Re: [darktable-dev] Contribution
On Monday, May 9, 2016 8:05:36 AM CEST johannes hanika wrote: > hi wolfgang, > > great to hear you're interested in this! > > On Mon, May 9, 2016 at 2:11 AM, Tobias Ellinghaus wrote: > > Am Sonntag, 8. Mai 2016, 15:07:49 schrieb Wolfgang Mader: > > > > [...] > > > >> Licence of the Matlab implementation of the Local Laplacian Filters > >> "Start quote" > >> Copyright (c) 2011 Sam Hasinoff > >> > >> Permission is hereby granted, free of charge, to any person > >> obtaining a copy of this software and associated documentation > >> files (the "Software"), to deal in the Software without > >> restriction, including without limitation the rights to use, > >> copy, modify, merge, publish, distribute, sublicense, and/or sell > >> copies of the Software, and to permit persons to whom the > >> Software is furnished to do so, subject to the following > >> conditions: > >> > >> The above copyright notice and this permission notice shall be > >> included in all copies or substantial portions of the Software. > >> "End quote" > >> > >> [1] http://people.csail.mit.edu/sparis/publi/2011/siggraph/ > >> matlab_source_code.zip > > > > That is the MIT license: https://opensource.org/licenses/MIT > > > > You can freely use that code in darktable. Nice, thanks. > > well, you can't because it's matlab.. maybe you can run it through > octave. there are also two more implementations by andrew adams: I should have mentioned that I have a full Matlab licence containing all tool boxes. So this is a good starting point for me. > > one in halide > > https://github.com/halide/Halide/blob/c699d819194fa5fe299476dd3dfa8a25dece7b > f2/apps/local_laplacian/local_laplacian.cpp > > and one in imagestack: > > https://github.com/abadams/ImageStack/blob/master/src/LocalLaplacian.cpp Thanks for these pointers. Cheers, Wolfgang > > cheers, > jo > ___ > darktable developer mailing list > to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org signature.asc Description: This is a digitally signed message part.
[darktable-dev] Re: Lua GUI box model
Am 08.05.2016 um 14:03 schrieb jeremy rosen: > Ok, multiple questions in here, i'll try to answer them as I can > > ABOUT LAYOUT OF BOXES > > what you are doing right now is not possible, the widget elements are new > in lua and everything is not exposed to lua yet. I add stuff on a need > basis and I didn't need fine-grained control on the layout so far. > > GTK (which is the real resplonsible for drawing the UI elements) has lots > of complicated options to control layout. I'm not sure exposing them to lua > would be simple/easy. I'll have a look at what's in there wrt layout and if > some controls can be easily passed to lua for users to play with That would be great. I guess, several scripts that are already in the repository could benefit from more control over the GUI elements as well. Since Darktable uses its own set of GUI elements, that would not mean exposing all GTK features/elements, but just what DT uses. I mean, as a long term perspective :-) > RELOADING LUA SCRIPTS > > with the way lua works, it is not possible. you have to restart. This is > because lua doesn't load a script, it runs some code. there is no generic > way to revert what code did. your solution of setting package.loaded[] to > nil is dangerous. if it works for you, that's fine, but the risk of it > causing unexplainable behaviour exists. especially if you end up crating > the same thing (widget element, preference, shorctu...) multiple time I work on dummy data/database when working on code, therefore I feel somehow safe. However, a possibility to safely unregister/delete a lua-created module would be great. > LUA DOCUMENTATION > > As was pointed, your best source of documentation is the user manual (for > tutorial-like info) and the lua-api reference. (for complete reference) the > lua API, in particular is automatically checked against the source code and > is guaranteed to be up to date. no hidden feature*. If you are working with > master, I recommand generating it yourself and not getting it from the > website, so it contains all the latest stuff I know the existing documentation, however, my problem is somehow different. Not being a real programmer (electrical engineer by training) I have problems to see myself where the limits for Lua scripts are. Maybe, something not exposed in the DT Lua api may be possible to access/work around with pure Lua somehow. > LUA SHORTCUTS > > you can create shortcuts in lua, see > http://www.darktable.org/lua-api/ar01s03s03.html.php for details. This will > create a new shortcuts that will call some lua code. Note that you can't > use lua to assign shortcuts, only to create them. You need to assign a key > to your shortcut via the preference menu Any possibility to add functionality to dynamically show, register and unregister shortcuts including the keybinding from Lua? The entire set of shortcuts could be handed over similar to the image lists that are available. What I would need is the possibility to temporarily store the original shortcut config, temporarily remove most of them, temporarily set new shortcuts and restore the original config afterwards. Maybe this is possible to circumvent by trapping the key presses, but unfortunately I have no clue how that could work. > Overall i'd be interested in seeing your scripts,we do have a script repo > that accepts external contributions and any chance to have new > contributions is welcomed Thanks, I know the repo, I was already able to contribute a little fix to a script. For my actual ideas (and unfortunately I have a lot of them) I have very little code. I started with the speed tagging idea from the GUI side since it would not make sense without GUI. I wanted to check that it is possible first, because I am not able to add the missing items on the darktable side to make it run. Unfortunately, there is a lot missing to implement my idea, namely * fixes for better box alignment * dynamic access to shortcut configuration * more control over widgets, especially sizes, alignment and styles * backend to write configuration from lua (this is a nice-to-have, one could start writing an individual config file) Maybe I start with a simpler idea first. It would be great if I could ask about the limits of the current Lua API wrt my ideas from time to time here or, if that would be better, on IRC. Thanks for your detailed answer! Best regards Chris > *except the april 1st view, which can be accessed via lua > > On Tue, Apr 26, 2016 at 11:41 AM, Christian Mandel wrote: > >> Am 13.04.2016 um 10:20 schrieb Christian Kanzian: >> >>> Hi, >>> >>> Be aware I'm not a developer so my comments can be wrong. >>> >>> Am 09.04.2016 um 15:22 schrieb Christian Kanzian: > > [...] > > like the usermanual. What's written there is possible. If it is not > there you > simple can't do it. > That's one of the problems: Of course I try to learn from the documentation (especially the Lua API
Re: [darktable-dev] Contribution
hi wolfgang, great to hear you're interested in this! On Mon, May 9, 2016 at 2:11 AM, Tobias Ellinghaus wrote: > Am Sonntag, 8. Mai 2016, 15:07:49 schrieb Wolfgang Mader: > > [...] > >> Licence of the Matlab implementation of the Local Laplacian Filters >> "Start quote" >> Copyright (c) 2011 Sam Hasinoff >> >> Permission is hereby granted, free of charge, to any person >> obtaining a copy of this software and associated documentation >> files (the "Software"), to deal in the Software without >> restriction, including without limitation the rights to use, >> copy, modify, merge, publish, distribute, sublicense, and/or sell >> copies of the Software, and to permit persons to whom the >> Software is furnished to do so, subject to the following >> conditions: >> >> The above copyright notice and this permission notice shall be >> included in all copies or substantial portions of the Software. >> "End quote" >> >> [1] http://people.csail.mit.edu/sparis/publi/2011/siggraph/ >> matlab_source_code.zip > > That is the MIT license: https://opensource.org/licenses/MIT > > You can freely use that code in darktable. well, you can't because it's matlab.. maybe you can run it through octave. there are also two more implementations by andrew adams: one in halide https://github.com/halide/Halide/blob/c699d819194fa5fe299476dd3dfa8a25dece7bf2/apps/local_laplacian/local_laplacian.cpp and one in imagestack: https://github.com/abadams/ImageStack/blob/master/src/LocalLaplacian.cpp cheers, jo ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Re: sRGB internal color profile
Am Sonntag, 8. Mai 2016, 17:40:51 schrieb jmsarat: > Ok, it was just a wild guess. If DT is using its own color management > for export when lcms2 is not selected, what are the differences or > benefits of using lcms2? It is faster, but somewhat limited, most obvious limitation is the lack of configurable rendering intent. And it doesn't work for all color profiles. For those not supported by our code we fall back to using lcms2. So the results _should_ be the same. If they aren't in your case that's probably a bug. > Jean-Marc > > Le 08/05/2016 à 16:08, Tobias Ellinghaus a écrit : > > Am Sonntag, 8. Mai 2016, 15:54:28 schrieb jmsarat: > >> Hi Matthias, > >> > >> With a clean $HOME/.config/darktable, the option lcms2 is deactivated by > >> default therefore the export function use lcms1-1.19. > > > > That is a misunderstanding. darktable might use some internal functions > > instead of lcms2, but it will never use lcms1. > > > > [...] > > > > Tobias > > ___ > darktable developer mailing list > to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org signature.asc Description: This is a digitally signed message part.
Re: [darktable-dev] Re: sRGB internal color profile
Ok, it was just a wild guess. If DT is using its own color management for export when lcms2 is not selected, what are the differences or benefits of using lcms2? Jean-Marc Le 08/05/2016 à 16:08, Tobias Ellinghaus a écrit : Am Sonntag, 8. Mai 2016, 15:54:28 schrieb jmsarat: Hi Matthias, With a clean $HOME/.config/darktable, the option lcms2 is deactivated by default therefore the export function use lcms1-1.19. That is a misunderstanding. darktable might use some internal functions instead of lcms2, but it will never use lcms1. [...] Tobias ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
[darktable-dev] Re: sRGB internal color profile
Am 08.05.2016 um 15:54 schrieb jmsarat: > Hi Matthias, > > With a clean $HOME/.config/darktable, the option lcms2 is deactivated by > default therefore the export function use lcms1-1.19. The resulting image > does not have defect in the blue shade but if the lcms2 option is activated > then the problem appears. I do not see a difference with lcms2 activated or deactivated with either sRGB or Adobe RGB. Matthias ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Fwd: Problem with darktable.collection in lua
Ok, bug fixed, it will be in the next stable release On Tue, May 3, 2016 at 7:52 PM, Jefferson Ferreira wrote: > > Hi, > > I have seen an issue at the collections in the LuaAPI of the darktable. > The collections should return only the visible images in the lightroom > view. Instead, the darktable.collection is returning all images from film > roll. > > For example ... > > - film roll > filmA (5 photos) > filmB (10 photos) > filmC (5 photos) > > if I select the filmB and print the darktable.collection (lua), my output > appears like: > > filmA/foto1 > filmA/foto2 > filmA/foto3 > filmA/foto4 > filmA/foto5 > filmB/foto1 > filmB/foto2 > filmB/foto3 > filmB/foto4 > filmB/foto5 > > the #darktable.collection (length of the table) is correct ( = 10 ), but > the list always starts with the first film (album) of the film roll. > > to try fix this problem, I'm using the darktable.database avoiding the > functions ... > > for _,image in (darktable.database) do > if (type(image) == "function") then > break > end > -- process the photo > end > > thanks in advance. > > ___ > darktable developer mailing list to unsubscribe send a mail to > darktable-dev+unsubscr...@lists.darktable.org > ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Contribution
Am Sonntag, 8. Mai 2016, 15:07:49 schrieb Wolfgang Mader: [...] > Licence of the Matlab implementation of the Local Laplacian Filters > "Start quote" > Copyright (c) 2011 Sam Hasinoff > > Permission is hereby granted, free of charge, to any person > obtaining a copy of this software and associated documentation > files (the "Software"), to deal in the Software without > restriction, including without limitation the rights to use, > copy, modify, merge, publish, distribute, sublicense, and/or sell > copies of the Software, and to permit persons to whom the > Software is furnished to do so, subject to the following > conditions: > > The above copyright notice and this permission notice shall be > included in all copies or substantial portions of the Software. > "End quote" > > [1] http://people.csail.mit.edu/sparis/publi/2011/siggraph/ > matlab_source_code.zip That is the MIT license: https://opensource.org/licenses/MIT You can freely use that code in darktable. [...] Tobias signature.asc Description: This is a digitally signed message part.
Re: [darktable-dev] Re: sRGB internal color profile
Am Sonntag, 8. Mai 2016, 15:54:28 schrieb jmsarat: > Hi Matthias, > > With a clean $HOME/.config/darktable, the option lcms2 is deactivated by > default therefore the export function use lcms1-1.19. That is a misunderstanding. darktable might use some internal functions instead of lcms2, but it will never use lcms1. [...] Tobias signature.asc Description: This is a digitally signed message part.
Re: [darktable-dev] Re: sRGB internal color profile
On Sun, May 8, 2016 at 4:54 PM, jmsarat wrote: > Hi Matthias, > > With a clean $HOME/.config/darktable, the option lcms2 is deactivated by > default >therefore the export function use lcms1-1.19. Where from are you getting this? I do not think dt ever used lcms1. > The resulting image > does not have defect in the blue shade but if the lcms2 option is activated > then the problem appears. > > I tried with a fresh DT 2.0.4 compiled on Mageia 5 with lcms2-2.7 and with > the latest dev version 2.1.0+1082~g6e50d7a. The problem is there with both > versions. > > Jean-Marc Roman. > > > > Le 08/05/2016 à 13:09, Matthias Bodenbinder a écrit : >> >> Hi Jean-Marc, >> >> I can not reproduce your JPG output. Not even with your xmp file. I am >> using DT 2.0.4, self compiled. >> >> Have you tried with a different user resp. with a clean >> $HOME/.config/darktable ? >> >> Matthias >> >> >> Am 08.05.2016 um 12:35 schrieb jmsarat: >>> >>> Hi Jo, >>> >>> With this clean xmp : https://1fichier.com/?rd1q8touto >>> >>> I obtain this image : https://1fichier.com/?3uqu4ijx0g >>> >>> No style on export and using the image parameters for color profile >>> (sRGB(web)) and rendering intent (perceptual). >>> >>> The raw image is there : https://1fichier.com/?5n9xbgf9ej >>> >>> The problem is visible in the blue shade but could be present in other >>> colors. >>> >>> Jean-Marc >>> >>> Le 08/05/2016 à 08:26, johannes hanika a écrit : heya, this is what it looks like for me: https://jo.dreggn.org/img_0001.jpg much like your not broken reference image? do you apply any styles on export maybe? if you have an exported jpg i could check the history stack in the xmp stored in that one.. cheers, jo On Sat, May 7, 2016 at 4:42 AM, jmsarat wrote: > > Bonjour Johannes, > > The problem occurs when exporting an image and using the internal sRGB > color > profile. The problem disappear if I select another sRGB color profile > (Sony > sRGB or RT sRGB). > As the source is a sony arw raw image, I do not specify a particular > icc > profile and use the standard matrix. I tried sony dcp profiles for > A6000 > converted to icc but if the result is different the problem has been > moved > to other places in the blue shades. > Attached the xmp corresponding to the image. This is a simplified xmp > with > the minimal options that create the problem in blue shades. > The arw file is there : https://1fichier.com/?8khnco34cy > > darktable version 2.1.0+1080~gfedb40 > > Jean-marc > > > > > Le 27/04/2016 à 15:06, johannes hanika a écrit : >> >> heya, >> >> i can't reproduce. is there anything else? can you share an .xmp >> sidecar file so we can precisely reproduce your steps? and just to >> double check: you are talking export, not display, right? so it can't >> be a display profile issue. are you using any particular icc profile >> as input colour space? >> >> -jo >> >> On Wed, Apr 27, 2016 at 2:06 PM, Moritz Moeller >> >> wrote: >>> >>> On 25/04/16 12:43, jmsarat wrote: [...] lower image result with an external profile (Sony Corp sRGB). [...] >>> >>> >>> Can you share the source of the Sony profile? >>> >>> I have some images that have always shown exactly the artifacts you >>> documented and which I could never develop through DT therefore. I >>> thought >>> it was some module inside DT causing this. If it is as simple as >>> changing >>> the profile, that would be really awesome. >>> >>> .mm >>> >>> >>> >>> >>> ___ >>> darktable developer mailing list >>> to unsubscribe send a mail to >>> darktable-dev+unsubscr...@lists.darktable.org >>> ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org >>> >>> ___ >>> darktable developer mailing list >>> to unsubscribe send a mail to >>> darktable-dev+unsubscr...@lists.darktable.org >>> >>> >> >> >> ___ >> darktable developer mailing list >> to unsubscribe send a mail to >> darktable-dev+unsubscr...@lists.darktable.org >> >> > > ___ > darktable developer mailing list > to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org > ___ darktable developer mailing list to unsubscribe send a mail to darktable-
Re: [darktable-dev] Re: sRGB internal color profile
Hi Matthias, With a clean $HOME/.config/darktable, the option lcms2 is deactivated by default therefore the export function use lcms1-1.19. The resulting image does not have defect in the blue shade but if the lcms2 option is activated then the problem appears. I tried with a fresh DT 2.0.4 compiled on Mageia 5 with lcms2-2.7 and with the latest dev version 2.1.0+1082~g6e50d7a. The problem is there with both versions. Jean-Marc Le 08/05/2016 à 13:09, Matthias Bodenbinder a écrit : Hi Jean-Marc, I can not reproduce your JPG output. Not even with your xmp file. I am using DT 2.0.4, self compiled. Have you tried with a different user resp. with a clean $HOME/.config/darktable ? Matthias Am 08.05.2016 um 12:35 schrieb jmsarat: Hi Jo, With this clean xmp : https://1fichier.com/?rd1q8touto I obtain this image : https://1fichier.com/?3uqu4ijx0g No style on export and using the image parameters for color profile (sRGB(web)) and rendering intent (perceptual). The raw image is there : https://1fichier.com/?5n9xbgf9ej The problem is visible in the blue shade but could be present in other colors. Jean-Marc Le 08/05/2016 à 08:26, johannes hanika a écrit : heya, this is what it looks like for me: https://jo.dreggn.org/img_0001.jpg much like your not broken reference image? do you apply any styles on export maybe? if you have an exported jpg i could check the history stack in the xmp stored in that one.. cheers, jo On Sat, May 7, 2016 at 4:42 AM, jmsarat wrote: Bonjour Johannes, The problem occurs when exporting an image and using the internal sRGB color profile. The problem disappear if I select another sRGB color profile (Sony sRGB or RT sRGB). As the source is a sony arw raw image, I do not specify a particular icc profile and use the standard matrix. I tried sony dcp profiles for A6000 converted to icc but if the result is different the problem has been moved to other places in the blue shades. Attached the xmp corresponding to the image. This is a simplified xmp with the minimal options that create the problem in blue shades. The arw file is there : https://1fichier.com/?8khnco34cy darktable version 2.1.0+1080~gfedb40 Jean-marc Le 27/04/2016 à 15:06, johannes hanika a écrit : heya, i can't reproduce. is there anything else? can you share an .xmp sidecar file so we can precisely reproduce your steps? and just to double check: you are talking export, not display, right? so it can't be a display profile issue. are you using any particular icc profile as input colour space? -jo On Wed, Apr 27, 2016 at 2:06 PM, Moritz Moeller wrote: On 25/04/16 12:43, jmsarat wrote: [...] lower image result with an external profile (Sony Corp sRGB). [...] Can you share the source of the Sony profile? I have some images that have always shown exactly the artifacts you documented and which I could never develop through DT therefore. I thought it was some module inside DT causing this. If it is as simple as changing the profile, that would be really awesome. .mm ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Contribution
On Friday, May 6, 2016 11:03:18 PM CEST Moritz Mœller wrote: > What is missing is porting one of the contemporary black and white > conversion plugins to be available in DT. > > The monochromes module is just very ... traditional. > > Some of the recent papers have C code/Gimp plugins. One of them should be > ported to DT. I nan send your some suggestions. Please do so. Cheers, Wolfgang > > .mm > > > ___ > darktable developer mailing list > to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org signature.asc Description: This is a digitally signed message part.
Re: [darktable-dev] Contribution
On Saturday, May 7, 2016 7:05:16 PM CEST johannes hanika wrote: > ooh, nice! papers to implement :) > > one that's been on my list for quite a while but i never got around to > do it is this: > > http://people.csail.mit.edu/sparis/#cacm15 > http://people.csail.mit.edu/sparis/publi/2015/cacm/Paris_15_Local_Laplacian_ > Filters.pdf > > which can separate coarse/fine detail without introducing halo effects > a little bit better than what we currently have in, say, the > shadows/highlights module. this is a building block that may replace > the bilateral filter in a couple of places. Thanks for the nice welcome and the reference. I read the paper and it sounds interesting to me. Under [1], Matlab source code of the filter is provided. I quote its licence at the end of this mail. As it is not exactly GPLV3, I kindly ask someone to check if I am allowed to read that code. I would offer to summarize the main points of the paper to bring everyone (including me) on the same page. We than can discuss details. Does this work for you? I am looking forward to work an this filter. Cheers, Wolfgang Licence of the Matlab implementation of the Local Laplacian Filters "Start quote" Copyright (c) 2011 Sam Hasinoff Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. "End quote" [1] http://people.csail.mit.edu/sparis/publi/2011/siggraph/ matlab_source_code.zip > > re:denoising. there is a little bit to be gained if we make use of > negative values, passed on from the input with black point subtraction > up to the denoiser. i'm not sure how well this would work with the > Anscombe transform we currently do for noise stabilisation (initial > results were only quite marginally better), we may want to add a Fisz > transform wavelet mode to the profiled denoiser (i have some code for > this in a branch). > > let us know what you fancy, then we can discuss details.. > > cheers, > jo > > On Sat, May 7, 2016 at 9:03 AM, Moritz Mœller wrote: > > What is missing is porting one of the contemporary black and white > > conversion plugins to be available in DT. > > > > The monochromes module is just very ... traditional. > > > > Some of the recent papers have C code/Gimp plugins. One of them should be > > ported to DT. I nan send your some suggestions. > > > > .mm > > > > > > > > __ > > _ > > darktable developer mailing list > > to unsubscribe send a mail to > > darktable-dev+unsubscr...@lists.darktable.org > > ___ > darktable developer mailing list > to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org signature.asc Description: This is a digitally signed message part.
Re: [darktable-dev] Re: Lua GUI box model
Ok, multiple questions in here, i'll try to answer them as I can ABOUT LAYOUT OF BOXES what you are doing right now is not possible, the widget elements are new in lua and everything is not exposed to lua yet. I add stuff on a need basis and I didn't need fine-grained control on the layout so far. GTK (which is the real resplonsible for drawing the UI elements) has lots of complicated options to control layout. I'm not sure exposing them to lua would be simple/easy. I'll have a look at what's in there wrt layout and if some controls can be easily passed to lua for users to play with RELOADING LUA SCRIPTS with the way lua works, it is not possible. you have to restart. This is because lua doesn't load a script, it runs some code. there is no generic way to revert what code did. your solution of setting package.loaded[] to nil is dangerous. if it works for you, that's fine, but the risk of it causing unexplainable behaviour exists. especially if you end up crating the same thing (widget element, preference, shorctu...) multiple time LUA DOCUMENTATION As was pointed, your best source of documentation is the user manual (for tutorial-like info) and the lua-api reference. (for complete reference) the lua API, in particular is automatically checked against the source code and is guaranteed to be up to date. no hidden feature*. If you are working with master, I recommand generating it yourself and not getting it from the website, so it contains all the latest stuff LUA SHORTCUTS you can create shortcuts in lua, see http://www.darktable.org/lua-api/ar01s03s03.html.php for details. This will create a new shortcuts that will call some lua code. Note that you can't use lua to assign shortcuts, only to create them. You need to assign a key to your shortcut via the preference menu Overall i'd be interested in seeing your scripts,we do have a script repo that accepts external contributions and any chance to have new contributions is welcomed *except the april 1st view, which can be accessed via lua On Tue, Apr 26, 2016 at 11:41 AM, Christian Mandel wrote: > Am 13.04.2016 um 10:20 schrieb Christian Kanzian: > >> Hi, >> >> Be aware I'm not a developer so my comments can be wrong. >> >> Am 09.04.2016 um 15:22 schrieb Christian Kanzian: >>> [...] like the usermanual. What's written there is possible. If it is not >>> there you >>> simple can't do it. >>> >>> That's one of the problems: Of course I try to learn from the >>> documentation (especially the Lua API documentation), but it is not >>> really very detailed and for me it is hard to understand where the >>> limits are. E.g., some problems may be possible to circumvent or may be >>> default Lua and therefore not documented in the API docs. >>> >> >> Yes I'm in the same situation. Here is the default Lua: >> https://www.lua.org/manual/5.2/ >> Because I'm already middleaged, sitting all day long in front of a >> screen I usually print this manuals for later reading ;-). >> Takes a lot of time for me to get simple things working. E.g. until now >> I still don't know howto reload a lua script without restarting darktable. >> > > I experimented a bit because to restart dt all the time was annoying for > me as well. Therefore, I found a partial solution. In my luarc file I > register a new shortcut like this to reload my module: > > dt.register_event( > "shortcut", > function() > package.loaded[ 'mymodule' ] = nil > require "mymodule" > print("module reloaded") > end, > "reload module" > ) > > With this I am able to reload my module, but it does not overwrite the old > one but adds a new instance. Therefore, I am able to reuse the session > about 10 times before I have to restart dt. I did not find a way to really > unregister a module (an inverse to dt's register_lib function), if someone > has an idea here ... > > Should be easy from a developers point of view. What's even more >> confusing is, that sometimes you need to know how darktable is supposed >> to work. e.g. darktable.gui.action_images. >> >> >>> Does your answer mean that everything I try to achieve is impossible at >>> the moment? >>> >> No, that was more related to the UI elements and all the formating. >> Maybe everthing is possible. >> >>> What extensions are planned, is it likely that these things >>> become possible? >>> I ask because I have a bunch of ideas of new >>> features/modules for workflow speed-up that would perfectly suit the >>> idea of Lua integration, all of them being little helpers of low >>> complexity. >>> >>> [description of speedtagging module idea] >>> >>> Such a modul will be really helpful for me too, therefore I responded to >> your mail. >> > > Thanks for the moral support :-D > > >> [Description of Christian K's scripts] >> >> Coming to your speed-tagging ideas: >> 1.) shortcuts >> ATM I can't imagine how shortcuts can be assigned via moduls. I'm not >> aware of any place in darktable expect
[darktable-dev] Re: sRGB internal color profile
Hi Jean-Marc, I can not reproduce your JPG output. Not even with your xmp file. I am using DT 2.0.4, self compiled. Have you tried with a different user resp. with a clean $HOME/.config/darktable ? Matthias Am 08.05.2016 um 12:35 schrieb jmsarat: > Hi Jo, > > With this clean xmp : https://1fichier.com/?rd1q8touto > > I obtain this image : https://1fichier.com/?3uqu4ijx0g > > No style on export and using the image parameters for color profile > (sRGB(web)) and rendering intent (perceptual). > > The raw image is there : https://1fichier.com/?5n9xbgf9ej > > The problem is visible in the blue shade but could be present in other colors. > > Jean-Marc > > Le 08/05/2016 à 08:26, johannes hanika a écrit : >> heya, >> >> this is what it looks like for me: >> >> https://jo.dreggn.org/img_0001.jpg >> >> much like your not broken reference image? >> >> do you apply any styles on export maybe? if you have an exported jpg i >> could check the history stack in the xmp stored in that one.. >> >> cheers, >> jo >> >> >> >> On Sat, May 7, 2016 at 4:42 AM, jmsarat wrote: >>> Bonjour Johannes, >>> >>> The problem occurs when exporting an image and using the internal sRGB color >>> profile. The problem disappear if I select another sRGB color profile (Sony >>> sRGB or RT sRGB). >>> As the source is a sony arw raw image, I do not specify a particular icc >>> profile and use the standard matrix. I tried sony dcp profiles for A6000 >>> converted to icc but if the result is different the problem has been moved >>> to other places in the blue shades. >>> Attached the xmp corresponding to the image. This is a simplified xmp with >>> the minimal options that create the problem in blue shades. >>> The arw file is there : https://1fichier.com/?8khnco34cy >>> >>> darktable version 2.1.0+1080~gfedb40 >>> >>> Jean-marc >>> >>> >>> >>> >>> Le 27/04/2016 à 15:06, johannes hanika a écrit : heya, i can't reproduce. is there anything else? can you share an .xmp sidecar file so we can precisely reproduce your steps? and just to double check: you are talking export, not display, right? so it can't be a display profile issue. are you using any particular icc profile as input colour space? -jo On Wed, Apr 27, 2016 at 2:06 PM, Moritz Moeller wrote: > On 25/04/16 12:43, jmsarat wrote: >> >> [...] lower image >> result with an external profile (Sony Corp sRGB). [...] > > > Can you share the source of the Sony profile? > > I have some images that have always shown exactly the artifacts you > documented and which I could never develop through DT therefore. I > thought > it was some module inside DT causing this. If it is as simple as changing > the profile, that would be really awesome. > > .mm > > > > ___ > darktable developer mailing list > to unsubscribe send a mail to > darktable-dev+unsubscr...@lists.darktable.org > >> ___ >> darktable developer mailing list >> to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org >> >> > > ___ > darktable developer mailing list > to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org > > ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] sRGB internal color profile
Hi Jo, With this clean xmp : https://1fichier.com/?rd1q8touto I obtain this image : https://1fichier.com/?3uqu4ijx0g No style on export and using the image parameters for color profile (sRGB(web)) and rendering intent (perceptual). The raw image is there : https://1fichier.com/?5n9xbgf9ej The problem is visible in the blue shade but could be present in other colors. Jean-Marc Le 08/05/2016 à 08:26, johannes hanika a écrit : heya, this is what it looks like for me: https://jo.dreggn.org/img_0001.jpg much like your not broken reference image? do you apply any styles on export maybe? if you have an exported jpg i could check the history stack in the xmp stored in that one.. cheers, jo On Sat, May 7, 2016 at 4:42 AM, jmsarat wrote: Bonjour Johannes, The problem occurs when exporting an image and using the internal sRGB color profile. The problem disappear if I select another sRGB color profile (Sony sRGB or RT sRGB). As the source is a sony arw raw image, I do not specify a particular icc profile and use the standard matrix. I tried sony dcp profiles for A6000 converted to icc but if the result is different the problem has been moved to other places in the blue shades. Attached the xmp corresponding to the image. This is a simplified xmp with the minimal options that create the problem in blue shades. The arw file is there : https://1fichier.com/?8khnco34cy darktable version 2.1.0+1080~gfedb40 Jean-marc Le 27/04/2016 à 15:06, johannes hanika a écrit : heya, i can't reproduce. is there anything else? can you share an .xmp sidecar file so we can precisely reproduce your steps? and just to double check: you are talking export, not display, right? so it can't be a display profile issue. are you using any particular icc profile as input colour space? -jo On Wed, Apr 27, 2016 at 2:06 PM, Moritz Moeller wrote: On 25/04/16 12:43, jmsarat wrote: [...] lower image result with an external profile (Sony Corp sRGB). [...] Can you share the source of the Sony profile? I have some images that have always shown exactly the artifacts you documented and which I could never develop through DT therefore. I thought it was some module inside DT causing this. If it is as simple as changing the profile, that would be really awesome. .mm ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org