To Keep or not to Keep .... * RESOLVED *

    If I'm not being trolled, or if anyone else stumbles on this thread and is curious about latency: Try using XLR cables instead of SPDIF, switch your DAW and interface to 96KHz, lower your sample rate until there are no more artifacts. This solved my laggy feeling input, which did bother me plenty.


    I had POD 1.0 and 2.0, latency was terrible, around 40+ms from input to output. At 96KHz, and a Focusrite 6i6 gen1, I get 7.90ms latency, total. If I switch to 44.1KHz, I see a value of 30+ms.


    The only thing I've ever had, even resembling a complaint about the Kemper, is no 96KHz output over SPDIF.


    The sound and feel is worth it though, once you get it worked out.

    Kemper Powerhead w/remote & Kabinet
    Focusrite 18i8 (2nd Gen) - Windows 10 - Ableton Live - Yamaha HS-8's - DT770 80 ohms

    Edited once, last by JSB ().

    Like 1


    Sonic, just to be sure we have covered all the bases for you...


    Is there any chance your KPA happens to be in an inertial frame that is moving at significant fraction of the speed of light (c), as compared to your stationary reference frame?


    This might explain the perceived latency in the signal you are receiving. If you are experiencing a latency of 5 ms, and my calculations are correct (which they are), your KPA is currently in a spaceship travelling at a constant velocity of 30,000 km/s.


    So, call the ship's captain, and tell him to slow the heck.


    :P


    EDIT 01/28/2017 -- In original post, I accidentally reversed the stationary (SonicExplorer) and moving (KPA) reference frames. Now corrected.

    Edited 3 times, last by Tritium: EDIT 01/28/2017 -- In original post, I accidentally reversed the stationary (SonicExplorer) and moving (KPA) reference frames. Now corrected. ().

    Like 9


    Because the engines kenna take any moar, cap'n! :D

    Well, I haven't resorted to derision or ad-hominem barbs yet, 'cause I'd like to get to the bottom of this. Besides, it's not my MO.


    Sonic, I'd like you to try monitoring directly and not through your interface, and if necessary to try reducing the Power Sag parameter and increasing Definition for a more-modern response.


    You obviously believe 2ms or even less is something you can feel whilst playing; if you've got the real estate to be able to move a foot or two closer, I'd also like to know how things feel after having done so.


    Failing all this and what others suggested earlier, I can't help but feel that your stratospheric requirements simply cannot be met by any modeller out there. Helix and Axe are both low-latency units, in the ballpark of the Kemper (although the Kemper is the most-immediate-feeling unit I've ever played), so even if you're able to tolerate the obvious downgrade in authenticity, especially given all those years of your tube-amp experience, you'll still be saddled with the latency issue.


    IOW, your only options are to make the Kemper work, or return to tube amps and live with the fact that your palette of tones will be highly-restricted by your budget and amount of storage space... and back.

    I very much want to make the KPA work, otherwise I would not have put so much time & effort into it.


    For those who provided constructive suggestions I very much appreciate it. To be clarify though, I wasn't searching for opinion on whether to keep the unit. I take partial responsibility for the confusion as I now reflect back on the title of the post. Notice the lack of a question mark. It was supposed to be taken as a lead-in to the real question, as posed in the initial entry in bold:


    Does anybody have any idea what the chances are of S/PDIF or latency improvements, or a KPA 2.0 on the horizon?

    Yes, Sonic, but you asked "what the chances... of latency improvements" were. You've received responses focussed on that aspect of the question 'cause folks find it hard to believe that you're having an issue with it.


    If I'm not mistaken, there may have been one shredder who claimed to have had an issue a few years ago; he received a similar response... unless I'm thinking of another forum. If indeed it was on another forum, you'd be the first to claim here that the Kemper lacks immediacy. If not, you'd be the second out of 31 260 members and tens of thousands of Kemper owners world-wide.


    Surely, given the ratio of pro and talented guitarists and engineers and producers to bedroom hacks (like me!) that the Kemper ownership pool boasts, you'd understand how quaint your declaration might seem to us. It's an object of fascination, really, which is why I want to get to the bottom of it.


    Chances of latency improvements? How low do you need it to go? If it takes at least 1.5ms right now, even a Kemper 2 ain't gonna shave 2ms off that, is it? My guess is that 1ms would be an admirable achievement, but why would the company even bother when only 1 or 2 end users have complained in 5 years? It wouldn't be a stretch to suppose that the company, if word had got to it of this apparent shortcoming, might've had a good laugh over a cup of coffee or German brew about this. I can hear the team now:
    "I know, let's install the latest-and-greatest SHARC processors, adding $500 to the RRP of the unit, in order to appease the two people who're not satisfied with our 2ms latency figure." [Cue copious sculling of more German brewskis...]


    I'm sure you see where I'm coming from. It's not that I don't believe you; I'm not egocentric enough to dismiss your claims, but this has to be viewed in the context of numbers, and unfortunately for you, the required weight of support you'd need in order to be able to prompt a reaction, design-wise, from the team, simply isn't there.


    So, it's back to edging physically-closer to your speaker cone/s, I'm afraid, and / or trying my suggestions made earlier (mentioned twice - still waiting for feedback on them).

    Mike, to try to answer the questions:


    1) Yes I do notice the issue even with S/PDIF although it is closer to the window of tolerance.
    2) Sort of. Because my interface can auto-detect the Master/Slave thing isn't necessarily an issue - for now. But it is still a limitation that baffles me and could present future issues if my configuration changed. The bigger issue at the moment is the sample rate as I have to work in 44.1K, 48K, and 96K at times. Mostly 44.1K and 48K. So I either have to go analog only, or have two inputs and remember to switch between them for any sessions not in 44.1K. Again, why am I having to do this in 2017 with a $2K device of this level of sophistication? Again, baffled.


    I am using an RME Fireface and doing direct monitoring.


    If I don't look at the fret board it helps a lot - but that's not how I record (most people for that fact). It's really the combined equation of all three things - the visual, the feel and the sound that just don't mesh together for me with the latency. Keep in mind, it is not fair to criticize somebody being bothered by latency, even if by 1ms difference (or even 0.5ms). There is a point where it can become bothersome. It isn't necessarily linear, there may simply be a threshold point.


    MonkeyMan, I did try various settings like definition and clarity, etc. They do help a bit but are not a solution, and it can really start to impact the tone.


    The latency deltas, while a bit variable on the KPA side appeared like the following, and please remember I'm an analog guy so I may not have this perfect - but I was able to reproduce this enough times to believe it to be accurate:


    Relative to the direct signal:
    KPA Analog out: 3.0 - 3.25 ms
    KPA S/PDIF out 1.75 - 2.0 ms
    POD Analog out 1.25 ms


    Keep in mind this is relative to a split direct signal going into all devices for the test. So another A/D conversion needs to be added in for total latency. So add another 0.5 or so to the above. Which brings even the KPA S/PDIF to the range of 2.25 - 2.50 ms. And lastly, on a total relative basis another 0.5 or so would have to be added in for the final D/A to my ears.


    I'm an engineer so when I encounter problems I like to understand them and try to fix them, not just gloss over them. There's nothing more behind things. Again, I really want to make the KPA work for me.


    I could probably try some tricks like using KPA direct out back into the POD and mixing some of the KPA signal with the POD to mask the latency when recording, while only recording the KPA signal. That's getting ridiculous for a $2K piece of gear. Why in 2017 should I have to mix a 15 year old device (worth 10% of the KPA) to try and overcome this?


    So again, while I welcome any constructive ideas to try, I am wondering Does anybody have any idea what the chances are of S/PDIF or latency improvements, or a KPA 2.0 on the horizon?


    Sonic

    Quote from Monkey_Man

    Chances of latency improvements? How low do you need it to go? If it takes at least 1.5ms right now, even a Kemper 2 ain't gonna shave 2ms off that, is it? My guess is that 1ms would be an admirable achievement, but why would the company even bother when only 1 or 2 end users have complained in 5 years? It wouldn't be a stretch to suppose that the company, if word had got to it of this apparent shortcoming, might've had a good laugh over a cup of coffee or German brew about this. I can hear the team now:


    "I know, let's install the latest-and-greatest SHARC processors, adding $500 to the RRP of the unit, in order to appease the two people who're not satisfied with our 2ms latency figure."

    MM, we cross-posted while I was typing up a reply so responding back...


    If you want to know my view, here it is:


    A $2K device in 2017 should not have both the latency AND the S/PDIF limitations as they are inter-related and impact tone and latency. It's the combination of these "limitations" that are presenting my concern.


    Asking things another way, I'm wondering is there any reason this cannot be improved without adding much to the cost? Is it a cost issue or is it a design constraint? Are there plans to resolve any of this (via SW, FW, KPA v2.0) ? Or has this already been dismissed? Sure I could certainly ask Kemper corp but I suspect the chances of gaining insight are probably quite slim due to potential business impacts. Therefore I wondered if someone on the forum might possibly have knowledge or history they can share that may provide some insight....

    Mate, if there were any plans, only the team would know about them; they're tight-lipped when it comes to this sort of stuff.


    Personally I don't foresee much change in the latency figure in the nearish future. Even if a Kemper 2 appeared, the addition of parallel amp Profiles (likely IMHO) and / or simultaneous stomp-profile playback, and possibly even a convolution engine for IR-'verb, would all contribute to an increased CPU load commensurate with any gains CPU-speed-wise made in the interim.

    im sure whenever there will be a KPA 2,
    which s in my opinion not necessary in the next time,
    sonic, you will be the first who will find things which you don't like
    and oops , all starts again ;(

    Quote from SonicExporer

    Relative to the direct signal:
    KPA Analog out: 3.0 - 3.25 ms
    KPA S/PDIF out 1.75 - 2.0 ms
    POD Analog out 1.25 ms


    Keep in mind this is relative to a split direct signal going into all devices for the test. So another A/D conversion needs to be added in for total latency. So add another 0.5 or so to the above. Which brings even the KPA S/PDIF to the range of 2.25 - 2.50 ms. And lastly, on a total relative basis another 0.5 or so would have to be added in for the final D/A to my ears.


    First, to address your overall question:


    As MonkeyMan said: only the team knows if and when a new unit will come out. NOBODY will know until it's announced. Could be a year, could be never.


    As far as improving latency: a big fat "very unlikely" would be my answer. You compare to the POD, but do not take into account the difference in DSP processing of the units. There's a lot more going on under the hood of the Kemper than the POD. That's probably the difference in latency right there. I don't think it's fair to see it as an "issue".



    So adding those numbers up you outline, we get (max):
    KPA analogue: 4.25 ms
    KPA SPDIF: 3 ms.


    I know that there are some people that are VERY sensitive to latency, and you might be among them - I appreciate that it might be a real problem for you,
    HOWEVER...
    4.25 ms of latency equals your head being 5 feet away from a real amp or monitor due to the speed of sound.


    I am guessing you haven't experienced these problems playing with real amps.


    I think there is either some expectation bias at work here, or some psychoacoustic stuff (headphones etc).

    I don't think that the Kemper team sees the latency as an issue, so I guess in their eyes (and in the eyes of all except very few (?) ) there's nothing to be resolved.


    You might want to go back to real amps if this bothers you.

    Am only comparing to an old bean POD because, well, that's the only other digital amp sim I own (besides a J-Station). Agreed, I'm sure there is much more DSP going on in the KPA. That's one of the implied questions...is the latency a design constraint? It may well be there is little room to improve anything all, I simply have no idea. In which case the S/PDIF constraints become even more relevant IMO.


    Live is another animal altogether from recording, the two just don't compare for a litany of reasons. None of which negates or justifies the existence of recognizable latency in a recording environment when dry tacking with cans.

    Being an engineer type I just had a thought...Granted it is sort of semi-hijacking my own thread but it is directly related to the issue being discussed and rather than open a new Feature Request topic wanted to bounce it off you guys for opinion (or abuse as the case may be... :P )


    What about a degraded mode output? Specifically, what if there was a mode selectable on an existing output (for example Monitor Out) to act as a degraded mode output such that the latency is reduced? This would be a software/OS update solution. Maybe there is a place in the DSP processing logic where such a signal could be tapped and routed to the degraded output and with no impact on the overall behavior of the regular DSP processing path. A user could then connect the degraded mode output (i.e Monitor out) to their interface for direct monitoring while recording the normal output from the KPA. This might even have applications in a live environment, although such escapes me at the moment. Anyway, a user could then select to monitor the degraded mode output for those situations having latency-sensitive performances, and easily switch back to normal monitoring as desired.


    This would be a solution for anybody having a sensitivity to latency, and while not ideal, at least it is something to work with (and possibly be spun as a marketing feature as well? ) Depending on how the DSP processing / algorithms work, this might be a very simple feature to implement. If so, certainly seems a worthy feature to consider at Kemper HQ.


    Okay, flame away.... :rolleyes:

    Quote from SonicExporer

    Live is another animal altogether from recording, the two just don't compare for a litany of reasons. None of which negates or justifies the existence of recognizable latency in a recording environment when dry tacking with cans

    AC/DC did it fine with playing in a room.
    http://www.woodytone.com/wp-co…BIB_studio_rehearse_1.jpg


    In all seriousness, I think either you have to decide to accept the kemper as it is and start using it and stop worrying - or just get rid of it.

    A logical idea, Sonic, but the same issue that I pointed out earlier would apply - why would Kemper expend any time and expense coding essentially for you?


    Have a look through the history of this forum and see how many latency complaints you can find; it might help convince you of my belief that this wouldn't even be on Kemper's radar.


    I've always been extraordinarily phase (cancellation / comb-filtering), pitch and latency-sensitive, so I know all too well how it feels when it seems a whole forum or indeed "the world" can't relate to what you're hearing. For this reason, I truly sympathise with you on this. It's also for this reason that I kinda feel bad that in this instance I'm totally-happy with what I like to call the "immediacy" of the Kemper's response. It literally feels alive to me when compared to all other "modellers", and by a very-long way.


    I'm still impressed, confounded and conflicted by your not being happy with it. So sorry I can't offer a magic bullet, man.

    ... there's "Direct Monitoring" or "Zero Latency Monitoring" in an audio interface ... and then there is "Direct Monitoring" as a feature in a DAW.
    Just sayin' ... I wouldn't be surprised if he's experiencing a much higher latency while believing hard it's the calculated 2-3ms he "feels".