Cab Changes effect tone with Cab off

This is because amp and cab (and mic) are not profiled individually. It's not like having these components modelled in single items.
Excluding the cab renders the "direct" sound through an algorithm.
Also consider that the sound of an amp changes depending on which cab it was connected to, this makes things more complicated since - again - there are no individual "blocks" but just one profile.

Which begs the question: Is it possible to profile the amp and cab separately and thus allow for completely separate and interchangeable units?

I know that the second phase of profiling, where chords are strummed and the sounds from a miked cab are fed back into the Kemper, is meant for 'fine-tuning' the profile. I guess I'm wondering out loud if the day might come when the amp and cab profiles can be separate and distinct. I suppose it might involve isolating the cab portion of the tone (a library of cab tones recorded at various frequencies might exist) and removing it from the amp profile leaving behind a neutral tone that could then be combined with the cab of your choice.

I'm afraid we'll never see such an option.
A power amp and a cab interact with each other. Each of them sounds differently because there's the other device, and it's that specific one.
So you should have a library of every possible cab connected with every possible power amp at different settings.

Also, as long as the profiled gear is connected in series (one device's output feed next one's input) there's no way to tell the exact contribution each of them gives to the overall sound.

Right, understood, but I was thinking the Kemper might be advanced to the point where it could get everything it needs from the amp alone (phase 1) so no cab connection need be in place. Then a neutral cab driver could be employed to profile the cab independently. Granted, hooking the two profiles together would not duplicate the sound of the amp and cab connected in the room, for the reason you mentioned, but I'm wondering how discernible that difference would be.

To have a separate amp and cab component in a profile, the profiling process would have to change. The amp would have to be profiled with a standardized dummy load then the cabinet would be connected to profile the speaker.

Currently the amp and speaker are profiled as a single system. Apparently the cab off feature is a simulation of what an amp might sound like without a cab. There are profiles that have been made from an amps output into a dummy load. These are accurate and many say they work well with a real cabinet.

Quote from elvisscottyandbill

Granted, hooking the two profiles together would not duplicate the sound of the amp and cab connected in the room, for the reason you mentioned, but I'm wondering how discernible that difference would be.

While the "in the room" part could be a bit misleading, the difference would be huge.
That's why a modelled amp + a modelled cab will never sound "real", unless the models take into account their complex and dynamic interactions taking place at different volumes and drives.
And that's why the profile of amp A with cab B off will sound different than the profile of amp A with cab C off, even at the same amp settings. This is one thing you might play around with, if you can, to hear the difference (tone setting has to be the same tho, and this is probably the hardest part, unless you take profiles by yourself or use profiles you know tone setting of).

The KPA Stack works like this:

--- Before-Filter ----- Distortion Generator ---- After-Filter ----

The first filter is called Amp the second Cab - it would be better to call them Before and After filter.

You can freely use the Before and After filter from any any other rig - but it's not the same as exchanging the amp and cab.

It's not possible regardless of the used technology by profiling both (amp and cab) at once.

A big part of the amp influence (power amp, presence, tone stack, ...) is stored into the the After-Filterilter ( = CabFilter).

But you can freely experiment and exchange these filters and some combinations may sound great for your ears - but they have nothing to do with the real amp and cabinet - they must be profiled at once and both filters have to be used at once.

That's why the KPA sounds best with FRFR.

Right. Understood that this is how things currently work. I was speculating on how things might be made to work in the future. Getting the technology to the point where individual components in the signal chain can be profiled independently and then virtually attached to one another to produce the sound of the physical amp, cabinet, eq and stomps in the room.

I understand that the Kemper is not designed to work this way. Merely wondering out loud if it's possible to get there at some point further down the road.

The difficulty appears to be in capturing the interplay between the electrical components hooked together in the signal chain. Not sure if it would even be noticeable. But I envision a day when a device like the Kemper would have interchangeable libraries of components that could be mixed and matched as the user requires. As opposed to depending on finding a cab that has already been profiled with another amp and thus picking up some of that amp's affect on the tone produced by that cab.

Yes, eg the AxeFx does component modelling .

The problem is modelling/profiling an amp and cab independently sounds different - since a lot of interaction happend between the amp and cab.

Have the observations discussed here regarding how people think the Amp/Cab/Tonestack profiling process breaks down been formally commented on by Kemper? If so, I must have missed it.

From what I've been able to tell (along with one particular interview with CK that talks a bit about Step 2) the profiling process goes something like this:

1) Briefly analyze the distortion of the amp throughout its range to find where the cleans end and make note of this. If there are no cleans to be found, give up.
2) Profile the EQ response of the system via a combination of impulse response and level-varying noise (all within the amp's clean range). This becomes the cab.
3) Discover (using CK's proprietary mechanism) the amp's naked waveshaping character and generate a harmonic resynthesis function from it. (The information gathered in Step 1 is probably also used to determine the test signal gain.) This becomes the amp.
4) Set the EQ to 0/0/0. This becomes the tonestack.

Now, EQ is EQ - barring some weird proprietary magic, it's difficult (if not impossible) to separate the amp's fixed tonestack, mic and cab EQ properties so I wouldn't be surprised if these all -do- end up in the cab section.

If those assumptions are true, then we could also assume that:

1) The "cab" section of an amp profiled speaker-out (no mic, no cab thru a DI box) would contain the amp's tonestack EQ and little else... unless the profiling process was smart enough to recognize "that can't possibly be a cab" and zero the cab out (while possibly migrating the tonestack EQ into the amp section.)
2) Changing the "cab" section of an amp shouldn't affect the Studio Monitor Out if the cab section is, indeed, turned off. I have not bench-tested this, but note: I do remember CK saying something about a bit of the original cab tone being 'leaked' into the Monitor Out for aesthetic (tonestack maybe?) considerations.

So it's Point #2 I'm concerned about - which is the question that started this thread. How much do we know for certain, really?

-djh