Support Head-Fi.org by
starting all of your
Amazon.com shopping by
clicking here.
____________________________________________________________________
Today's Featured Head-Fi Blog: Jude's Blog
____________________________________________________________________
Please help
support Head-Fi by becoming a Contributing Member
CLICK
HERE -- Contributing Members, thank you
for your generous support! --
We still don't have the full papers back from the U of T, but they did prepare a power point presentation for us as an overview of the testing results which I posted to the website.
We still don't have the full papers back from the U of T, but they did prepare a power point presentation for us as an overview of the testing results which I posted to the website.
I don't see where it says that is from the "U of T" in the powerpoint slides, all it says is it was authored by "Ryan". Who at the U of T did the test?
Also, the conclusions drawn are a bit strange...
Slew rates on those scales for audio apps are meaningless. Typical line level output from a CD player or DAC is 3V peek to peek. That means that the "regular" cable could do a full peek-to-peek swing in 0.1us, or 1/10,000,000th of a second, or 10MHz. 10MHz is a bit beyond what most speakers can reproduce. So even if you played that square wave, the extreme possible example, the movement of the speaker cones would be the same with every cable. They just can't move that fast, in fact a tweeter that can go up to 44100Hz is 266 times slower.
The noise ratings for power cords are very odd as well. The numbers are well above what can normally be measured in a cable. The differences in the measurements (handily expanding by the scale of the graph) are too small to matter, or to put beyond manufacturing variance or error.
__________________
Team W5000 - Team HEADFIVE - Team Onkyo - FLAC -> Onkyo SE-200PCI -> C.E.C. DA53 -> HEADFIVE -> ATH-W5000
Slew rates on those scales for audio apps are meaningless. Typical line level output from a CD player or DAC is 3V peek to peek. That means that the "regular" cable could do a full peek-to-peek swing in 0.1us, or 1/10,000,000th of a second, or 10MHz. 10MHz is a bit beyond what most speakers can reproduce. So even if you played that square wave, the extreme possible example, the movement of the speaker cones would be the same with every cable. They just can't move that fast, in fact a tweeter that can go up to 44100Hz is 266 times slower.
We manufacture digital cables as well as analog and for that application it definitely is important to have slew rate as high as possible, as to the rest, you can hear the difference in resolution. It's about reproducing the micro details of the wave, which requires a very high slew rate response. There any improvement is relevant, even at very short timescales.
If you want to argue audibility, go ahead. You won't get a very good argument going from me because I have heard the difference.
Perhaps you could answer the actual question posed by mojo: Who at the U of T did the test?
My guess would be a group of undergraduate EE students, probably as a term project.
Without seeing the complete paper it is difficult to know what to make of it except...
1) In several slides they mix variables, i.e IC s and power cables in slides 8,9,10 and 11. What for instance was the analog cable when the power cable was changed. What was the power cable for each of the ICs ?
2) The generic IC is not identified by name or photographed, unless it is the grey one at the bottom of slide 6
3) Slide 4 contains an unsupported assertion Faster Response time means better sound quality, perhaps this is true but a citation is de minimus here.
4) The presentation has no references whatsoever.
5) Was the "original" slew rate calculated or measured. This is important since we are dealing with finite rise times.
6) The DVD.....ADC.......Laptop was single (1 channel )connections ?
7) Statistical analysis - there is none !, we do not know if these numbers are at all significant, at least a one-way Anova is required. Also the difference between the original wave and the best other slew rate is far ( 0.0105 vs 0.0052) greater that the difference between the worst and best other slew rate. Also the differences in noise levels looks very small was this tested for significance ?. Could this have been random variation ?
8 The difference between worst(stranded) and best(genesis) peak to peak slew rate on a nominal 2V signal (Slide 10) would be 1.46 microseconds viz
78.80 - 77.34 i.e 0.00000146s, I wonder if this is detectable by human ears ?
9) They did average 5 samples from each combination ?
10) No scale on the square wave diagrams, a 37hz square wave has a period of 0.027s or 27 milliseconds, you hardly need a microsecond (10 ^ -6) time scale for that
11) There is no rationale at all for the choices they made or their methodology, why use a 37hz square wave when you could use a higher frequency wave that would be far more challenging to a disc player? Why choose a DVD player and not a CD player ?
12) As mentioned earlier these noise figures are enormous, what do they mean ? A CD player has a noise floor of -96db (give or take) a -25db noise level added on by a cable looks bizarre and would swamp most low level signals.
13) Slide 10 - scale misleading, 0.0476 looks very close to 0.02586, the loss between original and "genesis" is 0.02176 , the difference between genesis and worst case is 0.00048 over 200x times worse
I would be happy to run some stats on the raw data if Virtual Dynamics would like to supply it.
I hope the final paper fills in the gaps, I look forward to reading it.
Last edited by nick_charles; 06-17-2008 at 09:39 PM.
I'd like to know who performed this study. Was it sponsored by a department or was it just some students? Were the people who performed it compensated in any form?
__________________
UNCLE ERIK Vinyl, Tubes & Grado
Orbe SE -> SME IV -> Fi Yph -> Zana Deux
RS-1, HP-2, HF-1, K-1000, K-701, K-501, K-340, K-240DF, HD-650, HD-600, HD-414, DT48, DT880, MDR-SA5000, ATH-6, Aperio Alpha
Linear Mode
