Please give me your honest opinion about the *accuracy* of the Kemper

Orvillain

Rock Star
Edgelord
Messages
4,472
I did this video a while back. I think it shows clearly the inaccuracies of the Kemper when it comes to palm muted riff type stuff. All I've ever heard is "it doesn't matter" and "the profile sounds better than the amp" neither of which was the point, and then other comments trying to poke holes in the comparison by blaming the DI box or some weird metaphysical concept of impedance loading with line level signals!! The point is a technical and objective one - the Kemper isn't 100% accurate. I wouldn't even say it was 98% accurate.

It is quite off in many instances. This video demonstrates just one. Are people not able to hear this stuff? Genuine question.

 
I picked up a powered toaster a few years ago, the goal was two things. I wanted to capture my amps for easier home recording and secondary goal was to cure all future amp gas and just play profiles of my favorite amps. The only problem was I could never get it to sound like my amps. And something about the upper high mids sounded so digital ish to me.

My goal was direct captures so I didn't use a mic, instead I tried using my Suhr RL, Fryette PS-100, Waza TAE, I even got desperate enough to buy the damn Kemper di box thinking there must be some magic ingredient I'm missing but nope. I was never happy with my profiles overall. The only way it sounded decent was thru a tube power amp. I still have it and mess with it every 6 months thinking I'll finally get it.

I will say the only profiles I bought that I truly liked was the Friedman Naked from @BigHairyGuitars. Best profiles out there to me.
 
Great video @Orvillain, I think the Kemper sounds REALLY close to my ear especially after you tweaked the Tube Shape, I wouldn't be bothered by the last 2%.
The QC is probably just as good.

What they definitely CAN'T capture is bias excursion and 'swirl' that is a big part of a push-pull tube power amp, but they excel with preamp gain.
 
Last edited:
I think a Kemper through a guitar cab is a glorious thing. Especially if it is something with a Marshall/Vintage intention behind it. You can certainly sing praises to :satan but it's going to have that little bit of classic flair to it. And that's ok by me :guiness


Sorry; I don't think I answered your question :bag
 
I picked up a powered toaster a few years ago, the goal was two things. I wanted to capture my amps for easier home recording and secondary goal was to cure all future amp gas and just play profiles of my favorite amps. The only problem was I could never get it to sound like my amps. And something about the upper high mids sounded so digital ish to me.

My goal was direct captures so I didn't use a mic, instead I tried using my Suhr RL, Fryette PS-100, Waza TAE, I even got desperate enough to buy the damn Kemper di box thinking there must be some magic ingredient I'm missing but nope. I was never happy with my profiles overall. The only way it sounded decent was thru a tube power amp. I still have it and mess with it every 6 months thinking I'll finally get it.

I will say the only profiles I bought that I truly liked was the Friedman Naked from @BigHairyGuitars. Best profiles out there to me.
I agree with you 100%. I got a Kemper a couple years ago and couldn’t wait to make make some profiles. I dialed in my amp, mic’d the cab, and recorded some stuff. Then I set up the Kemper to make a profile and just having it “in the chain” boosted some weird upper mids and chopped off the top end. When comparing the source to profile, it was pretty close, but the “source” through the Kemper was vastly different than the track I recorded that was just amp and cab.

I talked to Michael Nielsen about my problem with the upper mids and top end reduction, and he confirmed that the Kemper just does that.

I’ve gone to shows where players are using Fractal or Helix and I couldn’t really tell from FOH, but I’ve seen several bands where I’ve heard that weird upper mid thing and found that they were using Kempers.

To be positive, the idea of a Kemper was a cool digital solution when it first came out and was a useful tool for a lot of players. The reality now, in my humble opinion, is that amp modeling has come so far since then that the Kemper tech is a bit outdated and the newer modeling units are far more accurate in apples to apples comparisons.
 
I agree with you 100%. I got a Kemper a couple years ago and couldn’t wait to make make some profiles. I dialed in my amp, mic’d the cab, and recorded some stuff. Then I set up the Kemper to make a profile and just having it “in the chain” boosted some weird upper mids and chopped off the top end. When comparing the source to profile, it was pretty close, but the “source” through the Kemper was vastly different than the track I recorded that was just amp and cab.

I talked to Michael Nielsen about my problem with the upper mids and top end reduction, and he confirmed that the Kemper just does that.

I’ve gone to shows where players are using Fractal or Helix and I couldn’t really tell from FOH, but I’ve seen several bands where I’ve heard that weird upper mid thing and found that they were using Kempers.

To be positive, the idea of a Kemper was a cool digital solution when it first came out and was a useful tool for a lot of players. The reality now, in my humble opinion, is that amp modeling has come so far since then that the Kemper tech is a bit outdated and the newer modeling units are far more accurate in apples to apples comparisons.
What are your thoughts on the Neural QC? Do you find that you are hearing similar weird frequencies or tones in the captures
 
What are your thoughts on the Neural QC? Do you find that you are hearing similar weird frequencies or tones in the captures
I'm curious on this too. If its actually accurate I would get one for sure just for capturing.
 
When I had the QC, I preferred the captures to the Kemper.

QC captures are closer to the amp, and I never heard any weird high frequencies. But I did feel a bit of latency, especially when using multiple captures at once.

Also, aside from the capture functionality, the unit just isn't MATUREEEEEEE enough right now, and I really didn't like what the unit did to the signal quality when used in 4-cable-method.

All in all... a bit of a miss for me. Maybe it'll get better in the next hundred yeaarrrrsssss.
 
I'll probably get one at some point, but realistically I'll probably wait for a hardware revision first. All I want to do is capture my recording chain of my boosted amps accurately... every time I tried with my Kemper it tells me to turn the gain down... :cautious:
 
I did this video a while back. I think it shows clearly the inaccuracies of the Kemper when it comes to palm muted riff type stuff. All I've ever heard is "it doesn't matter" and "the profile sounds better than the amp" neither of which was the point, and then other comments trying to poke holes in the comparison by blaming the DI box or some weird metaphysical concept of impedance loading with line level signals!! The point is a technical and objective one - the Kemper isn't 100% accurate. I wouldn't even say it was 98% accurate.

It is quite off in many instances. This video demonstrates just one. Are people not able to hear this stuff? Genuine question.


Nice video! It sounds to me like the higher frequencies on the palm mutes were actually closest to the reference with the original, un-refined capture. By the time you got to the last tube-shaped adjustment the upper frequencies had been noticeably attenuated. But in all cases, the low end of the palm mutes was indeed different from the ref. The Kemper had a fuller/boomier low end but less frequency separation. The real amp has less boominess in the lows and more articulation between the notes.

That said, if they weren't being A/B'ed one after the other, I doubt I'd ever be able to pick out any of these details unless I was very familiar with the real amp. Even then I'm not sure I could do it consistently.
 
What are your thoughts on the Neural QC? Do you find that you are hearing similar weird frequencies or tones in the captures
That’s a good question. I haven’t made any amp captures with the QC yet, but I have friends who say they can be very good.
I’m going to try it with several amps with different gain structures to see how it captures tightness/sag from different amps as well as how it responds when you roll off your guitar’s volume.

I think this list should cover enough ground:
‘77 50w Marshall JMP
Mesa Dual Rec Rev F
Matchless HC-30
Orange OR80
Marshall JCM Slash
Mesa DC-5
 
I've made a gazzilion profiles. Kemper's accuracy, or lack of, often depends on the amp itself and the way you do profile. Some amps are spot on. Other amps not so much. The signal chain and gain staging also has a big impact.
 
I've made a gazzilion profiles. Kemper's accuracy, or lack of, often depends on the amp itself and the way you do profile. Some amps are spot on. Other amps not so much. The signal chain and gain staging also has a big impact.
Yeah me too, and there are variances for sure. But very rarely did I ever get a profile where I couldn't hear or feel a difference. It is what made me sell it in the end and just bail on the whole eco-system.

I think the QC is better. But I sincerely hope Kemper come out with a 2nd unit with massive improvements and a more modern UI and user experience.
 
Surely you mean the opposite… right?
Nope. The Kemper, on all three takes, has something extra in the lowest bass frequencies that isn’t there on the real amp. It sounds almost like a low-freq cab resonance which is sounding continuously behind the palm mutes and making them sound less tight.

The real amp only has the tighter articulation of the notes without that droning low frequency resonance behind them. There’s also more detail in the grit of the low strings on the real amp vs. the Kemper.

It’s a good example of how you can match static (time-invariant) qualities really closely just by matching frequency balance with a match eq, but the dynamic traits (things that vary in time, like the way those palm muted notes attack, bloom and resonate) are much harder to represent accurately with a model or profile. This is generally true of math modeling in the engineering world, too — steady state behavior can often be represented with great precision, while dynamic behavior is far more complex to represent accurately. This is where neural networks can have a big advantage, because they aren’t constrained to the same mathematical complexity limitations, and can “learn” behavior that would take a ton of extra variables and functions to represent with traditional math.
 
Nope. The Kemper, on all three takes, has something extra in the lowest bass frequencies that isn’t there on the real amp. It sounds almost like a low-freq cab resonance which is sounding continuously behind the palm mutes and making them sound less tight.

The real amp only has the tighter articulation of the notes without that droning low frequency resonance behind them. There’s also more detail in the grit of the low strings on the real amp vs. the Kemper.

It’s a good example of how you can match static (time-invariant) qualities really closely just by matching frequency balance with a match eq, but the dynamic traits (things that vary in time, like the way those palm muted notes attack, bloom and resonate) are much harder to represent accurately with a model or profile. This is generally true of math modeling in the engineering world, too — steady state behavior can often be represented with great precision, while dynamic behavior is far more complex to represent accurately. This is where neural networks can have a big advantage, because they aren’t constrained to the same mathematical complexity limitations, and can “learn” behavior that would take a ton of extra variables and functions to represent with traditional math.
This is exactly how I perceive it too.
 
The QC captures amps differently for sure from a tonal perspective.

I find it handles boosted amps infinitely better, especially with RAT and HM2 style pedals.

Kemper will not capture them well for high gain. And by high gain I mean Death Metally type stuff like Entrails, Grave, Entombed, etc.

Nothing I’ve captured on either I feel is 100% accurate BUT… are always similar enough to where I’m very happy with the results AND… many times I’ve found I like the captures a tad better with those inaccuracies present.

I’m a weirdo though…


Note: please sub capture for profile in my Kemper statements, lest more sand creep into the eVag
 
Back
Top