Originally posted by Ruse Black magic and witcvhcraft.Where's your proof - double blind listening tests or other validated test methodology.The sort of 'analysis' you make above is similar to the high end hi-fi media.It may be true, but you demonstrate no proof. The placebo effect is powerful. If I buy a new SACD machine, drop my new $big SCAD in the tray, smoothly glide it into the soft lit front panel and press play - my god its going to sound good!
Originally posted by bryant I would go further and say that there is no chance of either of these formats having any kind of mass market. However, their development was not driven by customer's desires for multi-channel remixes of their favorite CDs (or their disgust with that wretched 16/44 sound). Their development was driven by the major label's demands for a secure format and their inability to make unrippable CDs. And I suspect that as soon as most new CD players will handle SACD and/or most DVD players will do DVD-Audio, then the phase-out of the CD will begin....assuming that the market puts up with that...
Originally posted by DarkAvenger Thinking further:I dunno how it is in the ultra high-end section, but usually digital amps are used in the the price regions I know. SACD can *never* sound better then DVD-A on such an amp due to the medium/technique! As I stated above the amp processes the signal digitally (mid-range: 48kHz at 16 or 24bit to high end 192khz at 24bit) the SACD will be converted DVD-A like, so if we can assume that it could be possible to connect DVD-A digitally the DA/AD steps will go away and we have a pure pcm transfer, DVD-A will sound more original than SACD.If SACD does sound better on a digital amp it is due to mastering. I don't think that that DA/AD will improve sound...So for me SACD makes absolutely *no* sense, and even sense of DVD-A is doubtful.
Originally posted by gdougherty ...The issue is really technical rather than purely being about the reproduction of nyquist frequencies. Part of it deals with the acuracy of reproducing given frequencies. At 44.1 you get 2.6 samples per 17Khz wav, at 96Khz you get 5.6 samples, at 192 you get 11 samples. Higher frequencies also inherently reduce the effects of jitter error...
As Ethan Winer uncovered in a semi-official test, most people can't identify the difference between 16 and 24-bit audio, unless dithering has been applied to the 16 bit audio.
Originally posted by DigitalMan So, there you have it, SACD and DVD-Audio crippleware. On top of it all, SACDs and DVD-Audio discs cost more than CDs ( US +). Tough to see a mass market for these as much as they are technically impressive. Remember that 24bits gives you a dynamic range of 144dB - not many audio systems have that kind of accuracy, and it really become pointless in a car/airplane or even a home with an air conditioner.Rant over.
Originally posted by Frank Klemm Dynamic range is calculated by 6.02 dB * bits + 1.76 dB + Gain
Originally posted by Garf Frank, I've been wanting to ask this before: where does the 1.76dB come from?
Originally posted by Garf I'm not sure if you badly worded this or if you are misunderstanding something.You present the number of samples for a given freq, and go on to say that 'higher frequencies also...', which implies that the number of samples in itself is an argument, which it isn't (and that's exactly what nyquist is about).
Originally posted by Garf This is wrong. Most people could not identify the difference between the 16 and 13 bit files. There were two who could distinguish between the 16-bit ones, and _those_ preferred the truncated file on _subjective_ grounds. (http://ff123.net/24bit/24bitanalysis.html)-- GCP
Originally posted by gdougherty More samples do mean greater accuracy and faithfulness to the original analog signal, I did mean that portion of my statement. Nyquist as I understand it is not about accuracy, but determining the cutoff that a given sampling rate can reasonable reproduce. IMO 88.1 or 96Khz are sufficient, 192Khz would theoretically be better for accuracy, but I believe it's a point of diminishing returns, and with 192 the returns would be extremely diminished.
My mistake, the results still fully support my argument.
Originally posted by Garf You have a very fundamental misunderstanding. At 44khz sampling rate, you can reproduce a 20Khz signal exactly as good as if you were sampling at 192Khz. This is the very essence of the nyquist theory. There is no gain whatsoever in increasing the sampling rate above two times the highest frequency that you want to record/reproduce.If this feels intuitively strange, read up on the theory of sinc functions.I agree with this. I made the correction because as written your original statement suggests it's better not to dither to get better accuracy in reproduction, which is false. (Not dithering may introduce additional distortion which can make the music sound better subjectively in some cases, which is what likely caused the curious result)Dithering is _good_. You don't need to dither, but you _want_ to do it. The reason why the result of the test is seemingly at odds with this is that the listeners were not comparing to the 24-bit original, but only comparing among the two 16-bit files alone. They had no idea of determining which one was more true to the original, so they had to make a completely subjective pick among them. In that case, it's possible they pick the more distorted file because it has more ' pizzaz ' or whatever. -- GCP
As to your other questions, higher sampling frequencies will not benefit us when it comes to jitter because the amount of jitter between like amounts of time will be the same. The error from 1 44,100th of a second to the next will be still x picoseconds, regardless of what the frequency derived for clocking purposes is.
Originally posted by Pio2001 At least recording a vinyl in 44.1 and 96 kHz 16 bits on the Marian Marc 2 soundcard, I sweared that the 96 kHz captured incredibly much more of the "analog" vinyl sound than the 44.1 kHz... until all audible differences suddenly vanished under the ABX hammer :cry2:
Originally posted by bryant obviously is useful for determining what is directly perceivable, it is well known that stimuli below the threshold of perception can still have an effect. For example, I have seen a study that showed that PC monitor flicker had a significant effect on people’s tested reaction time even when it was well below the threshold of being perceived directly. It doesn’t seem to me that unlikely that the auditory system might not have some similar characteristics.
You know, statements like this are so universal that it makes me question the absolute validity of ABX testing
Originally posted by bryant While it obviously is useful for determining what is directly perceivable, it is well known that stimuli below the threshold of perception can still have an effect.
Originally posted by Pio2001 When an ABX is positive, it does prove that the difference is audible, but when it's negative, it doesn't prove that there is no difference, it proves that the difference is "quite" inaudible.
Originally posted by Pio2001 It is often more difficult to ABX a difference than to hear it.One of the only people that managed to ABX a 16 bits dithered file vs a 16 bits truncated file (while some other heard a difference but couldn't ABX it) performed one try each day ! Making a difference 16 times in a row can be very difficult.