Kemper Profiler MK 2

[BenIfin]

Hey, just trying to be honest. I find him annoying as hell, but his purple plexi was really awesome.

I get it :) Actually a lot of his demo's of his packs sound really great. I am/was just more of a fan of Direct Profiles and most of his are Studio. I can take his rants for a couple of minutes, but after that its like c'mon man !!!!

He took the outcry over the Player Paid Levels -way- too personally, which was ultra weird.

Funny thing is, to my ears, the MBritt ones are just glorious - I really like the mid-forwardness of them, but I know that puts a lot of people off them - but all my stuff is clean to classic rock - "ACDC style" tone is as heavy as I get - most of my playing these days is edge of breakup to crunch - covers - so his stuff really suited me.

 

[Enchilada Jones]

If I’m understanding it correctly, Kemper profiles are some kind of complex eq mojo applied to a single static base amp model regardless of the target amp characteristics. All profiles are constructed with the same generic one size fits all foundation. And the results were considered amazing back in the day but now in light of more modern methods, NAM, machine learning etc., it is criticized as a weakness.

So if the new mkII method uses some modern tech to replace the ‘base model’ by reading the target amp output and modeling it to be closer to the target then the following steps of the Kemper eq mojo are working on a more accurate foundation suited to the target amp characteristics.

Then if their new speaker sim is as good as they say mkII is going to be a pretty good upgrade.
The cab situation with Kemper has always been where I thought they should improve things. Cab sims are a big part of any digital amp success.

 

[paisleywookiee]

I get it :) Actually a lot of his demo's of his packs sound really great. I am/was just more of a fan of Direct Profiles and most of his are Studio. I can take his rants for a couple of minutes, but after that its like c'mon man !!!!

He took the outcry over the Player Paid Levels -way- too personally, which was ultra weird.

Funny thing is, to my ears, the MBritt ones are just glorious - I really like the mid-forwardness of them, but I know that puts a lot of people off them - but all my stuff is clean to classic rock - "ACDC style" tone is as heavy as I get - most of my playing these days is edge of breakup to crunch - covers - so his stuff really suited me.

Yeah, I’ve got a friend who interviewed him for a podcast, and says he’s actually a really nice dude who just loves gear and tech like the rest of us. He definitely takes things personally though, for some reason, and his online persona is…..what it is.

 

[Sascha Franck]

Funny thing is, to my ears, the MBritt ones are just glorious - I really like the mid-forwardness of them, but I know that puts a lot of people off them - but all my stuff is clean to classic rock - "ACDC style" tone is as heavy as I get - most of my playing these days is edge of breakup to crunch - covers - so his stuff really suited me.

Every bit the same over here. The KPA by buddy lent to me had a whole bunch of the MBritt profiles, only slightly adjusted to suit his gigging needs, and I still remember how I simply just didn't want to stop playing them. I also found the options to tweak the tonestack perfectly acceptable (and that was *way* before liquid profiles came up), so I never understood how people would take so much issues about profiles only being snapshots of an amp. Sure, it defenitely has to do with them being low to mid gain as well, so they were reacting just fine to gain tweaks, might be totally different in the high gain realm, but I'm not exactly interested in those kinda tones.
Also, the KPA seems to take pedals just fine, so you could always go that way, too.
Anyhow, to these day, it's been some of the most satisfying out-of-the box playing experiences.

If I’m understanding it correctly, Kemper profiles are some kind of complex eq mojo applied to a single static base amp model regardless of the target amp characteristics.

I think they're using more than just one baseline model. I seem to remember someone posting a video of updating a KPA and it said something like "updating amp models" or so. I very obviously have no idea how the algorithm decides on which model to use, but it'd make sense to start with at least some variety of gain characteristics because even the most elaborated pre/post EQ-ing will only take you this far.
Maybe there's also some trickery happening regarding amp dynamics. As we all know, some amps clean up more when dialing down your guitar volume, some less so - no idea what could be done for these differences to find their way into a profile, but using input level as an additional gain modifier can take you a long way. I used that "trick" back in the days to make some Guitar Rig patches react more dynamically.

 

[Jarick]

So worship tutorials made a video yesterday about the Helix where they said there’s one amp modeling engine that has gotten two major updates over the years. It sounds like when they have new amps they sometimes have to expand the engine. So there seems to be one central model with a whole bunch of variables and parameters that does all the amp models.

I would bet Fractal is a similar approach because it would make more sense to create one big engine than 300 plus. Especially when there are big updates…do you think Cliff rewrites 300 amp models or updates one and then measures and updates parameters for each of the amps?

Kemper likely does something similar but automates the parameters based on measurements taken during profiling. It’s not magic. So I don’t necessarily think it’s an indictment to say there’s one underlying amp model. That’s probably true for all the digital products.

Lastly there’s a lot of talk about 100,000 point measurements. A standard IR is what 1024 samples? Higher res may be twice that? Let’s say Kemper takes 50 different measurements of 2048 samples at diffferent dynamic levels or something like that to capture the measurements. Then it crunches the numbers to put in the variables on the Kemper model and that’s your profile. That’s my guess.

 

[Orvillain]

I find it hard to believe your choice of an example ‘any other device is good enough’ was literal and wasn’t presented without the intent to challenge the validity of the premise of good enough.

Yes. You are smart. I was definitely challenging the premise of "good enough" - because I think "good enough" is actually code for "I don't care that it isn't as good as it really should be, I still luvs it"

You would have to believe the Kemper user truly would find no difference, dismissing the Kemper users preference and any reason for it, because accuracy is paramount and what it sounds and ‘feels’ should be unimportant to them.

So many Kemper users have told me that they don't hear a difference between their amp and the profile they make. I'm genuinely stunned that this is the case, but it appears to be so. Whenever I even talk about it, I get people like @OneEng implying that my hearing is shit, or that I'm making stuff up. Absolutely unreal.

Aside from that, I only think accuracy is important (read: not saying paramount) because of the decade long marketing claims on part of Kemper (and Neural and ToneX while we are at it!) coupled with my empirical observations that it just isn't true. Even NAM isn't 100% all of the time, but it is definitely way way way closer.

You can go back through all of my posts if you fancy. You won't find a single time where I've said anything like "X/Y/Z should be unimportant to A/B/C person" - because that just isn't how I think.

At the same time, I think it is completely valid to discuss and conjecture over Kemper's accuracy - past, present, and future.

That would be a very simplistic perspective for an extremely experienced engineer. It’s like a mechanical engineer arguing with an architect that the sweeping curve of the designed building should be altered to a simple cube shape because the strength of the cube is all important and the architect assertion that the curves are strong enough is wrong.

It would be a very simplistic perspective yes. Good job it isn't mine. Dodged a bullet there!

 

[Orvillain]

If I’m understanding it correctly, Kemper profiles are some kind of complex eq mojo applied to a single static base amp model regardless of the target amp characteristics.

I believe Christoph already said in the past that this isn't the case.

All profiles are constructed with the same generic one size fits all foundation.

No one knows if this is true or not.

And the results were considered amazing back in the day

This is false.

but now in light of more modern methods, NAM, machine learning etc., it is criticized as a weakness.

And this is also false.

So if the new mkII method uses some modern tech to replace the ‘base model’ by reading the target amp output and modeling it to be closer to the target then the following steps of the Kemper eq mojo are working on a more accurate foundation suited to the target amp characteristics.

This sentence doesn't mean anything.

Then if their new speaker sim is as good as they say mkII is going to be a pretty good upgrade.
The cab situation with Kemper has always been where I thought they should improve things. Cab sims are a big part of any digital amp success.

I don't share this view. Aside from anything else, it was often impossible to know with the Kemper what aspects were amp related, versus cab related. They're kind of inseperable. So ascribing certain negative sound qualities to one or the other wasn't feasible.

Saying that, we all know that the IR importing part of Kemper is woeful, and yes, that could do with a lot of improvement.

 

[JiveTurkey]

I always read there 7 or 8(?) base amp models?

 

[Orvillain]

I always read there 7 or 8(?) base amp models?

Cliff claimed that. Christoph denied it.

 

[Mongillo19]

Cliff claimed that. Christoph denied it.

I can't remember, were you a beta tester for Kemper back when?

 

[Orvillain]

I can't remember, were you a beta tester for Kemper back when?

I was for a bit - maybe a year. Nothing special happened back then. Just iPad bullshit.

 

[Sascha Franck]

Cliff claimed that. Christoph denied it.

Obviously he denied. Gotta preserve the fairy dust as long as possible.

 

[PorcinePicker]

Hmmm....I've limited my comments to the hardware changes, or rather, the decided lack thereof.

It appears that fw 13 beta is just an interim step, and that nobody has heard the Mk II processing yet. Amiright?

 

[Jarick]

Hmmm....I've limited my comments to the hardware changes, or rather, the decided lack thereof.

It appears that fw 13 beta is just an interim step, and that nobody has heard the Mk II processing yet. Amiright?

Correct. 13 added the fixed blocks from what I can tell.

 

[Lysander]

So there seems to be one central model with a whole bunch of variables and parameters that does all the amp models.

No, that's not how it works.

 

[Deadpan]

No, that's not how it works.

Pretty sure this is what CK said.

 

[Lysander]

Pretty sure this is what CK said.

He's talking about Helix, not Kemper.

The Kemper platform (and Neural DSPs, as far as we known) is indeed built on 8-10 base amp models.

 

[Deadpan]

He's talking about Helix, not Kemper.

The Kemper platform (and Neural DSPs, as far as we known) is indeed built on 8-10 base amp models.

Ah, I didn't see the complete post only your quoted part.

I do seem to recall CK saying the Kemper was one very complex model.

 

[Deleted member 472]

 

[Enchilada Jones]

Yes. You are smart. I was definitely challenging the premise of "good enough" - because I think "good enough" is actually code for "I don't care that it isn't as good as it really should be, I still luvs it"

Yes. Good enough means good enough. It’s not code.
You are the one who is projecting a need for “enough” to also cover “as good as it can be” and then becoming indignant when others don’t adopt your custom definition and push back at how you demand your perspective be the benchmark.