Harmony Central Forums
Announcement Announcement Module
Collapse
No announcement yet.

Bit rate

Page Title Module
Move Remove Collapse









X
Conversation Detail Module
Collapse
  • Time
  • Show
Clear All
new posts

  • Bit rate

    I was wondering why it is important to record at a high bitrate, if ultimately when you mixdown to a CD, the bitrate has to be 44100 anyways. Thanks,
    jared

  • #2
    Jared,


    You are confusing bit rate and sample rate. I'm happy to record at 44.1 KHz. and 16 bits - it sounds fine to me - but I may be in the minority.

    --Ethan
    The acoustic treatment experts
    Buy my DVD

    Comment


    • #3
      I guess my question is then why is it important to record at a high sample rate. Like my audio interface says it supports up to like 96 kHz or something, but I cant even use that on a CD, so whats the point?

      Comment


      • #4
        It's kind of like when you're working in photoshop - you want to work in the highest resolution possible and only compress or resize down and the very end to preserve detail.

        Comment


        • #5
          With 96 kHz, I don't believe that's the case. I'd go so far as to say that most of the time if you record at 96 and then sample-rate convert down to 44.1 it'll sound worse than it would if you'd just recorded at 44.1 from the beginning. Even keeping it at 96 the differences are so small and so rare it's almost never worth it (for me, anyway...why use twice the hard drive space?).

          Bit depth is a little different...in most cases 16 bits is fine, but with highly dynamic material those extra bits can make a difference, even when your final product is a 44.1kHz 16-bit CD.

          -Duardo

          Comment


          • #6
            Jared,


            It's not.

            --Ethan
            The acoustic treatment experts
            Buy my DVD

            Comment


            • #7
              cool, thanks a lot for the help guys, i understand this stuff a lot more now.

              Comment


              • #8
                Originally posted by jaredoco
                I guess my question is then why is it important to record at a high sample rate. Like my audio interface says it supports up to like 96 kHz or something, but I cant even use that on a CD, so whats the point?
                Two reasons: One, if you own a commercial studio, you'll run into all sorts of stupid clients who insist on recording 96k because some dumbass at Guitar Center (or a magazine who relies on advertising dollars from 96k gear manufacturers) told them it was better. You may find it easier to jump on the bandwagon than it is to explain why that's not necessarily true. And two, if you want to produce future-proof audio. It looks like DVDA and SACD will both hit the market at 96k. If you plan on supporting those formats, you'll want to start recording your stuff at higher sample rates now. It may be too soon to tell though, so my whole point is moot.

                Actually, none of this matters when kids are happy listening to MP3 files, which sound worse than cassettes IMO.

                By the way, I feel the same way about surround.

                Comment


                • #9
                  Fanatic Music-- you clarifyed the benefits of using high sample rate/bit rate. But you didn't really answer the question regarding putting the files onto an Audio CD. If i was to record at high sample rate and then reduce it, will that make the quality actually worse than if i record at 44.1 and then export at 44.1? I'm using DP3, so is there a feature that would allow me to decrease the sample rate well? Thanks,
                  Jared

                  Comment


                  • #10
                    Fanatic Music,

                    I understand (and teach!) the technological advantages behind higher sample rates, but none of that stuff means crap if you can't HEAR the difference. I liken it to Mac/PC debates where geeks start spouting benchmarks, and never discuss the day-to-day interfacing with the OS GUI. The only people I run into anymore who rave about 96k are the kids whose mommy sent them to recording school, and they were taught a bunch of technobabble, but weren't taught what sounds "good" or "bad". The truth is, if you put a bunch of people, including recording engineers, in a room and do double-blind listening tests (we've done plenty of em', with manufacturers present), and your gain staging is perfectly even across the board, most people will tell you they CAN hear a small difference, but they won't tell you 96k necessarily sounds BETTER. A million factors come into play, and in my personal experience, better cables (!) make a bigger difference in my sound than recording 96k does. But hey, your mileage may vary.

                    For all of you newbies present, if you sleep better knowing your stuff is at 96k, well then, hey, you won't hurt anything. Except your disk I/O, track count, plug-in count, DSP horsepower...

                    Sorry to be such a smartass. I have a toothache and it makes me cranky...

                    Comment


                    • #11
                      Recording at higher resolution is theoretically better,
                      however, if you want to pursue this, I would suggest
                      you make a few test recordings with your soundcard
                      at various sample rates and bit depths, then process
                      them down to 44.1/16 files to compare them. This
                      way you will hopefully be able to hear how good the
                      downsampling and dithering functions on your system
                      actually work.

                      In my case, I plugged my CD player into my analog
                      inputs, and recorded a short track from one of my
                      reference discs at the different sample rates and
                      bit depths, compared them, then converted them
                      all to 44.1/16 and compared them again. I found
                      I got better results recording in 24 bit, but the
                      higher sample rates when converted down to 44.1k
                      ended up sounding a hair worse. In essence the
                      downsampling function on my software negated
                      the advantages of recording at higher resolution,
                      so I feel that I'm getting the best results recording
                      at 44.1k at 24 bit. I'm sure people with other
                      software packages, downsampling plug-ins, or
                      external sample rate converters may be getting
                      better results.
                      "Consider everything, believe nothing" Elgar.

                      Comment


                      • #12
                        Originally posted by FanaticMusic
                        Btw...if the dentist's tool spins at 96000 rpm instead of 44100 rpm, then your teeth are ok again in more than half the time...
                        LOL! LOL! Ow, my tooth! LOL!

                        I'm interested in what kind of setup you're running. Aside from the double-blind tests we ran, I also had the chance to use a ProTools HD system over one weekend. My business partner and I use MixPlus systems and were interested if HD would really give us markedly better sound. We multed signals from an older (but really nice-sounding) CAD console into each system. Our Mix system was running Apogee AD-8000 converters with DA and Digi AMBUS cards and Aardvark clocking whereas the HD system utilized a 192 interface and its internal clock generator (the SYNC interface wasn't available at the time). We recorded female voice (Neumann U87 into Manley VoxBox) and acoustic guitar (AKG TLII and AT4051a into UA 2-610) into the Mix system at 44.1 and the HD both at 44.1 and 96k. We then ran each system back into the CAD for monitoring. The results? The HD system at 44.1 had a little smoother top end (VERY subtle), but the bass sounded tighter and the imaging was better on our Mix system. At 96k, the HD system was MAYBE 5% smoother (as if you could quantify sound quality) than at 44.1, and the bass and imaging sounded exactly the same. Interestingly enough, the 192k was about 5% again smoother, which didn't make sense, since you're splitting hairs even thinner at that point, but the bass still wasn't as defined and the stereo field was still a little smeared. Since we do primarily rock, hip hop, and R&B music, we actually preferred the Apogee/Aardvark 44.1k Mix system over the 96k HD system. Now of course, we were comparing apples to, well, granny smith apples, and our testing wasn't absolutely perfect, but it certainly saved us $15k.

                        And we decided to record everything at 24-bit, 44.1k. Want something to blow your gourd? At a friend's house, we did a listening test with Cool Edit Pro and a Hammerfall card into an Apogee Rosetta into a headphone amp (admitedly not the best setup). We thought THIRTY-TWO BIT, 44.1k sounded better than 24-bit, 96k. Weird!

                        Comment


                        • #13
                          Two reasons: One, if you own a commercial studio, you'll run into all sorts of stupid clients who insist on recording 96k because some dumbass at Guitar Center (or a magazine who relies on advertising dollars from 96k gear manufacturers) told them it was better. You may find it easier to jump on the bandwagon than it is to explain why that's not necessarily true.


                          That's a good reason. Probably the best.

                          And two, if you want to produce future-proof audio. It looks like DVDA and SACD will both hit the market at 96k. If you plan on supporting those formats, you'll want to start recording your stuff at higher sample rates now.


                          You don't have to record at higher sampling rates to be compatible with these formats.

                          A very important factor is TIMING. A wave starting between 2 samples is NOT exactly recorded. So making the mesh smaller gives a better Timing-representation of the original signal.


                          It is if it's within the Nyquist range. And since we're talking about the end result winding up on CD in this case, you're not picking up anything extra that will make it to the end.

                          Samplefrequency and Bitresolution are BOTH limitations.
                          While the SF gives the horizontal resolution limit, the Bitrate gives the vertical resolution.
                          Both are a tradeoff, and are somewhat linked together. For example it makes no sense to record at 192KHz@4bit, or at 1KHz@32bit.


                          I don't think anyone will argue with you on that one.

                          Unfortunately this is mostly theory. There are so many bad convertors with good specs. That leads us to some old 38KHz@16bit ADCs that "sound" better than today 192/24 ADCs.


                          Exactly. You can get great results reording at 16/44.1 with good converters and most of the time you'd gaine nothing by stepping up to higher sampling rates or bit depths.

                          While higher bitrates and samplefrequencies may cause more CPU-load on your computers, they are to be preferred for extensive editing. (Plugins, EQ, Fades, ...)


                          You don't have to record at a higher bit rate to process at a higher bit rate.

                          Converting the samplerate from i.e. 96KHz to 44.1KHz does NOT do any bad things to your material, IF it is done correctly!!! BUT Cheap SRCs make a cheap sound!!!
                          So for the typical homestudio demo lacking a pro-SRC, better start with 44.1/16 or 24. This means less trouble, and more free CPU-power.


                          Good recommendation.

                          Oh, and btw... people that think 44.1/24 is good enough, don't need to waste money on highend preamps too...


                          I disagree. Have you ever compared preamps recorded at 24/44.1? The differences are not subtle (well, sometimes they are, but if it's a good converter you can easily hear them).

                          If i was to record at high sample rate and then reduce it, will that make the quality actually worse than if i record at 44.1 and then export at 44.1?


                          In most cases, yes, the artifacts of the sample rate conversion will make your audio sound worse than if you'd just recorded at 44.1 in the first place. And even if you have a good sample rate converter, why would you want to take more than twice the bandwidth and storage space to get something that will sound at best just as good as recording at 44.1, but no better?

                          I'm using DP3, so is there a feature that would allow me to decrease the sample rate well?


                          Not that will sound as good as it would if you recorded at 44.1 in the first place.

                          If the SR-conversion to 44.1KHz is done properly, then the final CD should benefit from a higher recording SR.


                          No, it won't. It can benefit from recording at a higher bit depth (even though that too in most cases is unnecessary), but you won't gain anything by recording at 96 or 88.2 and then converting to 44.1.

                          I understand (and teach!) the technological advantages behind higher sample rates, but none of that stuff means crap if you can't HEAR the difference.


                          Amen to that.

                          A million factors come into play, and in my personal experience, better cables (!) make a bigger difference in my sound than recording 96k does.


                          Amen to that.

                          It's 100% true, that a higher sample-rate gives a more exact representation of the original signal. So the higher the sample-rate, the NEARER to the original. See, there is no room for personal taste yet.


                          Nobody is arguing with you on that...the point is, it doesn't matter if it's a more exact representation of the original signal. Why do you need to capture a more exact representation than what you're capable of hearing? Especially if you're going to convert it to 44.1 eventually, when all that extra inaudible "detail" you picked up is wiped out again?

                          In theory, the "best" preamp should amplify the incoming signal to the desired gain, and should be as exact as can get, without introducing any distortion or adding any artifacts.


                          Only a few of the "best" preamps out there do this. Most of what are considered to be the "best" preamps have a very distinct color to them, which is why so many studios have multiple preamps. But it's different with converters...they're almost always built with total accuracy as the goal.

                          The things is, if you have already a good acoustic room, best cables, great monitors, excellent consoles and mics, then...
                          you WILL hear, that higher SR are more exact and therefore better.


                          More exact is not necessarily "therefore better" if it's more exact than it needs to be. I've done plenty of listening tests with converters and have heard some where the higher sampling rates do sound a little better than the lower ones, but more where they don't, and I'm not convinced that the ones that sound better at higher sampling rates do so because of the rate so much as because of the design. Some of the best ones I've listened to don't sound better at all.

                          Also, one other thing I've noticed is that when comparing the converters to the original source, the differences between the various converters are much more subtle than the difference between any converter and the actual analog source.

                          At a friend's house, we did a listening test with Cool Edit Pro and a Hammerfall card into an Apogee Rosetta into a headphone amp (admitedly not the best setup). We thought THIRTY-TWO BIT, 44.1k sounded better than 24-bit, 96k. Weird!


                          Interesting. What D/A converter were you using to listen back? If you were using the Rosetta I'd expect them to sound about the same (when you're recording at 32 bits your'e still only recording 24 bits of audio information).

                          -Duardo

                          Comment


                          • #14
                            There are so many variables you really need to look at your whole chain and what your final requirements are. To me the sound much more important than the specifications. Many who swear by 24/96 probably have never tried a high quality 16/44 device. I use 24/88 on my M-audio because it sounds like crap at 16/44. Even my 16 bit DAT's sound as good as it does at 24/96. On the other hand I am quite satisfied with my 2408 and Q10 at 16/44 and rarely use them at higher settings. But it also depends on what I am recording too. Another point is the recorder only plays back what it hears and IMO most setups today are better thabn what they are hearing.

                            Comment


                            • #15
                              FM,


                              It's 100% true that you have little understanding of digital theory.

                              If people could hear above 20 KHz., and your mikes and loudspeakers could capture and reproduce past 20 KHz., then maybe we'd be having a different discussion.

                              --Ethan
                              The acoustic treatment experts
                              Buy my DVD

                              Comment



                              Working...
                              X