A/D Conversion & Bus Mixing Tests

Digi002 vs. Apogee AD-500 vs. ProTools LE
(Tested 16-bit A/D converters with 4-bus and 1-bus mixes,
with comparison in-box bus mix & bounce to disk)

OK, here's the deal: I set up the multitracks submixed to 4 stereo busses on the Mackie 24•8. The main TRS outs went back to the Digi002's line ins, and the the XLR main outs went to the Apogee AD-500 converter, then into protools via the SPDIF input.

(These are 320kbps mp3 files, nice and fat, and are about 45 seconds each. They're a bit low in volume,
as they are completely uncompressed - just turn up your system and enjoy some old-school amplitude variations.)

Then I routed all the tracks to the L/R bus of the board, bypassing the 4-bus submix. I recorded that mix again to the Digi002 and the Apogee.

(I'm aware of how ridiculous trying to gauge A/D conversion nuances from
mp3 files is, but 320kpbs is really quite good. There are AIF loops below, too.)

Then, just for good measure, I bussed the mix internally to a stereo track and recorded a mix straight back into the session. Finally, I bounced a mix to disk, and imported that back into the session. I lined up all six conversions with the zero point and started to listen.

inbox-bus.mp3 inbox-bounce.mp3

(The only other test I could think of was an in-box submix, but that's a whole 'nother setup.)

There are definitely differences in the mixes, although very very slight indeed. I actually preferred the single bus mixes to the 4-bus ones, and the apogee seemed to have a bit more clarity than the Digi002 converters. But the in-box mixes surprised me by sounding better than all the others, and most unexpectedly the bounce was clearly and positively distinguishable from the bus recorded mix!

I've made a 4-bar loop of the six conversions - you can line them up in Protools or Reason or whatever and solo back and forth to really check out the fine details. There are very slight volume variations you may wish to compensate for with small fader moves for each loop. I decided not to resample any files with a gain change, but they're pretty damn close already. Here's the collection (6.9 meg zip archive of six 16-bit stereo AIFF loops):


(These are the original AIF files - crank up your best monitors or get some
headphones and let me know what you think of each conversion!

A Visual View

Fun stuff, right? Now with all six mixes lined up to the same zero point, you can check out the comparison shots. Here you can see that the mixes look pretty consistent and aligned:

If we zoom in closer, you can see that there is almost no variation in phase, even between out-of-box and in-box bounces. Protools LE seems to be handling the latency issue in it's "magical" way.

Here's the view at the sample level - slight variations finally appear. Top to bottom, it's board-4bus, apogee-4bus, board-1bus, apogee-1bus, inbox-bus and inbox-bounce. Weirdly, the Apogee matches the inbox mixes precisely, and the mixes coming back in via the Digi002 are 3-4 samples behind!

Thanks for joining me on this episode of Astounding Adventures in Audio. See you all next time and remember: Stay Cool, Stay In School.

©2007 Dubtrain Records