
    Joined:
    Mar 26, 2008
    I've been at Line 6 in Calabasas quite a number of times. They actually have the amps (and guitars) that they model on site.
     
  20. dspellman

    dspellman Member

    Messages:
    8,058
    Joined:
    Mar 26, 2008
    They can and do, but they also have pretty distinctive readouts of the amps they have there to compare as well.
     

Share This Page

  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.
    Dismiss Notice