Timeline for Color degrades when using Copy As Bitmap
Current License: CC BY-SA 3.0
40 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 6, 2015 at 21:02 | comment | added | Mr.Wizard | @Oleksandr Happy Birthday from me as well. :-) | |
| Nov 6, 2015 at 13:56 | comment | added | Alexey Popkov | @Oleksandr Happy birthday! :) Since I observe similar behavior with all tested MMa versions, v.9 or 8 on Win2003 is sufficient. | |
| Nov 6, 2015 at 13:21 | comment | added | Oleksandr R. | @Alexey I will try this later (if I'm not too drunk--it's my birthday today). At the moment I am in the office and only have Windows 7, which I assume won't be useful to you. Note that Mathematica 10 can't run on Windows 2003. | |
| Nov 6, 2015 at 13:03 | comment | added | Alexey Popkov | @Oleksandr It was my mistake: from your first comment I assumed you have. But what about Windows 2003 and 8.1? It seems for me that the fundamental bug I describe in the UPDATE section is purely a bug in the FrontEnd and it should be present on any version of Windows (and is present in all MMa versions I checked). It also seems to be unrelated to Windows color management (unlike the original bug). | |
| Nov 6, 2015 at 12:45 | comment | added | Oleksandr R. | @Alexey I'm sorry, I don't have Windows 10 yet (I didn't update my laptop so far). I only have 2003, 7, and 8.1. | |
| Nov 6, 2015 at 11:45 | comment | added | Alexey Popkov | @Oleksandr Please check updated version of my answer. Do you observe the same behavior on Windows 10? | |
| Sep 8, 2015 at 20:49 | vote | accept | Mr.Wizard | ||
| S Sep 8, 2015 at 20:49 | history | bounty ended | Mr.Wizard | ||
| S Sep 8, 2015 at 20:49 | history | notice removed | Mr.Wizard | ||
| Sep 8, 2015 at 6:08 | comment | added | William | I can't tell at first but this seems to be related mathematica.stackexchange.com/questions/26105/… | |
| Sep 6, 2015 at 19:24 | history | edited | Mr.Wizard | CC BY-SA 3.0 | added 911 characters in body |
| Sep 6, 2015 at 19:19 | history | edited | Mr.Wizard | CC BY-SA 3.0 | added 911 characters in body |
| Sep 6, 2015 at 7:17 | answer | added | Alexey Popkov | timeline score: 13 | |
| Sep 6, 2015 at 0:56 | history | edited | Mr.Wizard | edited tags | |
| Sep 5, 2015 at 21:46 | comment | added | Sjoerd C. de Vries | Given that it's neither there on OSX nor on Vista suggest to me it is unintended. | |
| Sep 5, 2015 at 18:34 | comment | added | Mr.Wizard | @Sjoerd I was waiting for support of that position. I suppose here is a slight chance this is intentional? | |
| Sep 5, 2015 at 6:11 | comment | added | Sjoerd C. de Vries | BTW shouldn't we tag this with the "bugs" tag? | |
| Sep 5, 2015 at 6:06 | comment | added | Sjoerd C. de Vries | @Mr.Wizard I would expect that too, but since your question seems to be about bitmaps specifically the workaround you asked for would be simply keyboard copy 'n paste, wouldn't it? I checked ImageData of the bitmap I used and it doesn't change under that operation. | |
| Sep 4, 2015 at 23:52 | comment | added | Mr.Wizard | @SjoerdC.deVries I expect to be able to select any object or group of objects in the Front End and use Copy As Bitmap to make a verbatim bitmap copy. Image, Raster, Graphics, or otherwise. I do not expect that operation to change my colors. | |
| Sep 4, 2015 at 16:07 | comment | added | Sjoerd C. de Vries | @Mr.Wizard As I already noted above, copying a bitmap using keyboard copy-and-paste does not change the bitmap. This only happens when you use the copy-as-bitmap menu. So what workaround are you actually looking for? Or is the problem that non-bitmap graphics change colors when copied that way (I didn't test that)? | |
| Sep 4, 2015 at 16:00 | comment | added | Sjoerd C. de Vries | I should note I checked this on win 8.1 | |
| Sep 4, 2015 at 15:57 | comment | added | user8074 | @SjoerdC.deVries I cannot reproduce this in version 9.0.0.0 on Windows Vista. So, perhaps, VIsta is still ok. | |
| S Sep 2, 2015 at 17:31 | history | bounty started | Mr.Wizard | ||
| S Sep 2, 2015 at 17:31 | history | notice added | Mr.Wizard | Authoritative reference needed | |
| Aug 30, 2015 at 8:22 | history | tweeted | twitter.com/#!/StackMma/status/637903119557509120 | ||
| Aug 30, 2015 at 1:03 | history | edited | Mr.Wizard | edited tags | |
| Aug 29, 2015 at 23:58 | comment | added | Oleksandr R. | Okay. Well, all I can say is that I was able to set it as a monitor profile on my computer, so maybe something is different about yours, or I did it a different way than you did. I don't have either Mathematica 10 or Windows 7 to hand at the moment, so I can't really comment any further on that. As for the reason why, I don't know. It is probably due to copying in DIB format on Windows, whereas other platforms may use a different approach that does not involve color profiles, or does it differently so as not to expose the bug. | |
| Aug 29, 2015 at 23:52 | comment | added | m_goldberg | Should this question be tagged as a Windows problem? I can't reproduce it on OS X. After making six copies of a bitmapped gradient I see no degradation. | |
| Aug 29, 2015 at 23:39 | comment | added | Mr.Wizard | @Oleksandr Not all RGB profiles are recognized by Windows as valid monitor profiles; I believe the file must be .icm not .icc though I have not explored it further. Why should Mathematica be transforming on either copy or paste? I think the RGB triplets should not be touched, and the monitor profile should be used for display only should it not? This is how Windows itself works which is why Print Screen doesn't change anything, copy or paste. | |
| Aug 29, 2015 at 23:25 | comment | added | Oleksandr R. | The sRGB profile should be valid for monitors (although the other one isn't; it's for printing onto coated paper). I agree that it is probably a bug. I think Mathematica transforms one way (either on copy or on paste; clearly it is on copy according to you, and seemingly also in the linked thread as the problem is visible in the SEUploader) but fails to apply the inverse transformation. The same problem is what I suppose leads to the orange color of the interface. | |
| Aug 29, 2015 at 23:13 | comment | added | Mr.Wizard | @Oleksandr I was not able to set the ICC profile as monitor colorspace because it is not recognized as a valid monitor profile. However following your idea I set a different monitor profile and repeated the experiment. This resulted in degradation in a different pattern than the one shown above indicating this is related if not directly causal. Why would (or rather should) Mathematica be changing the data based on the monitor profile? This feels like a bug to me but I need confirmation. | |
| Aug 29, 2015 at 22:49 | comment | added | Oleksandr R. | @SjoerdC.deVries I cannot reproduce this in version 9 on Windows 2003. This is what makes me think it is due to color management (which is implemented much more simply on Windows versions before Vista; the same problem with slow startup also does not exist in XP/2003). | |
| Aug 29, 2015 at 21:57 | comment | added | Sjoerd C. de Vries | Copying a bitmap just by ctrl-C ctrl-V seems to leave the bitmap unaltered though. | |
| Aug 29, 2015 at 21:48 | comment | added | Sjoerd C. de Vries | On Windows Mathematica versions 8, 9 and 10 have exact the same problem. | |
| Aug 29, 2015 at 19:42 | comment | added | Sjoerd C. de Vries | This is probably the question OleksandrR. is referring to: my-mathematica-has-turned-orange-and-not-able-to-find-out-why | |
| Aug 29, 2015 at 18:06 | comment | added | Oleksandr R. | I am not sure if it is documented, because color management certainly does not work properly (it is the cause of the slow startup and annoying "formatting notebook contents" messages experienced by many, and also Nasser's bizarre orange problem). Anyway, if you search within the installation directory for *.icc, you will find there are a couple of them in there. These also exist in version 9 and the same slow startup problem exists in that, but the implementation is sufficiently limited that there are no other observable issues until version 10 as far as I know. | |
| Aug 29, 2015 at 17:54 | comment | added | Mr.Wizard | @Oleksandr I feel ignorant; I did not know that Mathematica had its own ICC profile. Please tell me more or give me the appropriate documentation link. | |
| Aug 29, 2015 at 17:36 | comment | added | Oleksandr R. | Probably one of the many problems with device color profiles in version 10. Have you tried applying Mathematica's ICC profile to your monitor and seeing if the problem persists? | |
| Aug 29, 2015 at 17:07 | history | edited | Mr.Wizard | CC BY-SA 3.0 | edited title |
| Aug 29, 2015 at 16:59 | history | asked | Mr.Wizard | CC BY-SA 3.0 |