MQA bails on Rocky Mountain Audio Fest*

"By Grabthar's Hammer, by the suns of Worvan you shall do the test!"

 

Strange...

I just tried to post a new thread with a summary of the MQA test results and as a thank you to all who contributed; similar to this message on Squeezebox forum:
Final MQA Round-Up...

I'm not trying to debate the merits or anything like that... Just a few links to the recent posts for those who wanted a summary and conclusion with my thoughts on this whole MQA thing. Poof... Can't find the thread...

 

Sure, you will see differences in the data from the different formats. But you won't be able to make absolute correlations about the data (bandwidth, slew rate...) leading to the reasons for:

All you can do is conjecture.

 

No

True

True. But EE's have spent a lot of time on this.

Find someone who has then, and post it here.

 

I read the articles.

First of all I want you to know that I appreciate the effort and the time you put into this project and I'm sure I'm not the only one who is thankful for your contribution.

Secondly: Would you care to expand a bit on the strange findings that the better the hearing (younger and more trained people) the better the equipment used - the less there is love for MQA. Isn't it a bit counter intuitive or is it hinting that MQA is indeed introducing artifacts of its own?

 

Did I miss the Galaxy Quest thread??

 

I have, DSD won it. No contest. I heard the same thing you heard.

 

Hi Claude,
Thanks for the note.

To be honest, I can't be sure of course plus chance plays a part statistically and we're not looking at highly statistically significant results. I think it is an interesting finding which if someone had the time and interest, could use as a followup with maybe young, trained listeners to tease out in a controlled setting.

Objectively, I think it is important to remember that MQA does have a lower bit-depth resolution than an actual high-res 24-bit file regardless of what they claim to be doing in the time domain. I'm certainly not saying that we *need* 24-bit audio but who knows, maybe in a small group of people when they tested the MQA decode vs. actual 24-bit 2L source, were able to show preference for the version with less de-correlated noise and dithering in the lowest bits.

 

What's hanging me up is that I'm not into Creedence and don't have their disc.

Now--if there's a revealing comparative moment on a Nat Cole disc that you mastered for SACD/CD, I'll do the test straight away!

 

OK, here's the result of my "test."

(I have to put the word in quotes because what follows is a very subjective impression. I couldn't isolate a particular moment of decay that would have made a comparison more certain. And the test was far from blind. I was doing the switching, and I always knew which layer I was listening to.)

The track: "Ballerina," on Disc 2 of the Gray/Hoffman-mastered "Nat King Cole Story."

The CD and SACD layer both sound fantastic. But I found myself preferring the SACD layer slightly, for intangible reasons. I listened to the track in its entirety each time, and went back and forth a few times. (So that I probably heard it 3 times apiece in each format.)

Here are the differences I perceived.

On the SACD layer, Cole's vocal just sounded slightly more dynamic. So did the band. There was more sense of "life" in both--even though the "spectral content" of the tracks sounded identical. (I.e., they sounded eq'ed exactly the same. Neither was brighter or duller than the other, when it came to the vocal or the band.) Cole's vocal just sort of jumped out of the speakers more on the SACD layer, even though it wasn't louder; internal dynamics seemed greater. And the trumpet section on the SACD layer sounded like it had more horns in it. The right number, as opposed to the CD layer, where it sounded a little like the A&R man on the date had hired one less horn than he should have.

 

DSD, especially on the higher end ADC's, to me sounds like "digital reel reel". There is a continousness and seamlessness
that to me betters PCM, and even more so at DSD128. Sort of the difference between a superb panel speaker and a multi driver
box speaker...

 

I was tempted to respond to that post, but I couldn't come up with a diplomatic way of putting it.

 

That's pretty much how I would describe a good SACD layer if I can give so much details. My words would've been "Everything just sounded more real."

 

The first time I heard this my self was about ten years ago comparing the CD and SACD layers of Dylan's Desire. The difference was
dramatic on an Ayre SACD player. Every track was more satisfying on the SACD layer.

 

I've decided to get a Oppo 205....can't wait to hear my SACDs on a state-of-the art player.

 

On the Sony SACDs, and the later Mobiles, it is. But yes, I am very careful when comparing, just in case the Redbook layer is
a different master.

 

I have not heard that player, but have always enjoyed Oppos.

 

No.

 

I trimmed the files I used in my previous ABX test to 45 seconds & posted them for anyone to try:

Fortunate Son (24/88)
Fortunate Son (16/44->24/88)
+
Fortunate Son (delta 24/88 & 16/44->24/88)


Here is the bit difference report for 2 track excerpts:

Code:

Differences found in compared tracks.
Zero offset detected.

Comparing:
"M:\AUDIO\CD\06 - Fortunate Son (2488) [excerpt].flac"
"M:\AUDIO\CD\06 - Fortunate Son (1644-2488) [excerpt].flac"
Compared 3969001 samples.
Differences found: 7936167 values, starting at 0:00.000000, peak: 0.0055344 at 0:18.717132, 2ch
Channel difference peaks: 0.0026886 0.0055344
File #1 peaks: 0.9727595 0.8991761
File #2 peaks: 0.9731759 0.8989024
Detected offset as 0 samples.

This is the audio statistics for tracks delta:

Code:

Creedence Clearwater Revival - Willy And The Poor Boys [DSD]

Peaks
=====
0.6

Peak Table
==========
 1|  0.6

Full Statistics
===============
 #|Chan|Peak %|  Peak dB|AvgRms dB|TotRms dB|MinRms dB|MaxRms dB| MinSample| MaxSample| DCOffset
--+----+------+---------+---------+---------+---------+---------+----------+----------+---------
 1|   1|   0.3| -51.4093| -70.4900| -69.7951| -78.2520| -60.6181|-0.0026886| 0.0026656|   0.0000
  |   2|   0.6| -45.1386| -67.3400| -66.0194| -75.7970| -55.5195|-0.0055196| 0.0055344|  -0.0000

0 dBFS: Sine
RMS Window Size: 50 milliseconds

I also re-ran ABX test with excerpts (once again unsuccessful):

Code:

foo_abx 2.0.4 report
foobar2000 v1.3.16
2017-10-15 10:56:59

File A: 06 - Fortunate Son (2488) [excerpt].flac
SHA1: 260205f084f5987027afd9365789a630401db041
File B: 06 - Fortunate Son (1644-2488) [excerpt].flac
SHA1: 461092b611733427b465e5dbdd15ad0bd085d527

Output:
ASIO : ASIO4ALL v2
Crossfading: NO

10:56:59 : Test started.
10:59:09 : 01/01
10:59:26 : 01/02
11:00:10 : 02/03
11:00:41 : 02/04
11:00:57 : 03/05
11:01:06 : 04/06
11:01:13 : 04/07
11:01:17 : 05/08
11:01:25 : 05/09
11:01:37 : 05/10
11:01:37 : Test finished.

 ----------
Total: 5/10
Probability that you were guessing: 62.3%

 -- signature --
1db9564371faa20795bc6f46e0642e72f649838a

Feel free to do comparisons yourself & post your successful ABX test reports (if any)...

 

DSD is not "hi-res", it's only one bit.

 

And, now, for DSD, which is only one bit.
DSD's waveforms look like
__--__-_-_

Only one stairstep: one at zero, one at 100%. That's the proof that DSD has got no resolution at all!
DSD is not Hi-Res.

 

Subjectively I agree that DSD is less high-res sounding that proper good well done PCM played back on a proper good well done PCM DAC. But your statement about DSD not being high-res just proves you have no understanding of how sigma-delta works. Sigma-delta and DSD absolutely are high-res. Even Dan Lavry would agree.

I have subjective differences with DSD and sigma-delta. I find DSD played back on my supremely awesome Schiit Multibit PCM DAC to sound overly softened like the sound has gone through a soft focus filter similar to how portrait photographs are done with an overly soft focus rather than the sharp focus that is normally used for landscape photographs. DSD sounds like I'm listening through a fog and soft focus. Even with those caveats, I consider DSD to be high-res. I can hear the high-res benefits listening to a DSD recording even on a PCM DAC. I do consider well done PCM to be better than DSD when played on my Schiit multibit DAC.