Higher Sample Rates really better?

Discussion in 'Recording/Live Sound' started by jjboogie, Feb 27, 2009.

  1. jjboogie

    jjboogie Member

    Messages:
    3,649
    Joined:
    Jan 25, 2008
    Location:
    Hotlanta Ga.
    I was just curious if recording starting at 96khz is really that much better in sound quality than 44.1khz........

    Would love for some real life experienced engineers to chime in!

    I was debating whether or not to from now on go ahead and eat up disc space with the hopes of having much better resolution BUT is that resolution really that much better in the final product?

    Inquiring minds want to know!!! :BOUNCE
     
  2. Somniferous

    Somniferous Member

    Messages:
    1,207
    Joined:
    Nov 1, 2008
    Location:
    Brighton, MA
    Is it better, Yes. Is it worth the extra space, No, unless you plan to release something at that resolution. I like to work with the sample rate the finished product will be in, which is 9 times out of 10 CD quality so I mainly record with 44.1kHz. That way I know for sure how it will sound once bounced.

    Bit depth makes a much bigger impact upon the quality of the sound than sample rate, and you normally want to record with a higher bit depth than the delivery medium (gives you a lower noise floor).
     
  3. chrisgraff

    chrisgraff Member

    Messages:
    2,649
    Joined:
    Oct 2, 2002
    Location:
    Nashville, TN
    for acoustic instruments and/or vocals, yes.

    For rock & roll, no...

    ...Kinda like watching the evening news on HDTV; too much information.
     
  4. Steve Dallas

    Steve Dallas Supporting Member

    Messages:
    8,267
    Joined:
    Feb 22, 2005
    Location:
    Surrounded by Idiots
    Unless you are recording something with a lot of dynamic information like an orchestra, it doesn't make that much difference. I typically record at 24/48 with great results.
     
  5. LSchefman

    LSchefman Member

    Messages:
    13,441
    Joined:
    Jan 4, 2002
    I record at 24/48 because that's the de facto for delivering broadcast spots for audio post. If I was creating CDs, I'd use 44.1.

    The sad thing is that people are moving to compressed-file based listening, instead of disc based listening. It almost makes this discussion academic.
     
  6. jjboogie

    jjboogie Member

    Messages:
    3,649
    Joined:
    Jan 25, 2008
    Location:
    Hotlanta Ga.
    I do production for both an acoustic rock band with two electric guitarist doing harmonies and female vocalist and then I also do production for a Hip hop group that has two female vocalists and two male vocalists who sing VERY WELL and rap too. We love the warmth and fatness of some of the modern "reggae" artists that have come out like Stephen Marley.....the production on that CD is SICK!!!!
     
  7. delirious131

    delirious131 Member

    Messages:
    4
    Joined:
    Nov 22, 2007
    I own a studio and can say that your sampling rate is important. If you can use the disk space (most can)... do. you will gain a some headroom that you do not notice very much but as you get experience, it will definitely be worth it. So.... I find dynamics a little better with a better sample rate. And yes you will be bringing it back down to cd quality (44.1) but in the end, you will end up with a better product. Just my .02
     
  8. jjboogie

    jjboogie Member

    Messages:
    3,649
    Joined:
    Jan 25, 2008
    Location:
    Hotlanta Ga.

    I am definitely on my next session amp up the sample rate. Should I go ahead do the highest I can possibly do? Shoot for 96khz?
     
  9. Zero Point

    Zero Point Member

    Messages:
    483
    Joined:
    Feb 21, 2007
    I'll be honest. Recording at 96 khz isn't really necessary unless you have an orchestra, strings, piano, or some other multi-range acoustic source. For most projects, 48 khz is all you'd ever need.

    Some plugins, like amp sims, sound better at 96 khz. Like Guitar Rig 3. That sounds better at higher rates because of the type os sim they use. Waves' GTR 3 sounds fine at any sample rate because they use a convolution modeling system and include the sample for each rate.

    Really though, unless you are using some awesome analog gear, like condenser mics, Neve preamps and EQ's, etc... you will not see a benefit from going the 96 khz route unless you need the extra headroom or if you use a lot of time stretching plugins and tools to alter tempo.

    I personally like 24/48 for digital recording. This gives you a bandwidth of 20 hz to 20 khz, with 144 db of dynamic headroom. Which includes 2nd and 3rd order harmonics for 'most' frequencies, and 4th order harmonics on those nice instruments like cello, bass and guitar.

    If I had the huge set of analog gear (other than the 1073 preamps I have already) I'd work in 96 khz. But just recording my own guitar and ideas, I'm fine at 48.

    -Z
     
  10. wolf9309

    wolf9309 Member

    Messages:
    407
    Joined:
    Feb 4, 2006
    Location:
    Boston
    I do mainly rock, and yes 96 khz is worth it.

    To clarify something a few people seem to be confused about, 96 k won't give you more headroom or dynamic range- that's all in in the bit depth, which does make a bigger difference than sample rate.

    The reason why sample rate makes a significant difference, even if it is eventually getting converted to 44.1 for a CD is because it does not directly cut off at 44.1 khz, or else you would have a clear picture up to 22.05 khz which would be fine- to avoid artifacts that would be caused by a sudden cutoff, they have anti-aliasing filters at the top of the range. If you're working at 44.1 khz, the filters cut off a significant amount of high end information from 18 khz and sometimes even lower. If you're sampling at a higher rate like 96, the filters can be much, much more gradual, so you lose less information.

    The hopeful idea is that in the end, you can simply go through a very high quality SRC once in mastering which will provide better conversion than recording directly to 44.1.

    For what it's worth, I can hear the difference between projects i've done at 44.1 and 96 in the final CD format- i haven't noticed as significant a difference going up to 192, which may just be my ears.
     
  11. Sunbreak Music

    Sunbreak Music Member

    Messages:
    1,086
    Joined:
    Sep 27, 2007
    Location:
    Scottsdale, AZ
    This needs to be repeated. :AOK

    My belief is that for CD, 44.1 is just fine if you're using latest generation converters. This is an ongoing debate, but there are so many variables it's next to impossible to have one side "win" or even determine a statistical outcome as to which sounds "better". I know classical engineers who say the same thing. DVD-audio is popular in that genre, so obviously you'd track at the appropriate rate for the destination format which can be as high as 192k.

    I don't do much recording these days, but I tracked for a project last week that was initially recorded at 96k, and the tracks simply didn't sound good (acoustic guitar). They were recorded with very nice gear, by the way. I couldn't live with them, so we retracked them at 44.1, and the difference was overwhelming. The basics of gain staging, signal path, mic placement and such are always much more important to the "sound"......

    In the big picture, there's probably something better to worry about. :)
     
  12. meterman

    meterman Supporting Member

    Messages:
    7,603
    Joined:
    Jun 24, 2002
    Location:
    ATL
    I've been wondering about this myself. I like the idea of capturing as much information as possible as long as disk space and processing power is not an issue. I've actually been using 88.2khz b/c I read that downsampling to 44.1 for putting on a CD is easier and less destructive, and that the difference b/t 88.2 and 96 is not very significant....I'm just starting out so I don't know for myself yet....There does seem to be some consensus that 192 is overkill....
     
  13. wolf9309

    wolf9309 Member

    Messages:
    407
    Joined:
    Feb 4, 2006
    Location:
    Boston
    This is so true. I've known so many engineers that get so caught up in the minute details that they end up missing the overall picture.
     
  14. Zero Point

    Zero Point Member

    Messages:
    483
    Joined:
    Feb 21, 2007
    It is Alchemy, not science ;)

    Although there does tend to be a whole lotta measuring going on! It's like magic with the structure of science...

    It's Hermetic Harmonies! :eek:

    -Z
     
  15. KennyM

    KennyM Supporting Member

    Messages:
    1,727
    Joined:
    Jul 24, 2003
    Location:
    Burbank, Ca.
    I disagree with anyone that says the higher sample rates are not worth the disk space or not a big enough improvement to use them.

    Anything and everything that you stick a mic in front of sounds much better at 96 or 88.1. Much better! I don't hear enough improvement to warrant working at 192K although I've heard some live 24 track orchestral recordings at 192 that were just phenomenal.

    The only thing I don't hear a real difference with the higher sampling rates is in the recording of music that uses a lot of sampled sounds such as drum machines, etc. Most of those samples are 16 bit, 48K and usually more grungy sounding anyways so not really any benefit there.

    Still, depending on your system and what plug-ins you use, many of them can process the signal at higher resolutions so it still may be of some benefit to be working at the higher rates.

    Another factor to consider is nobody really knows what's coming down the road audio standard wise. Why would you want to limit audio quality on a possible needed future high resolution mix just because it takes more storage space. Years ago I spent a grand on buying my first 1 gig hard drive used and the damn thing looked like a car battery. Now you can get a 1 Terrabyte drive for a hundred bucks. Hard drive space is about the cheapest piece of the puzzle now so might as well use it. I've never heard someone say the higher sample rates sounded worse.
     
  16. Sunbreak Music

    Sunbreak Music Member

    Messages:
    1,086
    Joined:
    Sep 27, 2007
    Location:
    Scottsdale, AZ
    I always appreciate a bold statement. :RoCkIn

    What do you attribute this to? If it sounds "much better", there must be a reason. At least one. This might get technical. ;-)
     
  17. jjboogie

    jjboogie Member

    Messages:
    3,649
    Joined:
    Jan 25, 2008
    Location:
    Hotlanta Ga.
    I think my initial question or thing I was mainly wondering about is if recording in a higher sample rate resulted in a warmer bigger better sound.

    If it was something noticeably different and if major studios do it for major artists.
     
  18. loudboy

    loudboy Member

    Messages:
    27,371
    Joined:
    Nov 3, 2003
    Location:
    Sedona, AZ
    Hopefully he can link to an ABX study...

    Because there's a pretty definitive one that proves otherwise. In April or May 2008 MIX, there was a column about a study that showed no audible differences between 44.1/16 and higher rates, on finished products.
     
  19. KennyM

    KennyM Supporting Member

    Messages:
    1,727
    Joined:
    Jul 24, 2003
    Location:
    Burbank, Ca.
    Well, from a technical standpoint since frequencies above 1/2 the sample rate are basically chopped off you basically change that cutoff from about 22K or 24K to 44K or 48K. Now before someone reminds me that none of us hear above 20K (probably even lower for us old guitar players) I would just say to google Rupert Neve and have a look at some of his tests illustrating the effects of chopping off these higher frequencies on just a simple sine wave. To summarize this test, every listener heard differences in what the audible sine wave actually sounded like even though the only frequencies being altered were much above the human hearing range. Even though we do not audibly hear these frequencies, we hear the effects these frequencies have on the lower frequencies we do hear.

    Alright, that's not really too technical of an explanation, but it is an example of technical tests that Mr. Neve conducted. Since every piece of gear that I have worked on since the 70's that Rupert has designed have been some of the best examples of state of the art audiophile engineering I've ever encountered, his opinions are good enough for me.

    If you look at it from just a resolution standpoint, you are comparing assembled audio from 48,000 snipets or miniscule views of the source happening in a second to 96,000 divisions of the source in that same second. Doesn't it stand to reason that audio just like a digital photo gains something from this extra resolution. And as for the value of this extra resolution, photos I take at higher resolutions certainly look better after I compress the living daylights out of it to email than one I took on my iphone.

    I don't know of the test you mention, but I'll dig out that issue and have a look. Sample rate aside, I can't believe that someone assembled a panel of listeners that could not hear the difference between 16 bit audio and 24 bit unless these tests were conducted at the Sir James Whitney School for the Deaf. The difference between these two bit rates is not subtle in the least and much more pronounced than comparing sample rates.

    All I can say is that it sounds better to me based on the tests I've personally conducted. I bought my first two channel digital recorder in 1991 (sadly still in my garage somewhere) and at this point have recorded more hours of digital audio than I care to even think of. I've been through 5 or 6 different systems since that time and the last upgrade (PT Mix to PTHD) was entirely based on hearing a big difference with music recorded at these higher sampling rates, even using the same higher end converters and clocks on both systems. Admittedly, I speak in generalities and personal point of view as opposed to your technical tests, but I do speak as someone with much experience in using all these technicalities and technology to accomplish an ends to the means. In most instances I simply much prefer the results from higher sample rates on the music I produce.

    So from my point of view, not from a test that I did not take part in, any music consisting of elements and sources you put a mic in front of, definitely sounds better recorded at higher sample rates to my ears as well as the ears of a lot of other people I know personally who also do this everyday as well. It's really only my tests that ultimately matter to me as it should be for you too. Set up a nice mic and mic pre and record a great acoustic guitar and vocalist at both sample rates. If it doesn't move you any more at 96K, then by all means record at 48K and save your hard drive space. MIX Mag tests aside, That's really the only test that matters.
     
  20. KennyM

    KennyM Supporting Member

    Messages:
    1,727
    Joined:
    Jul 24, 2003
    Location:
    Burbank, Ca.
    I wouldn't necessarily say warmer, but bigger - possibly yes. The fact that you actually get a lot more recorded high frequency content at higher sample rates might lead a listener to conclude that the sound is less warm.

    To me it just sounds more real as in you get more of the audio image as if it were happening live in front of you. More dimension. This usually seems bigger to me even though it isn't necessarily.

    The main advantage I find with this more 3D source is that I tend to use less eq, compression, etc. to get things to sound right and have the desired impact.
     

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice