The implementation details might (or might not) result in a performance penalty for additional local adjustments, therefore I'd prefer to prioritize mine. I think the team is well aware that having everything localizable is a desirable thing, but if (!) there is a penalty, I'd prefer WB and NR to be top of the list. Creative sharpening is, I feel, important as well, but it might be a separate item (or not) from the local adjustments.
I agree that all adjustments should be localisable. A performance penalty should only occur for those using computing intensive adjustments.
If it has to be a subset of controls, creative sharpening and HSL control are ones that should be high up the list (The latter has been suggested as an idea already).
Although I'd even like to be able to apply camera calibration profiles locally, and lens corrections, and even color noise reduction, I realize some things may require bigger design changes and trade-offs than others. I'm OK with *some* things being global if "necessary", so:
Priorities:
-----------
1. Lum. Noise Reduction *and* Sharpening Parameters (not just amount) (I now see these as parts of the same whole).
2. Color Adjustment (e.g. HSL, WB)
3. Tone Curve
There is zero lag for me when using the adjustment brush - very spunky - I can't brush faster than Lr can keep up...
Biggest part of fix was new motherboard - although motherboard faster, I think a bigger factor may have been the extensive system re-configuration and new drivers that occurred in the course of installing it.
Rob, I bet if you keep on adding adjustment brush strokes, at some point things will slow down. It has always been like that for me and I've been told that it never had been any different, even in old ACR versions.
Interactive performance should not go down with the number of local image adjustments, but it does.
TK, I'll take notice. But, I do a fair amount of brushing and haven't noticed it yet, granted I wasn't looking for it either. In any case, its 2-orders of magnitude faster now than before, straight out of the gate, so a bug is no longer biting... - at least some of the problem for some people is due to "abnormal" behavior. If that is the case for you or anybody else, it would be good to get past it if possible...
Rob, do you also see a considerable improvement in interactivity when you reduce the LR window size?
I love to work in full-screen mode (the only proper way of using LR, AFAIC). I also like to see the image I'm working on as big as possible, so I collapse all panels, except the develop panel. This means, however, that if I move e.g., the exposure slider, the updates of the image rendering lag behind a little. If I move the slider by a large amount, the image rendering catches up in a jerky way.
If I leave full screen mode and reduce LR's window size then suddenly updates to the image are practically immediate! The editing experience becomes truly interactive then.
I cannot believe that my hardware isn't powerful enough to cope with displaying a bigger image. I believe this is an age-old ACR problem, isn't it? Do you still observe it on your system? Is it justified to create a bug report?
How I wish I had the responsiveness of the "small window"-LR in full-screen mode!
Performance at full-size is *about* the same as small-window, for me. - not quite, but so good that I never even think about it - I use almost exclusively in "full screen" mode... I'm guessing its more likely a (display?) driver issue than a horse-power issue, (assuming its not just a difference in perception of the same thing) but I'm a bit out on a limb again... - I think this has been an ongoing issue in the other forum, e.g. "Lr 3.3 Performance Feedback", but it may be worth seeing what kind of response you get here.
Hello,
i would like to use all adjustments with the "local adjustment brush", not only the few offered; like the color correction, noise, sharpen and so on.
BR
Sascha
For performance reasons we can't reasonably do ALL adjustments on a local basis, at least not with the current design. We have to focus on what makes sense to localize the most, i.e., what would be the most useful, provide the most bang for the buck. Photographically, for example, it makes sense to want localized exposure, but probably not localized keystoning (though certainly would be an interesting effect ...)
I appreciate your feedback on this topic Eric, especially regarding to your comment about the "current design". In the longer term (but not too long ;^)) I would like to see a redesign. In the short term, the items I would like to see added, if feasible, in order of importance to me:
I have situations with long lenses (600mm) showing both blue and red CA where LR can only correct for one or the other but not both. If I could apply CA locally then I could completely remove the CA. BTW, nikon software can completely remove the CA.
ok, if i have to select my top ones, it would be color and sharpening
concerning the performance, is it not possible to have two local adjustment versions, the current one, and an advanced version, this then triggers an internal copy of the image (on the HD) where you make all changes persistent, so you do not have to calculate all on the fly?
so fake it for the cost of disc space ... as role back to original still possible
sascha, you can already do local sharpening in LR2 and LR3.
We're not keen on making intermediate rendered versions on disk, partly because writing huge temp files is also a big performance hit, and partly because that eliminates the non-destructive aspects of the workflow.
Rory, you could, but frankly local CA correction is very suboptimal (you wouldn't want to be doing this on every image, separately). Have you tried using lens profiles instead?
Eric, I have tried creating a lens profile, but I found it challenging to do with a 600mm + TC14. The profile I created did not resolve the CA issues. I agree with you that localized CA is a kludge (suboptimal). I was just looking for a quick hit.
As I think about this more what I really want, in the short term, is a fringe removal brush, that would deal with fringing, tough CA and halo issues. This would be extremely useful for CA, fringing and cleaning up after auto mask.
I would be very opposed to them designing a tool to clean up after the failings of another tool. I'd rather they spend that time fixing the failings!
Fringing (longitudinal CA) is something that's really hard to correct properly. I wonder if there's a simple way to be more effective than the current fringe tool which seems to have too small of a radius.
I agree that fixing the shortcomings of existing tools is important and can often lead to improved workflow. The problem with just making things local is that local adjustments are inherently hard to sync meaningfully across a bunch of images.