Harmony Central Forums
Announcement Announcement Module
Collapse
No announcement yet.

Bit rate

Page Title Module
Move Remove Collapse







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

  • #16
    You guys are literally arguing about a very small increase in sound quality - and you're not looking at the overall big perspective. That is - who are you making the music for? Is it for recording engineers and acutely trained musicians? If so, then spend all the money, space and CPU with higher bit and sampling rates. But if you're recording for the public, they probably couldn't tell the difference between even a stereo SACD and a mp3 at 128, at least not on the consumer speakers that they will be listening on. I know my parents couldn't. My best friend can't. So what's the point, unless it really makes you happy. They will, however, hear the difference in good mixing and mastering, so spending your time worrying about that is probably a better use of your time.

    Comment


    • #17
      I did NOT talk about the frequency-theme, I did talk about the TIMING-domain! So what has the timing-precision todo with people's hearing limits ?

      The average person can distinguish an interaural delay of about 15 microseconds. Some people can even hear 3-5 microsecs.
      A CD with 44100 samples per second means that there are 22,67 microsecs between EACH sample.


      That's a common misconception, and it does make sense. The truth is, even though there are 22.67 microseconds between samples, the stuff (below the Nyquist frequency) that happens between those samples is perfectly reconstructed upon D/A conversion, timing and everything. That's not a matter of opinion.

      -Duardo

      Comment


      • #18
        Originally posted by FanaticMusic
        And again, for homerecordings, make a song and dont worry. For the rest, 44.1KHz is NOT enough...
        Why, because of specs? Specs have little to do with REAL sound quality-- The sound quality you hear with your ears, not the sound quality you read on a spec sheet with your eyes. The vast majority of my peers haven't bought into the 96k hype, and they all produce fantastic, professional-level (and quite a few major-label) mixes.

        Comment


        • #19
          FM,


          Sorry, I agree that was a little harsh. But I'm just so sick of arguing this same nonsense over and over.

          Nobody can hear much past 20 KHz. - not me, not you, and not the mostly highly paid mastering engineer on the planet. So using disk space and CPU resources to capture more than 20 KHz. is a complete and utter waste. Any difference that can be shown repeatedly in double blind testing is guaranteed to be due to something other than the higher bandwidth.


          Okay, so what's your name?


          Which is a high enough resolution for audio.

          --Ethan
          <div class="signaturecontainer"><a href="http://www.realtraps.com/experts.htm" target="_blank">The acoustic treatment experts</a><br />
          <a href="http://www.ethanwiner.com/video/" target="_blank">Buy my DVD</a></div>

          Comment


          • #20
            "Nobody can hear much past 20 KHz. - not me, not you, and not the mostly highly paid mastering engineer on the planet."

            I have no desire to get into an argument with anyone, but Rupert Neve among many many other people have been proven to hear information at 50khz and higer.

            I think the harmonics and overtones are a large part of what makes music sound good. You can't put a number or a spec on what sounds good. It sounds good, or it doesn't.
            <div class="signaturecontainer">It sounds to me like it's about time you stop being a consumer, and start being an engineer. -Fletcher-</div>

            Comment


            • #21
              Wow- what a thread111
              I am a Film/Tv/Music Mixer. 10 years as a mixer-not an assistant, but actual mixing. I have to weigh in with this-
              Audacity-Here's where Higher sample rate/Bit depth is important-with 200+ tracks running in your sessions. Increasing the dynamic range is really important, especially when you're not mixing in the Box. I also use a PT Mix3 system, clocked with an Aardsync-A track by track comparison yield the results you were experiencing. But when you do a mix, whole different story. A 192 interface is a great piece of engineering, and sorry, blows away a Mix System. and your test is slightly flawed-you are probably hearing headroom issues in the different Mixer Models within PT. A better test is to run an Analog signal, Mult it into 3 separate signals, run each through the system on input, and compare to the original analog sound-That's a telling test, with HD far in front.

              Ethan- FM is correct in what he is saying-Although we cannot hear past 20k, we CAN "sense" Higher overtones. Also, a Higher sample rate/Bit depth will allow for a more accurate representaion of the Audio wave, thus eliminating errors that are occuring in the audible region due to interpolation.

              To the original poster-
              Sample rate is how many times per second the audio signal gets sampled. In the case of 44.1, it's sampled 44,100 times per second. at 96k it's 96000. so a 96k converter (let's assume quality of converters ins the same for a sec) takes twice as many "Snapshots" of the audio in a given period of time.
              Bit depth is how many values can be represented within that snapshot-each "Word" has a set number of "Value", with larger bit rates having more "values" available.

              And finally, as many people here are saying, don't sweat it. Just record, learn your craft, then worry about the equipment.
              -todd A.
              <div class="signaturecontainer">My Rig:<br />
              <br />
              EAW LA325's (2)<br />
              EAW LA400's (4)<br />
              QSC PLX 3402 (2)<br />
              QSC PL1.8<br />
              QSC RMX 2450<br />
              Ashly ProTea 3.24c<br />
              Ashly ProTea 4.24G<br />
              EAW LA212 Monitors<br />
              JBL PRX-512M Monitors<br />
              Whole Lotta Cables <img src="http://img3.harmony-central.com/acapella/ubb/smile.gif" border="0" alt="" title="Smilie" class="inlineimg" /></div>

              Comment


              • #22
                Sorry little friend.

                What is that supposed to mean?

                What you say is ABSOLUTELY WRONG. Really.

                If you read all my posts, you will see that i said, that nyquist's theorem is ONLY valid for sinoids AND unlimited bitrate.

                I did read all of your posts, but I still don't agree. Nyquist's theorem is valid for all frequencies that are less than half the sampling rate. Any complex waveform can be broken down into sine waves, and if the "smallest" (or highest-frequency) component of that complex waveform is more than half the sampling rate it doesn't meet the theorem. In a 44.1kHz A/D converter anything that's above about 22 kHz will have already been filtered out by the anti-aliasing filters.

                As for the bitrate, adding more bits just extends your dynamic range. If the sound you're recording has a dynamic range of, say, 70 dB, then 16 bits are more than sufficient to capture that. You won't gain any more "resolution" or "detail" or anything useful by using more bits. Those same twelve bits that are required to reproduce that signal perfectly will be just as accurate if they're part of a 24-bit word, or a 32-bit word, or 64...

                So at unlimted bitrate it is just enough to have 2 points of a sinoid at max. half the samplefreq to reconstruct it 100%.

                It's enough at any bitrate so long as it's enough to contain the entire dynamic range of the signal.

                I would say 16 or 24 bit rate IS a limitation, isn't it?

                In most cases, no. Not for what we're talking about here. Sure, if you're recording an orchestra you may actually record a signal with a true 100 dB dynamic range. Even if you're not using up that whole dynamic range, with 24 bits you can leave yourself plenty of headroom. But for what we're talking about here, in most cases 16 bits isn't a practical limitation, and 24 certainly isn't. If there are limitations they're in the design of products we use, not the number of bits themselves.

                So that vertical quantization (bitrate) cuses also a small horizontal quantization aka PHASESHIFTS aka changes in the Timing-Domain.

                No, the quantization adds random noise to the signal. The more bits, the lower down in amplitude this quantization noise goes. Again, in most cases 16 bits is fine, and with 24 bits quantization noise is so much lower than the noise of any of the analog gear we have these days it's not an issue.

                This means INcreasing the samplefreq DEcreases the horizontal quantization.
                in short: higher samplefreq -> better timing-representation (even of sinoids)

                No. How did you draw that conclusion? The phase and amplitude of all signals below the Nyquist frequency are perfectly recreated down to the level of quantization noise. All you get with higher sampling frequencies is representation of higher-frequency components of complex waveforms that our ears filter out anyhow.

                It's easy to understand.

                What you say does make sense intuitively, but the whole idea behind the Nyquist theorem is that even though you'd intuitively think that the more "dots" you have to connect the more accurate of a picture you'll have, the truth is that because of the limitations of the theorem the dots can only be re-connected upon D/A conversion in one unique way which will recreate the waveform at is proper amplitude, with all frequencies intact, and in phase. You'd think "well, maybe the line between these two dots curves up instead of down or is straight or comes to a point" or whatever but that doesn't happen. Based on the surrounding "dots" there's only one way to connect them.

                If you don't believe ME, believe Ed Foster.
                Edward J. Foster is president of Diversified Science Laboratory and technical editor of Pro Audio Review.

                Give us a link to what he's said, so we can read it in his own words rather than as interpreted by you.

                And people with much more impressive credentials than his have made totally erroneous assumptions about digital audio as well.

                I have no desire to get into an argument with anyone, but Rupert Neve among many many other people have been proven to hear information at 50khz and higer.

                My point exactly. He knows analog a lot better than he knows digital.

                -Duardo

                Comment


                • #23
                  Here's where Higher sample rate/Bit depth is important-with 200+ tracks running in your sessions. Increasing the dynamic range is really important, especially when you're not mixing in the Box.


                  How are you increasing the dynamic range by recording at a higher bit depth? The only way to increase the dynamic range is to record a signal with a louder dynamic range. A converter alone won't do it.

                  Also, a Higher sample rate/Bit depth will allow for a more accurate representaion of the Audio wave, thus eliminating errors that are occuring in the audible region due to interpolation.


                  In a good converter (such as your 192) there won't be audible interpolation errors.

                  Sample rate is how many times per second the audio signal gets sampled. In the case of 44.1, it's sampled 44,100 times per second. at 96k it's 96000. so a 96k converter (let's assume quality of converters ins the same for a sec) takes twice as many "Snapshots" of the audio in a given period of time.
                  Bit depth is how many values can be represented within that snapshot-each "Word" has a set number of "Value", with larger bit rates having more "values" available.


                  What needs to be added that is that as we add bits we increase the dynamic range. That's it. There's a common misconception that as we add bits we increase the resolution over the entire dynamic range, and that's simply not true. All we do is push the level of quantization noise lower and lower, and all that extra "resolution" we get is for the most part so low in amplitude we don't even take advantage of it.

                  -Duardo

                  Comment


                  • #24
                    Originally posted by Duardo


                    How are you increasing the dynamic range by recording at a higher bit depth? The only way to increase the dynamic range is to record a signal with a louder dynamic range. A converter alone won't do it.


                    Most film production work has an extreme dynamic range; deeper bit depth merely captures this better. Also, the self noise of a 24 bit system is nuch cleaner.


                    Originally posted by Duardo

                    In a good converter (such as your 192) there won't be audible interpolation errors.

                    agreed, to a point.....Any converter will have interpolation errors; ther's just a point at which we can no longer hear them. Also,, many people here don't have access to that kind of hardware


                    Originally posted by Duardo

                    What needs to be added that is that as we add bits we increase the dynamic range. That's it. There's a common misconception that as we add bits we increase the resolution over the entire dynamic range, and that's simply not true. All we do is push the level of quantization noise lower and lower, and all that extra "resolution" we get is for the most part so low in amplitude we don't even take advantage of it.


                    I should have also added that this depends heavily on the recording level; in a controlled environment, a much hotter recording can be achieved than out in the field....
                    -todd A.
                    <div class="signaturecontainer">My Rig:<br />
                    <br />
                    EAW LA325's (2)<br />
                    EAW LA400's (4)<br />
                    QSC PLX 3402 (2)<br />
                    QSC PL1.8<br />
                    QSC RMX 2450<br />
                    Ashly ProTea 3.24c<br />
                    Ashly ProTea 4.24G<br />
                    EAW LA212 Monitors<br />
                    JBL PRX-512M Monitors<br />
                    Whole Lotta Cables <img src="http://img3.harmony-central.com/acapella/ubb/smile.gif" border="0" alt="" title="Smilie" class="inlineimg" /></div>

                    Comment


                    • #25
                      I should have also added that this depends heavily on the recording level; in a controlled environment, a much hotter recording can be achieved than out in the field....


                      The thing is, with today's good converters there really isn't the need to get a hotter signal for a hotter signal's sake. Again, using the example of a signal with a 70 dB dynamic range (which is very wide), with a good 24-bit converter it doesn't matter if you're peaking at -1 dBFS, -10 dBFS, or -20 dBFS...you're capturing that signal with the exact same precision. Not that there aren't other reasons to do so...especially if you're mixing through an analog board...but not to "maximize the bits" or "get more resolution" or "use more steps".

                      Most film production work has an extreme dynamic range; deeper bit depth merely captures this better. Also, the self noise of a 24 bit system is nuch cleaner.


                      Deeper bit depth does not capture it better unless the dynamic range is so wide that it exceed the limitations of the 16-bit system. If the 16-bit system can capture it accurately, the 24-bit system won't capture it "better". And the self noise of the system won't be an issue unless the dynamic range of the signal approaches the limits of the signal. With that same theoretical signal with a dynamic range of 70 dB, if you record it peaking at, say, -2dBFS, you're capturing with all the accuracy you need and the self-noise of the system is still low enough you won't hear it. If you record it at 24 bits you're not gaining anything, except maybe the ability to set your levels a little more conservatively and peak at -10 dBFS or lower so you don't have to worry about accidental overs or anything like that.

                      -Duardo

                      Comment


                      • #26
                        Originally posted by The Chinese
                        Wow- what a thread111
                        I am a Film/Tv/Music Mixer. 10 years as a mixer-not an assistant, but actual mixing. I have to weigh in with this-
                        Audacity-Here's where Higher sample rate/Bit depth is important-with 200+ tracks running in your sessions. Increasing the dynamic range is really important, especially when you're not mixing in the Box. I also use a PT Mix3 system, clocked with an Aardsync-A track by track comparison yield the results you were experiencing. But when you do a mix, whole different story. A 192 interface is a great piece of engineering, and sorry, blows away a Mix System. and your test is slightly flawed-you are probably hearing headroom issues in the different Mixer Models within PT. A better test is to run an Analog signal, Mult it into 3 separate signals, run each through the system on input, and compare to the original analog sound-That's a telling test, with HD far in front.
                        Yeah, the reason I mentioned our methods wasn't to justify my findings as much as it was to A. admit our testing procedure wasn't the best (plus, we had been drinking! ), and B. That there are people with professional rigs who know them and use them well who couldn't give the tiniest turd about 96k in 2003.

                        I agree that a stock HD system will blow away a stock MIX system, even with an Aardsync. The three main reasons behind this, I think, is that HD employs #1. Drastically higher quality converters (both in analog circuitry before A/D conversion and after D/A conversion, and in implementation), #2. Digi's new loop clock, and #3. Increased headroom on the mix bus (which you mentioned). I don't think it has all that much to do with higher sample rates, though Digi I'm sure would lead everyone to believe otherwise, because sticking with obsolete gear sure doesn't fuel the economy. We don't use any Digi hardware at all, aside from the DSP cards and an ADAT bridge|24 to link a Mackie D8B optically-- So our system sounds dramatically better than a stock MIX system-- Apogee converters (#1), Aardvark clocking (#2), and we route each track out discretely into an external mixer and use it as a mix bus instead. I think the HD system sounds really good, but for the type of music we record, even with 40+ tracks of background vocals, 96k isn't even close to necessary, since we have everything else covered tightly.

                        It reminds me of a story: When the Alesis Masterlink came out, we did a pretty exhaustive test at multiple sample rates, both with its stock converters and a Prism Sound converter borrowed from a client. We noticed a difference right away with acoustic guitar tracks at 96k... UNTIL... we discovered either by fault of Alesis, or a damaged unit, the 96k track was about 1.5 dB louder. When we compensation for the gain discrepancy, the audible difference all but disappeared, even with $6000 converters. Without jumping to conclusions, we joked about how Alesis did that on purpose just to convince people to buy a Masterlink so they could record at 96k.

                        Oh, and this is off topic, but the first NAMM show I went to a few years ago, Roland introduced their DS-90 studio monitors and had a shootout between them, the Event 20/20bas, and the Mackie HR824s. Now I've owned the Events in the past, and have used the Mackies now for over four years, and I KNEW there was something wrong from the start. There's no way the Roland DS-90's could make my Mackies suddenly sound tubby and one-dimentional. When I looked back to check and see how everything was set up, someone from Roland stopped me. A month later, when the DS-90s came in, we did the exact same shootout at my store. The result? The Rolands sounded like ass in comparison. I was pissed off at that company for a long time after that. What's my point? Uh, I forget. Oh yeah, companies will skew listening tests to sell product. All those Digi Tours touting 192k where everyone claimed to hear a noticable difference? Okay, how come no one else I've talked to can recreate those differences in their own private shootouts? And even if your average joe's tests are the most exacting or clinical, they're definitely real world, and the only place where sound quality should matter (unless you're performing scientific studies or some equally non-creative endeavor).

                        I definitely believe 24-bit recording will give almost everyone a noticable boost in sound quality (and it is here where the connection to headroom lies, and not in sample rate), but if, with professional gear, we can't hear but the tiniest bit of difference between 44.1 and 96k, it's definitely my place to defend the easily corruptable newbies when someone tells them 44.1 has no place in professional audio. Hundreds of world-class facilities would strongly disagree.

                        But for film/post and DVD authoring, I guess I could see where 96k is appropriate. I mean, you guys have million-dollar budgets!

                        Comment


                        • #27
                          Just to throw a spanner in the argument...

                          I read somewhere (somewhere on this site in fact) that it is better to record at 88.2kHz if you are mixing down to 44.1 since the conversion is easier for the processor to do when it is exactly half. Is this the case? It sounded possible but unlikely to me but it could be the case, I don't know. I figured a good program would be as good at converting at any sampling rate.

                          Also, what happens if you burn CDs at the wrong sampling rate, do they not read or do they play at the wrong speed?
                          <div class="signaturecontainer">proud member of the HCBF <br />
                          <b>Tube Bass Amp Club</b><br />
                          <br />
                          less proud member of compulsive ebayers anonymous<br />
                          <br />
                          Ibanez EDB-600 / Novation Bass Station<br />
                          Inta-Audio custom laptop running audiomulch<br />
                          Ibanez PD-7<br />
                          Trace Elliot V8<br />
                          Peavey 115</div>

                          Comment


                          • #28
                            I'd also like to say that I reckon there's a noticable difference in sound between 16/44.1 and 24/96 but maybe not everyone hears it, it's certainly small though more obvious in certain frequency ranges. Remember though there are a lot of people who can't hear the difference between audio and MP3s so how much of an effect the step up will have in the ears of the masses I don't know.
                            <div class="signaturecontainer">proud member of the HCBF <br />
                            <b>Tube Bass Amp Club</b><br />
                            <br />
                            less proud member of compulsive ebayers anonymous<br />
                            <br />
                            Ibanez EDB-600 / Novation Bass Station<br />
                            Inta-Audio custom laptop running audiomulch<br />
                            Ibanez PD-7<br />
                            Trace Elliot V8<br />
                            Peavey 115</div>

                            Comment


                            • #29
                              I read somewhere (somewhere on this site in fact) that it is better to record at 88.2kHz if you are mixing down to 44.1 since the conversion is easier for the processor to do when it is exactly half. Is this the case? It sounded possible but unlikely to me but it could be the case, I don't know. I figured a good program would be as good at converting at any sampling rate.


                              Why does that sound unlikely? It's a very simple process mathematically...you basically just have to drop every other sample and you're there. Converting to 44.1 from 96 is much more complex mathematically, and that's why there are plainly audible artifacts on cheaper converters or poorly-written code.

                              -Duardo

                              Comment


                              • #30
                                OS,


                                That simply is not true. He reported that a circuit oscillating at 50 KHz. produces artifacts within the audible band. That's an entirely different thing.


                                You miss the fact that loudspeakers don't reproduce frequencies that high!

                                --Ethan
                                <div class="signaturecontainer"><a href="http://www.realtraps.com/experts.htm" target="_blank">The acoustic treatment experts</a><br />
                                <a href="http://www.ethanwiner.com/video/" target="_blank">Buy my DVD</a></div>

                                Comment

                                Working...
                                X