Timothy OConn wrote:
Im certain those artifacts are caused by the beta nature of ACR/LR...Are they also present on the C1 or DPP conversions of this image?
They are not as apparent with DPP but they are still there, although DPP seems to smooth over smooth areas a bit more and its interpolation also tends to break solid lines into dotted lines so it looks a little less directly like mazing but it is definitely still messed up with DPP compared to how a 20D/40D/50D/5D2/1D3/etc. look. And the dotted lines and stray artifacts from DPP exactly overlay with the maze pattern from ACR so it is not something that ACR 5.5 is making up.
It is a little hard to compare converter to converter since all the sliders work differently and you need to try to keep noise and detail, not quite sure waht to say about C1v5 yet. It might be possible to get at least reasonably close to full detail while getting rid of mazing, not sure yet, haven't put the time into it you'd need to really find out for sure. C1 in some ways seem to act quite a bit like DPP, but it does have more and finer controls.
I don't like how either DPP or C1 leave bright white and dark black dots sprinkled all over, ACR seems to filter those out a lot better.
Anyway there seems to be at least some signs of it with all the converters so far, although ACR 5.5 makes it most obvious and C1 perhaps least.
In some ways the artifacts left behind might be more apparent as ISO goes up, with C1.
There are some hopeful signs from ACR 5.6Beta, although it is a little hard to be sure just judging by a jpg compression that someone else made with who knows what settings, the DPR samples are much softer than my conversions which could hide things, although if I dump a ton of sharpening on both the DPR samples still appear to be free of mazing. Maybe ACR 5.6 will be the one that finally can give really good quality out of the 7D under any circumstance without any reservations or catches. I hope they release it soon and that is does work well.
ACR 5.6beta sample on DPP also appeared to be more free of moire and stuff than C1.
i didnt have time to grab and upload another DPP convert like that but i can say it shows some mazing garbage too like ACR 5.5 if a bit more brokemn up, DPP does not handle it perfectly, not free of it like 5.6B above.
easy to see maze on 5.5 but not at all on 5.6B even with extreme overprocessing (granted the 5.6 did show a lot less detail as they posted, and averaging away detail is one way to get rid of these artifacts, but I did try to pre-sharpen it somewhat to even it up before applying the same amount of extra oversharpening to both, anyway it is a hopeful sign but it is still possible that the jpg compression or some NR they did or the new algorithm simply filtered away some detail along with maze, but there seem to be hopeful signs that acr 5.6 will deliver at least 50D detail while avoiding mazing and then we'd be set since the detail seemed like it might be pretty close once i sharpened it up to match)
now another major overshapening of ACR 5.5 just to show that it messes up even areas of detail, not just smooth areas, gives lots of fine 'detail' that is not really real at all as this makes very clear: http://skibum4.smugmug.com/photos/706954036_UH2xb-O.jpg
all those straight little lines are not actual fuzz detail on the dog or on those little fuzzy things, it's all fake artifacts that just look like super fine detail. didnt bother to waste time putting up caps from 5.6 but with that you can see where the real fur lines actually go and just how much some of the 5.5 and even DPP detail is really just pure artifacts.
Anyway, don't use ACR 5.5 or any third party commercial or open source converter for the 7D that does nto specifically do things to handle the 7D (not needed to handle other DSLRs), I don't think DPP does it that well either myself, it definitely leaves garbage behind.
Hopefully ACR 5.6 will do it very well. C1v5 not sure yet. And by do it well I men get rid of mazing without actually washing away microcontrast and giving up some of the MP count (DPP does seem to show it somewhat if you equalize detail captured with ACR 5.5). C1v5 on default settings gets rid of it but definitely gives up some fine crispness, but it seems like it just might be possible to bring that back and still avoid mazing by playing around with the sliders enough.
It looks like there is a real chance that ACR 5.6 might pull it off though and really make the 7D the best CROP DSLR bar none.
on another thread i think it was suggested that more than a G1/G2 imbalance the amp simply fluctuates a bit much from column to column even within G1 alone compared to any previous DSLR release making standard converters mess things up (maybe 5.6 is trying to read the mask area and compensate a bit?). A typical converter doing things as they typical did with previous cameras does tend to look messed up with the 7D to varying degrees though so the complaints have had some basis.
anyway, i've had enough wasting time on this today, going out.
if I had a nickle for every time I said that....well I'd only have .20c but still.