Closed johnhaddon closed 8 years ago
Any reason this has been moved to this milestone? I put it in 3.03 originally, where it has a chance of getting done on time - it's a pretty big one and it's not gonna get done by this Friday.
No reason... moved back now...
Is it on the books to look at the performance of the pixel sampling too?
I've just noticed that the bigger the bounds of the image in screen space (e.g. if i maximise my session rather than have it in a small floating window) the slower the sample values update when scrubbing around the image.
Also - I think the pink area sample box slows down the bigger the box is.
Feels like some extraneous work is going on there somewhere....
Could you give us exact steps to reproduce the first one? I don't see that here.
As for the area sample box, it is currently expected that that would slow down when increased in size as it has to average more pixels. We should be able to deal with that pretty easily by using an internal computation which caches the statistics for whole tiles, so we only need to keep recalculating the stats for partial tiles...
I'll take a look at that if I have time in this milestone, but depending on how the other stuff goes, I may not get to it.
I've tried reproducing the first note on my workstation in the office, and had no luck. Previously I had been running on my laptop - so possibly issue is related to my other graphics problems.
Will shout if I see it again on a machine with reasonable hardware
See also #245.
Have we put this on hold for a little bit now that we have some of the speed upds? If so, I'd propose switching it to m3.06 so it doesn't appear overdue.
Sounds good to me.
Closing, since the colour picker is in the right space now, we have an HSV view, and we've removed the stats for now.
Currently the min/max/avg stats are computed before colour transformation (as they should be), but the "colour under the mouse" display is not. This needs fixing. We should also :