Running Kemper at 96k

  • Most serious professionals in the music industry aren’t mathematicians, don’t understand digital audio and don’t have any interest in true blind tests (why would they? They’re busy making hits and money). Marry that to the human condition of “more = better” and you’ll get a lot of engineers working at high sample rates, convinced that it just sounds better.


    For the record, I’m a professional and work with digital audio every day. My job includes production, mastering, composition, sound design and broadcasting.

    Indeed Sammy, the more=better thing's relevant here, and it starts with the companies producing the hardware. They have to devise features that distinguish their wares from the pack and sample rates are the lowest-hanging fruit, given that pre and op amps have progressed far beyond the point of diminishing returns development and ROA-wise.


    It stands to reason that the "pro's" therefore, working in swanky high-end studios that have the commensurate budgets, would expect to have such equipment at their disposal. I've heard Will say here many times that his peers use the high rates, but I'd expect no less given the factors I just outlined. If any of them reverted to lower rates he or she would incur a contraction of clientele numbers and / or a loss in hourly, chargeable revenue.

    Seriously, if you haven't already, record a project at a higher sample rate and bit depth, close your eyes and listen to it vs a 44khz 16bit project.

    Firstly, this is about rates, so I don't think it's fair to include the dropping of 8 bits, and besides, "everyone" uses 24 bit whilst recording and mixing.


    Secondly, conducting this experiment using a single model of interface isn't an apples-to-apples comparison. Every interface has a sweet spot, a rate at which it performs best. The difference in sound "quality" between a given interface's best and worst-performing sample rates is huge in relative terms and would surely mask any perceptible SR-based differences, if indeed there are any to begin with.

    Now it could be multiple other things, like the quality of my hardware converters, signal chain etc, maybe they don't work as well at lower sample rates. Or maybe they accentuate the difference at higher rates.

    Bingo.


    If it makes you feel any better, even Rupert Neve was sucked into believing the high-SR hype following a 3-rate test (44/48 - 88/96 - 192kHz) he participated in back in the '90s... using the same high-end interface.


    As for Massenberg et al, these peeps have a barrow to push, one filled with moolah. As I suggested, for the time being at least, SR rates have been the low-hanging fruit. Might as well pick it whilst it's ripe in order to gain or maintain market share.

  • The one advantage to going 96KHz or even higher that I can think of that nobody's mentioned is for the sound-design purpose of extreme downward-pitch adjustment. In theory, every time you double the sample rate you'd buy yourself an extra higher-quality lower octave. 96KHz positively affects 1 further octave, and to get to 2 you'd need to go 192kHz and 3, 384KHz and so on.


    Crazy CPU and storage demands for the sole purpose of achieving incrementally-smoother downward shifts, but in mission-critical situations where the main focus is these transpositions, I can understand why someone might switch to these rates. Unlikely that entire projects would revolve around such things 'though, so in-practice the rate might be increased in order to record a single track that requires this extreme treatment, the audio transposed and rendered and the project's rate returned to its normal settings thereafter.

    Just just saw this video by Justin Colletti (posted 2 weeks ago) and highly recommend it.


    He joked about bats and cricket mating calls, but essentially could only think of 1 incontestable reason to use higher rates, the same one I thought of.


    Skip to 27:35 if you want to hear the only situations he can think of where it might help. The rest of the video is an excellent reality check IMHO, and it's well-delivered too... by a mastering engineer.


    External Content youtu.be
    Content embedded from external sources will not be displayed without your consent.
    Through the activation of external content, you agree that personal data may be transferred to third party platforms. We have provided more information on this in our privacy policy.

  • do almost every record I make at 96k.
    so do most of the other professionals I know.
    George Massenburg argues for 192k!

    And then sell them as MP3 files or in the past burned them on CDs at 44/16. I have lots of CDs that sound amazing and I have heard 96K files that sound like crap. I'm one to believe it's more your recording techniques than sample rates. I have tons of albums that were done on ADAT and they sound just as good as anything today. IMO above 48K is just cork sniffery unless you're using it for compatibility, not sound. I guess that makes me an "amateur" then.

    Nobody on God’s green Earth is going to hear the difference between 44 &96 because our ears just aren’t that sensitive. The most common range of human hearing tops out around 20K If your hearing can't reach anything higher than 22.05kHz, then the 44.1kHz file will outresolve the range of frequencies you can hear.

  • Firstly, this is about rates, so I don't think it's fair to include the dropping of 8 bits, and besides, "everyone" uses 24 bit whilst recording and mixing.

    I don't. Another audio quality myth is that 24-bit audio will unlock some sort of audiophile nirvana because it’s that much more data-dense, but in terms of perceptual audio, any improvement will be lost on human ears. Capturing more data per sample does have benefits for dynamic range, but the benefits are pretty much exclusively in the domain of recording not human listening.

  • I don't. Another audio quality myth is that 24-bit audio will unlock some sort of audiophile nirvana because it’s that much more data-dense, but in terms of perceptual audio, any improvement will be lost on human ears. Capturing more data per sample does have benefits for dynamic range, but the benefits are pretty much exclusively in the domain of recording not human listening.

    I agree: 16-bit resolution gives 65,536 values to represent samples. 24-bit resolution gives 16,777,216 possible values. This is huge improvement and since most (probably all) plugins are using floating point arithmetic, which has limited precision - it makes computation more precise if you start doing it from more precise values. You can export final product to 16-bit and you probably would not hear much difference, but recording and processing in 24-bit depth is a must, in my opinion.


  • This gentleman is absolutely right.

    But I have to affirm his arguments even more.


    Making recordings for bats or other scientific purposes is the only reason to use higher sample rates. People doing that, instincltively care also for the frequency response of the analog recording equipment (microphones etc.).


    World class producers simply forget to inquire this aspect before setting a high sampling rate. Anyone out there?


    Ringing (or preringing of is often mentioned as an unwanted side effect of anti aliasing filters is often mentioned. But this is only observed by people watching an oscilloscope. But you cannot hear this ringing, even if it was much deeper in the audible range.


    The latter would happen if you pitch down a sample, a track by a fifth or an octave on your DAW, or your guitar by the Profilers pitch shifter.

    While the downpitching is a widely used argument for using higher sample rates, I have on the other hand never seen anyone demonstrating the shortcommings of pitching down from 44.1 kHz vs. a higher sample rate. The critical experts simply forget to do this crucial test.


    Todays AD or DA converters don't utilize your sample rate for converting, they work with sample rates in the megahertz region, no matter what sample rate you feed it. There is no way to make your audio interface sound bad on standard sample rates. The converters don't let it happen.


    Talking about experts:

    There is a Steve Vai rig rundown video out there (cannot find it back) where he shows his actual analog rig going through an Axe Fx being used for effects only. He states that he never would use modeling amps because he can perceive its latency.

    Not noticing that he is creating the same (not perceivable) latency through the Axe Fx with its amp modeling defeated. That's hilareous.



    There is one reason to use high sample rates: When plug-ins on your DAW sound better with that.

    But then you should ask yourself and the makers of those plugins, why they sound worse at standard sample rates.

    They should do their homework and use any means to make their plugins sound equally good on any sample rate.

    One easy method for them is (depending what plugin it is) to use a higher sample rate internally for the plugin. And making this even optional with a Hi-Res switch.

    This way they don't force you to eat up half of your computer and double track storage, just to make these one or two plugins sound better.

  • I don't. Another audio quality myth is that 24-bit audio will unlock some sort of audiophile nirvana because it’s that much more data-dense, but in terms of perceptual audio, any improvement will be lost on human ears. Capturing more data per sample does have benefits for dynamic range, but the benefits are pretty much exclusively in the domain of recording not human listening.

    That's what I said 'though. This is what you quoted from me:


    Firstly, this is about rates, so I don't think it's fair to include the dropping of 8 bits, and besides, "everyone" uses 24 bit whilst recording and mixing.


    The discussion is about recording and mixing rates, so I stand by what I said. At any rate(!), 24 bit is invaluable, starting with feeding one's interface with appropriate levels. One no longer needs to try to get as close to clipping as possible; the -20 -> -10dB range now serves as an ideal, wide target that inherently also protects us from unexpected overs with 10 -> 20dB headroom for such things.

  • Making recordings for bats or other scientific purposes is the only reason to use higher sample rates.


    it’s mostly the armchair and online hobbyists who love to tell each other that “it’s unnecessary” and how stupid the professionals are.

    So I guess CKemper would be considered a "armchair and online hobbyist" then like me. Good to see I'm in good company.

  • This is an interesting read. Rightly or wrongly I record at higher rates because years ago I settled on them in blind tests.

    But, not wishing to upset anyone, just as an analogy or think piece...

    4:3 looked great on crt interlaced tvs 20 years ago with 520lines or whatever. 4k now looks infinitely superior. But 35mm film before video captured things brilliantly, there was a cross over where only in recent decades digital video has raced ahead.


    The beatles retained their 4 track pre bounced tapes which allowed them to remix and preserve their work 40 years later on multitrack, what foresight. Yet many would say the final 4 track or stereo (or mono) was the definitive article for many years. Point being that while those tapes may have seemed to be without purpose at the time the potential in them was released later on.

    Many bands have gone back and remastered those early CDs to unlock the improvements of the current gen, even in the early days aad seemed better to my ears than add.


    Who's to say there won't be some incredible leap in Dolby atmos type sound design that finally releases this, and yes I've been listening, I know our hearing range etc has limits. It could be that speaker cones will react more favourably to higher rates, resonant frequencies might be used to create movement imperceptible to ears but make your hairs on your arms sway to the music.

    As someone who now suffers with permanent tinnitus it's ridiculous for me to say, but maybe we will even evolve to hear higher frequencies in the next millennia. Haha

  • He’s a great designer. Not a record producer.


    I’m happy to be in George Massenburg’s ‘company’


    by all means, do what works for you and if you do t hear a difference then you don’t.
    but no one should dismiss the people who do.


    meanwhile, every record I make for major labels is being archived at mastering as 24 bit 96k files no matter the source.
    Sterling saves everything at 96k

  • There’s a whole gearslutz and YouTube cottage industry out there in “debunking” what ‘those stupid elitist professionals in audio’ do or say.


    Talk about a money grab


    that’s who has a vested interest in this


    no one pays me, or George, because of my opinions in sample rates.
    When I walk into a studio I set the sample rate with my assistant engineer.
    the client doesn’t even know.


    Believe me, George has nothing to ‘gain’ from his position. It’s just what he hears and thinks.

  • It could be that speaker cones will react more favourably to higher rates, resonant frequencies might be used to create movement imperceptible to ears but make your hairs on your arms sway to the music.

    In fact, the ultra-high frequencies reproduced through recording at 96kHz and beyond cause distortions audible in the normal frequency bandwidth that we hear, by speaker drivers. Today's tech can't even emit them accurately as SPL's and not by a long shot.


    If the delivery files are downsampled to 44/48kHz, there's no problem, but so-called hi-def offerings present playback-distortion challenges that can only be prevented by modern monitors' low-pass filtering in order to restrict the upper frequency response to around the low-20kHz's, thereby preventing intermodulation distortion. This is what those high-end monitors do, by the way.


    Oh, and another very-good reason for this low-pass filtering is the resonances exhibited by most hard-domed tweeters between 20 and 30kHz.


    In a nutshell, you don't want to feed an amp and speakers with signals they cannot reproduce; such scenarios always result in compromised playback integrity / clarity / accuracy.

    As someone who now suffers with permanent tinnitus it's ridiculous for me to say, but maybe we will even evolve to hear higher frequencies in the next millennia. Haha

    Not going to happen, I'm sorry to say brother.

  • meanwhile, every record I make for major labels is being archived at mastering as 24 bit 96k files no matter the source.
    Sterling saves everything at 96k

    In my scenario, I couldn't perceive a significant difference between tracking at 48k or 96k in the final mp3 files. Consequently, I didn't feel the bang for the buck was there in terms of the cost in processing power and disk usage, so I went with 48k.


    I understand your perspective that higher sample rates are better than lower ones. With that in mind, I have to wonder why you're working at 96k and not 192k.

    Kemper remote -> Powered toaster -> Yamaha DXR-10

  • The scariest part of this discussion is the fact that I've spent all evening on .gearapace and youtube watching videos, reading discussions and now have probably two days of recording and a-b ing projects ahead of me. On the video from Fabfilter, who's mastering suite and delay I love, I thought there was a very obvious difference between the 44, 48, 96 with and without oversampling. Rightly or wrongly I felt there was some air or something I found pleasing at 96, but the music I do would work well with the 48 and plugins oversampling, I like that sort of mix.

    Considering my hearing is completely shot and I can barely hear anything above 7khz I can't get my head around the roundness and space I perceive above 48khz. Obviously youtube isn't the best place to listen so a weekend of discovery awaits.

  • There is one reason to use high sample rates: When plug-ins on your DAW sound better with that.

    And another reason would be in situations where latency is crucial (live sound e.g. in broadcast or even concerts). You just don't want a bunch of oversampling linear phase "plugins" in your signal chain considerably adding to the overall latency (and potentially even severe phasing issues between "parallel" busses. So you rather use 96kHz (if you can) instead of 48kHz on your desk and keep most of the nasty stuff happening in low latency "plugins" well above the audible spectrum ... and have the DA stage take care of that.