Lightroom: LR4 Clarity2012 weaker than ACR 6.7 RC Clarity2012

The release notes of LR 4.1 RC2 state that this problem is fixed now, quote: "The following issues have been fixed in the Lightroom 4.1 release candidate 2 (April 26, 2012): [...] Lightroom 4 and ACR 6.7 render files differently."

But as it seems this is NOT fixed: The situation is exactly as before: http://forums.adobe.com/thread/974963

Does it work on a completely new file that LR and ACR have never touched, in case it was an error in the metadata, not the rendering engine?

Or maybe it is fixed in ACR 6.7 that we haven’t seen, yet?

The 6.7 RC has a bug in its Clarity implementation. The final release of 6.7 will match the Lr 4 behavior.

ACR 6.7 final has just been released and now images with large clarity look nearly the same in LR4.1 RC2 and when edited in Photoshop CS5.

However, there still seem to be very subtle differences in brightness.

Have you tried doing a Difference operation on two 16-bit TIFs of the same RAW that have had Clarity applied, one saved from LR as ProPhoto and one saved from ACR as ProPhoto, and doing the comparison with of a pair of Clarity 2010 TIFs as well as a pair of Clarity 2012 TIFs?

The former being a control for detecting subtle differences in the underlying profile conversions across different platforms.

Meaning, if you don’t see any difference between Clarity 2010 TIFs but do see differences between Clarity 2012 TIFs then there is some issue, still.

I just tried that. I used clarity +100 for both, and additionally for PV2010: Recovery 40, Fill Light 30, and for PV2012: Lights -50, Shadows +55, White -30, Black +60.

Then exported two versions of the file, the first directly from LR, and the second from Photoshop (after "Edit in PS" in LR). Loaded both back into PS into two layers with mode "difference", added adjustment layer for tone control with white point set to minimum (3). Unfortunately, I did the whole experiment in sRGB, but that should not really matter

Result for PV2010:



Practically no differences (RGB values vary around 0...1).

Result for PV2012:



Differences are there. You can see the differences also when comparing the two images side by side.

EDIT: Here are the original pictures in PV2012, first exported from LR, second saved from PS (scaled down for the forum using IrfanView, you don't need the full resolution to see the differences):





As I said, the differences are very subtle, but they are there.

P.S. Side note: At present, a comparison does not really make sense IMHO. A comparison should be done using LR 4.1 final (!) and ACR 6.7 (or ACR 6.7.1/6.8 if or when the issues with clarity artefacts and missing defringe are resolved - or ACR 7.x final).

With the PV2010 control added to the mix, the output colorspace shouldn't matter all that much.

One more question/variation, can you do the PV2012 difference w/o any lens-geometry corrections occurring--the section turned off or all the corrections turned off?

There is a current bug in LR4.1RC2 and ACR6.7 where the pre-corrected image edges seem to be used for the boundaries of the various PV2012 toning operations, in case that is really where the problem lies.