Be aware of the digital sonic artifacts (aliasing) in your KPA!

    Quote from ckemper

    We have a big number of beta testers.
    The flaw that I am fixing was not even heard by Hadley ;)


    On the other hand:
    A user that opens a thread pointing out a topic, after we have been in direct contact with him and promised to work on this topic doesn't really qualify as a beta tester.

    I appreciate your efforts very much, though I don't consider myself as a beta tester, nor had you suggested in direct contact that I was one, I'm willing to help in any way I can. I don't mind using my 30 years production experience if it can benefit this excellent device.

    Quote from ckemper

    No, by physics the distortion aliasing level is a function of gain.
    Hodgsons sine sweeps were high gain only.
    I can easily create noticable aliasing on many modelers with sine sweeps.


    Some stated to hear aliasing with clean guitar sounds. Yet to be demonstrated.

    Here are sound files that that I made that demonstrate the digital artifacts in clean sounds that I mentioned earlier, and there are 2 to consider here:
    I made these files by profiling a Grace M101 mic pre DI. The first file (Grace-KPA Clean Test 1) has two groups of high descending notes; I am A/B-ing the Reference amp, and Kemper amp straight to ProTools via the KPA's analog outs. The first group of notes are the Reference amp, and are very clear, the second group are from the Kemper amp, and to me it sounds similar to what I hear in the high gain aliasing... though how could that be? Perhaps it's more of the "burst" effect than the "wind howling", you be the judge.


    Grace-KPA Clean Test 1


    The next file (Grace-KPA Clean Test 2) demonstrates what I will refer to here as 'light distortion' on clean sounds. I find this artifact undesirable, and since no one seems to be noticing it, I figured it was time to reveal--it is apparent enough that I intend to refrain from any pure-clean profiling until it's fixed. If it's related to the aliasing, I just don't know.


    Grace-KPA Clean Test 2


    No stomps were used at all.

    Quote from Jimmyno

    Well, i agree. I absolutely love how KPA sounds, and i love it even more with the 1.09, since it doesn't need to add "compression" and "clarity" in the amp section to sound less spongy and mode liquid.
    But those digital artifacts are definitely there, we all can hear them from the clips. It is honest to say that you'll NEVER hear them in a mix (studio or live).
    Of course an address of those little issues would be a plus.
    But other features like Perform mode and more FX are way more important IMHO.


    Unfortunatelyy I've been way too busy to really test the current version beyond checking aliasing for this thread with that one patch :( , but are you saying the sound has changed? Can CK confirm whether anything has changed that would make it more "liquid" (and therefore in my mind less like the original amps and more like an amp sim)? This is a little worrying.

    For the record, I don't go around looking for aliasing when I play - if it's *way in the background* it doesn't bug me too much. With the high gain Stomp ODs, I heard it immediately & obviously, so I started posting about it. I will not even have a chance to try 1.0.9 until our current tour is over (Dangit!) :pinch:


    I am very happy to see that Kemper is moving ahead on the various issues being raised, and believe the KPA is going to have it's greatest advances ahead!


    Christoph: I love the new time-saving features and gates you added in 1.0.9 :thumbup:

    All modelers known to man 8o

    I can definitely hear it on the 1st clip (KPA part) but I think most of the people will not hear it. I find that the part were it is easier to spot would be in 2nd note you play with the KPA just after the note attack.


    I also can hear that light distortion (or compression?) you mention on the 2nd clip.

    Quote from Per


    Unfortunatelyy I've been way too busy to really test the current version beyond checking aliasing for this thread with that one patch :( , but are you saying the sound has changed? Can CK confirm whether anything has changed that would make it more "liquid" (and therefore in my mind less like the original amps and more like an amp sim)? This is a little worrying.


    Don't get me wrong: i feel that the whole has become less spongy and more natural in the attack. I don't know if there's something about the new gate, about latency, about anything different in the algorithm, but i like it more than before.
    I cannot exclude that it's just a "120% more real" syndrome :thumbup:

    I can definitely hear it on the 1st clip (KPA part) but I think most of the people will not hear it. I find that the part were it is easier to spot would be in 2nd note you play with the KPA just after the note attack.



    MadH, I can't help it if some people can't hear it, it's the ones who can that matter most to me. I'll bet any pro-studio engineer would recognize the artifacts in my sound files.


    3 things are needed:


    1. An understanding of the noise we are examining.
    2. A good pair of monitors and/or 'phones.
    3. A keen sense of hearing.


    If these artifacts were masked enough that they were virtually unnoticeable, I could easily live with it, but that's not the case for me.
    Granted, it's even harder to hear it in clean sounds than high gain sounds, but, I wanted to provide CK and others an example.


    The noise is there--tonal purity compromised--though the degree is not especially high.

    Edited 4 times, last by miles ().

    Thanks for those! It is pretty audible in both soundfiles through active monitors.
    (Ha! No need to go play Jazz now for me! :D ).


    Seriously - it's there - and clearly audible in both files.


    I gotta say though - as clearly as those samples are - in the heat of playing - I haven't noticed these when I play mine. Must be in a different state of mind when I play... 8o

    yeah thats easy to spot..when its set out like that. and quite noticable too.
    almost seems the profiling has gotton confused, and put a clean signal via a distored gain stage or 2. - (sounds like tubes are broken and on the way out LOL)


    What guitar pickups were used for this test? - and do you have any other guitars to test with to see if you get the same level..
    as it might just be default gain stage set a bit HOT in the KPA (clutching at straws here)


    i just tried this, and dont get that problem, - or maybe its to do with SPDIF traffic? bad cable? interference.. anything.. again not based on fact just trying to eliminate everything possible so we can work towards an exact problem..


    Is this test you done noticble while playing pre-record? - or record only?


    Are your clean and distortion sensi at low levels? - and did you adjust these and did it make any differnce?


    Do you get the same results on every pickup position?


    have you tried a differnt guitar cable? and/or SPDIF cable?


    if you turn the volume down to half on your guitar does it still appear?


    Just tryng to figure out more.. so excuse the random questions...


    A

    Quote from miles


    MadH,
    I can't help it if some people can't hear it, it's the ones who can that matter most to me. I'll bet any pro-studio engineer would recognize the artifacts in my sound files.


    Sure, I did not mean it that way. It does not matter how many people notice it in order to take action to fix it. I was just pointing out that I thought that a lot of people would not be able to notice it (unless they used proper equipment like you suggested). My first audition of the 1st clip was using computer speakers and even if I noticed something off I could not hear clearly what it was. When I plugged in my studio headphones it was more obvious.


    How does your guitar sound when played direct without the KPA? (Also does changing the clean sens etc setting or rolling down the volume on your guitar pot help or change things at all?).


    On the second one it sounds to me more like soft limit than aliasing.


    On the first though sounds more like what's now being described as aliasing, which is a compresison artifact caused by too quick compression attack. Is there any compression in the first rig, or is it just the amp itself?

    Quote from ckemper

    Today, 6:17am
    We have actually found a flaw in 1.09 that causes loud aliasing in the amp at high gain settings. We will have an alternative 1.09 soon.
    Thanks!
    CK


    Below is a quote from the "FW. 1.0.9" thread, which states this newer 1.0.9 version addresses a "stability issue".
    So, I am guessing there will be yet another 1.0.9 version, that CK refers to in his quote above. Yes/No?


    Quote from G String

    Today, 12:33pm
    We've released a new public beta for version 1.0.9 just now.
    The new version is called 1.0.9.7426 - It is identical to version 7415 but addresses a stability issue introduced in the first release of 1.0.9


    have a good weekend, gs

    Quote from Per


    How does your guitar sound when played direct without the KPA? (Also does changing the clean sens etc setting or rolling down the volume on your guitar pot help or change things at all?).


    On the second one it sounds to me more like soft limit than aliasing.


    On the first though sounds more like what's now being described as aliasing, which is a compresison artifact caused by too quick compression attack. Is there any compression in the first rig, or is it just the amp itself?


    The guitar sounds perfectly clean when played direct without the KPA. Changing the clean sens or guitar volume doesn't make the noise go away, but turning down either one makes the noise relatively less noticeable until you crank the monitors and hear that it's really still there.


    No compression, but the limiting I added makes it a bit less painful to listen to loud--good for applying the aural microscope. ;)

    The guitar in the sound file is a Strat with a Duncan Custom bridge pup. I also tested with another Strat with Fralin Split Blade Blues pups and though the noise is less, it's due to the lower output of the single coil, which doesn't solve the problem of wanting to use humbuckers, really clean. I have no doubt that all my guitars and basses will have the same end result.


    The noise is nowhere in my recording chain.


    Clean and dist sens are both low, but increasing them does seem to bring the noise forward relative to the amount on gain.


    Cables not an issue.


    One more thing I noticed, when I had the Grace sitting idle, and the KPA in profiling mode, after matching levels between the Grace and KPA, I would A/B back and forth and I could hear the increase of the noise floor of the Kemper amp vs. the Reference amp, but that's to be expected of any digital or analog gear.

    Quote from MadH

    Sure, I did not mean it that way. It does not matter how many people notice it in order to take action to fix it. I was just pointing out that I thought that a lot of people would not be able to notice it (unless they used proper equipment like you suggested). My first audition of the 1st clip was using computer speakers and even if I noticed something off I could not hear clearly what it was. When I plugged in my studio headphones it was more obvious.

    Hey MadH, no sweat, I wasn't referring to you, but rather to the doubters. Peace. :thumbup:


    Once saved the profile have less noise floor than in profiling mode..but still it's a bit higher than the original..I said that from the beginning and even posted a clip :)

    Quote from Jimmyno

    Don't get me wrong: i feel that the whole has become less spongy and more natural in the attack. I don't know if there's something about the new gate, about latency, about anything different in the algorithm, but i like it more than before.
    I cannot exclude that it's just a "120% more real" syndrome :thumbup:


    Thanks !
    But we have not applied any changes to the core sound engine since the release of the profiler.
    Must be your imagination. But it's a good imagination!


    In three month or so somebody will say 'I have gone back to 1.09 because the new firmware sounds too fizzy. The old one is much fatter'. ;)


    In out sister company Access Music (Virus Synthesizer) we have a 15 years history of compatible sounds. You can load a sound from 1997 into the latest Virus and it'll sound the same.


    We have no plans or ideas to improve the core sound engine of the profiler in terms of getting closer to something.
    If we ever applied changes to that in the future, we will still make existing profiles and rigs sound the same.


    My plan is to let your profiles and rigs sound the same in 15 years, as you have created it today. No matter what hardware, software or firmware you run it on. Of course there will be improvements and paradigm shifts. But no technical innovation can algorithmically improve the sounds that you have created by 20% or whatsoever.
    If there is an innovation, you will have to add it to your sound by hand and in purpose.


    Of course, if there is noise etc. to reduce, we will try our best. But I don't consider this to be part of the sound.


    CK

    Edited 2 times, last by ckemper ().