SACD vs. CD - Unfair competition?

[Thomas_A]

I have read this report, nothing new here. Did these numerous tests at this high end studio include listening to signals downconverted from higher sampled sources. Or did the test just transfer a analog signal to 16/44.1khz and compared from there? You make mention of these test, but you don't even follow your own example. Do you have a link that outlines how the test was implemented? Was this published? Do you have a link to the white paper?

Most of the tests are published in the Journal from the Swedish Audio-Technical Society, but in Swedish. They are not peer reviewed. As I said before I have participated in other tests, and they always contain blind tests with statistics of the number of correct choices. That the journal exist can be seen from small number of translated articles, although only of general interest subjects, here:

http://www.sonicdesign.se/amptest.htm
http://www.sonicdesign.se/tooleinw.htm
http://www.sonicdesign.se/optimum.html
http://www.sonicdesign.se/subplace.html

A PDF-link to a test between media:

http://www.lts.a.se/artiklar/SACD-praktik.pdf

It's one of the tests published in "Musik & Ljudteknik" no 1/2002. It's in Swedish.

The test included a high-quality analog musik signal with a bandwith of 46 kHz recorded with high-speed DAC (fs=96 kHz) with two Earthworks microphones. This was the original signal. Signal 2 was the same but filtered analog with 3rd order butterworth at 25 kHz. Signal three, the original was converted to normal DAT standard (48 kHz).

Two different tweeters were used, both which were ±0.5 dB up to 20 kHz ±30°. The first falls soft above 26-28 kHz, the second continue 20-80 kHz ±9dB. If the peak at 28 kHz is exlcuded it was within ±4 dB.

There were no signifcant audible differences between the three program material used with any of the tweeters.

I won't go through the program material 4 & 5, since I have no time for translation of the whole text.


When it comes to 16/44.1, other persons at the Studio have made conversions and listening tests and also made recording in directly in different formats. Their conclusion was that 16/44.1 is sufficient as long as you don't play as loud that you hear the dither noise.Thus when listening to music at realistic levels, there was no audible difference. I gave a link to the discussion between CD, SACD and DVD-A and it is mentioned that the CD standard is much more transparent than people believe. It is of course based on the various tests made. Here they are again:

http://sound.westhost.com/cd-sacd-dvda.htm
http://www.fivechannels.com/artiklar...A-followup.htm

You will probably now say say that these are irrelevant, but then you have probably not read all of it.

And as you claim above, the HF content does not matter. It is i) not heard because our ≈18 kHz limits, ii) it will be masked by music and noise. That means I can do the tests at home right? No need of supertweeters. And there should be an clear audible improvement of the sound if I sample at 24/96 than at 16/44.1?

And again, can you or cannot you provide any data from your listening tests. ABX, blind, double-blind, who was the test leader and how many correct choices? Since you claim to have positive results according to AES standards it should be data, right? Negative results will be series like 5/10, 6/10 4/10 etc. Positive will be 9/10, 10/10 9/10 etc.