Kemper Profiler MK 2

[BenIfin]

Regardless of whether null tests are accurate or not that’s now logarithmic math works

True - just a simple raw way to highlight the very small differences .. and b.t.w, the % figure is accurate even for logarithmic number comparisons.

But like I said above .... the "answer" to the question about "Integrated [or other] LUFS" figures for qualitative audio capture comparisons is here => https://www.izotope.com/en/learn/what-are-lufs#integrated

 

[LP59]

It is pretty simple - null residual energy is not equivalent to perceptual difference. Two residuals can have the same LUFS but very different audibility. Null tests work well for linear time-invariant systems; which guitar amps and guitar amp models are not.

It's only a momentary LUFS that would require time-invariance. As long as it is integrated, discrepancies in a time-variant system are absolutely captured in the residual. Yes, there are cases where audibility differs for identical LUFS (and vice versa), most notably when there is poor phase alignment, but Leo Gibson has discussed in the past how he takes measures to ensure alignment is correct.

Again, as I mentioned above, there is empirical evidence of the success of his methodology. There is no need to go hunting for explanations for a problem which does not exist.

 

[Lysander]

Regardless, the key question anybody who doubts null tests must answer is: if null tests are so useless, why did they work so well in his Kemper V2 tests? His results align very well with people's subjective impressions about the profiling accuracy.

 

[Sascha Franck]

im not spamming videos,im just trying to open your ears

LA Lead tones haven't been an issue for digital stuff anymore since a decade or so.

 

[Sedaxel]

Did Kemper added capturing of compressors in V2? How about the "feeling thing"?

Just asking because it seems that some smell of "Kemper is now on par to QC" is floating over here, and it seems it´s based on null tests again...



QC/NC users, me included, raved about the V2 update NOT BECAUSE of a little dBs improvement in f-----g null tests. Anyone suggesting that is just spreading BS. We raved because We compared both versions and saw a real improvement, and We saw compressors brilliantly captured too, which wasn´t been done with this tech before.

Oh, and by the way, I´m one of those that think that a 10 year old Zoom would perfectly do the trick live (or even recorded). And I also think that in the forum We are veeeeeeery picky about tone, and discuss differences that no listener out of here would care about. But... that´s what this forum is about. So, even when "corazón Espinado", or any other video clip anyone wants to post here, can sound perfectly fine with a Kemper, as it would with a Zoom G1 Four, We are not talking about that here.

What We are talking about is that Kemper grabbed their very same old hardware, made a revision of its 15 year old profiling tech which seems to be a little more accurate, more easy to use, and all seems to indicate that it´s now a slightly better product. And We´re also talking about that it´s more than probable that it could be done with MKI machines and they decided to play a "not good looking game" by saying MKI units can´t benefit from the updated code. And We´re also trying to quantify the improvement, and so far it doesn´t seem they´ve reached none of the competitors (QC, ToneX, NAM). We still need more trusted opinions, tests and shootouts to get a reliable veredict... but no, so far it doesn´t look like Kemper has made anything groundbreaking at all, which is not surprising, since they use ancient hardware.

 

[Ed DeGenaro]

can you please stop spamming videos now?

Why? His videos are the most enjoyable time pass in this thread.

 

[Orvillain]

It's only a momentary LUFS that would require time-invariance. As long as it is integrated, discrepancies in a time-variant system are absolutely captured in the residual. Yes, there are cases where audibility differs for identical LUFS (and vice versa), most notably when there is poor phase alignment, but Leo Gibson has discussed in the past how he takes measures to ensure alignment is correct.

Again, as I mentioned above, there is empirical evidence of the success of his methodology. There is no need to go hunting for explanations for a problem which does not exist.

But that still only measures energy, not perceptual significance.

Residual A

• broadband fizz
• masked by distortion

Residual B

• narrow mid spike at 1.5 kHz

Both could integrate to −30 LUFS, but B will sound far more obvious.

 

[MirrorProfiles]

Our ears are pretty terrible at perceiving phase differences, but those will affect a null test quite drastically. Shove some all pass filters over a guitar tone and do a null test if you want to try for yourself. You won't hear a difference but they won't null

The only reason one would persist with doing null tests and using LUFS here is through a lack of understanding the massive limitations of what they can show.

And of course, it totally depends on what the input signal is doing too.

It really doesn't matter how many times each pitfall of this method is explained, the people who believe in what they show will continue to believe it and post more videos doing null tests.

 

[MirrorProfiles]

Here's an example. 3 files.

2 are identical, one has a single all pass filter on it. The 2 identical files null to minus infinity, as expected. Here are the waveform stats for the phase rotated one:

That is a LOT of residual in the null. Now listen to all 3 files:



Is it possible to hear which one has the phase rotation?

Now imagine that there isn't just one phase difference, but several. How do you quantify what is a meaningful difference in how it sounds based on how the numbers look? and we are only talking about phase here. There is also distortion characteristics (which are non linear), as well as all kinds of other factors that will skew things. It's so null and void, which is why its so painful that people not only persist with these tests, but they fail to understand the drawbacks.

 

[Orvillain]

@MirrorProfiles a perfect demonstration of one of the principles at play here, well done!

 

[Richard_G]

Another thing that irritates me is that the "don’t get hung up on accuracy"-crowd also tends to throw everyone who cares about accuracy into the same pool as the "make null tests and then rank different capturing tech by the residuals measured by integrated LUFS"-crowd.

 

[nicolasrivera]

Lets have some fun

Just to put some perspective and data around this .... Leo always uses his Soldano for Null Tests so we have consistency at least at the stimulation point for Capturing / Profiling.

Now lets assume that Null Tests tell us everything we need to know about the quality of relative Profiling methods ... which of course is total bullshit ... but regardless, lets play a round of digital amp tech sites newest favorite Gameshow "Null Tests Give Me The Horn" :)

From Leo's own video's and test result:-

=> V2 Mk2 Kemper is -33.8 LUFS

=> Q/Cortex V2 is -34.6 LUFS

=> Literally 0.8 LUFS / %2.3 difference.

Not even dear old Sky Daddy will hear that difference !

But of course all the spreadsheet-tone-chasers hear will have better than Jesus ears .. so that's reassuring

B.T.W ..... plenty of big names, Kenny Wayne Shepherd, Pete Thorn and numerous others are using the QC live and from various Rig Rundowns, they are all almost always using their own Amp Captures ..... KWS for example has the biggest collection of Dumbles in the world and is using his own Dumble QC Captures on his current EURO Tour ..... but w.t.f would a bluesy-tone-feel player like him know that the math-tone-chasers here know better.

Can me crazy, but these "results" make the KPA V2 as good as the QC V2 .. if you worship at the Alter of Null Tests

Not bad given everyone was "blowing their juju" when QC V2 was released just recently.

I'll leave it there.

B.T.W ... I've had a Kemper Player for over 6 months and haven't even turned it on or taken it out of its box ..... once V2 settles and MBritt releases some packs - which I prfefer - I'll give it my personal shit testing methodology ..... how good or bad does it sound and feel and respond etc.... to me ..... you know, all that irrelevant crap that doesn't matter.

Remember ... Tone is in the Spreadsheet !

I need to get my morning coffee

Yes, but still 2.3% inferior to the QC.

 

[BenIfin]

Question

Since posting this Izotope Link => https://www.izotope.com/en/learn/what-are-lufs#integrated I've gone back and read the whole page 3 times during the day to better understand it ..... and the more I read it ..... the clearer it is becoming to me that using the Integrated LUFS -or similar- to determine the qualitative and quantitative differences between 2 x Guitar Amp Capture Results -vs- a real amp .. and saying one is "better" / "more accurate", is, if not mostly, of no sensible value or meaning (?)

Am I reading this right (?) Or am I missing something glaringly obvious (?)

 

[MirrorProfiles]

Question

Since posting this Izotope Link => https://www.izotope.com/en/learn/what-are-lufs#integrated I've gone back and read the whole page 3 times during the day to better understand it ..... and the more I read it ..... the clearer it is becoming to me that using the Integrated LUFS -or similar- to determine the qualitative and quantitative differences between 2 x Guitar Amp Capture Results -vs- a real amp .. and saying one is "better" / "more accurate", is, if not mostly, of no sensible value or meaning (?)

Am I reading this right (?) Or am I missing something glaringly obvious (?)

Genuine question - When was the first time someone told you that null testing for comparing amp modelling accuracy is bullshit?

 

[MirrorProfiles]

I put the test on YouTube too, but used a different source file and shuffled them up (so if anyone cheats by saving the previous files, they'd have to get this right too to prove me wrong):



What % difference is -infinity and -12.6 LUFS?

 

[[Nathan]]

Another thing that irritates me is that the "don’t get hung up on accuracy"-crowd also tends to throw everyone who cares about accuracy into the same pool as the "make null tests and then rank different capturing tech by the residuals measured by integrated LUFS"-crowd.

Yeah its crazy. I'm not on an accuracy quest or anything but when I had constant issues with ToneX v1 then accuracy is a problem. People had constant issues with Kemper therefore the accuracy was a problem. That whole time fanbois from each camp would just gaslight and try time wasting workarounds when the tech just had issues (the worst part is they would know the tech had issues but they just had agendas and would do these workarounds themselves.... thus further proving that the tech had issues or they wouldnt need to do these things in the first place).

To me the topic of capture accuracy isn't born out of nerd LUFS tests, it's genuine problems that are obvious to the ear. When the tech improves and it hits a point of "no BS, I can barely tell a difference between the capture and the amp + it feels great to play" then the accuracy discussion just gets boring to me.

I can hear issues with Kemper v1/v2. ToneX v1 was bad for high gain, v2 was a massive update and is now acceptable/great. NAM is great. I havent had hands on experience with QC v2 but it seems accpetable/great.

Staring at charts and constantly trying to pigeonhole these things into LUFS tests I kind of just dont believe in any of it. Show me real world A/B's its pretty much all we need to hear to derive results. Would be great if there was a series of tests you could do and have a bunch of scores but I guess that doesn't really exist in a meaningful way.

 

[texhex]

Wow, check this out.

This is from the guy that made the old Kemper Editor software (which was actually quite good) that worked over MIDI before Kemper released their own Editor in Rig Manager.

OS14 / Profiling 2.0 Public Beta Discussion - Kemper Profiler Forum

Please Discuss in here. Here's a link to the beta version: IMPORTANT: We value first-hand experiences in this thread. Please avoid posting second-hand information, rumors, or opinions based solely on what you've heard on YouTube or other…

forum.kemper-amps.com

 

[BenIfin]

Genuine question - When was the first time someone told you that null testing for comparing amp modelling accuracy is bullshit?

I was initially sucked in by Leo's tests when he started and I assumed he was right - there seemed to be a methodology at play and I had zero understanding of what LUFS or any type really were.

As time passed and I followed key people here and at TOP, it started to make less and less sense

Hence I searched for an found the Izotope link above which - whilst I don't fully understand it - seems to me, to objectively confirm that using LUFS to prove which is better is, a total waste of time and of zero value and not the right tool for the job..

Cant be more honest than this.

Was just seeking some confirmation from those here that understand this *way* better than me that my reading of the article is right.

Honestly don't know what else I can say ?

 

[MirrorProfiles]

I was initially sucked in by Leo's tests when he started and I assumed he was right - there seemed to be a methodology at play and I had zero understanding of what LUFS or any type really were.

As time passed and I followed key people here and at TOP, it started to make less and less sense

Hence I searched for an found the Izotope link above which - whilst I don't fully understand it - seems to me, to objectively confirm that using LUFS to prove which is better is, a total waste of time and of zero value and not the right tool for the job..

Cant be more honest than this.

Was just seeking some confirmation from those here that understand this *way* better than me that my reading of the article is right.

Honestly don't know what else I can say ?

The issue isn’t just using LUFS, it’s that null tests are useless besides from telling you if something is identical or not.

My example above shows files sounding identical, but which have underlying differences that result in a poor null. How much it nulls doesn’t really tell us anything useful at all, especially in this context. There is literally nothing of value, Leo continually using this methodology and (even worse) framing it as “scientific” needs to stop.

 

[2dor]

Yeah its crazy. I'm not on an accuracy quest or anything but when I had constant issues with ToneX v1 then accuracy is a problem. People had constant issues with Kemper therefore the accuracy was a problem. That whole time fanbois from each camp would just gaslight and try time wasting workarounds when the tech just had issues (the worst part is they would know the tech had issues but they just had agendas and would do these workarounds themselves.... thus further proving that the tech had issues or they wouldnt need to do these things in the first place).

To me the topic of capture accuracy isn't born out of nerd LUFS tests, it's genuine problems that are obvious to the ear. When the tech improves and it hits a point of "no BS, I can barely tell a difference between the capture and the amp + it feels great to play" then the accuracy discussion just gets boring to me.

I can hear issues with Kemper v1/v2. ToneX v1 was bad for high gain, v2 was a massive update and is now acceptable/great. NAM is great. I havent had hands on experience with QC v2 but it seems accpetable/great.

Staring at charts and constantly trying to pigeonhole these things into LUFS tests I kind of just dont believe in any of it. Show me real world A/B's its pretty much all we need to hear to derive results. Would be great if there was a series of tests you could do and have a bunch of scores but I guess that doesn't really exist in a meaningful way.

Yeah I mean that's dead on the money. If I would have had good results with QC or ToneX (like their v2 flavors offer today), I probably wouldn't have started down the "dark path" of NAM lol.

I do think that there's still something going on for the null tests when done right though but yeah - it's obvious that the community's in need of a better approach for gauging profile accuracy to the source.