NotMyFault

Hi Niki, used flood selection tool for lower and uper area to create masks added two rectangles, added gradients, Have fun. two_gradients.afphoto

Hi, sorry to hear about this issue. Would it be possible for you to share an example document (afphoto file)?

I've recorded a video to explain the issue.Please excuse for some noise from the mouse, and my bad english.

Hi, i would expect all tools working in RGB32 / HDR mode to deliver 32 bit color results, or at least 16 bit. It seems merge visible reduces to 8-bit color depth, which is not expected. When using RGB32 with a HDR capable display, i can positively check that 10 bit (1024) grey values are displayed. Have a look zommend in to 800% or more. Look in the area bewteen gradient and black rectangle, left from red box, between "unlimited" and "256". The gradient has 4 vertical areas: top: original below: 3 posterized boxes to 256/128/64 levels to make simulated / real banding visible. However, when you "merge visible", the merged pixel layer is ditherd and reduced to 8 bit color depth. To reproduce, you need a "real" HDR 10-bit capable display (real 10-bit, 8+2 with dithering may show wrong results). In perfomance settings, dither gradient must be un-checked. 10 bit before merge: 8 bit after merge visible: levels.afphoto

Hi hobie, i would say yes. Using a levels adjustment with 2% output black and 96% output white fave me 245 for white and 5 for black - spot on. levels.afphoto

Hi GeeJay, You only need to rasterize if you have to deal with an image layer. To be able to answer your original question, It would be great if you can share which tutorials you are referring too, and upload the document you are working on. Regards, Timo

Hi Ron, thank you for commenting, but please allow me to emphasize my point this is buggy. did you watch the video i’ve uploaded? Especially the final test using red outline? This contradicts your argument, which may work within tight constraints, like grey color and fully transparent background, but not in general. The video clearly shows completely different behavior of the fill, depending on the outline yes/no, and non expected behavior. Mixing the fill within red (color of outline) is definitely not expected when reducing fill opacity, as no red outline should be inside the fill area. Coming back to your argument: opacity (means: alpha channel) is completely independent from RGB channels, but: what you and the color picker sees is the result of blending all layers factoring in the alpha channel. Using grey 50 at 100% opacity is not the same as using white with 50% opacity, even if you get the same color picker values. When you place e.g. a green fill layer beneath, you can see the difference. 50 gray with 100% alpha stays grey, while white with 50%alpha lets the green fill shine through. Best regards, Timo

I have experienced photo as a tool for creatives, much more capable as photoshop in many areas, and not as a policing tool to blindely restrict you to those things photoshop can do. Degrading the more capable functionality of prior versions until 1.8.5 to much more limiting capabilities of photoshop does not look like a good move. And there is an absolute difference bewteen rotating by 0.1 degree steps with excellent control over how small or big these steps are, to the hit-or-miss approach of the crop tool. The old way a super efficient and effective tool to correct a certain amount of mis-alignment and perspective distortion, e.g. to straighten handheld photos of rectangular shapes, even a vertain amount of tilt-shift correction. These capabilities have been reduced by at least a factor of 2,5 (rotation) and ~5 (hor/vert). Most important, it was possible to correct 4 different types of mis-alignment in one tool. Now these corrections are seperated into several tools. I there is a good technical reason that rotating above 10° would cause any issues in Photo, i would accept. But then, please remove rotation completly from lens correction. I don't expect any lens exists on earth which suffers from a "rotation" imperfection. Rotation correction is required to correct mis-rotated cameras, and not to correct lenses flaws. Similar, hor/vert correction is mainly required to correct mis-alignment of the camera position with respect to the chosen object, and not to correct any lens faults. And again, if Affinity would provide the former "lens correction" fucntionality in photo persona, with the old UI from developpersona, but extended to +/- 360° rotation, unrestricted hor/vert correction, unlimited scale correction, i would be happy to give up all of this in develop persona. Just rename them to "perspective correction" instead of "lens correction". Regards Timo

Hi, the dust & scratches filter does not work in RGB32 mode. 2021-02-13_19-15-50.mkv

Hi Andrew, i see another post from you where you shared a layer stack of many seaside photos. So using "New >stack" with "Automatically Align Images" seems the first choice for your type of images. Unfortunately, with waves it might not work ideally. And please keep in mind that masking in a stack works different (only 0 or 1, no gradual masking). So after using the stack to get the images aligned, simply remove the stack before adding masks. Even a focus merge may be useful. For focus merge, you may need to stack / align the images first, and then export every image isolated as TIFF-16, as focus merge expects pre-aligned images. Regards, NMF

Hi, since Release 1.9, the effect of lens correction seems to be very limited: rotation to -10..+10, and you can't enter higher values As i use lens correction rotation regularly as it provides a much finer control (0.1 degree) compared to crop or layer transformation, this is a step back from +/-25 % before (until 1.8.5) Hor/vert shows value of +/+100, but effect is extremley limited. I have many picture which require far more room for adjustment. If i could get a life filter in photo persona which would allow these transformation with reasonable range (rotate +/-180 degree, hor/vert to get at least 45 degree angle), i could live with the "reduced" state in develop persona.

Hi Tony, have a look at the menu: View->Studio->History It should be checked. You may click "reset studio" to get back to the defaults. Regards, Timo

Hi Andrew, this is possibly a bug. To find out, can you please - inspect the layer by clicking Alt-left Mouse (on Windows) the layer in the layer panel? Does the layer content is displayed correctly? Then we have only a minor "display glitch". Depending on how often this occurs any annoys you: ignore it or try to reproduce and report bug (see next point) If the layer content is displayed incorrect, i would assume a major bug. Please try to capture all relevant data, including the affected file saved with history enabled, and report the bug in the dedicated part of the forum https://forum.affinity.serif.com/index.php?/forum/76-report-a-bug-in-affinity-photo/ Regards, NMF

I've uploaded a video with updated settings to make the bug even more obvious: Fill in blue stroke in black outline in red 1. When adjusting opacity of fill, fill uses color of outline effect (if outline is used) 2. When reaching 0% opacity for fill, outline is displayed as "center" instead of "outside" 2021-02-13_12-15-27.mkv outline opacity issue V2.afphoto

Please ignore the "97%" image. i tried to edit the post and delete the image, but failed.

Hi Carl, you are right, the right shape has a stroke too, but this has no impact. you can remove the stroke. I share a few screenshots with different opacity settings. 100% 80% 50% 10% 0%: The outline extends inwards, fill becomes transparent

Ok, that is a different animal 😅 Then you may try one of the stacking tool: new stack, or new HDR merge. These have the option to automatically align pictures. This works really well even with handheld shots. In case of HDR merge, you do not need to tone map, just uncheck the box, and only use the alignment and cloning functionality. The big plus: With HDR merge, you can automatically remove ghosts, and manually adjust with clone tool if the automatically removal did not work perfectly. If your mages are too different from each other, this manual process might help. It uses helper shapes to get a perfect perspective adjustment for stacking assuming you use 3 images (pixel layers). Image 1 as base/ reference on bottom, image 2 and 3 to be aligned, above layer 1 in the layer stack On every picture, use the pen tool to create a shape and connect 4 points (corners) in the images with reference points. Use separate shapes, 1 per image. Rename layers to 1,2,3 to stay organized. Use black stroke of about 2 pixel, no fill Hide all layers except the 3 shapes, 1 lowest, 3 on top add a live perspective filter and nest it to shape 2 adjust the perspective until the corners match perfectly repeat same process with shape 3 when perspective is ready, move the perspective filters from the shapes to the pixel layers hide shapes, enable pixel layers add masks (black) on pixel layer 2 and 3 Paint with brush tool in white to reveal wanted areas. Work from bottom to top. ic case i missed your question again, it would be great if you can upload example files and a draft version of what you want to achieve.

I might be wrong or expecting to much. When using a document with 16 or 32 bit color depth, and 16-bit colors in color panel, i would expect beeing able to enter 16-bit values to RGBHEX. But it will only accept 00-FF (8 bit) and always replace the "high" byte with 00. The "resulting" color code from all 3 channels will simply double/repeat the given 8-bit hex value. This is at least confusing, or a bug. HSL or LAB do not offer any option to enter 16-bit values. And last but not least, the info panel is completely unable to show 16-bit colors. All you get is either 8-bit or less than 7-bit if using % (0-100%). Only option to enter 16-bit is using RGB or greyscale. Only option to display 16-bit color value is to use color picker, and read out result in color panel.

Hi Andrew, could you explain how you try to align the images? Are you using the "New Panorama" and "stich Panorama"? The official tutorial may provide some hints to optimize the stitching with manual interventions and masking. Regards, NMF

Hi GeeJay, Can you share a screenshot of your layer panel? if possible the area you want to use the in-painting brush, too? If it shows a image layer, you need to rasterize this to an pixel layer first (e.g. by "merge visible"). Please check the pacity and flow, use 100% to start with Depending on your pixel layer, inpainting might not work well, and you may swicth to clone brusch, or patch tool instead, to have more control over what parts of the image are used to replace unwanted areas. Regards, Timo

Hi, i assume a bug: when you use a vector shape, fill it with white, add a black outside outline (layer effects) reduce the fill opacity The fill gets darker and finally black (not expected). Expected is fill gets transparent, but stays white. Instead, it gets the color of the outline. When reaching 0% opacity, fill gets transparent (expected), but the outline is extended to the inside (wrong). Spotted by my son in 1.8.5 and 1.9.0.932 Example file shows two linked vector shapes, one using stroke behaving correctly, one with outline mis-behaving. Linked so opacity affects both, and you can see the issue. outline opacity issue.afphoto

Hi Ian, to definitely get rid of all the history, just copy the merged pixel layer, and „create new from clipboard“. Then save the new file. If you used 32 color depth, the reduce to 16 bit. Timo

And you may start cmd.exe dxdiag Which will provide more details about your GPU.

There is a post in the beta subforum:

Hi, when using a RGB document and then a curves layer with LAB color space, affinity becomes totally slow. I added a DOF filter like James Ritson in his 1.9 Youtube today, when you mode the center it takes about 1 second to update. rainbow lab curves slow.afphoto