Find Holes In My Logic

[Empty Planet]

Hello all.

I worked out a method of testing my converters and I'm wondering if there's some giant logical point that I'm missing.

Lowly up-and-coming slut that I am, I have a years-old Darla24 (24 bit/96khz). Problem: Are my Darla's wanging my sound before I even begin mixing? Test: Run a line from one of Darla's outs to one of its ins, then run multiple generations of audio, always using the last-imported audio to feed out and then back into the converters. Like a photograph endlessly copied on a Xerox, I'm reasoning, the nature and degree of degradation should become apparent.

Assuming -- for the sake of the argument -- no flaws in my monitoring system, are there holes in my logic?

Briefly, looking for high frequency content, I used separate samples of a bell tree, a triangle, a clave hit (which I ran through a giant reverb to check degradation of the tail) and then just for the hell of it a snare and a beefy kick. Matching levels was a bit of a chore, as Cubase's buss meters are hardly the best; I wound up using RMS Buddy from DestroyFX. I ran twelve generations (at 24 bits/44.1 khz), and finally compared the last to the original itb samples which had never seen my Darla.

Result? Yup, some degradation, (in two or three of the sample hits) but much less than I would have expected from twelve runs through the Xerox, and nothing I feel that I would notice in a mix (of course the cumulative effect of multiple tracks should be considered, but usually just about a quarter or third of my tracks require the conversion). But that was after twelve generations, which actually is more like 24 conversions, counting all the ins and outs, while my usual tracking method requires I just hit the converters once.

I realize it all comes down to what I personally can live with, and I may well invest in superlative converters in the coming year or two, but I'd love to hear y'all's thoughts on the subject. thanks.

[u b k]

here's what i get from your tests: your a/d is mostly capable of capturing what your d/a is capable of producing.

the test is self-limited, in that you're not determining the absolute limits of the converters, only the relative ones. if you could listen to a sweet cd thru your d/a's toggling with a pair of sublime converters, you might be shocked. then, if you took the output of the sublime converters and ran *that* thru your a/d, you might be even more shocked.

if you took a beautifully played acoustic, mic'd up nicely in a sweet room, and ran that thru your a/d's as well as a pair of lavry's, and played the tracks back thru the lavry d/a's, you would then get a sense of what your converters are (and aren't) doing.

does that make sense?


gregoire
del ubik

[Empty Planet]

Brilliant. It's a closed system. Exactly the kind of thing some part of me was suspecting.

Doesn't mean the converters are crap, but definitely points out the limits of the test.

Thanks!