"Blend" or "cascade" two profiles into one

  • While recently there have been requests for a hardware update (i.e., a "Kemper 2.0") with two independent signal chains, one main objective of these requests is the possibility to have two different amps in parallel (see also the ongoing QC thread in "other gear").


    While two amps in two parallel signal chains do not seem feasible for the Kemper due to DSP restrictions (I think this has already been addressed many times), there might be a "workaround feature":


    How about "blending" two amp profiles, i.e., choosing two different amp profiles as "input", and have a "blended" (or mixed) profile of the two amps as "output", that could then be used in the amp section. I understand that the desire for two parallel amps is mainly for practicing, rehearsal and live purposes (rather than studio where reamping separately would normally be the better option). So for these purposes, the blended profile could at least get close to the parallel setting, even if the post effects would be applied to the blended signal.


    The blending could be controlled via the editor (but the computations could be run on the Kemper hardware).

    There could even be options for setting a "mixing ratio" other than 50:50. Maybe there could even be something like a frequency-dependent mixing, i.e., the output profile is based on Amp1 for low frequencies and based on Amp2 for high frequencies (preferably, with an adjustable transition or splitting frequency).


    And then, one profile could also be used as an input into another profile (such that it would not be a mixing, but the output would be the profile of the full chain of the two input profiles one after another). This could be used for combining drive pedal profiles with amp profiles. So far, if one profiles a drive pedal, either it is "baked in" into an amp profile, or there is no amp profile in a rig. For baking it in, however, one needs to have both the amp and the pedal physically at the same place.

    You cannot profile a friend's pedal at a rehearsal or after a gig and then "cascade" it into your favourite amp profile - so far. But with the proposed cascading option, one could even add the pedal to a running profile via a simple remote switching in a performance.


    I guess these could be really useful features for many live oriented players.

    And though the frequency dependent option for blending and also the pedal into amp cascading might be more advanced to implement, it might well pay off.

  • I proposed this in 2018 but later realised that the idea had been floating around since 2014.


    Here's my response to cmboowns earlier this year here:


    ------------------------------------------------------------------------------------------


    Hey, not far-fetched at all, mate. I started a thread in which there was a deep discussion about this about 2 1/2 years ago.


    I remember asking for the ability to blend two Rigs to-taste. The idea was along the lines of what you're suggesting, where the "invisible" parameter values of two Profiles could be added or subtracted from one another en-masse, expressed as an overall percentage value of one Rig to another, adjusted to-taste and then saved as a new Profile and subsequently played.


    Not realtime. No RT Morphing involved, just offline combining of two Rigs' Profiles. There'd be a small delay after each adjustment of the percentage value due to said offline combination of data, so at-worst you'd think that that playing whilst adjusting might be a little "skippy".


    The discussion was quite-lengthy but unfortunately we didn't hear from Kemper.


    Just found the thread. Here's my OP. Link follows:


    We know the CPU can't render two Profiles at once. We know Profiles are only a few kb in size.


    Bearing these facts in mind, shouldn't it be possible to load two Profiles into the edit buffer and render a single resultant amp / cab or both based upon an interpolation of the raw (Profiled) data of each? IOW, you choose how much of each DI, Merged or Studio Rig you'd like to hear via a "blend" parameter, and this resultant set of parameters, which would be no different in size or scope than that of a single Rig, could be rendered through the Profile-playback engine... in realtime.


    Further possibilities employing the same principle spring to mind:


    Blending Cabs.

    Awesome possibilities of morph-pedal assignment. The mind boggles!

    Splitting the single blend parameter up into hitherto-unheard-of parameter sets that allow different characteristics of two amps or Cabs to be blended.

    [Insert plethora of imagined possibilities here]


    To recap:

    Profiles are small files that contain a specific set of parameter values determined during the Profiling process.

    I propose that the KPA should in theory be able to be programmed to allow morphing / blending between any two of these sets because it doesn't have to involve rendering two Profiles at once - just a basic, low-CPU-overhead, user-determined blend / interpolation / morph of the raw data that differentiates one Profile from another, just as morphing between two parameter sets within a single effect doesn't require two instances be calculated at the same time. The resultant data values could then be treated as if they represented a "regular" Profile.


    In theory. Unless the K-Team™ has a spanner to throw in the works...


    ------------------------------------------------------------------------------------------


    Additional Links:


    Mine

    Amp & Cab Morphing Without the Impossible CPU Hit


    b_ryan in 2020

    Combing two rigs


    Lokasenna first proposed this, unbeknownst to me when I created my 2018 thread, back in 2014 here:

    First mention in 2014

  • I have good reason to believe that morphing between 2 different profiles is absolutely possible. I will be highly surprised if we don't get it within the next year.

    From what we've been told previously (and of course, now I can't find the relevant reference from CK to corroborate my statement), morphing between two profiles requires both profiles to be loaded and active simultaneously. Current hardware is not (and cannot be) set up to do this.

    The only way to do this now that I'm aware of is two KPAs, or an entirely new device. The latter requiring an overhaul of the KPA OS as well as features in RM not even hinted at.

    “Without music, life would be a mistake.” - Friedrich Nietzsche

  • From what we've been told previously (and of course, now I can't find the relevant reference from CK to corroborate my statement), morphing between two profiles requires both profiles to be loaded and active simultaneously. Current hardware is not (and cannot be) set up to do this.

    The only way to do this now that I'm aware of is two KPAs, or an entirely new device. The latter requiring an overhaul of the KPA OS as well as features in RM not even hinted at.

    If using the sound of two different profiles - i.e., "dual amping" - right now would be feasible in any way, the feature probably would already have been implemented, as it seems to be a highly popular feature, judging by the requests.


    So it is most likely not feasible in an efficient way (i.e., without sacrificing the reverb tails etc.) for the hardware.


    And that's just why some workaround to at least approximate this dual amping via blending the two amps offline into a new profile and then using it for amping could be a very useful feature, as it might be the closest the KPA can get to dual amping.


    And in particular, as the blending would be done offline, with the option of a frequency dependent blending one could generate an "output" profile with the chiming, sparkling top end of a (insert your favourite soloing lead amp here), but still maintaining that thick and tight low end of the (insert your favourite rhythm amp here).


    I definitely think it would be worth it.

  • And in particular, as the blending would be done offline, with the option of a frequency dependent blending one could generate an "output" profile with the chiming, sparkling top end of a (insert your favourite soloing lead amp here), but still maintaining that thick and tight low end of the (insert your favourite rhythm amp here).


    I definitely think it would be worth it.

    my only issue with that is that you would be creating the new blended profile in the dark. How would you decide what the correct blend is as you can only hear one rig at a time not both simultaneously you have no reference to tweak the blend to taste.

  • Yes, and in that situation you can hear the result before you commit to a single profile. If it was to be done by an algorithm in Rig Manager as people seem to be asking for, I can't see any way to be able to audition the effect of the blend until after the dual amp rig has been created. That could entail an awful lot of trial and error creating profiles based on guesses.

  • I covered that several times in my thread back in 2018, Alan. An example, here:


    Let's take, for example, Parameter 01 from each Profile:

    It would be "blended" mathematically first, producing a single parameter which is treated as if it was only ever the only Parameter 01. If Parameter 01 in Profile A is 1, and it's 10 in Profile B, the blend range would be from 1 to 10, representing 100% A, 0% B through to 0% A, 100% B.


    We know that Rigs can be switched in a matter of milliseconds, so the realtime feasibility of morphing or manual blending via a percentage knob of the two Profiles should theoretically be "instantaneous"... and gratifying, even if each parameter update requires the entire Profile be recalculated.


    IOW, there may be a tiny amount of stair-stepping / stuttering whilst adjusting the blend and listening in realtime, but when you think about it, as long as there aren't total dropouts of signal involved, we'd not even hear the "stepping" 'cause similar signals are being joined to each other / crossfaded just as they are when auditioning in RM, but more-smoothly in this case due to the incremental changes (similarities in sound) involved as opposed to switching to entirely-new Rigs.

  • No mate; that was the whole point of the thread:


    To suggest that two Profiles could be interpolated "offline" 'cause they occupy very-little space (only a few kilobytes). As one turns a knob to blend them in various proportions, the resulting set of values ("new"Profile) could be instantiated again and again with each adjustment. I based this conclusion on the fact that RM allows you to switch between Profiles very-quickly (a few ms). Obviously you can do this on the units too.


    To put it another way:

    You'd simply be altering hidden Profile parameters mathematically and hearing the result as soon as the Kemper were able to make it active, which only takes a few milliseconds.


    HTH, bud.

  • You can blend 2 amps into a single profile during profiling. It entails using 2 mixers. One from the output from the Kemper to split the signal to the 2 amps. Another mixes the sounds from the mics on each individual amp. You can mix them any percentage you like, just like a multiple mic profile.

    While for more professional users this might be no problem, as they simply mix down two of their favourite amps in their studio, others might just not have these possibilities, or just would like to blend an amp with a profile of another amp they do not own.


  • I am not sure how quickly it can be done.


    As far as I understand the procedure of profiling, you have some sort of Wiener-Hammerstein topology where you have two frequency response curves sandwiching a nonlinearity such as a hyperbolic tangent. At least, that is what their patent says.


    As the frequency responses are sandwiching the nonlinearity, you cannot simply "morph" their parameters for blending

    (its A1*NL*B1 + A2*NL*B2, you cannot "morph" a parameter between A1 and A2 for blending, not even offline).

    So the calculus might be more involved, and not that instantaneous, there would highly likely be a switching gap.


    Still, it could be done at the editor, and if the gap is a second or so then there is not really a problem for "trial and error", IMO. You select a blending ratio, have the output profile, play it, change the ratio, wait a second, and play it again.

    It's not like a reboot, it can be done, I'd say.

  • I hear you mate. Interesting info; thanks. :thumbup:

    Still way above my pay grade Nicky but I’ll bow to your superior knowledge :)

    Well you know what they say about false Gods and idolatry, Alan. :pinch: Way above my pay grade too brother.