Harry Lavo wrote:



For what it is worth, I did a summary of all five units that John
provided
links for, comparing noise levels relative to ordinary 16/44.1 CD at
40hz,
3khz, 10khz, and 20khz. Results as follows:

5 DVD-A players: -8db, -12db, -12db, -9db
3 SACD players -10db, -11db, +-0db, +11db

The averaging of results for 40Hz makes very little sense. In a lot of
players, the low frequency noise floor is dominated by line-related
spurs, which have nothing to do with the architecture, and could be
measurement related. In other words, the same spurs may be present at
the same level whether that player is playing 14/44.1 or 24/96 or DSD.


Only one of these had that problem, and it happened on DVD-A, not CD. So I
averaged the high and low points of multiple spurs in and around 40 hz to
get the effective noise floor.

If you want to compare the architectures' (or the DACs') noise floors,
do not count the line spurs.



While the +11db at 20khz looks horrendous, it was still at an average
level -93db below 0db
reference so it should not be an audible problem. The lack of
complicated
or brickwall filtering seems more to make up for a good sounding treble
than
any inaudible contribution of noise detracts.

Are you aware that in LPCM/DVD-audio players, these filters can be made
phase linear and perfectly flat in the audio band? And that the
"brick-wall" can be pushed out way into the supersonic region, via
oversampling?


Yes I'm aware and it has helped a lot! But CD still requires a lot of
manipulation to get around high end problems that SACD simply ignores.

But you were saying that SACD is better than DVD-A in this respect. Even
in the case of redbook CD, this filtering problem has been solved,
although there still are players with poor implementations.


The same brickwall filters apply to DSD. The sampling theory applies to
DSD as well as LPCM!


Sorry, the SACD has a gradual filter as part of it's spec, from what I
understand (and see).


I was referring to the filtering applied to 24/96 or 24/128 LPCM.
Whether it is LPCM or DSD, you still need the anti-aliasing filter prior
to the A/D's. It is the same requirement. If you sample at 192KHz, you
have to stop signals above 96KHz from getting through.

In any event, the problem of preserving phase linearity and flatness in
the audible band, and cutting off before 48 KHz (or 96 KHz) is a solved
problem. There is absolutely no basis for believing that the filter
cannot be implemented without degradations in the audio band in high
resolution LPCM.


In other words, your claim that "The lack of complicated or brickwall
filtering seems more to make up for a good sounding treble than any
inaudible contribution of noise detracts" has no technical leg to stand
on. I guess you started with a faulty premise that SACD's have a better
sounding treble than DVD-A, and you were just trying hard to support
that premise. Didn't you say that it was the bass and midrange that
SACD's really are superior?


I said that there were other facts than extended frequency response that
seemed to work in SACD's favor, and I postulated that the greater lower and
mid-frequency quieting might be contributing, versus CD.

We agree that SACD's can have a better S/N ratio, below, oh, around 10KHz.


I also said that DVD-A had cleaner highs and sounded smoother than CD in the
treble, which it does.

This is an unverified observation. Whether the better S/N translates
into cleaner highs or a smoother sound has not been proven, since CD's
can have ruler flat frequency response up to 20KHz. My take is that
recording and mastering have a much larger effect on the smoothness of
the sound.


I also said that the "noisy" treble of SACD didn't strike me as a practical
problem, which it doesn't.

Others would posit that the high ultrasonic noise can be bad for
downstream electronics.


It is precisely this level of technical knowledge exhibited in the Audio
Asylum that makes it a very noisy place to hang out.

Well, if you have such low tolerance for us "mere audiophiles" perhaps you
should spend more of your time on rec.audio.tech. Last time I looked, this
was a hobby-based forum, not strictly a technical forum.

That is why I stay away from Audio Asylum.