While I tend to agree with most of the comments that were made, a bit more looking also reveals a bit more to the story. in particular, they seem to have at least as obvious of an anti-Sony (or perhaps pro-Canon) bias as any I've seen yet.
For example, they've recently tested both the Sony 35/1.4G and the Canon 35/1.4L. I've combined their resolution testing on the Canon 5D and the Alpha 900 into a single chart (attached).
Wide open, the Sony is a bit better at the edge, and the Canon is a bit better at the center.
At f/2, the Sony is better at both the center and the edges.
From f/2.8 to f/8, the Sony is better at the center, but the Canon is better at the edges.
At f/11, the Sony is better at the center, and they tie at the edges.
At f/16 and f/22, the Sony is again better at both the center and the edges.
In terms of astigmatism, the Sony does quite a bit better than the Canon as well (no numeric rating given, but looking at their pictures, I'd say the Canon looks about twice as bad in this respect.
Likewise, the Sony does better in terms of corner shading. Neither is perfect by any means, but the Sony shows 2.6 stops of light drop-off wide open, vs. 3.5 stops of light drop-off for the Canon wide open.
The Canon did score better in one respect: it does have lower lateral CA. Personal experience with a Minolta 35/1.4 (pre-G designation) has shown that its CA is quite correctable though. Since it is correctable, the CA isn't usually going to be a problem in itself. Better still, it's probably the source of a great deal of the lack of sharpness they observed at the edges (their test procedure apparently does NOT include correcting for CA before testing resolution). Correcting the CA might easily allow the Sony to equal or beat the Canon for resolution across the board.
The two are essentially tied in a few respects (e.g., distortion and Coma).
Overall, the objective testing seems to indicate that the Sony is generally the better lens -- the only respect in which the Canon clearly wins is also the most easily corrected, and doing so probably gives the Sony a clear win in resolution as well.
The LensTip.com commentary doesn't reflect anything of the sort though. Quite the contrary, based on the commentary, you'd have a hard time concluding that the Canon was anything less than drastically superior in essentially every possible respect. Nearly everything they have to say about the Canon is almost gushingly positive, with a few sample quotes below:
Nearly everything they have to say about the Sony is at least equally negative:"...it is truly outstanding, as it befits an expensive L-series!"
"brilliant frame center picture quality at all apertures range"
"very good frame edge picture quality"
The objective testing does not seem to show the drastic difference in quality that would be necessary to support the difference in the commentary."In both cases, the performance at wide aperture is weak. A level a bit over 20 lpmm is far from the result of 30 lpmm, which we consider acceptable."
"When we look at the picture below, we see that there are even more troubles in the frame edge. "
"That’s why, while working on A900, the lens never achieves the decency level..."
"Taking into account the results of this test, I am surprised that Sony continues to produce this lens. It would be perhaps better if the lens was buried and forgotten with the late Minolta company. A new lens should be designed - a simpler, smaller and cheaper instrument with 2/35 parameters, which would please the whole crowd of small and big sensors owners. Let’s hope it will happen soon and with this hope we can end for now. The further description of Sony 1.4/35 performance, taking into account its exuberant price, would be like kicking an underdog…"
P.S. I should also note that I've ignore the testing on crop-sensor cameras for the moment -- mostly because I can hardly imagine anybody buying either of these exclusively for crop-frame use, and partly because it doesn't seem to be drastically different from the full-frame results.
