Kemper is so much better than Fractal

  • I could be wrong but I believe the KPA was released in 2011. The Axe FX II Mark I & II were also released in 2011 and Fractal Audio actively releases new firmware updates for them.

    Well here is what we know. Christoph is the creator of synths through Access with, I think, 3 major products, the oldest over 16 years old, and all still supported. Now he has the Kemper company (is Access the parent company?) that has this product and it's got a few incantations (lunch box, rack, differing colors) but essentially the same in terms of no differences with how the software runs on them or support for them all.


    Compare this with Fractal, the AXE and ULTRA are no longer supported. Mark I/II aren't backwards compatible with presets because of HARDWARE limitations, unlike the newer XL and XL+ versions. So we see the chopping block coming at some point perhaps.


    Now, nothing wrong with this in terms of how the market does electronics and Fractal in particular. BUT... I have a choice, and there being a choice, I certainly am concerned with longevity and so.. I choose Kemper on this account in a big way. But not the only reason, just a big one.

  • Well here is what we know. Christof is the creator of synths through Access with, I think, 3 major products, the oldest over 16 years old, and all still supported. Now he has the Kemper company (is Access the parent company?) that has this product and it's got a few incantations (lunch box, rack, differing colors) but essentially the same in terms of no differences with how the software runs on them or support for them all.
    Compare this with Fractal, the AXE and ULTRA are no longer supported. Mark I/II aren't backwards compatible with presets because of HARDWARE limitations, unlike the newer XL and XL+ versions. So we see the chopping block coming at some point perhaps.


    Now, nothing wrong with this in terms of how the market does electronics and Fractal in particular. BUT... I have a choice, and there being a choice, I certainly am concerned with longevity and so.. I choose Kemper on this account in a big way. But not the only reason, just a big one.

    Remember that I bought a Kemper as well, but it had nothing to do with the longevity of the product. That said, you're comparing two different types of product(synth vs. modeler), and it doesn't logically follow that a product made by a specific company will have the same upgrade lifecycle as every other product they make.


  • Remember that I bought a Kemper as well, but it had nothing to do with the longevity of the product. That said, you're comparing two different types of product(synth vs. modeler), and it doesn't logically follow that a product made by a specific company will have the same upgrade lifecycle as every other product they make.

    Yes, it does follow the same logic, because Kemper have repeatedly stated that they intend to apply their business model, update policy, backwards compatibility, etc. to the Profiler just the same as to the Access Virus line.

  • The only thing about Kemper that worried me, is what impact down the road is it going to have on future development of Tube Amp makers? Part of it already seems a little dirty, kinda like Napster back in the 90's. Modeling your own amps is one thing, but buying models. I worry that will come back to haunt us one day.


    Isn't it reasonable to pay like $250 per amp model with the revenue going to the amp manufacturer? Today, I can buy 10 amps for less than $50 bucks. Don't get me wrong, I did, and I use them, but something is just not right about it.

    Hi All,


    Just to put a little marketing and engineering spin on things here.....


    From a marketing perspective, neither Kemper or Axe II are "cheep". I am just talking how much quan I have to lay out to get my guitar to make music here, not overall pedals, cables etc needed to do the same thing.


    Since it isn't "cheep", and there are still a wealth of decent sounding 1x12 or 2x12 combo's that can be had for considerably less, the sales of Kemper and Axe II will remain the province of the high end crew .... for now.


    From an engineering perspective .....


    First, my disclosure here ... I am a very seasoned engineering manager (and used to be a damn fine engineer some years ago ;) ).


    The Kemper does what it does with less than half the processing power of its nearest competitor (Axe II Fx) (from what I have read). I have to tell you guys, that this is simply AMAZING. The efficiency of the internal firmware design and DSP must be a thing of beauty. This is the real Intellectual Property (IP) that Kemper holds IMO.


    Given a super computer and terabytes of RAM, there are a number of people that could likely do what the KPA has done. Doing it with so little ...... man, it is a thing of beauty.


    Now, this won't last forever. DSP and micro-processors are improving rapidly. Furthermore, every electrical engineer (EE) studies the basic concepts that CK has used in the KPA. Eventually, it will be possible for those with less engineering chops than the fine engineers at Kemper to duplicate the feat they have done since it will be much easier with hardware that is .... say 1000 times as capable.


    This will eventually bring the capabilities of the KPA into the sub $500.00 range (in a foot controller as well). Once this happens ..... well, tube amp sales will have a problem.

  • Yes, it does follow the same logic, because Kemper have repeatedly stated that they intend to apply their business model, update policy, backwards compatibility, etc. to the Profiler just the same as to the Access Virus line.

    That's their intention. Whether it happens or not is another story, though hopefully it will. If we're comparing apples to apples though, the Kemper hasn't endured a longer upgrade lifecycle than the Axe FX II(nor the original Axe FX for that matter). That said, five years of support is actually quite good when compared to the upgrade path of most modelers. I certainly don't consider it a peeve by any stretch, but that's me.

  • I don't see the Kemper as hurting Amp sales for some time, and not solely. Maybe as a trend-setter for the modeling industry that itself will eventually hurt amp sales.


    Amps are here to stay for some time. They are the sources of the sounds Kemper uses so they won't go bye-bye, nor should we want them to.


    I think though, that down the road, as either complete modeling mimics an amp, or multiple-profile-blending mimics an amp, and then the technology gets put into an AU/VST/etc Plugin format, this will then hurt amp-sales.


    But we are talking some time into the future. 5-10 years? More? It depends on the competition. With Kemper and Fractal being the top 2 dogs, unless Line6 gets as good as these products, and they have no incentive to do more than they already have for their target market, I expect it will be slower than technology might predict because it IS solely based on human ability to program these things. And that takes brains and inspiration and hard work. I can say that, the cat is out of the bag. ONCE people KNOW it's possible, then more people find the way on their own. So expect more "bias"-like emulations.


    Putting all three (genius, inspiration, hard work) together takes either a huge company (not enough money to be worth it like for Apple, or Microsoft) or people like Cliff and Christoph pushing pushing their technology to new grounds. They can't push so fast they make themselves or their current product obsolete before making back the money they put into the growth of their businesses. There are many pressures pushing back against the speed of increasing technology, so that hardware aspect alone won't dictate the time-table. Nor the price. The Price isn't the cost of hardware, but the software cycle and continued free support. The continued support is why the old unit have equity too.


    (On another side-note, look to warranties to show continued support for an item. Kemper's 3-year-warranty is no slouch. Remember when Yamaha offered 10-year warranties on the Clavinova? It was a large part of why people first bought them, and THEN their sound reputation made the brand.)


    Amps aren't sweating. And from my conversations, amp makers aren't sweating either. In fact, they laugh at the notion. Maybe they ARE dinosaurs, but it's still quite a few years before the big meteor.

  • we are talking some time into the future. 5-10 years? More? It depends on the competition.

    In my opinion, it may be 15 or 20 years... possibly more. I think Kemper's technology is currently the most authentic sonic representation of a real amplifier, though the primary limitation is in reproducing an amplifier in whole rather than in part. While Fractal Audio's technology attempts to model amps in whole, it doesn't reproduce a perfect model of a vacuum tube, as note by Cliff Chase(owner and head guy at Fractal Audio), "to perfectly model a vacuum tube you would need to model the electron flow through the tube which would be extraordinarily complex and require a super-computer to do in real-time." Thus, the bottleneck appears to be processing power. So, unless there's a major breakthrough within the next few years and/or until the proper technology becomes commonplace, modelers will continue to pose no real threat to amp sales, in my opinion.

  • In my opinion, it may be 15 or 20 years... possibly more. I think Kemper's technology is currently the most authentic sonic representation of a real amplifier, though the primary limitation is in reproducing an amplifier in whole rather than in part. While Fractal Audio's technology attempts to model amps in whole, it doesn't reproduce a perfect model of a vacuum tube, as note by Cliff Chase(owner and head guy at Fractal Audio), "to perfectly model a vacuum tube you would need to model the electron flow through the tube which would be extraordinarily complex and require a super-computer to do in real-time." Thus, the bottleneck appears to be processing power. So, unless there's a major breakthrough within the next few years and/or until the proper technology becomes commonplace, modelers will continue to pose no real threat to amp sales, in my opinion.

    Ok, lets dissect the LOGIC of Cliff's statement, if he in fact said that.


    1) "to perfectly model a vacuum tube" means he doesn't model the actual tube perfectly. So he's just a "slimmed-down" version of the "perfect model of a vacuum tube".
    2) "require a super-computer to do in real-time", which the Fractal, and all home PCs, are incapable of.


    BUT, the Kemper does model the sound of many tube amps entirely authentically. And it is not modeling them "perfectly" from the bottom up, nor is the Kemper a "super computer" with it's weaker little sharks. (lol)


    So we KNOW you can have a slimmed-down model with variables that are input by an actual probing of the real amp to get authentic emulations. Hence it IS possible to have a slimed-down version, that isn't perfect, running on a non-super computer, that gets the job done, right? Right.


    Cliff would have you think it's an impossible task awaiting some far future date (2040?) to be cost-effective and consumer possible. Don't even ask for it! It's not scientifically possible for you under normal circumstances! What rubbish. That is a Straw Man if I ever saw one. Create an impossible situation to cover for your flaws.


    Say, folks, I'd love to make you the PERFECT ice cream but.. I'd need the vacuum of deep space, salt from Europa's oceans, and unicorned-space-cows to do that perfectly, so it's just gonna be chilled local dioxin-milk for ya now.


    Look, we know it's 100% possible, to-day, with old hardware. And if Cliff could, you know he would, probe those amps and input those variables and turn his slimed-down un-perfect tube model into something more authentic. But he isn't, because he hasn't found the way to do it.. yet.


    It really does come down to intellectual property. Not hardware. Not cost. Imagination, inspiration, genius & hard work. He's crazy smart, hard working, imaginative, but inspiration come largely on the back of Lady-Luck, not budget nor time.


    It seems to me that Fractal takes the "brute force calculation" approach to playing chess while Kemper takes the "pattern matching" approach human brains use to solve problems. And eventually you can win with brute force IF you have the super-computer, a team of programmers (and somebody to tell you when to stop playing) But there is nothing stopping Cliff from using a bit of both, because, by his own admission, a non-perfect model IS less brute-force and more pattern-matching already, which he employs in the Fractal units. So why not give in a bit more and come closer to the side with the chocolate cookies?

  • In my opinion, it may be 15 or 20 years... possibly more. I think Kemper's technology is currently the most authentic sonic representation of a real amplifier, though the primary limitation is in reproducing an amplifier in whole rather than in part. While Fractal Audio's technology attempts to model amps in whole, it doesn't reproduce a perfect model of a vacuum tube, as note by Cliff Chase(owner and head guy at Fractal Audio), "to perfectly model a vacuum tube you would need to model the electron flow through the tube which would be extraordinarily complex and require a super-computer to do in real-time." Thus, the bottleneck appears to be processing power. So, unless there's a major breakthrough within the next few years and/or until the proper technology becomes commonplace, modelers will continue to pose no real threat to amp sales, in my opinion.

    That's based on YOUR criteria of what the genre needs to be. Not everyone has the same criteria. I don't for one. Kemper as is was good enough for me to get rid of all but my Dlx Reverb. I know quite a few others that have made the switch to Kemper or Axe or Helix completely as well. More pros, PARTICULARLY under a certain age point are switching to digital, or never went tube to begin with.


    I also heavily believe, it isn't the technology, right now, that are keeping more people from using high end digital gear.


    If you were to actually ask those out there who use only tube amps, you will get 3 primary reasons, for the vast majority. to whit:


    1. Too complex - They want to just plug in and play, they say. Now mind you, some of these have a large pedal board, and do anything but just plug and play. But they see all those buttons on an Axe or a Kemper, and think complex to use. And the majority of this perception, is just plain ignorance. If they truly knew how easy, by and large the KPA is.


    2. Too expensive - Axe FX II or a KPA is a large outlay of cash. But, what most of the folks with this objection are not factoring in is the value of that outlay. They don't see it as equal to price of one boutique amp, let alone that it can be all of those boutique amps. They also don't see that their current, 5 or 6 low or medium priced amps cost as much or more than the KPA or Axe, and could all be replaced by either. Many of them have also bought and sold (at a loss from purchase price) quite a few amps, of which said process would stop with the KPA or Axe.


    3. Too fake or Digital - MOST that have this are basing their opinion on low end modelling tech or old tech. When you ask if they have tried a KPA or an AXE in an FRFR rig, the answer is often, in large majority, no. They are basing their opinion on Fender Mustangs, Pod beans, etc.


    Ask people out there. I have, and the above is really the major reasons given.


    I have also seen, so many times, myself included, people being illuminated or educated about the fallacies of the above 3, and then actually TRIED an Axe or KPA rig, switch. Those that aren't of the ever decreasing minority that will be tube amp period, no matter how good digital is.


    Last, you have mentioned something similar to "I think Kemper's technology is currently the most authentic sonic representation of a real amplifier, though the primary limitation is in reproducing an amplifier in whole rather than in part." above multiple times in this thread.


    It's stated as fact, as in this is something everyone thinks is a limitation. You obviously do. But not everyone does. I assume you refer to the fact that a single profile/rig doesn't duplicate exactly how that amp sounds at every single possible setting of it's EQ/GAIN/VOLUME knobs. I still conjecture, that is a limitation that doesn't have to really matter. How often, when playing live, does one go and tweak their amp during the set? If they have need of drastically different tones, they will run multiple amps, use pedals to achieve, or multi channels, or a mixture of all. Generally they will set the amp to, yep, a sweet spot, and make minor adjustments for that room. Pretty much, the concept behind the individual RIG. And, what the Kemper can do, now, as far as the EQ stack is more than enough to tweak for a room, or even volume setting and stay accurate to that 'sweet spot'.


    In the studio, perhaps, there is more EQ changing, etc. But, that is why most profilers, profile the same amp at various different settings. In the studio, it really is no more difficult, to switch a profile, than to dial in an amp for different EQ, settings etc.


    Put those in a performance, and, when looked at, at the performance level, you have captured the amp 'as a whole' for anything that really matters in the practical world. And live, this actually gives you a lot more ease to change the amp for different tones, easier, imho, that tweaking that analog amp.

  • Having been one, I have a few guesses on why someone would keep to their amp:
    1) it CAN be quite easily cheaper, amp depending. Not everyone needs 3-5 amps or a boutique to do their thing.
    2) they know how to manipulate that/those amps based on years of experience. Why lose that comfort level?
    3) they KNOW it sounds good "in the room" and it's hard to initially know how to setup a Kemper to compete on that level.
    So why go through all the fuss?
    4) they know the amp will be there 20-50 years from now. Amps can be fixed indefinitely. Electronics will eventually be obsolete.
    5) people in the audience know an amp when they see one. sometimes folks are like 'where is that sound coming from?!?"
    6) and many are just plain convinced the real amp sounds better and no one can convince them otherwise.
    7) software based things like the Kemper represent not only complexity, but more needs. You need a computer, you need a USB stick, you need different types of wires, you need speakers to hear it out, you gotta figure out WHICH speakers you want (FRFR?, Cabinet?) There IS something to be said of just going home, turing on the tube amp and play. No software updates, no having to reformat your USB stick, bad USB wire, synching with PM taking forever, etc, etc.


    I totally get that and those are totally valid reasons.


    If I had the space and the easy budget, I'd own both. The Amps for their ability to originate tones and the Kemper for that ease of using many-in-one box without the sacrifice my tone.


    (ps, paults knows why I didn't list number 8 )

  • Ok, lets dissect the LOGIC of Cliff's statement, if he in fact said that.


    1) "to perfectly model a vacuum tube" means he doesn't model the actual tube perfectly.


    No current modelers do, nor do companies that make them claim to. Cliff is very forthright about it and has stated such on the Fractal Audio forums. The rest of us his post is as follows:


    "No one has a perfect model of a vacuum tube yet. I like to think we have the best models. Our models are not the crude waveshapers used in most other products but actual mathematical models of tubes. There are a variety of tube models out there: Leach, Reynolds, EXCEM, etc., etc. None of them are identical to a real vacuum tube. I have tweaked our models using a combination of measurement and listening.


    To perfectly model a vacuum tube you would need to model the electron flow through the tube which would be extraordinarily complex and require a super-computer to do in real-time. As it is using mathematical models (as opposed to waveshapers) is extremely CPU intensive and barely runs real-time. If you've ever used SPICE to do a transient simulation you'll know that these simulations run at a tiny fraction of the speed necessary for real-time applications. If you simulate a tube circuit the results vary considerably depending upon the model you use for the tubes. I've tested dozens and dozens of different tube models to come up with what I consider the most accurate given the hardware constraints."


    BUT, the Kemper does model the sound of many tube amps entirely authentically. And it is not modeling them "perfectly" from the bottom up, nor is the Kemper a "super computer" with it's weaker little sharks. (lol)


    So we KNOW you can have a slimmed-down model with variables that are input by an actual probing of the real amp to get authentic emulations. Hence it IS possible to have a slimed-down version, that isn't perfect, running on a non-super computer, that gets the job done, right? Right.

    There's a reason the company refers to the Kemper as a profiler rather than a modeler. Because it's not a modeler. The Kemper doesn't require a super computer because it doesn't model vacuum tubes and as a result has its own limitations. The amp emulations are superior to anything else on the market, though the caveat is that because it's not performing component level, the limitation is an emulation of an amp at a specific setting. While component level modeling currently yields inferior results by comparison, it's inevitable that the technology will ultimately improve as processing power increases.


    The entire point has nothing to do with which modeler is superior in any respect but rather which criteria will need to be fulfilled in order for amp manufacturers to worry.

  • ColdFrixion,


    I'm not sure that component modeling needs more power per se. I think it needs a better understanding of it's own integration. It could very well be that at the 2 meter level it fails at the 7,000 meter level because there are characteristics that can't be understood in a schematic. (component layout interactions, speaker cabinet wood, the effect of the speaker pushing back on the power amp's circuitry, etc)


    So you can come bottom up and miss the big picture as Fractal does. And you can come top down and miss manipulation details as I'm sure Kemper has.


    The first question is: overall, do you sound like the amp
    The second question is: can you manipulate the sound like THAT amp does (with it's dials and tube choices)


    Kemper excels at the first, isn't really capable except in separate profiles of the second WITH enough manipulation that puts you into the ballpark outside of that individual profile.


    Fractal puts you in the ballpark of the amp and has excellent manipulation to get you much closer. And if effects are a key measure of the sound you are looking for, you can get to a point where both boxes achieve similarity.


    You lost similarity as you go straight up amp, no effect.


    It's all a very academic and inserting discussion. Few understand it at the level Cliff and Christoph do. We're playing with gun parts in the sand here. But that don't make it less fun! lol

  • That's based on YOUR criteria of what the genre needs to be.

    Yes, hence the words "I think" and "in my opinion" sprinkled throughout the post.

    I also heavily believe, it isn't the technology, right now, that are keeping more people from using high end digital gear.

    I respect your opinion, though I disagree in part.


    You referenced "the vast majority". Approximately how many people have you asked?


    Last, you have mentioned something similar to "I think Kemper's technology is currently the most authentic sonic representation of a real amplifier, though the primary limitation is in reproducing an amplifier in whole rather than in part." above multiple times in this thread.


    It's stated as fact, as in this is something everyone thinks is a limitation.

    Note the words "I think".

    I assume you refer to the fact that a single profile/rig doesn't duplicate exactly how that amp sounds at every single possible setting of it's EQ/GAIN/VOLUME knobs. I still conjecture, that is a limitation that doesn't have to really matter. How often, when playing live, does one go and tweak their amp during the set?

    I haven't taken a national survey so I can't speak for all gigging musicians. That aside, there's a reason amps have more than one knob. It's the same reason commercial vendors profile amps at various settings.


    In the studio, perhaps, there is more EQ changing, etc. But, that is why most profilers, profile the same amp at various different settings. In the studio, it really is no more difficult, to switch a profile, than to dial in an amp for different EQ, settings etc.



    Put those in a performance, and, when looked at, at the performance level, you have captured the amp 'as a whole' for anything that really matters in the practical world. And live, this actually gives you a lot more ease to change the amp for different tones, easier, imho, that tweaking that analog amp.

    Creating multiple profiles at various settings no more captures an amp as a whole than taking multiple pictures of specific colors on a color wheel captures the entire spectrum. In my opinion, there are far more non-gigging hobbyists that own tube amps than vice versa. Profiling your own tube amps is great, though that still requires a tube amp.

  • "Modeling" is always going to be more intense than profiling IMO.


    Lets just look at the two approaches:


    Modeling - produce a model for EACH AND EVERY electrical and mechanical component of the amp, then make all that logic string together in a massive series - parallel system. Each bit of input is put through this massive ringer and sent to the output.


    Profiling - Measure the response of the entire SYSTEM that you want to emulate, put it in a can, and then put every bit that goes into the system through that can and send it to the output.


    In electrical engineering the latter is called "the system response" frequently refereed to in math as H(X) (Read "H of X").


    Modeling attempts to build up H(X) from all the little h(x) for each component. This will be many times as computationally intensive, and any error in any component h(x) will be amplified as error throughout the system.


    To be more accurate about what CK has done, the architecture of the KPA consists of 4 pre-processing steps that are in series. This feeds into 3 major system response transfer functions for AMPLIFIER, EQ, and CABINET. The output of this stage is fed into 4 post processing engines.


    The really cool thing about the KPA is its ability to capture those 3 system response transfer functions so accurately and easily, and implement them with so little processing.


    I am not giving away any huge secrets here. Any decent EE with a signals and systems background would come to the same conclusion about the KPA. It is that last step where things get really hairy. Getting it done accurately, easily, and in a way that takes very little processing power.


    It's kinda like Colonel Sanders special recipe for guitar amps ;)

  • I'm not sure that component modeling needs more power per se. I think it needs a better understanding of it's own integration. It could very well be that at the 2 meter level it fails at the 7,000 meter level because there are characteristics that can't be understood in a schematic. (component layout interactions, speaker cabinet wood, the effect of the speaker pushing back on the power amp's circuitry, etc)

    I wouldn't be surprised if there are at least some characteristics that aren't well understood. For characteristics that developers like Cliff do understand and feel incapable of modeling with 100% accuracy due to a lack processing power, I think faster processors will ultimately prove useful. In my opinion, and this is just me, I feel fairly confident that at some point modeling will achieve perfect accuracy, not just in terms of sonic reproductive accuracy but the full gamet of parameter control. I can't prove that of course, but I tend to believe it.