OK then: a test suggestion for Bybee and other magic dust
Hello to all.
Based on the replies made to the suggestion we investigate or develop methods similar to the ones Acuity ran for Nordost, and see if the Bybees show any blips with similar tests, well... I'd like to stir the pot some more.. even if I do "doubt the bybee", I wanna know for sure.
BTW... before I get constructive, and believe me, I will... some tongue in cheek to make an important point... Please feel free to skip ahead if you're feeling a little roasted...
SIR!
I have no desire to "insult the Audio Engineering Society" by implying they maybe missed out on something... (wow, I didn't know the AES had become a Holy Orthodoxy that needs defending from heretic free thinkers...)...
but I AM going to suggest that we need to take another look at testing with non-linear content, musical and otherwise, and I would like to suggest a few ideas for testing this way. Why? Because it seems clear to many of us that we are still hearing things that don't show up in the most commonly applied tests. Even if the "Holy Audio Orthodoxy" says the earth is flat, some of us somehow still free to suspect otherwise.
Remember THD was king until Matti Ottala showed us all about TIM? Remember how BAD some of those POS transistor amps sounded?
Oh wait, then the doubly blind men showed us we can't possibly tell the difference... and we suddenly couldn't hear any other forms of distortion because we don't have measurements for them... Yeah i know, we just can't hear cables... uh huh.
OK, so here's an actual suggestion to test for audible non linearities:
1) Start with something like bwaslo's audible diff tester. With it's time correction features, that method looks promising indeed.
And you know what? With some additional processing, we might be able to extend its functionality and generate measurable output. How? Well, use a really high resolution storage scope to measure diff-maker's output and apply boatloads of averaging to see/measure/listen to it with huge s/n ratio. ("Averaging?" I can already hear them say: Gadzooks man, if we have much time variance that won't work! well, as stated above, bwaslo already took care of some/much of that, but then add #2 and #3 and you have even mo' resolution)
2) Use a really good D/A. (try Antelope Audio Zodiac+).
reduce jitter and time errors by syncing everything up with a good external clock or syncing to the Zodiac's clock
3) Do NOT use a CD or a hard disk for the signal source! Store a variety of demanding HiRez 192/24 (go to 384/32 when they become available) musical test files on a good Solid State Drive and use that as a source instead of... anything with moving parts - eliminates a bunch of mechanical and jitter/timing variables and makes the tests a LOT more repeatable relative to time.
4) Using #3 above, because the signal may be MUCH more repeatable in time, it may now become possible to select very demanding musical signals... play them hundreds/thousands of times... and measure, store and average the output from the diff-maker... you get much higher test resolution by averaging many hundreds of tests together...if the time sig is repeatable you can go waaay beyond 24 bits...
and non-linear anomalies that affect the audio band may show up more clearly, (even if said anomalies are well beyond the audio band, if they are audible, they may show up more clearly with a good high s/n non linear diff test such as proposed)
So, again, refer to suggestion #3 .... from what I've been able to find, SSD's would exhibit much less of this time variance, also they are a much lighter load on the 'puter's power supply, a non trivial issue when compared to an optical drive or hard drive. And they seem to sound much cleaner than Hard drive... I can hear the differ...(uh-oh, another controversy brewing.)
Other considerations:
5) Maybe reduce power supply and ground plane noise issues by using battery power where practical. Use freshly charged batteries with way overspec'ed current capacity.
6) Test inside and outside a D'Arsonval cage to determine differences. It's entirely possible some of these bybees/cables/terminators/magic devices are simply filtering some RF artifacts, and it would be good to see what happens if we put 'em near a strong rf signal and can hear/test a difference.
7) Check the effect of isolation transformers between DAC outputs and the amps being used for testing. Once upon a time, I thought a DAC I was auditioning sounded really bad, then I found out it was reacting to the class D amp it was driving... when I put a good Jensen isolation transformer in between 'em the DAC sounded very good indeed.
8) To test cables and bybees use a variety of loads, and make sure at least one of them is a real speaker that presents truly nasty low impedances and highly reactive loads.
Test them at very low, middle level and extremely high power ranges. At high volumes some speakers show some really big impedance changes, it'd be interesting to see if any of these magic bullets help at high levels vs low. (Yes, I doubt it also, but test first, then talk trash.) Use earplugs.
9) Use a variety of test amps. Determine their audible diff characteristics.
10) test the cables being used and determine their audible diff characteristics Listen and also look at the differences using a storage scope, run many tests with the same track and average the time signals (if they are not jittery) and measure the output of the diff-maker.
Again: Yeah, I know, the Bybee is probably bogus, but if we SAY we're gonna be fair, let's actually BE fair, OK?
Methinks the preceding methodology will be much more useful in helping to determine what makes some cables/capacitors/inductors sound different from others... but IF the Bybees also show a blip here, well, wouldn't that just be a real embarrassment to some of us cynics!
In the meantime, please look over these suggestions and see if they hold up to scrutiny. I think they may work very well indeed, but I do not have the equipment, nor the time to do this. Any takers?
|
