I’m no PhotoFlow expert but it feels nicer than before. As a rare user of PhotoFlow I have a few suggestions. In no particular order but numbered anyway. Please take the suggestions with a pinch of salt.
The top left buttons do not respond to the width of the panel. Including getting cut off as there is no minimum width. Some sort of reflow or scaling would feel better and a min width that prevent the buttons from being cut would be great.
The first and second row of buttons are different types but follow the same visual logic. The first row is what typically goes in a File menu.
The export to raster button has both a position and icon that is hard to find whilst being one of the most important buttons. Perhaps move it to the position of the “exit” button?
The camera info lacks padding around the text but has excessive line-height placing the whitespace (greyspace in this case) in such a way that the information is visually scattered. The center alignment could perhaps also be reconsidered in favour of left aligned?
The “new layer” button becomes aligned with the mask, sticky, info icons. Some sort of vertical border splitting the tools off perhaps. Or a different vertical alignment.
I’ve always felt that PhotoFlow and in particular the new layer dialog would benefit from a dmenu type hud. Using type ahead to filter the list of possible layers hit enter to create.
Enable spacebar (gimp) for panning and double click for toggle 1:1 <> zoom to fit
I agree with everything you have said. I either mentioned it before or have experiened the same qualms.
1-3. The appearance, design and placement of these icons need some consideration. I avert my eyes until I need to use them.
4. Moreover, I am beginning to think that it may be a good idea to make it into a tab after histogram and samplers since it is static information. Would save vertical space.
5. More padding would help. Nothing else, besides the camera info, is that tight. Consistency is important. Mask, sticky and info should be on the right side.
6. My frequently asked request. Needs a search function!
7. Yes, better panning and zooming, and edge examination is important to me.
You are absolutely right, and that was part of the “glitches” of the new UI layout that I still had to work on. Fixed in current git.
I am indeed planning to implement a proper File menu, which will replace the first row of buttons. So I prefer to keep things like this until the menu will be ready…
See above: it will go into a File menu.
Improved in current git:
Still requires additional work, and more information displayed. To save vertical space, it is now moved into an additional tab, after “histogram” and “samplers”
I’ll keep that in mind. Will look into that once the File menu is ready.
Pan is buggy to me. If I click and release while accelerating in one direction, it scrolls for a distance. If I hold that click and move the mouse across the screen, the preview moves back and forth along the x or y axis but not gradually and ultimately to whatever direction the mouse is guiding it.
I do think the camera info could have even less line-height. Looks like near 2em at the moment. For info such as that 1.2em or even 1em could work. (1em is a relative measure equal to the point size of the type)
The new arrangement of the new layer icon and the rest looks way cleaner!
Unless there’s a reason I’d consider skipping the ctrl for the “viewport” and copy the behaviour from existing software people are likely to have used which one is the best I can’t say. I suggested gimp for oss familiarity.
I have just introduced a couple of interesting usability improvements in the image export process, namely:
a dialog that informs the user if the destination file already exists, and asks whether to overwrite it or not. Selecting “NO” brings you back to the export dialog, where the output file name can be changed
a warning dialog the informs the user if the export fails, although it does not yet provide hints about the reasons of the failure
Here is a little update about what’s going on in PhotoFlow’s development.
Recently, after a short break, I have been working on the integration of LibRTprocess into PhF.
At the moment, only the Amaze demosaicing is taken from the librtprocess code, the rest still uses PhF internal code. However, this is a first step towards the integration of a common raw demosaicing and processing library, that will allow PhF to stay up-to-date with other projects using the same code.
The change should be completely transparent to the user. Nevertheless, I had to modify a number of source files, so if you encounter any new issue/bug, please report it here, I will take care of it as quickly as possible!
Hello DrRaw,
How are you my Dear? I see you work yet very hard in this summer…
Very good news… demosaicing!
On my linux, I haven’t yet the last Photoflow release because I have difficulties with the Dariusz Duma ppa and my Mint 18.1 serena…but soon all would be resolved, I hope.
Latest continuous win10 release. I cannot open t/sharpening-techniques-lizard-subject/13427. PF is stuck at loading null and the window is not interactive. When I maximize window, I see two windows. If PF cannot load, make it stop gracefully and alert the user about it.
Ok I will check it this week-end, my Friend! I live at my wife’s house but I have my computer at my home. But recently the door of my apartment was forced, I brought my 1080 Ti graphic card back to Marie-Claire’s house so I can’t use my computer, I must re-install Mint 19,1 at home…I will check it.
Me, I am ok but now I live in Ardennes (Givet) et not at Lille in north of France. I changed since my Marie-Claire’s discovery… I am gardener now because my house has one and my landscape is very beautiful. Ardennes is a forest country with little villages: I love!
On Windows, cannot save the PFI of 20190820_Südpark_Blüten_8112_B.DNG using photoflow-w64-20190807_2112-git-simplified-pipeline.
PS Another thing I noticed is when negative clipping is on in raw developer and I export the image into another colour profile I still get negative values. Edit Perhaps include an option to clip negatives and overflow at this stage as well.
PPS I find that RT’s auto-CA performs better than PF’s one. Any chance it could be included? It is a part of librtprocess now.
PPPS This is unsubstantiated but I find myself never using lens correction (distortion) anymore. (Edit: and vignetting doesn’t seem to do anything.) For some reason, I feel that it is accentuating the distortion rather than fixing it. Let me explain. When the correction is an outward barrel, it looks like it is stretching the corners too much; and vice versa. Could it be that the correction is reversed or that the correction already happened? In any case, it doesn’t feel that it is doing it right, whereas RT and dt appear okay. Thoughts?
Speaking of lens correction, I really appreciate RT’s auto-CA. Now that it is in the librtprocess, would you mind porting it over to PF?
Enough with the PSs. This next issue is really bothering me.
Whenever PF is processing something, it does so until completion; however, if it is an intensive operation or I want to do something else, it either makes the app unresponsive to other input or it becomes unstable and / or crashes. E.g., I had to close PF just now because Gamut Mapping in the Colorspace Conversion module was taking a while and I proceeded to toggle Black Point correction, but I couldn’t because the GUI was unresponsive and the console seemed to be in a long (or infinite) loop.
Let me wrap up all what you observed and reported, and do some tests. I hope I will be able to finally get back to you in some days with some concrete solutions!
I admit I have been awfully side-tracked over the past weeks, and did not pay enough attention to the bug reports. I hope to catch up with my homework now that the local laplacian experiments are concretizing a bit…
Sorry for being vague when it comes to these reports. Hope they help. My mind, body and soul have been melting due to taking care of ill parents 24/7/365.
Oh and closing a raw image tab still crashes the app…
I have just merged the new-caching branch into stable.
The new code does not introduce any new tools, instead brings an improved managing of layer caching. The new caching mechanism is based on a global pool of tiles, the size of which can be adjusted in the options:
,
,
This next issue is really bothering me.
