Home |
Search |
Today's Posts |
#81
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
Audio_Empire wrote:
On Monday, October 28, 2013 5:27:08 AM UTC-7, Andrew Haley wrote: Audio_Empire wrote: On Sunday, October 27, 2013 6:39:53 AM UTC-7, Andrew Haley wrote: Surely, if a transparent DAC can be made for dollars, it makes sense to assert that any DAC which its not transparent is defective. I think I'd be quite happy to suggest that if two DACs can be distinguished by listening then at least one of them must be defective. Of course, I could be proved wrong by a properly- conducted blind test. Andrew. The problem here seems to be with the term "transparent". I'll define what I mean: any ADC/DAC that cannot be distinguished from a piece of wire in a listening test is transparent. Yes, that is the accepted definition. However, I can tell you, we aren't "there" yet. While that would be an impossible test to actually perform on a D/A converter, Eh? Surely it's easy. the very fact that all D/A converters sound slightly different from each other It's not a fact; not one that has been esatblished, anyway. tells us that we aren't at that point (unless we can agree that different pieces of wire "sound' different from each other - something that I dare say that none of us is prepared to assert. And it may be that all current DAC chips are themselves so close to Yes, but techniques exist in the world of instrumentation to reduce to reduce artefacts way below thresholds of hearing. Audio is easy when compared with, say, measuring nanoamps in a noisy electrical environment, as in an atomic force microscope. Good point. And while many modern electrical components do have specs that put artifacts below the threshold of hearing (the LM49710 and equivalent family of op-amps come to mind here), not all devices made from such components have ALL of their artifacts banished to the realm of inaudibility -yet. I'm also not convinced that ALL DAC designers are striving for complete transparency, but are striving for designs that they think "sound good". I don't doubt that for a moment. Maybe, maybe not. The only way to try is to listen. I will note, however, that discrete construction is a spectacularly bad way of making a hybrid component like a DAC: it's much harder to get the accurate matching of components that you need. I think that companies who build their own DACs from scratch are playing on a naivet? on the part of the audiophile who might remember when ICs for audio simply weren't very good for the application (709 op amps, anyone?). Exactly. It's the desire to be different, and to rise above the hoi polloi. The irony of that is that the only way to do so with a DAC is to make something that's worse, objectively speaking. I certainly remember when a decent microphone mixer used resin-potted, discrete component op-amp modules to get the low noise demanded by recording studios simply because Integrated Circuit op-amps couldn't cut the mustard. Then again, maybe companies like MSB and dCS make their own DACs these days simply to justify their astronomical prices. Whatever the reason, I suspect that hand selecting from a Burr-Brown or ESS SabreDAC (etc.) IC product would give at least just as good of a performance as hand-tweaked, hand-selected discrete component D/A converter modules. Almost certainly, I would have thought. Getting the low noise of a really good IC is going to be very hard with discrete hybrid construction. And how much of such a converter really is discrete, anways? I bet most of the digital stuff isn't. The problem is that while I'm sure that audio groups, both formal and informal, the world over have conducted DBT tests to determine the audibility of digital to analog conversion, there seem to be no definitive studies that once can cite. Why should there be? It's not as if the distortion caused by converters is special. There's no reason to believe that DAC distortion will [be] audible at thresholds below distortion caused by any other component. I don't think that's the point. The point is that there are two equally passionate schools of belief with regard to DACs. Once school says that all DACs are so different that one needs to choose one based on how they sound in one's system. The other school of thought says that they all sound alike because any criterion which would contribute to a DAC having a sound of its own is, today, in modern designs, well below the threshold of audibility. This second school maintains that there is no reason to spend more than a few dollars ( I believe $40 was mentioned) on any DAC for that reason. One of those two camps is wrong. Some definitive DBT results would do much to put this debate to bed. No, they wouldn't, because certain passionate people would use special pleading in an attempt to show that audio is, somehow, immune from double-blind testing. Yet audio researchers such as Lipschitz and Vanderkooy, or Meyer and Moran have not undertaken to address this question. Indeed, a search of AES papers yields no results that address this question in any direct way. That means while people in the strictly objective camp can CLAIM (and some have done on this forum) that DACs must sound the same because all artifacts have been rendered below audibility, but they can produce no more proof that this is the case than can the most dyed-in- the-wool subjectivist that All DACs sound different and that these differences aren't subtle. It's not up to them, because it's almost impossible to prove that something doesn't exist. Surely it's up to those who believe in this mysterious effect to provide some evidence that it does. But what would be the advantage to them? It wouldn't increase sales. Andrew. |
#82
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
On Tuesday, October 29, 2013 6:04:26 AM UTC-7, Andrew Haley wrote:
Audio_Empire wrote: On Monday, October 28, 2013 5:27:08 AM UTC-7, Andrew Haley wrote: Audio_Empire wrote: On Sunday, October 27, 2013 6:39:53 AM UTC-7, Andrew Haley wrote: Surely, if a transparent DAC can be made for dollars, it makes sense to assert that any DAC which its not transparent is defective. I think I'd be quite happy to suggest that if two DACs can be distinguished by listening then at least one of them must be defective. Of course, I could be proved wrong by a properly- conducted blind test. Andrew. The problem here seems to be with the term "transparent". I'll define what I mean: any ADC/DAC that cannot be distinguished from a piece of wire in a listening test is transparent. Yes, that is the accepted definition. However, I can tell you, we aren't "there" yet. While that would be an impossible test to actually perform on a D/A converter, Eh? Surely it's easy. Easy? One of us is smoking' something, surely! Please tell me how you compare a DAC directly against a straight wire? the very fact that all D/A converters sound slightly different from each other It's not a fact; not one that has been esatblished, anyway. I think that's the POINT of this entire exchange. tells us that we aren't at that point (unless we can agree that different pieces of wire "sound' different from each other - something that I dare say that none of us is prepared to assert. And it may be that all current DAC chips are themselves so close to Yes, but techniques exist in the world of instrumentation to reduce to reduce artefacts way below thresholds of hearing. Audio is easy when compared with, say, measuring nanoamps in a noisy electrical environment, as in an atomic force microscope. Good point. And while many modern electrical components do have specs that put artifacts below the threshold of hearing (the LM49710 and equivalent family of op-amps come to mind here), not all devices made from such components have ALL of their artifacts banished to the realm of inaudibility -yet. I'm also not convinced that ALL DAC designers are striving for complete transparency, but are striving for designs that they think "sound good". I don't doubt that for a moment. Maybe, maybe not. The only way to try is to listen. I will note, however, that discrete construction is a spectacularly bad way of making a hybrid component like a DAC: it's much harder to get the accurate matching of components that you need. I think that companies who build their own DACs from scratch are playing on a naivet? on the part of the audiophile who might remember when ICs for audio simply weren't very good for the application (709 op amps, anyone?). Exactly. It's the desire to be different, and to rise above the hoi polloi. The irony of that is that the only way to do so with a DAC is to make something that's worse, objectively speaking. I certainly remember when a decent microphone mixer used resin-potted, discrete component op-amp modules to get the low noise demanded by recording studios simply because Integrated Circuit op-amps couldn't cut the mustard. Then again, maybe companies like MSB and dCS make their own DACs these days simply to justify their astronomical prices. Whatever the reason, I suspect that hand selecting from a Burr-Brown or ESS SabreDAC (etc.) IC product would give at least just as good of a performance as hand-tweaked hand-selected discrete component D/A converter modules. Almost certainly, I would have thought. Getting the low noise of a really good IC is going to be very hard with discrete hybrid construction. And how much of such a converter really is discrete, anways? I bet most of the digital stuff isn't. That we don't get to know. I've never seen the inside of a dSC Scarlatti but I have seen the circuit board of an MSB DAC IV Diamond, it has four copper colored closed, potted modules (that I take to be dual differential DAC modules). We'll never know what's actually inside of them. But like you, I'd be very surprised if much of the circuit wasn't digital ICs. The problem is that while I'm sure that audio groups, both formal and informal, the world over have conducted DBT tests to determine the audibility of digital to analog conversion, there seem to be no definitive studies that once can cite. Why should there be? It's not as if the distortion caused by converters is special. There's no reason to believe that DAC distortion will [be] audible at thresholds below distortion caused by any other component. I don't think that's the point. The point is that there are two equally passionate schools of belief with regard to DACs. Once school says that all DACs are so different that one needs to choose one based on how they sound in one's system. The other school of thought says that they all sound alike because any criterion which would contribute to a DAC having a sound of its own is, today, in modern designs, well below the threshold of audibility. This second school maintains that there is no reason to spend more than a few dollars ( I believe $40 was mentioned) on any DAC for that reason. One of those two camps is wrong. Some definitive DBT results would do much to put this debate to bed. No, they wouldn't, because certain passionate people would use special pleading in an attempt to show that audio is, somehow, immune from double-blind testing. I have seen some very smart and very well known audio mavens put forth that very notion. People who don't really have a stake in whether DBTs work for audio or not (famous mastering engineer Stan Ricker comes to mind here). Yet audio researchers such as Lipschitz and Vanderkooy, or Meyer and Moran have not undertaken to address this question. Indeed, a search of AES papers yields no results that address this question in any direct way. That means while people in the strictly objective camp can CLAIM (and some have done on this forum) that DACs must sound the same because all artifacts have been rendered below audibility but they can produce no more proof that this is the case than can the most dyed-in- the-wool subjectivist that All DACs sound different and that these differences aren't subtle. It's not up to them, because it's almost impossible to prove that something doesn't exist. Surely it's up to those who believe in this mysterious effect to provide some evidence that it does. But what would be the advantage to them? It wouldn't increase sales. That it wouldn't. It's hardly a mysterious effect though. A DAC is complex bit of kit made up of a number of disparate circuits doing different things. There are power supplies, proprietary filters, analog stages, etc., any which could (and probably do) affect the sound. Then there are the companies who purposely stray from transparency simply to differentiate their products from the rest. Remember, most of the audiophile community doesn't care what the manufacturer does to make his product sound "better", as long as it does. That's ostensibly why some very expensive "cables" have boxes built into their length. They're full of passive components which cause the cables to act as (mostly) non-adjustable filters. They sound different because they are screwing with the passband. |
#83
Posted to rec.audio.high-end
|
|||
|
|||
Audio and "Special Problems"
Audio_Empire wrote:
On Tuesday, October 29, 2013 6:04:26 AM UTC-7, Andrew Haley wrote: Audio_Empire wrote: On Monday, October 28, 2013 5:27:08 AM UTC-7, Andrew Haley wrote: I'll define what I mean: any ADC/DAC that cannot be distinguished from a piece of wire in a listening test is transparent. Yes, that is the accepted definition. However, I can tell you, we aren't "there" yet. While that would be an impossible test to actually perform on a D/A converter, Eh? Surely it's easy. Easy? One of us is smoking' something, surely! Please tell me how you compare a DAC directly against a straight wire? Well, you need an ADC as well, obviously. That doesn't invalidate the test, as long as the time delay isn't very long. Andrew. |
Reply |
Thread Tools | |
Display Modes | |
|
|