Willakan
1000+ Head-Fier
- Joined
- Dec 26, 2010
- Posts
- 1,039
- Likes
- 109
/begins chanting
Test It! Test It!
(pulls out war drum)
Test It! Test It!
(pulls out war drum)
Well, a properly conducted one anyway! (Removes personal biases, that is.)
vandaven, excellent tests. I'm surprised they nulled that close -- -110 dB is impressive. That's below the noise floor of 16 bit audio (obviously you were testing with 24 bit audio). Given the quantization error of jitter, however, you'll of course never get them to null perfectly - and the uDAC 2 doesn't have the best jitter performance anyway (but not enough to make a difference to our ears, just to the measurements).
I've seen RMAA measurements for the noise of the uDAC 2 at 24 bit audio in the range of -96 dB -- I think that closer nulling than that should indicate that the noise is signal dependent - jitter, probably. Still, -96 dB is in no way audible under normal listening conditions; nor is a -110 dB null. Bob's goal (he achieved it) in the infamous Carver Challenge, for example, was a -70 dB null. -110 dB is 1000 times less difference!
What you see in the measurement is the SPDIF signal provided by the uDAC2, not the analog one. The audio data is delivered via USB stream to the NuForce uDAC2, clocked, and then delivered to the Mbox2 audio interface via SPDIF digital (clocked by the uDAC2).
IF the USB cable would make any difference in the signal chain, it would show up, and that more clearly than thru any ADC-analog chain, where the performance of analog components, approximation and quantization.
Cheers.
Ahh, okay, that makes more sense! I missed where you said that - see it now.
Yeah, it wouldn't make a whole lot of sense to introduce all the quantization error and analog performance issues by actually running through the DAC - the S/PDIF output is as close to perfect as we'll have short of reading the direct bit stream.
We'll still have the the jitter in the clocks then - but that shouldn't make a difference (when it's staying in the digital domain) unless there's enough clock drift to cause bit slip, or does it?
^ when money is involved people become stupid.
People can also not believe there aren't differences because marketing tells them there is so they hear them and refuse to question them.
If it's banal (like a power cable or USB cord) and costs lots and lots of money, it must be good because it's expensive, yes?
Psychosomatic/placebo effect is always an interesting thing, especially when it coincides with a field as subjectively opinionated as audio.
The problem is that in the analog realm, there are differences between cables. There is always a top or bottom line, and that clearly depends on the application (required length, required operating temperature etc.). A poorly manufactured cable will cost less than a well manufactured one, that's for sure. And people are used to that.
It sounds good so far - but you need to make sure you level-match all your testing. Also, for headphone cables, you need to control the potential for feeling the difference between them in weight or feel as they brush against your head.
And you shouldn't get zero guesses right (out of a reasonably large sample)! You would be expected to average about 50%, but the probability of a particular distribution depends on the number of samples. That's why you have to do a significant number of samples.
I'd say more, but this really outlines what you need to know and do to properly set up an ABX test (although yours won't be double blind out of practicality):
http://www.hydrogenaudio.org/forums/index.php?showtopic=16295
Also, at this point wouldn't it be prudent to move this to the Sound Science forum since that's what's being talked about.