|
|
Post by mshump on Oct 8, 2015 8:10:31 GMT -5
Last night out of curiosity I went into J river and changed the settings so that all my 44.1 CD rips would output at 96k. I wanted to see if I could hear a difference. This is by no means a scientific A vs B, or a double blind test. It was just me and my memory. What I did notice was that there was a difference going from 44.1 to 96. The 96 was a little more revealing, there seemed to be a little more precision in the sounds, especially in the mid-range and upper frequencies. It also seemed to loose a little bit of smoothness, the music just seemed a little more edgy. The sound at times was better and sometimes fatiguing depending on the original source (rips from Audio Fidelity and Mofi discs were a little better)
I then from curiosity switched the J river from 96k to 192k. I heard no difference between the 2 settings.
The differences I heard were 44.1 to 96 were noticeable but not overly significant. The major conclusion I made was that if the original rip is from a badly mastered disc, 96k will reveal more of its bad sound. I grew up on vinyl so I wonder if a more revealing sound (not as warm per say like on older vinyl) has clouded my perception of what is the proper/best sound ??
IMHO the original mastering makes all the difference in the world regardless of the settings.
|
|
|
|
Post by yves on Oct 8, 2015 8:15:03 GMT -5
The oppo 105 (Plus XSP-1 - important!) was the finest DAC I've heard in real life. So I can completely understand your description of the fantastic saber dac. So have you heard the yggy in real life? If so, what did you think? Did you find it veiled or something? I'm curios too, please share your views on the yggy Yes, I have heard it, but no, the Yggdrasil doesn't sound veiled or muffled... at least not to my ears. A lot has to do with what you expect from a DAC. Like I said, it is no secret that SABRE DACs put out ungodly amounts of detail. So much so, it certainly is no surprise that some people think it is possible to get a sound signature in which there is just too much clarity. (If you have watched the video that is being discussed here in this thread, surely you will have remembered the part of this video that touches this particular subject). Well, the thing about the DAC Supreme is... it has that kind of "non fake" detail (i.e., detail, as opposed to brightness that can very easily be misinterpreted as detail...), not just in the low mids and the mids, but all the way from the sub bass to the upper regions of the spectrum. The Yggdrasil has it too. At its price point, however, I would cry you a river if it didn't! Which brings me to my point. The Yggdrasil does nothing audibly wrong. It's an impressive DAC. Just that, in my own subjective, personal experience, the DAC Supreme trounces it in terms of realism, or "lifelike" presentation. Call it musicality if you will. For the most part, it's probably a matter of synergy combined with personal preferences—including the specific type of sound your ears have been adapting to over a prolonged period of time, your room, and your listening habits. I don't really like to talk about different traits of sound and how they might compare to eachother. Instead, I think a lot of people should probably let the sound itself do most of the talking for them. Because, by following a similar type of advice (...and I hope you don't mind my posting the link again), I learned how to *properly* cook a Steak Friet: |
|
|
|
Post by monkumonku on Oct 8, 2015 8:58:58 GMT -5
Last night out of curiosity I went into J river and changed the settings so that all my 44.1 CD rips would output at 96k. I wanted to see if I could hear a difference. This is by no means a scientific A vs B, or a double blind test. It was just me and my memory. What I did notice was that there was a difference going from 44.1 to 96. The 96 was a little more revealing, there seemed to be a little more precision in the sounds, especially in the mid-range and upper frequencies. It also seemed to loose a little bit of smoothness, the music just seemed a little more edgy. The sound at times was better and sometimes fatiguing depending on the original source (rips from Audio Fidelity and Mofi discs were a little better) I then from curiosity switched the J river from 96k to 192k. I heard no difference between the 2 settings. The differences I heard were 44.1 to 96 were noticeable but not overly significant. The major conclusion I made was that if the original rip is from a badly mastered disc, 96k will reveal more of its bad sound. I grew up on vinyl so I wonder if a more revealing sound (not as warm per say like on older vinyl) has clouded my perception of what is the proper/best sound ?? IMHO the original mastering makes all the difference in the world regardless of the settings. Maybe I am not understanding this correctly but if the source is at 44.1/16, I don't see how going to a higher bit rate can improve anything because it is still subject to the limitations of the source. That would be like playing an mp3 file recorded at 128 or 256 and then upscaling it to 96/24. You'd still be at the mercy of the original mp3 file limitations, wouldn't you? |
|
KeithL
Administrator 
Posts: 10,273
|
Post by KeithL on Oct 8, 2015 9:05:24 GMT -5
OK, I really need to explain something technical here....... (your results are interesting - but NOT in the way you think).
The digital audio on a CD is recorded at a sample rate of 44.1k . Therefore, if you RIP that CD at 96k, you are NOT getting one iota of extra detail off the disc; if it isn't there to begin with, then you can't get it, no matter what you do. In fact, if the program you used made a "perfect" upsampling from the 44k audio that you're reading off the disc to the 96k audio that you asked it to deliver to you, they would sound identical. (Although you cannot add information, it is certainly possible that you could get a "perfect" conversion where nothing was lost or changed - at least audibly. If you were converting an analog signal, like from a vinyl album, to digital, then the higher sample rate might in fact preserve more detail from the original signal; however, when you start with a digital signal, you are "hard limited" by what's there.)
However, I'm NOT suggesting that the difference you're hearing is imaginary; there could in fact be a real audible difference. However, if there is a difference, it is due either to imperfections in the conversion process (all sample rate conversion involves digital filtering, and the algorithms used by many programs are not audibly transparent), or to the fact that your DAC sounds slightly different when playing 96k files than when playing 44k files (which can happen for a number of reasons). The important point, though, is that you were NOT comparing a 44k file to a 96k file; rather you were comparing a 44k file played at 44k to a 44k file converted to and played at 96k, which isn't at all the same thing; because what you're playing is still subject to the fidelity limitations of the original 44k file.
If you want to actually compare the sound quality at two different sample rates, you need to start with an original at the HIGHER sample rate (which presumably has extra information), then convert DOWN to the lower sample rate, in which case the higher sample rate file will (presumably) contain extra information that was necessarily discarded when the sample rate was reduced during the conversion process. (The conversion process itself will still potentially alter the sound slightly, but that alteration will be added to any "legitimate" difference due to the lower sample rate, which might also still be audible.)
(The only single exception to what I said is that there are certain very specific encoders - including the Dolby Pro Encoder - which do in fact upsample audio and deliberately alter it along the way - with the intent of "repairing" flaws introduced by the encoding process with which they were originally produced. However, in that situation, if you hear a difference, it is due not to the different sample rate but because the upsampler is deliberately altering the audio. Neither jRiver, nor any "regular" upsampling program will do this; it's a big deal, and they'll make a point of telling you if they are.) Incidentally, I agree entirely on your final point  Last night out of curiosity I went into J river and changed the settings so that all my 44.1 CD rips would output at 96k. I wanted to see if I could hear a difference. This is by no means a scientific A vs B, or a double blind test. It was just me and my memory. What I did notice was that there was a difference going from 44.1 to 96. The 96 was a little more revealing, there seemed to be a little more precision in the sounds, especially in the mid-range and upper frequencies. It also seemed to loose a little bit of smoothness, the music just seemed a little more edgy. The sound at times was better and sometimes fatiguing depending on the original source (rips from Audio Fidelity and Mofi discs were a little better) I then from curiosity switched the J river from 96k to 192k. I heard no difference between the 2 settings. The differences I heard were 44.1 to 96 were noticeable but not overly significant. The major conclusion I made was that if the original rip is from a badly mastered disc, 96k will reveal more of its bad sound. I grew up on vinyl so I wonder if a more revealing sound (not as warm per say like on older vinyl) has clouded my perception of what is the proper/best sound ?? IMHO the original mastering makes all the difference in the world regardless of the settings.
|
|
|
|
Post by mshump on Oct 8, 2015 9:06:41 GMT -5
Last night out of curiosity I went into J river and changed the settings so that all my 44.1 CD rips would output at 96k. I wanted to see if I could hear a difference. This is by no means a scientific A vs B, or a double blind test. It was just me and my memory. What I did notice was that there was a difference going from 44.1 to 96. The 96 was a little more revealing, there seemed to be a little more precision in the sounds, especially in the mid-range and upper frequencies. It also seemed to loose a little bit of smoothness, the music just seemed a little more edgy. The sound at times was better and sometimes fatiguing depending on the original source (rips from Audio Fidelity and Mofi discs were a little better) I then from curiosity switched the J river from 96k to 192k. I heard no difference between the 2 settings. The differences I heard were 44.1 to 96 were noticeable but not overly significant. The major conclusion I made was that if the original rip is from a badly mastered disc, 96k will reveal more of its bad sound. I grew up on vinyl so I wonder if a more revealing sound (not as warm per say like on older vinyl) has clouded my perception of what is the proper/best sound ?? IMHO the original mastering makes all the difference in the world regardless of the settings. Maybe I am not understanding this correctly but if the source is at 44.1/16, I don't see how going to a higher bit rate can improve anything because it is still subject to the limitations of the source. That would be like playing an mp3 file recorded at 128 or 256 and then upscaling it to 96/24. You'd still be at the mercy of the original mp3 file limitations, wouldn't you? Yes you are correct. One of the things dr. Aix said was that some of the Hi-Res re-masters were just basically up converted, so I was trying it myself to see if there was a difference. Maybe I should've been a little more specific in my post ? |
|
|
|
Post by mshump on Oct 8, 2015 9:15:19 GMT -5
OK, I really need to explain something technical here....... (your results are interesting - but NOT in the way you think).
The digital audio on a CD is recorded at a sample rate of 44.1k . Therefore, if you RIP that CD at 96k, you are NOT getting one iota of extra detail off the disc; if it isn't there to begin with, then you can't get it, no matter what you do. In fact, if the program you used made a "perfect" upsampling from the 44k audio that you're reading off the disc to the 96k audio that you asked it to deliver to you, they would sound identical. (Although you cannot add information, it is certainly possible that you could get a "perfect" conversion where nothing was lost or changed - at least audibly. If you were converting an analog signal, like from a vinyl album, to digital, then the higher sample rate might in fact preserve more detail from the original signal; however, when you start with a digital signal, you are "hard limited" by what's there.)
However, I'm NOT suggesting that the difference you're hearing is imaginary; there could in fact be a real audible difference. However, if there is a difference, it is due either to imperfections in the conversion process (all sample rate conversion involves digital filtering, and the algorithms used by many programs are not audibly transparent), or to the fact that your DAC sounds slightly different when playing 96k files than when playing 44k files (which can happen for a number of reasons). The important point, though, is that you were NOT comparing a 44k file to a 96k file; rather you were comparing a 44k file played at 44k to a 44k file converted to and played at 96k, which isn't at all the same thing; because what you're playing is still subject to the fidelity limitations of the original 44k file.
If you want to actually compare the sound quality at two different sample rates, you need to start with an original at the HIGHER sample rate (which presumably has extra information), then convert DOWN to the lower sample rate, in which case the higher sample rate file will (presumably) contain extra information that was necessarily discarded when the sample rate was reduced during the conversion process. (The conversion process itself will still potentially alter the sound slightly, but that alteration will be added to any "legitimate" difference due to the lower sample rate, which might also still be audible.)
(The only single exception to what I said is that there are certain very specific encoders - including the Dolby Pro Encoder - which do in fact upsample audio and deliberately alter it along the way - with the intent of "repairing" flaws introduced by the encoding process with which they were originally produced. However, in that situation, if you hear a difference, it is due not to the different sample rate but because the upsampler is deliberately altering the audio. Neither jRiver, nor any "regular" upsampling program will do this; it's a big deal, and they'll make a point of telling you if they are.) Incidentally, I agree entirely on your final point Last night out of curiosity I went into J river and changed the settings so that all my 44.1 CD rips would output at 96k. I wanted to see if I could hear a difference. This is by no means a scientific A vs B, or a double blind test. It was just me and my memory. What I did notice was that there was a difference going from 44.1 to 96. The 96 was a little more revealing, there seemed to be a little more precision in the sounds, especially in the mid-range and upper frequencies. It also seemed to loose a little bit of smoothness, the music just seemed a little more edgy. The sound at times was better and sometimes fatiguing depending on the original source (rips from Audio Fidelity and Mofi discs were a little better) I then from curiosity switched the J river from 96k to 192k. I heard no difference between the 2 settings. The differences I heard were 44.1 to 96 were noticeable but not overly significant. The major conclusion I made was that if the original rip is from a badly mastered disc, 96k will reveal more of its bad sound. I grew up on vinyl so I wonder if a more revealing sound (not as warm per say like on older vinyl) has clouded my perception of what is the proper/best sound ?? IMHO the original mastering makes all the difference in the world regardless of the settings. Keith, maybe my post wasn't written as well as it could've been. In the video I posted Dr. Aix at one point stated that some of the Hi-res files available were just up-converted. That was the point I was trying to make or hear if there was a difference letting J river up convert it. |
|
|
|
Post by monkumonku on Oct 8, 2015 9:21:45 GMT -5
Maybe I am not understanding this correctly but if the source is at 44.1/16, I don't see how going to a higher bit rate can improve anything because it is still subject to the limitations of the source. That would be like playing an mp3 file recorded at 128 or 256 and then upscaling it to 96/24. You'd still be at the mercy of the original mp3 file limitations, wouldn't you? Yes you are correct. One of the things dr. Aix said was that some of the Hi-Res re-masters were just basically up converted, so I was trying it myself to see if there was a difference. Maybe I should've been a little more specific in my post ? Thanks, I understand now what you were getting at. Keith posted a good explanation just now. If you do hear a difference then it is due to imperfections in the upconversion process, not anything having to do with the actual music file. By "imperfection" I mean something that interferes with a 100% accurate transfer of the signal (which imperfections may make it sound better or worse to your ears). |
|
|
|
Post by garym on Oct 8, 2015 9:36:51 GMT -5
I thought the takeaway point in Waldref's talk was: You can't get more information from a studio master tape (or other source) than was recorded on it. That point should be kept in mind when buying hi-rez releases of anything originally recorded on analog tape, especially recordings made 40-50 years ago (e.g., the Rolling Stones, Beatles, Pink Floyd, etc. etc.). Such material will have been recorded on Ampex or Studer recorders that had an upper frequency response of 15K (+/- 3db) and a dynamic range of perhaps 60-70 db, per the manufacturer's specs. A 44.1/16 CD will easily accommodate that recording, and adding more "empty bits" will add nothing to the music delivered.
(I do agree with Keith's point that a new mastering can sometimes be better, but it will only be as good as the original tapes, and if originally recorded on analog machines in the 50s, 60s, 70s would sound just as good at 44.1/16).
Also agreed with his point about the new "Hi-Rez" logos --- a recording should not carry that claim unless it was recorded at 96/24 (minimum) from start to finish, and all the analog components used in the recording. from microphones to amps, can manage that frequency and dynamic range.
The scam of upsampling CDs to "Hi-Rez" was especially outrageous.
|
|
KeithL
Administrator
Posts: 10,273
|
Post by KeithL on Oct 8, 2015 11:53:31 GMT -5
You need to read the actual claims made for each particular release very carefully..... because there are actually a bunch of things that could have been done - or not..... (some of which do in fact make it possible for a new re-master of even a mediocre original tape to actually be significantly better.) 1) They could have started with a digital master, intended for a CD, and made at 44k, and done a straightforward up-sample to it. If that's the case, then simply converting that 44k copy to 96k or 192k isn't going to add any information; and so, if it sounds different, it must be either because of some slight alteration introduced by the conversion itself, or because your DAC just plain sounds slightly different at different sample rates. 2) They could have started with a digital master, intended for a CD or DVD, and made at 44k or 48k, and done a " FANCY" up-sample to it that involves some processing. One example of this would be to upsample a CD or DVD audio track using Dolby's Pro Encoder, and invoking their "apodizing filter" option. This option is intended to "correct" some of the "errors" introduced by early A/D converters. (It takes in a 44k or 48k original and outputs a MODIFIED 96k result.) In other words, it really does alter the sound in ways beyond simply upsampling it. If that's the case, then the 96k version might sound better - not because of the upsampling, but because of the other stuff they're doing. Likewise, they could have adjusted the EQ or other basic settings as well as simply changing the sample rate (in other words, maybe they made some other changes in their editing console besides the sample rate). If so, then it could sound better, worse, or the same (which means that it really could sound better). 3) They could have started with the same two-channel analog master tape that was used to make the CD and simply done a new analog to digital conversion at a higher sample rate. If that's the case, then whether there is any difference will, of course, depend on whether that original tape was of high enough quality that 44k wasn't "good enough" to properly contain all the information on the tape without some loss. This is probably the most common situation, and the one where, if the tape is somewhat old, or simply not of the absolute best quality, you shouldn't expect much if any difference. Even with a superb quality tape, depending on the microphones and other equipment used, there might simply not be any information present above what can be reproduced perfectly by an ordinary 44k CD - in which case re-converting it at a higher sample rate isn't going to make any difference. 4) They also could have started with the same multi-track analog master tape, and entirely re-mixed it - which used to be the implied meaning of "re-mastering". In that case, even though the individual tracks on the tape may be of limited quality, it's still possible that they've done a much better (or worse) job re-mixing them. The tracks may be mixed very differently, and various other modifications could have been made - ranging from substituting modern versions of specific effects that existed on the original (perhaps using a better modern reverb plugin on a track where a poor quality one was used originally), to making actual repairs and corrections to individual tracks (such as correcting speed variations or other aberrations present in the original tapes). Depending on what was done, the end result could be quite different, and possibly better, than the original - even if the original master tapes weren't of the best quality to begin with. (So, in this situation, the fact that the original master tapes were of relatively limited quality doesn't preclude the possibility that a new re-mastered version of the album could still be better than the original one - possibly even good enough to benefit from a higher sample rate.) One example of #4 is the remastered versions of the Grateful Dead studio albums. Even though the original master tapes of many of the albums were not of especially good quality, during the remastering process all sorts of repairs and alterations were performed on some of them. Special processing was applied to actually fix speed variations and dropouts present on some of the original master tapes, and the mixing was re-done. So, in that particular case, the "high-res remasters" are actually quite different, and of arguably much better audio quality, than even the original master tapes (they sound very different and I, for one, find the difference to be an improvement). My point here is that you really need to evaluate the claims and expectations for each release separately. You definitely shouldn't assume that a high-resolution reissue is going to be better. However, you also shouldn't make a blanket assumption that it can't possibly be better because the original master tapes simply weren't good enough. (Unfortunately, while a lot of documentation was actually provided about what was done to remaster the Grateful Dead albums, and is with many other albums as well, many releases are simply offered with few details, or none at all, about what was done - and, in those situations, it certainly does make sense to be skeptical about expecting an improvement.) Sadly, at the moment, "high-resolution" is about as meaningful as "new and improved" - for pretty much the same reasons. I thought the takeaway point in Waldref's talk was: You can't get more information from a studio master tape (or other source) than was recorded on it. That point should be kept in mind when buying hi-rez releases of anything originally recorded on analog tape, especially recordings made 40-50 years ago (e.g., the Rolling Stones, Beatles, Pink Floyd, etc. etc.). Such material will have been recorded on Ampex or Studer recorders that had an upper frequency response of 15K (+/- 3db) and a dynamic range of perhaps 60-70 db, per the manufacturer's specs. A 44.1/16 CD will easily accommodate that recording, and adding more "empty bits" will add nothing to the music delivered. (I do agree with Keith's point that a new mastering can sometimes be better, but it will only be as good as the original tapes, and if originally recorded on analog machines in the 50s, 60s, 70s would sound just as good at 44.1/16). Also agreed with his point about the new "Hi-Rez" logos --- a recording should not carry that claim unless it was recorded at 96/24 (minimum) from start to finish, and all the analog components used in the recording. from microphones to amps, can manage that frequency and dynamic range. The scam of upsampling CDs to "Hi-Rez" was especially outrageous. |
|
|
|
Post by lionear on Oct 8, 2015 11:54:53 GMT -5
This has been a very interesting thread. I see that @keithl has addressed Dr AIX's "heard from a friend, who heard from a friend" story. Am I correct that @keithl has also addressed Dr AIX's mistake in converting a digital file to a higher sampling rate, when he should have take a high sampled file, and converted it to a lower sampling rate? If so, can we agree that Dr AIX is one of the worst sources of information on audio? In academia, a career can be compromised if one makes ONE big mistake in a published article. Dr AIX has made two in the context of this thread alone. When I read stuff on Dr AIX's website, I'm struck by the number of things that (to use a phrase coined by mathematicians to criticize a flawed argument) "aren't even wrong". Sometimes I want to write about the flaws in Dr AIX's arguments, but they're so bad that I don't even know where to start. On the subject of sampling rates, in the 1980's (or may be early 1990's), the late Doug Sax (very famous mastering engineer) complained in an interview in Audio Magazine about the sampling rate being too low in the CD standard. He said he had heard digital music sampled at 2 million times a second - and it sounded better. Simply increasing the sampling rate might not be the answer, either. Meridian has a new approach: www.theabsolutesound.com/articles/beyond-high-resolution/I think it's the same idea they had with HDCD. Metadata is added to music data, and this is acted on by a DAC to shape the music waveform to address certain issues. I've heard Sony's waveform shaping technology on the HAP-Z1ES and it made MP3 files sound very, very good. |
|
|
|
Post by lionear on Oct 8, 2015 12:40:32 GMT -5
I thought the takeaway point in Waldref's talk was: You can't get more information from a studio master tape (or other source) than was recorded on it. That point should be kept in mind when buying hi-rez releases of anything originally recorded on analog tape, especially recordings made 40-50 years ago (e.g., the Rolling Stones, Beatles, Pink Floyd, etc. etc.). Such material will have been recorded on Ampex or Studer recorders that had an upper frequency response of 15K (+/- 3db) and a dynamic range of perhaps 60-70 db, per the manufacturer's specs. A 44.1/16 CD will easily accommodate that recording, and adding more "empty bits" will add nothing to the music delivered. (I do agree with Keith's point that a new mastering can sometimes be better, but it will only be as good as the original tapes, and if originally recorded on analog machines in the 50s, 60s, 70s would sound just as good at 44.1/16). Also agreed with his point about the new "Hi-Rez" logos --- a recording should not carry that claim unless it was recorded at 96/24 (minimum) from start to finish, and all the analog components used in the recording. from microphones to amps, can manage that frequency and dynamic range. The scam of upsampling CDs to "Hi-Rez" was especially outrageous. Sometimes you CAN get "more" out of something than you put in. The secret to Ansel Adam's photography was his realization that a photo negative didn't have the "range" of real life, and that photographic paper had a higher "range" than the photo negative. There were also a few non-linearities with negatives and photo paper and chemicals. Adams exploited these non-linearities to get more out of the final print (so that the result was closer to real life) than what the individual parts could deliver if used in a purely linear fashion. (When it comes to digital photography, when I look at a straight RAW file, most of the time, my reaction is "Ugh!". But after manipulating a few settings, the result is much better. The RAW file is more "accurate" - but it's not aesthetically pleasing, and it's not a true reflection of the aesthetics of the image - of what I saw in my mind when I took the photo. Making the adjustments allows me to compensate for the limitations of the computer screen and "get it back". When I make a print from the image, there's another round of adjusting to compensate for the limitations of the printer and the paper.) I think the same is true of the analog music process. Recording engineers will, for example, set things so that the tape will not reach its magnetic saturation point. They're not worried about much else. If you listen to a master tape that was recorded for LP production, you might not be all that impressed with the sound. But the interesting thing is that you can "get it back" during the manufacturing process. This might extend to the pressing process. After you cut the master (the "mother"), you have to create the "father" and then do the actual vinyl pressing with the father. All that changes the grooves in a certain way.... and perhaps that can influence the way the mother is cut in the first place. One of the areas that can be controlled is the shape of the RIAA curve used for the cutting head. There are also the characteristics of the cutter's amps that could be exploited (for beneficial effects) or compensated for (for undesirable effects), via the actual cutting and stamping process. (Aside: someone from the Grateful Dead band used to go to the pressing plants and listen to the records, to make sure that the LP was being made correctly.) I think one issue with digital music is that the process does change the sound (no process is "perfect" - that's a no-brainer), but it's tough to figure out how it changes the sound, so that we can apply the inverse or take it into account so that we can "get it back" at playback. Digital music is very accurate in the frequency domain. But is frequency fidelity that all we need to consider? (Meridian seems to be addressing some new digital stuff with their new format.) |
|
|
|
Post by monkumonku on Oct 8, 2015 13:13:22 GMT -5
I thought the takeaway point in Waldref's talk was: You can't get more information from a studio master tape (or other source) than was recorded on it. That point should be kept in mind when buying hi-rez releases of anything originally recorded on analog tape, especially recordings made 40-50 years ago (e.g., the Rolling Stones, Beatles, Pink Floyd, etc. etc.). Such material will have been recorded on Ampex or Studer recorders that had an upper frequency response of 15K (+/- 3db) and a dynamic range of perhaps 60-70 db, per the manufacturer's specs. A 44.1/16 CD will easily accommodate that recording, and adding more "empty bits" will add nothing to the music delivered. (I do agree with Keith's point that a new mastering can sometimes be better, but it will only be as good as the original tapes, and if originally recorded on analog machines in the 50s, 60s, 70s would sound just as good at 44.1/16). Also agreed with his point about the new "Hi-Rez" logos --- a recording should not carry that claim unless it was recorded at 96/24 (minimum) from start to finish, and all the analog components used in the recording. from microphones to amps, can manage that frequency and dynamic range. The scam of upsampling CDs to "Hi-Rez" was especially outrageous. Sometimes you CAN get "more" out of something than you put in. The secret to Ansel Adam's photography was his realization that a photo negative didn't have the "range" of real life, and that photographic paper had a higher "range" than the photo negative. There were also a few non-linearities with negatives and photo paper and chemicals. Adams exploited these non-linearities to get more out of the final print (so that the result was closer to real life) than what the individual parts could deliver if used in a purely linear fashion. (When it comes to digital photography, when I look at a straight RAW file, most of the time, my reaction is "Ugh!". But after manipulating a few settings, the result is much better. The RAW file is more "accurate" - but it's not aesthetically pleasing, and it's not a true reflection of the aesthetics of the image - of what I saw in my mind when I took the photo. Making the adjustments allows me to compensate for the limitations of the computer screen and "get it back". When I make a print from the image, there's another round of adjusting to compensate for the limitations of the printer and the paper.) I think the same is true of the analog music process. Recording engineers will, for example, set things so that the tape will not reach its magnetic saturation point. They're not worried about much else. If you listen to a master tape that was recorded for LP production, you might not be all that impressed with the sound. But the interesting thing is that you can "get it back" during the manufacturing process. This might extend to the pressing process. After you cut the master (the "mother"), you have to create the "father" and then do the actual vinyl pressing with the father. All that changes the grooves in a certain way.... and perhaps that can influence the way the mother is cut in the first place. One of the areas that can be controlled is the shape of the RIAA curve used for the cutting head. There are also the characteristics of the cutter's amps that could be exploited (for beneficial effects) or compensated for (for undesirable effects), via the actual cutting and stamping process. (Aside: someone from the Grateful Dead band used to go to the pressing plants and listen to the records, to make sure that the LP was being made correctly.) I think one issue with digital music is that the process does change the sound (no process is "perfect" - that's a no-brainer), but it's tough to figure out how it changes the sound, so that we can apply the inverse or take it into account so that we can "get it back" at playback. Digital music is very accurate in the frequency domain. But is frequency fidelity that all we need to consider? (Meridian seems to be addressing some new digital stuff with their new format.) But upsampling or upconverting a digital music file is not going to in itself improve anything and I believe that is the point Mark Waldrep is trying to make - that you can't get something from nothing. If you do other procedures like remixing or remastering the original tape or alter the sound in some way, then that can result in an improvement but if the original recording at 44.1/16 is upsampled to 96/24 without doing anything to modify the bytes on the original file, then maybe technically it can be called hi-res but there is no real improvement over the lower-res version. |
|
|
|
Post by geebo on Oct 8, 2015 13:31:39 GMT -5
Sometimes you CAN get "more" out of something than you put in. The secret to Ansel Adam's photography was his realization that a photo negative didn't have the "range" of real life, and that photographic paper had a higher "range" than the photo negative. There were also a few non-linearities with negatives and photo paper and chemicals. Adams exploited these non-linearities to get more out of the final print (so that the result was closer to real life) than what the individual parts could deliver if used in a purely linear fashion. (When it comes to digital photography, when I look at a straight RAW file, most of the time, my reaction is "Ugh!". But after manipulating a few settings, the result is much better. The RAW file is more "accurate" - but it's not aesthetically pleasing, and it's not a true reflection of the aesthetics of the image - of what I saw in my mind when I took the photo. Making the adjustments allows me to compensate for the limitations of the computer screen and "get it back". When I make a print from the image, there's another round of adjusting to compensate for the limitations of the printer and the paper.) I think the same is true of the analog music process. Recording engineers will, for example, set things so that the tape will not reach its magnetic saturation point. They're not worried about much else. If you listen to a master tape that was recorded for LP production, you might not be all that impressed with the sound. But the interesting thing is that you can "get it back" during the manufacturing process. This might extend to the pressing process. After you cut the master (the "mother"), you have to create the "father" and then do the actual vinyl pressing with the father. All that changes the grooves in a certain way.... and perhaps that can influence the way the mother is cut in the first place. One of the areas that can be controlled is the shape of the RIAA curve used for the cutting head. There are also the characteristics of the cutter's amps that could be exploited (for beneficial effects) or compensated for (for undesirable effects), via the actual cutting and stamping process. (Aside: someone from the Grateful Dead band used to go to the pressing plants and listen to the records, to make sure that the LP was being made correctly.) I think one issue with digital music is that the process does change the sound (no process is "perfect" - that's a no-brainer), but it's tough to figure out how it changes the sound, so that we can apply the inverse or take it into account so that we can "get it back" at playback. Digital music is very accurate in the frequency domain. But is frequency fidelity that all we need to consider? (Meridian seems to be addressing some new digital stuff with their new format.) Ah, but if you took a picture of an Adams print with a modern high resolution high dynamic range digital camera would that in itself make the picture better? Of course it wouldn't. What Ansel did was more like post processing. |
|
|
|
Post by yves on Oct 8, 2015 13:56:23 GMT -5
I thought the takeaway point in Waldref's talk was: You can't get more information from a studio master tape (or other source) than was recorded on it. That point should be kept in mind when buying hi-rez releases of anything originally recorded on analog tape, especially recordings made 40-50 years ago (e.g., the Rolling Stones, Beatles, Pink Floyd, etc. etc.). Such material will have been recorded on Ampex or Studer recorders that had an upper frequency response of 15K (+/- 3db) and a dynamic range of perhaps 60-70 db, per the manufacturer's specs. A 44.1/16 CD will easily accommodate that recording, and adding more "empty bits" will add nothing to the music delivered. (I do agree with Keith's point that a new mastering can sometimes be better, but it will only be as good as the original tapes, and if originally recorded on analog machines in the 50s, 60s, 70s would sound just as good at 44.1/16). Also agreed with his point about the new "Hi-Rez" logos --- a recording should not carry that claim unless it was recorded at 96/24 (minimum) from start to finish, and all the analog components used in the recording. from microphones to amps, can manage that frequency and dynamic range. The scam of upsampling CDs to "Hi-Rez" was especially outrageous. The "a 16bit 44.1kHz CD will accomodate that recording" remark is just an old myth that should simply have died a full quarter of a century ago. www.aes.org/e-lib/browse.cfm?elib=17497 |
|
|
|
Post by yves on Oct 8, 2015 14:03:16 GMT -5
Sometimes you CAN get "more" out of something than you put in. The secret to Ansel Adam's photography was his realization that a photo negative didn't have the "range" of real life, and that photographic paper had a higher "range" than the photo negative. There were also a few non-linearities with negatives and photo paper and chemicals. Adams exploited these non-linearities to get more out of the final print (so that the result was closer to real life) than what the individual parts could deliver if used in a purely linear fashion. (When it comes to digital photography, when I look at a straight RAW file, most of the time, my reaction is "Ugh!". But after manipulating a few settings, the result is much better. The RAW file is more "accurate" - but it's not aesthetically pleasing, and it's not a true reflection of the aesthetics of the image - of what I saw in my mind when I took the photo. Making the adjustments allows me to compensate for the limitations of the computer screen and "get it back". When I make a print from the image, there's another round of adjusting to compensate for the limitations of the printer and the paper.) I think the same is true of the analog music process. Recording engineers will, for example, set things so that the tape will not reach its magnetic saturation point. They're not worried about much else. If you listen to a master tape that was recorded for LP production, you might not be all that impressed with the sound. But the interesting thing is that you can "get it back" during the manufacturing process. This might extend to the pressing process. After you cut the master (the "mother"), you have to create the "father" and then do the actual vinyl pressing with the father. All that changes the grooves in a certain way.... and perhaps that can influence the way the mother is cut in the first place. One of the areas that can be controlled is the shape of the RIAA curve used for the cutting head. There are also the characteristics of the cutter's amps that could be exploited (for beneficial effects) or compensated for (for undesirable effects), via the actual cutting and stamping process. (Aside: someone from the Grateful Dead band used to go to the pressing plants and listen to the records, to make sure that the LP was being made correctly.) I think one issue with digital music is that the process does change the sound (no process is "perfect" - that's a no-brainer), but it's tough to figure out how it changes the sound, so that we can apply the inverse or take it into account so that we can "get it back" at playback. Digital music is very accurate in the frequency domain. But is frequency fidelity that all we need to consider? (Meridian seems to be addressing some new digital stuff with their new format.) and I believe that is the point Mark Waldrep is trying to make - that you can't get something from nothing. If you do other procedures like remixing or remastering the original tape or alter the sound in some way, then that can result in an improvement but if the original recording at 44.1/16 is upsampled to 96/24 without doing anything to modify the bytes on the original file, then maybe technically it can be called hi-res but there is no real improvement over the lower-res version. As a matter of true fact, Bob Stuart of Meridian Audio *is* a smart enough guy for him to be able to get something from nothing. www.dolby.com/us/en/technologies/dolby-truehd-encoder-white-paper.pdf |
|
|
|
Post by monkumonku on Oct 8, 2015 14:26:08 GMT -5
But upsampling or upconverting a digital music file is not going to in itself improve anything and I believe that is the point Mark Waldrep is trying to make - that you can't get something from nothing. If you do other procedures like remixing or remastering the original tape or alter the sound in some way, then that can result in an improvement but if the original recording at 44.1/16 is upsampled to 96/24 without doing anything to modify the bytes on the original file, then maybe technically it can be called hi-res but there is no real improvement over the lower-res version. As a matter of true fact, Bob Stuart of Meridian Audio *is* a smart enough guy for him to be able to get something from nothing. www.dolby.com/us/en/technologies/dolby-truehd-encoder-white-paper.pdfAdmittedly I am not well versed in this technology but in reading the white paper you referenced, there appear to be two components: upsampling from 48 to 96, and introducing what they call an "apodizing filter." It is this latter item that supposedly improves the recording. My question would be if this filter can be applied without upsampling. If so, then it is not really a function of higher resolution, but the filter itself that changes the file. Even if upsampling is required, they are not really getting something from nothing. What they are doing is removing something to make it nothing. I think there is a substantial difference between these two conditions. |
|
|
|
Post by yves on Oct 8, 2015 14:34:58 GMT -5
Sometimes you CAN get "more" out of something than you put in. The secret to Ansel Adam's photography was his realization that a photo negative didn't have the "range" of real life, and that photographic paper had a higher "range" than the photo negative. There were also a few non-linearities with negatives and photo paper and chemicals. Adams exploited these non-linearities to get more out of the final print (so that the result was closer to real life) than what the individual parts could deliver if used in a purely linear fashion. (When it comes to digital photography, when I look at a straight RAW file, most of the time, my reaction is "Ugh!". But after manipulating a few settings, the result is much better. The RAW file is more "accurate" - but it's not aesthetically pleasing, and it's not a true reflection of the aesthetics of the image - of what I saw in my mind when I took the photo. Making the adjustments allows me to compensate for the limitations of the computer screen and "get it back". When I make a print from the image, there's another round of adjusting to compensate for the limitations of the printer and the paper.) I think the same is true of the analog music process. Recording engineers will, for example, set things so that the tape will not reach its magnetic saturation point. They're not worried about much else. If you listen to a master tape that was recorded for LP production, you might not be all that impressed with the sound. But the interesting thing is that you can "get it back" during the manufacturing process. This might extend to the pressing process. After you cut the master (the "mother"), you have to create the "father" and then do the actual vinyl pressing with the father. All that changes the grooves in a certain way.... and perhaps that can influence the way the mother is cut in the first place. One of the areas that can be controlled is the shape of the RIAA curve used for the cutting head. There are also the characteristics of the cutter's amps that could be exploited (for beneficial effects) or compensated for (for undesirable effects), via the actual cutting and stamping process. (Aside: someone from the Grateful Dead band used to go to the pressing plants and listen to the records, to make sure that the LP was being made correctly.) I think one issue with digital music is that the process does change the sound (no process is "perfect" - that's a no-brainer), but it's tough to figure out how it changes the sound, so that we can apply the inverse or take it into account so that we can "get it back" at playback. Digital music is very accurate in the frequency domain. But is frequency fidelity that all we need to consider? (Meridian seems to be addressing some new digital stuff with their new format.) Sometimes you CAN get "more" out of something than you put in. The secret to Ansel Adam's photography was his realization that a photo negative didn't have the "range" of real life, and that photographic paper had a higher "range" than the photo negative. There were also a few non-linearities with negatives and photo paper and chemicals. Adams exploited these non-linearities to get more out of the final print (so that the result was closer to real life) than what the individual parts could deliver if used in a purely linear fashion. (When it comes to digital photography, when I look at a straight RAW file, most of the time, my reaction is "Ugh!". But after manipulating a few settings, the result is much better. The RAW file is more "accurate" - but it's not aesthetically pleasing, and it's not a true reflection of the aesthetics of the image - of what I saw in my mind when I took the photo. Making the adjustments allows me to compensate for the limitations of the computer screen and "get it back". When I make a print from the image, there's another round of adjusting to compensate for the limitations of the printer and the paper.) I think the same is true of the analog music process. Recording engineers will, for example, set things so that the tape will not reach its magnetic saturation point. They're not worried about much else. If you listen to a master tape that was recorded for LP production, you might not be all that impressed with the sound. But the interesting thing is that you can "get it back" during the manufacturing process. This might extend to the pressing process. After you cut the master (the "mother"), you have to create the "father" and then do the actual vinyl pressing with the father. All that changes the grooves in a certain way.... and perhaps that can influence the way the mother is cut in the first place. One of the areas that can be controlled is the shape of the RIAA curve used for the cutting head. There are also the characteristics of the cutter's amps that could be exploited (for beneficial effects) or compensated for (for undesirable effects), via the actual cutting and stamping process. (Aside: someone from the Grateful Dead band used to go to the pressing plants and listen to the records, to make sure that the LP was being made correctly.) I think one issue with digital music is that the process does change the sound (no process is "perfect" - that's a no-brainer), but it's tough to figure out how it changes the sound, so that we can apply the inverse or take it into account so that we can "get it back" at playback. Digital music is very accurate in the frequency domain. But is frequency fidelity that all we need to consider? (Meridian seems to be addressing some new digital stuff with their new format.) Ah, but if you took a picture of an Adams print with a modern high resolution high dynamic range digital camera would that in itself make the picture better? Of course it wouldn't. What Ansel did was more like post processing. Remember that almost every modern DAC unit relies on heavy upsampling, in order to both factually and measurably improve the *accuracy* of the analog output of the DAC unit. |
|
|
|
Post by geebo on Oct 8, 2015 14:39:44 GMT -5
Ah, but if you took a picture of an Adams print with a modern high resolution high dynamic range digital camera would that in itself make the picture better? Of course it wouldn't. What Ansel did was more like post processing. Remember that almost every modern DAC unit relies on heavy upsampling, in order to both factually and measurably improve the *accuracy* of the analog output of the DAC unit. That has nothing to do with taking a 44.1/16 file and converting it to 192/24 and doing *nothing else* and claiming it now sounds better. |
|
KeithL
Administrator
Posts: 10,273
|
Post by KeithL on Oct 8, 2015 15:36:31 GMT -5
I thought the takeaway point in Waldref's talk was: You can't get more information from a studio master tape (or other source) than was recorded on it. That point should be kept in mind when buying hi-rez releases of anything originally recorded on analog tape, especially recordings made 40-50 years ago (e.g., the Rolling Stones, Beatles, Pink Floyd, etc. etc.). Such material will have been recorded on Ampex or Studer recorders that had an upper frequency response of 15K (+/- 3db) and a dynamic range of perhaps 60-70 db, per the manufacturer's specs. A 44.1/16 CD will easily accommodate that recording, and adding more "empty bits" will add nothing to the music delivered. (I do agree with Keith's point that a new mastering can sometimes be better, but it will only be as good as the original tapes, and if originally recorded on analog machines in the 50s, 60s, 70s would sound just as good at 44.1/16). Also agreed with his point about the new "Hi-Rez" logos --- a recording should not carry that claim unless it was recorded at 96/24 (minimum) from start to finish, and all the analog components used in the recording. from microphones to amps, can manage that frequency and dynamic range. The scam of upsampling CDs to "Hi-Rez" was especially outrageous. Sometimes you CAN get "more" out of something than you put in. The secret to Ansel Adam's photography was his realization that a photo negative didn't have the "range" of real life, and that photographic paper had a higher "range" than the photo negative. There were also a few non-linearities with negatives and photo paper and chemicals. Adams exploited these non-linearities to get more out of the final print (so that the result was closer to real life) than what the individual parts could deliver if used in a purely linear fashion. (When it comes to digital photography, when I look at a straight RAW file, most of the time, my reaction is "Ugh!". But after manipulating a few settings, the result is much better. The RAW file is more "accurate" - but it's not aesthetically pleasing, and it's not a true reflection of the aesthetics of the image - of what I saw in my mind when I took the photo. Making the adjustments allows me to compensate for the limitations of the computer screen and "get it back". When I make a print from the image, there's another round of adjusting to compensate for the limitations of the printer and the paper.) I think the same is true of the analog music process. Recording engineers will, for example, set things so that the tape will not reach its magnetic saturation point. They're not worried about much else. If you listen to a master tape that was recorded for LP production, you might not be all that impressed with the sound. But the interesting thing is that you can "get it back" during the manufacturing process. This might extend to the pressing process. After you cut the master (the "mother"), you have to create the "father" and then do the actual vinyl pressing with the father. All that changes the grooves in a certain way.... and perhaps that can influence the way the mother is cut in the first place. One of the areas that can be controlled is the shape of the RIAA curve used for the cutting head. There are also the characteristics of the cutter's amps that could be exploited (for beneficial effects) or compensated for (for undesirable effects), via the actual cutting and stamping process. (Aside: someone from the Grateful Dead band used to go to the pressing plants and listen to the records, to make sure that the LP was being made correctly.) I think one issue with digital music is that the process does change the sound (no process is "perfect" - that's a no-brainer), but it's tough to figure out how it changes the sound, so that we can apply the inverse or take it into account so that we can "get it back" at playback. Digital music is very accurate in the frequency domain. But is frequency fidelity that all we need to consider? (Meridian seems to be addressing some new digital stuff with their new format.) I agree with a lot of what you said.... however I think there's more going on. Yes, it's true that any recording process will probably change the sound to some degree, and it's also true that the type of changes you get with digital recordings are unfamiliar to most people, which makes them more difficult to quantify and understand. (It also encourages some people who are obsessed with only certain specifications to claim that, as long as something measures well in those specific parameters, that it must sound good - and that anyone who claims to hear something else must be imagining it.) However, I think the biggest problem with how people perceive the accuracy of digital recordings is more due to what I would call the symmetry of the means of comparison we have available. Here's what I mean... I've heard any number of situations where someone will make a digital recording of a vinyl album, then play it back and compare it to the original vinyl album; with the inevitable result that they hear at least tiny differences; which they then quote as proof that "digital isn't as good as analog because I can hear a difference". However, what they don't (and can't) do is to make a vinyl recording of a digital original, and see how much difference the vinyl recording process has introduced. Neither can they make both a digital copy and an analog copy of the actual original and compare how great the differences are between them. Therefore, they never get to see a direct comparison of the imperfections of the analog recording process as against the imperfections of the digital process. (So, what they're really doing is to compare a copy to a copy of a copy, which provides no information whatsoever about how close to the original the first copy was to begin with.) As for the possibility that various adjustments in the recording process could actually end up delivering something closer to the real original performance than the master tape - I agree entirely. I rather suspect that a lot of the complaints people have about "digital recordings just not sounding natural" are simply the result of the fact that most modern recordings are miked rather closely, which tends to eliminate a lot of the room ambience (in studio terms they are "very dry"); the vinyl recording and playback process, as well as a lot of tube equipment (which is often also claimed to sound "more natural"), then adds a bit of second harmonic distortion, and some other audible imperfections, which substitutes reasonably well for the missing ambience, and so you end up with a recording that sounds "closer to the real thing than the master itself". However, the fact that this occurs with a vinyl recording and not a digital one is simply a fortuitous accident.... and, once we agree that we want such an alteration to be applied to our master, we could also apply it to a digital recording. (It's still distortion, which makes it a bad recording, unless that distortion is intentionally added by the mastering engineer to achieve a specific sound - in which case it is simply part of the recording.) I'd have to say that, for the most part, everything I've read by Dr Aix seems to be "heading in the right direction", and that I agree with his basic premise that there's an awful lot of hype in the "high-res market" today, which we would all be better off without. He's also probably right that at least a few recordings being sold as high-res really are just copies of 44k files that were directly upsampled (in which case any difference or improvement you hear is really just an artifact of the particular conversion software used). However, I don't think that this is a major problem with most of the reputable vendors. (I recall a major scandal when HDTracks sold a 24/96k version of a certain Fleetwood Mac album which was later determined to simply be an upsampled copy of a 44k original; it was a scandal; it was claimed to have been an error; apologies were made; proper 96k copies were given to the purchasers. And I'm sure they and all the other vendors prefer very strongly to avoid a repeat of that performance.) I also agree that many remasters are simply re-sampled copies made from original tapes whose quality is such that it isn't going to matter. (However, that's not the same thing as lying about it; even if all you're getting is a better fidelity recording of the ultrasonic noise on a master tape with no audio on it above 15 kHz.) In other words, while I think he's sometimes a bit sloppy, and sort of blurs a few lines, his overall premise that a lot of high-res recordings really aren't significantly better than their 44k counterparts is true... and so his basic claim that we shouldn't assume that they are is quite valid. He's also got a point that it's best to have high-res recordings with proper provenance, and proof that they were recorded and mastered using "a fully high-res workflow". (And, as its the case with those Grateful Dead albums, if some special processing was done on otherwise middling masters to justify reissuing them in high-resolution, then it should be well documented.) I'm not sure that I agree with garym's stand about "only calling something high-res if it was mastered that way". After all, plenty of vintage VHS tapes have been reissued on DVDs, and then again on Blu-Ray discs, where the actual quality of the content was clearly inferior, and it was really just a matter of reissuing them "on the latest format" pro-forma. Personally, I feel that providing the actual information is sufficient to allow the purchaser to make an informed decision; I know not to expect full HD quality on a Blu-Ray reissue of a 1965 movie, and likewise, I know not to expect full modern high-resolution quality from an album that was mastered in 1955. However, just as it's the responsibility of the producers to inform me if they've carefully restored that old videotape, it's up to the music vendor to convince me that their high-res reissue of that old album has really been improved or restored in some way... and, if they don't , then I probably won't be buying it. What you're saying about "frequency fidelity" is quite true, although that's not usually the term used... Digital processes, including those employed internally by DACs, operate in both frequency and time. Many digital filters in particular may produce errors where the correct amount of energy is delivered at the proper frequencies, but some of it is delivered at the wrong times. (This often appears as pre-ringing and post-ringing.) As a result, may DACs have a very flat measured frequency response, and very low distortion (because very low levels of extra harmonics are generated), but produce waveforms that are incorrect when presented with transient signals. These errors are not shown by traditional frequency response and THD measurements, but are quite obvious when you examine transient response. (The problem is that people who are used to the measurements used for analog devices tend to believe that measuring frequency response, S/N, and THD, tell you everything you need to know about a device - which is somewhat true for analog devices but not for digital devices.) As for the new Meridian compression.... it isn't actually similar to HDCD (as a lot of people in various forums seem to think). With HDCD, the actual signal processing used is a compander (the audio is range-compressed when you record it, and expanded back to its original dynamic range during playback). However, the biggest problem with this type of system is that the compression and expansion fail to track well. HDCD solves this issue by actually recording the compression setting used, from second to second, in a sort of metadata track. This information is then used to control the expander, thus ensuring that it tracks the original compression perfectly. (The metadata is then encoded and embedded in the LSB of the digital audio.) Meridian's new system is quite different (according to their description - which I'm heavily paraphrasing below)..... If you start with a 24/192k audio file, that file will be able to store audio from 20 Hz to about 90 kHz, and from 0 dB all the way down to about -140 dB. If you look at that overall range graphically, you can divide it into areas, and notice that some of the areas matter, while others don't. For example, we all agree that the range between 20 Hz and 20 kHz, and between 0 dB and -96 dB is important (that's the range you can store on a Red Book CD). Meridian is also convinced that the range between 20 kHz and 50 kHz or so, and between 0 dB and some lower bound is important. And we all probably agree that the range between 20 Khz and 90 kHz, and between -90 dB and -140 dB ISN'T terribly important. According to Meridian's description, their new CODEC basically uses lossless compression to compress the significant ranges there, then they discard some of the ranges "that contain only noise and no audio", and "fold" the other ranges into that now-available space. What this means, in English, is that their overall CODEC is in fact lossy, but it operates by losslessly encoding the important frequency and amplitude ranges, and discarding the "useless" ranges entirely to make space to store them. (They refer to this whole thing as "audio oregami - to go with the references to cutting and folding.... and they seem to manage to avoid the term "lossy" in their descriptions..... which seems somewhat misleading to me.) The basic upshot of all this is that they give you most of the useful extra information that you would get in a lossless high-res file, including what they claim is somewhat more than you would get in a normal 24/192k file, and squeeze it all into a file not much bigger than a Red Book CD quality file would normally use. (And, in case you noticed that I sound cynical about their phrasing of the explanation, it's only fair to say that I DO trust Meridian's technical chops, and so I'm sure it actually works very well.) It also seems most likely that the savings in file storage space isn't going to be enough to attract most file-download fans to the format, although that will somewhat depend on whatever other downsides it might have (like licensing costs), but that the new CODEC will be quite attractive to high-res streaming services (Tidal is reported to be on board with it). |
|
|
|
Post by yves on Oct 8, 2015 16:16:13 GMT -5
Admittedly I am not well versed in this technology but in reading the white paper you referenced, there appear to be two components: upsampling from 48 to 96, and introducing what they call an "apodizing filter." It is this latter item that supposedly improves the recording. My question would be if this filter can be applied without upsampling. If so, then it is not really a function of higher resolution, but the filter itself that changes the file. Even if upsampling is required, they are not really getting something from nothing. What they are doing is removing something to make it nothing. I think there is a substantial difference between these two conditions. Perhaps the paper that is still easiest to read and understand for someone with limited knowledge regarding this topic is the one linked below. www.ayre.com/pdf/Ayre_MP_White_Paper.pdfTo summarize, to be able to use an anti-alias filter that has a slower roll-off, you need to increase the sampling frequency because if you don't increase it, then audible aliasing artifacts will result from using this slower roll-off filter. The slower roll-off is what makes it technically possible for "ringing" (i.e., both pre-ringing and post-ringing) to be significantly reduced. By replacing Linear Phase filters with Minimum Phase filters, pre-ringing can be reduced to almost zero, albeit at the expense of adding more post-ringing (...and at the expense of introducing another artifact, called phase distortion). By combining Minimum Phase filter behavior *and* slower roll-off filter behavior, well... I think you get the picture. That said, increasing the sampling frequency *deteriorates* accuracy of the individual sampled values. It means there exists an optimum tradeoff between this type of deterioration and the improvement that results from the slower roll-off. However, due to how a Sigma Delta Modulator works (it being inherently noise-shaped), an additional improvement can be obtained from the use of ultra high sampling rates. This also helps to explain why, for example, the SABRE ES9018 chip uses (after upconverting everything to 32-bit first) 8 × upsampling. |
|
