Looking at our results:
9 people in total responded.
4 stated they could not hear any differences.
5 stated they could hear differences, BUT
4/5 of us made incorrect pairings on BOTH songs. In other words we never identified a sonic difference that was due to the mod. It's not just we wrongly thought stock was mod and vice versa, we thought stock paired with mod in all cases rather than correctly pairing stock with stock and mod with mod, to say nothing of identifying which was which.
1/5 of us got one correct pairing (for one song) and one wrong pairing (for the other song).
The way the test works, for each song we could either correctly pair the two stock samples together (which means we therefore also paired the two modded samples together), or we could match up stock with modded (if we match one set of stock and modded, the remaining pair must also consist of one stock and one modded). There are in fact two ways to do an incorrect pairing and one way to pair correctly, meaning if we picked randomly we had a 1/3 chance of getting each song right:
A-B (wrong): tonedeaf
A-C (wrong): Will_S (by default if B-D pair), notsofast, rowjimmytour
A-D (right): Brian Skallinder
E-F (wrong): rowjimmytour
E-G (wrong): Brian Skallinder, Will_S, notsofast, tondeaf
E-H (right): nobody
So the numbers pretty much speak for themselves. I don't think there's an audible difference between a stock FR2LE and a Busman transparency modded deck UNDER THE RECORDING CONDITIONS OF THIS TEST. Other conditions, and/or other mods, might be a different story. And the short clips, from different points in time, are a definite drawback. But I think this test was a good representation of the conditions I record under, and would expect a significant difference to remain audible even under these less than ideal testing conditions. Personally, I don't plan on getting my deck modded, but encourage you to listen to the samples and decide for yourself.
With such a small sample size, looking for any patterns in our answers is a dubious exercise as spurious relationships can easily come about by chance. That said, since we were all over the map on song 1, it seems reasonable to conclude that not only were we unable to tell stock from modded, but there was no clear difference in tonal quality due to external factors (eg, mix adjustments by the FOH engineer) that were misleading us into lumping unrelated tracks together. The most popular grouping was A with C. The actual temporal sequence in which trackes were recorded was D, A, (switch gear), B, C. So, it seems unlikely that there was an FOH adjustment that would have led us to group A and C together.
On song 2, which actually came earlier in the soundcheck, one might wonder about the fact that so many of us lumped F-H and E-G together (although tonedeaf and [probably] Brian disagreed with notsofast and myself as to which sounded better [reading between the lines to infer a preference for Brian]). In fact, if we were simply picking randomly, there is about a 4% chance that 4/5 of us would pick the same choice out of three options (google binomial probability if you want to see how I calculated that). Arguably, most statisticians would say when the probability of such a result coming about by chance is so small, we can reasonably talk about alternative hypothesis to explain why we identified a difference. But, significance is arguable in this case, because really we should be doing a significance test on the number of correct pairings across both songs, and overall we clearly did no better than would be expected by chance. By doing a second test on just a subset of our data, we are perhaps "data mining" and need to use a more stringent standard for statistical significance.
What strikes me as a possible explanation for so many of us pairing G and H together - given that this was the earlier song in the soundcheck, and the temporal sequence of samples was F, (switch decks), H, E, (change decks), then G - is that the FOH changed the mix slightly before sample E, thus causing most of us to identify F and H as linked. So perhaps this was not an ideal test, but it does seem clear that the sonic impact of the mod (again,
in this recording scenario) was small enough to be completely overshadowed by (possibly) a slight tweak at the board. You can listen for a change in the mix yourself here:
http://homepage.mac.com/satterwill/FR2LEComparison/CompTransition.flac(consists of H, the intervening period, and E)
So I suppose it's possible Brian could in fact hear the difference between decks on the first song, and was misled into not listening more closely to the second song due to changes in the mix. MAYBE, because as digifish_music mentioned with enough people "guessing", someone should get it right by chance. Clearly the rest of us were not hearing the effects of the mod. So maybe 1 out of 9 people can hear the effects of the mod under these conditions.
Here are the actual sources of each track:
A: Stock
B: Mod
C: Mod
D: Stock
E: Mod
F: Stock
G: Stock
H: Mod
But as I said, under other conditions the differences may be audible. For instance, maybe you can pick out mod vs. stock in these tracks, provided by page. Hopefully he will explain what we are listening to at some point:
http://homepage.mac.com/satterwill/BmodLab/P-ayurveda-TRS.flachttp://homepage.mac.com/satterwill/BmodLab/Q-ayurveda-TRS.flachttp://homepage.mac.com/satterwill/BmodLab/P-miller-TRS.flachttp://homepage.mac.com/satterwill/BmodLab/Q-miller-TRShttp://homepage.mac.com/satterwill/BmodLab/P-no-cords-just-TRS.flachttp://homepage.mac.com/satterwill/BmodLab/Q-no-cords-just-TRS.flac