Home |
Search |
Today's Posts |
#121
Posted to rec.audio.high-end
|
|||
|
|||
DAC Differences
On Tuesday, December 11, 2012 9:43:23 AM UTC-8, Doug McDonald wrote:
On 12/10/2012 9:34 PM, Audio_Empire wrote: Agreed. I don't know if this is true or not, but I have read that phase-shift at or near the Nyquist cutoff frequency can have an effect in the audible range, although no one has ever been able to explain to me WHY this might be so. Sure they have ... I have, right here, and you agreed! The answer is intermodulation distortion in tweeters. Changing ringing characteristics can change this distortion. I've been able to hear it, without a double blind test, very easily, using several crappy tweeters at high levels. But in really good tweeters, no, I didn't. This was many years ago ... I could redo the tests, but I'd have to run the brick wall down to say 11 or 12 kHz. Doug McDonald I agree that phase shifts can cause distortion in tweeters, but I still don't see how 180 degree+ phase shifts up around 20 KHz (or higher) can cause AUDIBLE IM distortion in the passband, |
#122
Posted to rec.audio.high-end
|
|||
|
|||
DAC Differences
On Tuesday, December 11, 2012 3:56:43 AM UTC-8, Dick Pierce wrote:
Audio_Empire wrote: On Monday, December 10, 2012 3:13:23 PM UTC-8, Sebastian Kaliszewski wrote: You're ignoring the fact that part of that power is beyond the reach of human ears. Squara wavee will produce more heat than same frequency sine wave -- but that does automatically mean that additi0nal power would be audible. Yeah, but that's more due to the duty-cycle of the output devices than it is power. wouldn't you say? Two points: the "duty cycle" of a square wave, by any reasonable definition of duty cycle, is the same for a square wave as a sine wave. Second, by definition, a square wave of a given RMS value will produce EXACTLY the same heat as a sine wave (or ANY waveform of ANY kind) of the same RMS value. We're throwing a lot of pseudofacts around without apparently understanding what we're saying. What adds nice squareness for 7kHz wave on an osciloscope is well beyond human hearing range. Hence that curvy thing at 7Khz which contains only 3rd harmonics (at 21KHz) will sound the same as something squarelike, containng umpteen harmonics. Those higher harmonics won't be audible to any human. Agreed. I don't know if this is true or not, but I have read that phase-shift at or near the Nyquist cutoff frequency can have an effect in the audible range, although no one has ever been able to explain to me WHY this might be so. What phase shift? I routinely look at complete A/D and D/A systems (and by routinely, I mean several times a day) whose phase shift across the entire audio band is within a VERY small range. For example, I'm currently measuring one that's within 10 degrees to 20 kHz, and the only reason it's not that good at low frequencies is because of the relatively small coupling caps that result in it being 3 dB down at 12 Hz. In the early days of CD (as I understand it), the analog filter on the output was multi-poled. It had to be in order for the signal to be 40 dB down (or whatever. I don't remember exactly what the roll-off rate had to be to satisfy Nyquist - Shannon) at the sampling frequency. Analog filters that steep have more than 180 degrees of phase shift. Today these filters are digital and can do 40dB/octave without the phase shift of multi-pole analog filters (at least, so I've read. I have no experience with digital filtering). Look, people keep claiming this phase shift myth. It's not a question as to whether large phase shifts near the Nyquist frequency might be audible: you get to have that discussion complete apart from reality, because the reality is these phase shifts you're so worried about simple DO NOT EXIST. Would it not exist in the vintage, first generation CD players we've been discussing? I'm not saying that such phase shift would be audible, just that such a phase shift would have to be the result of using very steep sloped compound analog filters. After all a gentle 6 dB/octave from a single-pole filter wouldn't be sufficient, right? |
#123
Posted to rec.audio.high-end
|
|||
|
|||
DAC Differences
On Tuesday, December 11, 2012 8:02:57 AM UTC-8, Dick Pierce wrote:
Audio_Empire wrote: On Sunday, December 9, 2012 7:58:24 PM UTC-8, ScottW wrote: The point is that square waves will show characteristics of an op-amp. That's why most op-amp data sheets have lots of pictures (usually not photos) depicting square-wave performance. One further point, it should be noted that when I went and loked up the slew rate performance of the various op amps in my earlier post, not a single one of these figures was derived from any pistcure or photos of square waves. Every one of the slew rate figures was cleverly and carefully derived through a special, proprietary algorithm I've developed that I call, "looking at the slew rate specification." The data is disguised in a very subtle fashion by the manufacturers, and you have to know exactly where to look. For example, in the spec sheet that Signetics published for the 5534, you'll see it hidden in the line that says: Slew rate: 13 V/us Funny man! I don't remember specifically saying that square waves are used in op-amp data sheets to show slew rate. As you say, it's usually in the data sheet specification table. But square waves are used on op-amp data sheets to show other things such as rise and fall times. |
#124
Posted to rec.audio.high-end
|
|||
|
|||
DAC Differences
I just said:
Now, run a simple frequency response ("fr" to the uninitiated): It will measure like crap: every 500 Hz, there will be a deep null in the frequency response. Listen to this system with some unusual signal like, oh, say, music: it will sound wierd beyond description. Sorry, I meant to type that there would be a deep null at every odd multiple of 500 Hz, e.g., 500, 1500, 2500, ... -- +--------------------------------+ + Dick Pierce | + Professional Audio Development | +--------------------------------+ |
#125
Posted to rec.audio.high-end
|
|||
|
|||
DAC Differences
Audio_Empire wrote:
On Tuesday, December 11, 2012 8:02:57 AM UTC-8, Dick Pierce wrote: Slew rate: 13 V/us Funny man! I don't remember specifically saying that square waves are used in op-amp data sheets to show slew rate. As you say, it's usually in the data sheet specification table. But square waves are used on op-amp data sheets to show other things such as rise and fall times. Ah, that's because the rise time specs are similarily disguised! For example, from the very same 5534 spec sheet from Signetics, we have the very mysterious and cryptic statement: Rise time: 20 ns Overshoot: 20% WHatever could they mean by that, one wonders. The point of my ironic sarcasm is that, in fact, pictures of square waves in spec sheets ARE NOT there to inform us of slew rate, rise time, overshoot or other elements of performance. Such pictures are useless as quantitiative informers of such performance specs. The PRINCIPLE reason square waves have taken on such significance in audio "performance" measurements is because it's easy to generate, easy to see, but pretty useless quantitatively. It tells you ABSOLUTELY nothing that a properly done Bode plot will tell you and, in fact, potentially tells you MUCH less. And, as a stimulus, it if wholely unrelated to the actual conditions of operation relevant to audio performance. In another thread years ago on this very same topic, I concocted an interesting gedanken to prove these points. Consider a completely linear system consisting of two signal paths through a system: one direct, one delayed by exactly 1 1 ms. The gains of both paths are identical and their outputs are summed with an amplifier that itself has zero distortion in both the amplitude and time domain. Now, take a 1 kHz square wave, put it through such a system, and observe its output: the result will be, within the limits of the measurement device and the human's ability to assess the picture on t6he scope, PERFECT square wave performance. Now, run a simple frequency response ("fr" to the uninitiated): It will measure like crap: every 500 Hz, there will be a deep null in the frequency response. Listen to this system with some unusual signal like, oh, say, music: it will sound wierd beyond description. Yet, it will have perfect square wave response under a number of conditions. So what? Well, what it tells you is that square waves have extremely limited utility AT BEST for characterizing a system. There are plenty of reasons beyond what we discussed here, some of them obvious, some of them subtely complicated and technical. But those engineers who actually are engaged in designing real high-quality, wide bandwidth equipment either do not rely on square wave testing very much or simply do not use it at all. There's MUCH more informative stuff around. But, in a consumer add or to an "engineer" designing some overpriced gold-plated piece of audio jeweelry that's more black magic than science, they sure do look cool. -- +--------------------------------+ + Dick Pierce | + Professional Audio Development | +--------------------------------+ |
#126
Posted to rec.audio.high-end
|
|||
|
|||
DAC Differences
Audio_Empire wrote:
On Monday, December 10, 2012 3:15:33 PM UTC-8, Arny Krueger wrote: I was part of a DBT where two Hafler preamps were built as kits. One was built totally stock. The other was built (by me) replacing every Mylar capacitor in the signal chain with "Wonder Caps" and the two were compared. Documentation? Why would there be documentation? This was an informal test between a bunch of audiophile buddies. I think that this points to language as one of the reasons why there is so much misunderstanding here. When you talk about DBTs, people think about a fairly formal kind of test with bias controls, statistical analysis, and so on. A friend of mine and I both bought Hafler DH-101 preamp kits at the same time. He built his stock, I substituted the coupling and bypass caps in the kit with Wonder Caps of the same value. At a meeting of our little informal group of audio nuts, it was decided that we would do a DBT of the two being fed the same signal (from a CD player) and then feeding the same power amp (a MOSCode 600, IIRC) through one of those Switchcraft surface-mount switches with the three pairs of stereo inputs and one output. The speakers were a pair of Magnaplanar Tympani-3C (all eight panels). Levels were matched using an audio generator at 400 Hz and a Radio Shack digital SPL meter and a VTVM. After a dozen of so tries everyone agreed that there was definitely a difference and that unit #1 sounded significantly less transparent than #2. No one knew which preamp was which (you could only tell by taking the covers off the units). But how could you stop them from knowing when Unit 1 was playing and when Unit 2 was playing? And how did you prevent collusion? Andrew. |
#127
Posted to rec.audio.high-end
|
|||
|
|||
DAC Differences
On Dec 11, 2:22*pm, "Arny Krueger" wrote:
"Scott" wrote in message ... On Dec 10, 3:14 pm, "Arny Krueger" wrote: "Scott" wrote in message ... On Dec 8, 7:30 am, "Arny Krueger" wrote: Given your stated position in this very thread on the lack of need to control for same sound biases in ABX DBTs Quotes? "The above ignores a well known aspect of human nature which is to strive for a positive result from any activity that requires substantial effort. An ABX test requires substantial effort on the part of listeners so the listeners will *follow their human nature and strive for a positive result simply because they are expending the effort and don't want the outcome tobe futile." Your conclusion does not follow from the quote. Sure it does. We've already been over this. Not only is the above quote factually wrong it was used as an excuse for the failure to use bias controls in ABX tests against same sound bias. *I simply pointed out The existence of a bias to strive for positive results that I have observed in real life. *Perhaps you would prefer that I not make an honest and complete report? Your report is quite anecdotal and filled with your own biases on the subject. What you "pointed out" is in fact erroneous information on the nature of human bias. And why did you "point it out?" was it an arbitrary thought unrelated to the issue of same sound biases? It was either that or an inference that same sound biases were defeated by this erroneous claim about human nature. Anyway, it's already been covered. and given the well known biases of some of the folks who were involved in those tests *Quotes? You need quotes for that? Tell you what, you are the one with all the back issues of Stereo Review. I have no such thing. I don't believe that I have even one paper copy of SR in my posession, nor do I know a priori where I might find such a thing. 1. So you don't have a copy of the article in question on the ABX DBTs that Stereo Review did? The article you cited as evidence in regards to the sund of those old CDPs? 2.Looks like we will have to go from memory. And as my memory serves me, Stereo Review never used ABX DBTs in any of their product reviews and yet always made the claim that CDPs and amplifiers and preamplifiers all sounded the same. Their listening tests were all done under sighted conditions for their product reviews. It's pretty obvious that their biases were towards same sound. Cite one review of any amp or CDP where they do not make the claim that they all sound the same. Where is the reliable evidence that the CDPs being discussed in SR at that time actually sound different? Why ask such an irrelevant question? Clearly people with a bias towards different sound will detect non existent differences. Given that Stereo Review did their auditions for their product reviews and always claimed same sound for CDPs it is pretty clear they did not have a bias toward different sound. We also know from the mountain of evidence in psychoacoustic research that there is no such thing as a person who has no biases. So the logical deduction is that they had a bias towards same sound. I mean really, they didn't even do level matched comparisons and we know the output levels of different CDPs are, if nothing else, not exactly the same from player to player. Still under sighted conditions they never ever thought they heard a difference. The bias is pretty clear in this case. And of course we have their own testimony as to what they believed...Maybe JH secretly disagreed with everything he wrote in Stereo Review? I don't think so. |
#128
Posted to rec.audio.high-end
|
|||
|
|||
DAC Differences
"Scott" wrote in message
... On Dec 11, 2:22 pm, "Arny Krueger" wrote: "Scott" wrote in message ... On Dec 10, 3:14 pm, "Arny Krueger" wrote: "Scott" wrote in message ... On Dec 8, 7:30 am, "Arny Krueger" wrote: Given your stated position in this very thread on the lack of need to control for same sound biases in ABX DBTs Quotes? "The above ignores a well known aspect of human nature which is to strive for a positive result from any activity that requires substantial effort. An ABX test requires substantial effort on the part of listeners so the listeners will follow their human nature and strive for a positive result simply because they are expending the effort and don't want the outcome tobe futile." Your conclusion does not follow from the quote. Sure it does. We've already been over this. We disagreed then, we disagree now! Not only is the above quote factually wrong How can it be wrong? It relates something that actually happened. It relates a well-known property of humans. it was used as an excuse for the failure to use bias controls in ABX tests against same sound bias. No it wasn't. It was used to address a specific statement that I disagreed with. You seem to be forgetting my very strong accomplishments in the area of bias controls relating to ABX tests. Just to jog your memory, it has long been my position that there should be a listener training program for each ABX test that is specific to that particualar test. My recommended program is a series of tests involving the same basic technical difference as the final test, but with the technical difference augmented so that in the initial form, it is so easy to detect that a person would have to be disqualified as a listener if they couldn't hear it. The technical difference is then reduced in logical steps down to the final level that is really the one that is most important. If any listener is somehow biased against hearing the technical difference in question, it will become obvious during the analysis of the various test runs. His sensitivity will be far less than the majority of the listeners. |
#129
Posted to rec.audio.high-end
|
|||
|
|||
DAC Differences
On Wednesday, December 12, 2012 6:17:58 AM UTC-8, Andrew Haley wrote:
Audio_Empire wrote: On Monday, December 10, 2012 3:15:33 PM UTC-8, Arny Krueger wrote: But how could you stop them from knowing when Unit 1 was playing and when Unit 2 was playing? And how did you prevent collusion? Andrew. That was fairly easy. At the time, the speakers were in my "listening room" and the stereo components were in another room. The person doing the switching was out of sight, and even she didn't know which preamp was connected to which switch input. The listeners couldn't see the two preamps or the person doing the switching, but they did know that one of the two preamps had stock capacitors and the other had Wonder Caps. So it was, I believe, truly double-blind. |
#130
Posted to rec.audio.high-end
|
|||
|
|||
DAC Differences
On Wednesday, December 12, 2012 6:21:47 AM UTC-8, Scott wrote:
On Dec 11, 2:22=A0pm, "Arny Krueger" wrote: =20 "Scott" wrote in message =20 =20 ... =20 On Dec 10, 3:14 pm, "Arny Krueger" wrote: And of course we have their own testimony as to what they =20 believed...Maybe JH secretly disagreed with everything he wrote in =20 Stereo Review? I don't think so. Actually, Julian Hirsch was of the "Yes, we have no opinion" school. His reviews tended to consist of a physical description of the unit in question= .. A discussion of the unit's feature set, and finally a series of measurement= s aimed at making sure the unit under test met its published specs. That's=20 generally all. If he listened to the unit, it was merely to confirm functio= nality. His conclusion was ALWAYS the same for every review and became famous: "This (whatever piece of electronics he was testing), like all modern XXXXX= Xs,=20 has no sound of it's own." The only time that he strayed from that non-opin= ion was when he was testing speakers, then he would give listening "opinions", = but even they were predictable: "The (speaker in question) had little bass belo= w about 60 Hz as it's frequency response test indicated." |
#131
|
|||
|
|||
Quote:
I think the 16 bit DAC feature of this model may not be a good feature for SONY while designing the product. What do you think, if this will be a 14 bit encoding then it may sounds well? |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Big differences between 44.1 and 96Khz. Why? | Pro Audio | |||
Differences between EL 84 and EL 34 ...? | Vacuum Tubes | |||
u87 differences | Pro Audio | |||
u87 differences | Pro Audio | |||
RME 8di Pro Vs DS.. Differences? | Pro Audio |