2. digital
  3. 1030318323

Stereophile "confirms" Moncrieff's SACD comments ?

You folks remember a thread about SACD that mentioned J. Peter Moncrieff's comments about SACD being "junky" at higher frequencies ? Well, if you read the September issue of Stereophile's review of the $16,500 Accuphase SACD player, you will see that John Atkinson somewhat confirms Moncrieff's statements. Here is a direct excerpt from the test results that JA published in that specific Stereophile review ( pages 115-116 ):

"Again, the player's excellent dynamic range is revealed, at least in the low treble and below. Note, however, the rise in the noise floor above 2 KHz, this due to the aggressive noise-shaping used by the SACD's DSD encoding. By comparing fig 4 with fig 3, you can see that SACD has less inherit dynamic range above 10 KHz than CD, though this is largely academic, i feel."

Since you can't see the graphs without looking at the actual magazine, i'll try to sum it up. CD shows a rise in noise above appr 2 - 3 KHz. The slope climbs at a gradual rate as frequency rises. On the other hand, SACD shows the same rise in noise at about the same frequency point, but the slope is much faster and sharper. By the time we get to 20 KHz, standard "redbook" CD is actually about 15 db's quieter in terms of the noise floor and increased dynamic range.

Besides all of the above, which some "might" say justifies Moncrieff's opinions of poorer high frequency performance on SACD, JA goes on to show the spectral analysis well beyond the 20 KHz range. The rising noise level that begins at about 2 - 3 KHz continues to rise until we hit appr 70 KHz. Using a dithered 1 KHz tone as a reference, the noise level climbs to a point that is PHENOMENALLY high i.e. appr 80+ dB's noisier than it is at 1 KHz !!!

While i don't know if this phenomena is directly related to the Accuphase design being used or can be found in all SACD players due to the wave-shaping taking place, it makes me wonder if this is what has given me a headache aka "listening fatigue" on a few occasions when listening to some SACD's ??? Is it possible that the level of ultrasonic noise and ringing is high enough to the point that it can ruin what might otherwise be a pleasureable experience ?

As a side note, the jitter on this machine is PHENOMENALLY high. JA measures it at 4.26 nanoseconds of peak to peak jitter while running in redbook format. He comments that this is "more than 20 times higher than i have found in the best cd players and processors". He then goes on to "feed a signal into the DAC section of the player via the DP-85's S/PDIF data input with 16 bit data of the same signal, the measured jitter level dropped to a respectable 311 picoseconds." As such, the phenomenally high level of jitter is directly related to how they are transferring signal from the transport into the DAC. For a "lowly" $16.5K, you would think that they might be able to do a little better. Even the "respectable" 331 picoseconds of jitter is quite high in my opinion. Sean
>

sean

Showing 2 responses by sean

sean OP
6,231 posts
 
Djjd, sorry for the phenomenally slow response. Depending on what test JA was doing, the noise level within the lower registers of the audio passband measured anywhere from very well to at least quite good. As noted, as frequency would climb, so would the noise floor. With that in mind, the noise floor anywhere near the accepted audible range was somewhere around -90 dB's or better. When it hit 70 KHz, it was somewhere around -50 dB's or so.

According to theory, something like that should be completely and utterly inaudible. However, Ben Duncan has done research in this area and published an article about noise floors within a specific series of Integrated Circuits and the results were quite interesting. Not only did he show that so called "identical" IC's with the same part numbers from different manufacturers measured quite differently ( some were literally twice as noisy as others ), there was a difference in the "blackness of background" when going from one IC to another. I would highly recommend checking out that article if you can.

Besides all of that, even if it was 50 dB's down at 70 KHz, with my "golden ears", i can still easily hear that : ) Sean
>

PS... It doesn't matter where you are or how far apart we are, i can hear you laughing behind my back : )
sean OP
6,231 posts
 
Djjd: the article appeared in Stereophile.

As a further note, i was digging through some old Stereophiles trying to find a review of an older piece of gear. While i did not find what i was looking for, i did find a review of the $1300 Audio Alchemy ACD-Pro transport. In direct comparison to the $16,500 Accuphase SACD player, the results were pretty staggering.

The ACD-Pro had a measurable jitter of 17 ( SEVENTEEN ) picoseconds. In comparison, the transport on the Accuphase was directly responsible for FOUR THOUSAND ( !!! ) picoseconds of jitter. When performing the same 1 KHz dither tests on the ACD-Pro as JA did to the Accuphase, the 5+ year old ACD-Pro was 40 dB's quieter at the same 70 KHz measuring point. So much for "bigger & better", "new & improved" and "increased technology", huh ?!?!?!

I would like to add one more thing to this. I am NOT slamming SACD as a whole here. I have heard SACD and previously commented that i did find it to sound quite appealing in many ways. As was also noted, both i and my brother have received headaches while listening to SACD's. The headaches went away shortly after we went back to listening to redbook discs.

As such, my comments are primarily aimed at this player and what your $16,500 is going to buy you in terms of measurable performance. I need to go back and find some other reviews / technical measurements of SACD players that Stereophile has performed and compare them to the Accuphase. This would give us a better idea of whether these results are an isolated case having to do with this specific product or if the situation is something that all SACD players are going to have to deal with. I want to use Stereophile reviews only though, as other magazines may use slightly different methods of testing and / or be done under slightly different conditions. Using one source of references for all of the testing would allow the most "apple to apple" results that we as consumers could hope to achieve. Sean
>
More to discover