Thread: What does "best sampling rate" sound like?

Results 1 to 20 of 31

  1. #1
    In a world of ones and zeros...this one's the latter. Breast reduction surgery candidate
    Join Date Sep 2008
    Location Sunny San Juan, P.R.
    Posts 562
    Rep Power 469762112

    Zero Amperes What does "best sampling rate" sound like?

    Hi, before you finish rolling your collective eyes this is not the original best post, this is a tribute to the original best post:

    http://thewombforums.com/showthread....ht=compression

    What I would like is something like that.
    What I would like is some pointers into the matter of determining just how can I take a given piece of AD/DA converter hardware and determine the optimal sampling rate. "Optimal" we can roughly translate as "sounds best" for the discussion. The implied idea is that for the given hardware 44.1KHz just "sounds better" or "takes you to a superior final mix" than 88.2KHz for example.

    But what kind of things would one look for (listen for) when trying to determine this? Keep in mind that, A/Bing is kind of awkward and/or impossible in the sense that switching the sampling rate in your Session and/or in the software driver is not as instantaneous as muting so... Also keep in mind that it'd be more productive to forget about the other reasons why one might opt for a given sampling rate (storage, cpu-hogging, final media format etc.)... forget that, let's just talk about best sounding.

    I have memories of peoples' preferences around these parts. Your personal preference is great, thanks, but "I always go for 96KHz" is not so helpful.

    What would be great for us the understudies is something along the lines of "I just noticed that my converter at 44.1KHz was muddier in the low-mids" (for example) "and you can really tell by playing Moombathon Cowbell samples at 240BPM ... notice how the cowbell lacks the sparkle that can be heard at 96K, well that's why"

    I'm gearing up to start recording and mixing again and this matter is something that has been gnawing at the back of my mind all this time.

    Tangential fun opportunity for tangential contributions by you (yes you!) here too: IIRC consensus was that 192KHz was borderline terrible compared to a solid 96KHz performer... all these years later, is this still the case?

    Thanks in advance!
    -M
  2. #2
    Join Date Nov 2006
    Location Long Island, NY, USA
    Posts 11,623
    Rep Power 536871428

    Default Re: What does "best sampling rate" sound like?

    Your computer has to be able to run a higher sample rate w/o crashing, for starters, if that's what you want to do.

    But I'd noiticed that with better converters it's like I can hear things I couldn't hear before even at the same sampling rate as the cheaper converters. Like I'll actually be able to hear an instrument I didn't even know was on the record.

    Regardless of what the insulation pimp says
    I'll leave any discussion of 'clock' and 'jitter' to people who know and care more about such things than I

    I have my DAW set to default 96 kHz for new projects, figuring plugins work better at that rate but most of what I'm working on are older projects which are 24 bit 44.1 kHz.

    I don't spend a lot of time thinking about it, tbh.
    Man! You have GOT to try a hit of this RANGE SUNSHINE!

    IMTBO = In My Thoroughly Biased Opinion
    CMIIW = Correct Me If I'm Wrong
    Never underestimate the amount of contempt a failed musician has for those of us who are still trying.
    If the party's good enough, you can actually suck to a remarkable degree.

    Greedle
  3. #3
    In a world of ones and zeros...this one's the latter. Breast reduction surgery candidate
    Join Date Sep 2008
    Location Sunny San Juan, P.R.
    Posts 562
    Rep Power 469762112

    Default Re: What does "best sampling rate" sound like?

    Thanks for your contribution Nobby,
    but to be honest
    everyone knows orange sunshine is best sampled at 384KHz so...

    Hey did you know Otek posted in the record-a-band-live-help thread!!?
    He's back!!

    AAaaanyway, I cognitively lean towards you know the fundamentals so to speak, in this regards: yes the talent on the mic must be on-point, yes your monitoring environment has to be on-point, the mic+preamp-channel has to be on-point
    But my heart just wants to say, all things being equal, I'd rather record in the given hardware's best performing setting...

    And I remember some prominent members of this community having strong preferences in this regard ... and they had solid reasons...

    I wanna know I wanna know

    Someone once in here mentioned how while working in the Graphic Design industry, as a greenhorn he'd look at a poster and he'd think it looked great but the more veteran designers later showed him how to examine or pinpoint details which made them conclude that the same poster was not so good... I have a hunch that sampling rates on converters would be something like that.
  4. #4
    Join Date Nov 2006
    Posts 5,361
    Rep Power 536871219

    Default Re: What does "best sampling rate" sound like?

    I would go for a highest SR your aggregate system (AD/DA+Computer+DAW+Plugins) allows without glitches.
    When in doubt, mumble!

    EVERYTHING SOUNDS LIKE SHIT IF YA LISTEN LONG AND HARD ENOUGH.
  5. #5
    Plays in Winger cover band Born To Be Mild!
    Join Date Sep 2007
    Location Toronto, cananda
    Posts 722
    Rep Power 536871055

    Default Re: What does "best sampling rate" sound like?

    As long as you have the storage for it, why not?

    I always just tend to record at 44.1/48. Limited with adat in that way. I might try printing things at a higher sample rate, tho.

    As far as what it looks like (waveform), I don't think it really matters. Once recorded bass and the waveform had all these strange downward spikes all over it. Wasn't jitter, it was some weird phase distortion thing, but ended up sounding great.
  6. #6
    In a world of ones and zeros...this one's the latter. Breast reduction surgery candidate
    Join Date Sep 2008
    Location Sunny San Juan, P.R.
    Posts 562
    Rep Power 469762112

    Default Re: What does "best sampling rate" sound like?

    I would go for a highest SR your aggregate system (AD/DA+Computer+DAW+Plugins) allows without glitches.
    But I remember Bob for example, mentioning that more than once he'd find the sound of a lower sampling rate more pleasant than the higher one. That the hardware just performed better that way. So the takeaway back then was that some best-seller converter tended to have a sweetspot in the 48KHz and another one had it at 96KHz instead, for example.

    And at least back then there was consensus that 192 KHz was useless.

    But I don't remember asking, how to notice? which is this thread.
  7. #7
    In a world of ones and zeros...this one's the latter. Breast reduction surgery candidate
    Join Date Sep 2008
    Location Sunny San Juan, P.R.
    Posts 562
    Rep Power 469762112

    Default Re: What does "best sampling rate" sound like?

    As long as you have the storage for it, why not?
    Because another lower setting might yield better sound; it's the gist of all this. But nothing against higher SR, just that I wish to understand better the ways in which one can say a higher or lower SR "sounds better".

    I always just tend to record at 44.1/48. Limited with adat in that way. I might try printing things at a higher sample rate, tho.
    I mostly have used 88.2 KHz because I figure there should be no calculation rounding errors implied in downsampling to 44.1 for the output, since, IN THEORY an efficient programmer would use a simple bitshift "<<" "division by 2" operator and just not introduce anything that wasn't there (although, c'mon we're spitting hairs about stuff that in theory happens at super close to the bottom of the db range, but why give up on it if you don't have to).

    The thing is, that's all a brainy/math driven decision, not "ears" driven, which is what I'd like to do now.

    As far as what it looks like (waveform), I don't think it really matters. Once recorded bass and the waveform had all these strange downward spikes all over it. Wasn't jitter, it was some weird phase distortion thing, but ended up sounding great.
    Right on, totally agree. The visual thing was just an example of how someone mentioning a delicate detail can turn your focus to it and open the door to a whole new perception. That idea isn't about waveform drawing at all, it was more of a comparison of audio engineering with graphical design, in the observations department.

    keep em' coming
    this is interesting
    thanks in advance
    -M
  8. #8
    Join Date Nov 2006
    Location Long Island, NY, USA
    Posts 11,623
    Rep Power 536871428

    Default Re: What does "best sampling rate" sound like?

    Because another lower setting might yield better sound; it's the gist of all this. But nothing against higher SR, just that I wish to understand better the ways in which one can say a higher or lower SR "sounds better".

    That's a tough one.

    It's like, what is the best size stick for beating a dead horse?

    The larger stick delivers more energy per blow, but you can deliver more blows with the lighter stick before your arm gets tired.

    And in neither case does the horse become substantially deader.
    Man! You have GOT to try a hit of this RANGE SUNSHINE!

    IMTBO = In My Thoroughly Biased Opinion
    CMIIW = Correct Me If I'm Wrong
    Never underestimate the amount of contempt a failed musician has for those of us who are still trying.
    If the party's good enough, you can actually suck to a remarkable degree.

    Greedle
  9. #9
    Has Many Personal Intergritys Expensive Boat Anchor
    Join Date Nov 2006
    Location New York
    Posts 10,666
    Rep Power 536871395

    Default Re: What does "best sampling rate" sound like?

    I'm going to go out on a limb and say that there isn't a comvertor or situation anywhere in the world in which 88.2 sounds better than 96.

    88.2 has no real world reason to exist. It's voodoo nonsense.


    Back in the real world, the better sampling rate will sound deeper and clearer and 'harder' with more defined edges and 3D solidity around sounds. You should hear the space around elements and not feel like the sound stage is flattened.


    I work at 96k unless the session is so huge that the computer can't deal, or unless it's a video project. In those cases I do 48k.

    There is no reason to record a multitrack at 44 or 88. Ever.
    Last edited by weedywet; February 19th, 2017 at 06:31 PM.
  10. #10
    Join Date Sep 2009
    Location Quake City
    Posts 11,777
    Rep Power 536871398

    Default Re: What does "best sampling rate" sound like?

    As long as you have the storage for it, why not?

    I always just tend to record at 44.1/48. Limited with adat in that way. I might try printing things at a higher sample rate, tho.
    Not much point to that - once it's gone it's gone.
    http://www.johnnyoklahoma.com/

    Originally Posted by Bob Ohlsson
    Everything is some mixture of awesome and suck. We simply want the awesome to be highlighted sufficiently that it distracts listeners from the suck.
    Originally Posted by Bob Ohlsson
    The appropriate role for science is the study of observed phenomena to gain an understanding. It is not dictating what people ought or ought not to be observing.
    Hey, if I'm Grumpy, where the hell is Snow White????
  11. #11
    Join Date Sep 2009
    Location Quake City
    Posts 11,777
    Rep Power 536871398

    Default Re: What does "best sampling rate" sound like?

    But I remember Bob for example, mentioning that more than once he'd find the sound of a lower sampling rate more pleasant than the higher one. That the hardware just performed better that way. So the takeaway back then was that some best-seller converter tended to have a sweetspot in the 48KHz and another one had it at 96KHz instead, for example.

    And at least back then there was consensus that 192 KHz was useless.

    But I don't remember asking, how to notice? which is this thread.
    Converter technology has improved.
    http://www.johnnyoklahoma.com/

    Originally Posted by Bob Ohlsson
    Everything is some mixture of awesome and suck. We simply want the awesome to be highlighted sufficiently that it distracts listeners from the suck.
    Originally Posted by Bob Ohlsson
    The appropriate role for science is the study of observed phenomena to gain an understanding. It is not dictating what people ought or ought not to be observing.
    Hey, if I'm Grumpy, where the hell is Snow White????
  12. #12
    Join Date Nov 2006
    Posts 5,361
    Rep Power 536871219

    Default Re: What does "best sampling rate" sound like?

    But I remember Bob for example, mentioning that more than once he'd find the sound of a lower sampling rate more pleasant than the higher one. That the hardware just performed better that way. So the takeaway back then was that some best-seller converter tended to have a sweetspot in the 48KHz and another one had it at 96KHz instead, for example.
    You need to have Bob'd ears to identify a sweetspot reliably =)
    To me the key differences are plugin performance and AD. Plugins will almost always sound better at higher rate (the exception would be cases when there's a something broken in the code so it only works correctly at 44.1 or 48). ADs usually have much easier job to do recoding at higher rates. Remember that virtually all ADs on the marker now are delta-sigma type so that the front end always works at something like x128 rate anyway.
    As far as DA on my converters it's really tough to say and almost impossible to test. Though it is better to listen at the native rate of the material, or to upsample/downsample using good SRC algorithm. The "on the fly" resampler at least on some drivers is not always perfect.
    And at least back then there was consensus that 192 KHz was useless.
    There probably is an optimum sample rate for hardware. Usually specs gets worse at 192. But I would love to try 192 for mixing in the box, although space and CPU requirements grow too much.

    But I don't remember asking, how to notice? which is this thread.
    Can tell you only about mixing in the box. It's just easier. Things tend to blend better. Also compressors tend to dull the attack less.
    When in doubt, mumble!

    EVERYTHING SOUNDS LIKE SHIT IF YA LISTEN LONG AND HARD ENOUGH.
  13. #13
    Plays in Winger cover band Born To Be Mild!
    Join Date Sep 2007
    Location Toronto, cananda
    Posts 722
    Rep Power 536871055

    Default Re: What does "best sampling rate" sound like?

    Not much point to that - once it's gone it's gone.
    Even if you're mixing down using hardware? Ive only tried a handful of times, but it sounded good to begin with. Maybe I'm fooling myself.
  14. #14
    Join Date Nov 2006
    Posts 5,361
    Rep Power 536871219

    Default Re: What does "best sampling rate" sound like?

    I mostly have used 88.2 KHz because I figure there should be no calculation rounding errors implied in downsampling to 44.1 for the output, since, IN THEORY an efficient programmer would use a simple bitshift "<<" "division by 2" operator and just not introduce anything that wasn't there (although, c'mon we're spitting hairs about stuff that in theory happens at super close to the bottom of the db range, but why give up on it if you don't have to).
    Xcept this is a wrong way to downsample. Irrespectively of what is the ration between your initial fs and your final fs, you HAVE to filter out EVERYTHING above half of your final fs. Just make a sketch of high frequency sine wave, then drop half of the points, then see what you got.

    PS this filter process would dominate the processor load, so in reality it doesn't matter much as to what is the ratio between rates.
    When in doubt, mumble!

    EVERYTHING SOUNDS LIKE SHIT IF YA LISTEN LONG AND HARD ENOUGH.
  15. #15
    Join Date Oct 2011
    Location Baltimore, MD
    Posts 800
    Rep Power 184549530

    Default Re: What does "best sampling rate" sound like?


    There is no reason to record a multitrack at 44 or 88. Ever.
    Really? I will have to try another session at 96. I usually record at 44 because it's just easier on my computer, though it's been a good while since I've tried anything else. I did however used to do mostly 'lo-fi' stuff, and have - over the last few years - tried to upgrade my gear with the desire to go for something more 'hi-fi', so maybe I would hear the difference more now.

    I will admit that I am a little skeptical that 44.1 should be avoided, but I will certainly revisit that with an open mind if you really think the difference is that significant.
  16. #16
    Join Date Sep 2009
    Location Quake City
    Posts 11,777
    Rep Power 536871398

    Default Re: What does "best sampling rate" sound like?

    Even if you're mixing down using hardware? Ive only tried a handful of times, but it sounded good to begin with. Maybe I'm fooling myself.
    Using a lower sampling rate imposes a brickwall lowpass filter at a lower frequency that using a higher sampling rate. Once you do that the higher frequencies that were filtered out are GONE. You can't get them back. Resampling to s higher rate is just fooling yourself - you still are stuck with the sonics of the lower sampling rate.

    It's kind of like converting an MP3 back to WAV - it still has the sound quality of the MP3 file because what has been discarded by the encoding process is GONE.

    Using hardware has nothing to do with it. It's not magic.
    http://www.johnnyoklahoma.com/

    Originally Posted by Bob Ohlsson
    Everything is some mixture of awesome and suck. We simply want the awesome to be highlighted sufficiently that it distracts listeners from the suck.
    Originally Posted by Bob Ohlsson
    The appropriate role for science is the study of observed phenomena to gain an understanding. It is not dictating what people ought or ought not to be observing.
    Hey, if I'm Grumpy, where the hell is Snow White????
  17. #17
    Has Many Personal Intergritys Expensive Boat Anchor
    Join Date Nov 2006
    Location New York
    Posts 10,666
    Rep Power 536871395

    Default Re: What does "best sampling rate" sound like?

    Even a jump to 48k will sound much better in my opinion.

    and if your computer can handle the processing, I find most plug ins sound a lot better at 96k


    Really? I will have to try another session at 96. I usually record at 44 because it's just easier on my computer, though it's been a good while since I've tried anything else. I did however used to do mostly 'lo-fi' stuff, and have - over the last few years - tried to upgrade my gear with the desire to go for something more 'hi-fi', so maybe I would hear the difference more now.

    I will admit that I am a little skeptical that 44.1 should be avoided, but I will certainly revisit that with an open mind if you really think the difference is that significant.
  18. #18
    Join Date Oct 2011
    Location Baltimore, MD
    Posts 800
    Rep Power 184549530

    Default Re: What does "best sampling rate" sound like?

    Even a jump to 48k will sound much better in my opinion.

    and if your computer can handle the processing, I find most plug ins sound a lot better at 96k
    Right now I'm using a mac mini, before that a G5. I don't actually think I've tried to record anything at 96 with this computer; I think the reason I settled on 44.1 was because the sessions got huge, and seemed to run a little slower overall.

    I also don't like to use a lot of plugins, but I do use a few. Though if that was the only deciding factor, I think I'd stop using plugins before I bogged down my sessions if my computer couldn't keep up.

    My main concern is the quality of the capture, and whether or not the difference is worth the inconvenience. It seems like you're saying it is, so I will keep that in mind and see if I can convince myself of that as well. Thanks for the insight.
  19. #19
    Join Date Sep 2009
    Location Quake City
    Posts 11,777
    Rep Power 536871398

    Default Re: What does "best sampling rate" sound like?

    Right now I'm using a mac mini, before that a G5. I don't actually think I've tried to record anything at 96 with this computer; I think the reason I settled on 44.1 was because the sessions got huge, and seemed to run a little slower overall.

    I also don't like to use a lot of plugins, but I do use a few. Though if that was the only deciding factor, I think I'd stop using plugins before I bogged down my sessions if my computer couldn't keep up.

    My main concern is the quality of the capture, and whether or not the difference is worth the inconvenience. It seems like you're saying it is, so I will keep that in mind and see if I can convince myself of that as well. Thanks for the insight.

    Why do your sessions get so huge?
    http://www.johnnyoklahoma.com/

    Originally Posted by Bob Ohlsson
    Everything is some mixture of awesome and suck. We simply want the awesome to be highlighted sufficiently that it distracts listeners from the suck.
    Originally Posted by Bob Ohlsson
    The appropriate role for science is the study of observed phenomena to gain an understanding. It is not dictating what people ought or ought not to be observing.
    Hey, if I'm Grumpy, where the hell is Snow White????
  20. #20
    In a world of ones and zeros...this one's the latter. Breast reduction surgery candidate
    Join Date Sep 2008
    Location Sunny San Juan, P.R.
    Posts 562
    Rep Power 469762112

    Default Re: What does "best sampling rate" sound like?

    Great Replies by everyone,
    Thank you so much.

    I think we can put this matter to bed now.
    Here's the lazily abstracted answer to the original question(s):

    (96 KHz vs lower ones, higher sampling rate side)
    will sound deeper and clearer and 'harder' with more defined edges and 3D solidity around sounds. You should hear the space around elements and not feel like the sound stage is flattened.

    (Processing side)
    ADs usually have much easier job to do recoding at higher rates.

    (Human factor side)
    You need to have Bob'd ears to identify a sweetspot reliably =)

    (Internets un-"wisdom" about 88.2 math advantage)
    filter process would dominate the processor load, so in reality it doesn't matter much as to what is the ratio between rates.

    (Further Reading)
    I kept digging on the Search results to find the source of my original positions and found a good discussion about all this here:
    http://thewombforums.com/showthread....=sampling+rate

    If you're short on time, just read this one reply in there by Ollhy (#10):
    http://thewombforums.com/showthread....380#post340380

    But you'd miss out on Weedy calling 88.2 dumb. Lol that actually made my day. Bahahaha and I felt dumb. But in a good way. God I love this place.

    The Search may be the best resource for someone getting into this world on the internet, period. I ... was fantasizing about turning access to the Search page, or maybe what my librarian friends called a PathFinder (curated, indexed, links page) into some sort of subscription service in order to solve the current abandonment threat... I got nowhere with it... makes me a bit sad. Whoa I drifted there.

    Anyway as Nobby said, dead horse. Pack it up for the cryo-freeze.
    Grateful,
    -M

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts