toodee

I may have spotted where my problem is: I use an OCIO config in Affinity Photo. Can it be that the profile conversion is not working right as long as an OCIO config is enabled? After a reset of Affinity Photo and I did another conversion from image 02 to image 03b I get this result. I hope this is more likely to what I should see and expect.

Hi, you said you can see image 02 fine on your display? I cannot. I see a dark halo around the star and a light halo closer to the center of the star. I cannot display an image with BT.2020 primaries, because I would need a laser projector/display to do so. I was expecting a different result by the conversion to to sRGB operation. I can confirm in Foundry/Nuke that a JPG export from image 01, 02 and 03a and 03b end up as the same image. As far as I understand it, this should not be like this? I can also convert the Rec.2020 document to sRGB and set the rendering intent to perceptual and saturation and still end up at the same result as image 01. It could be that something is not working properly in my Affinity Photo?

Thanks for your reply. I think I misunderstood the rendering intents absolut and relative and how they work. I was expecting a different result.

Hi, I am writing a blog post about exploring a simple made up image in different colorspaces. https://www.toodee.de/?page_id=5496 I set up a document in sRGB. Next, I assign this document the Rec.2020 primaries that I obviously cannot display on my iMac. Then I convert the document back to sRGB trying different rendering intents. 01 - looks right 02 - looks broken because I cannot display it 03a and 03b look okay again, but shouldn't the two conversions end up differently? Thanks for your help. Daniel 01_start_sRGB.afphoto 02_assign_Rec2020.afphoto 03a_convert_to_sRGB_relative.afphoto 03b_convert_to_sRGB_absolute.afphoto

Hi, I found already some similar questions related to this topic but not a simple answer. For this demonstration I use the ACEScg colorchart from here: https://github.com/colour-science/colour-nuke/blob/master/colour_nuke/resources/images/ColorChecker2014/ACEScg_ColorChecker2014.exr I use a tool like Nuke NON-Commercial, read in the file above (assuming all colorspace settings are correct in Nuke and Affinity Photo) and use a sRGB view transform. Then I use the Digital Color Meter from MacOS and read the marked grey patch value "on" the screen. Next I export a JPG and review the file in Preview and read again the value of the same patch. These are images 01 and 02. The I repeat the same steps in Affinity Photo and don't get the same result. These are the images 03 and 04. In my expectations both tools should end up with the same result. Why the the exported chart in 04 on the right side so much brighter? Do I have a wrong setting somewhere in Affinity Photo or is this a bug? Best regards Daniel www.toodee.de

Best Daniel

Hi kirk, yes, I am aware of that tool. We tested it a while ago back, but we were not really happy with the results. And for single images, I prefer to see the image in a raw converter and can check out settings. And as far as I remember it also kills all metadata from the raw images, so it makes it harder to use them in PTGui for example. But its a good hint as I should check it out again. Thanks Daniel

Hi Justin, thanks for your quick answer. Yes indeed, on one of my two machines the preference was set to Absolute C. and on the other to Relative C. I got no idea why this was set like this because I never touch these settings. When it is set to Relative C. then it behaves like it used to be. Thanks a lot for your help. Daniel

Hi, I noticed a strange behavior with the release version of Affinity 1.7 that I don't recall with the last Beta versions (that I already deleted) and for sure not with the version 1.6.x. I use the Serif RAW converter to convert canon raw photo files to 32-bit linear files. When the prefs are set to 32-Bit sRGB linear, then after the raw conversion I convert the format to ACEScg. When the prefs are set to 32-Bit ACEScg, then I don't need to do anything else after the raw conversion. In both cases I export EXR (half-float files) without the check box "Color profile from name". When I loaded both files in Nuke (from Foundry in an ACES environment with each the right colorspace settings) they were looking (nearly) identical. And this made sense to me. I checked with older files EXR files (that were converted with an older version of Affinity Photo) and the two exports are visually identical. There are very small differences in the colors, this might be rounding errors. The linear sRGB file is actually ever so slightly warmer then the ACEScg file. When I do this now with the new release of Affinity Photo, the ACEScg converted file got much warmer. The two attached images (IMG_7330) were converted with version 1.7 : Prefs 32-Bit sRGB linear, Serif Engine (no tone curve), the lin sRGB was then directly exported as EXR, the ACEScg was converted with Document/Convert Format/ICC Profile to ACEScg and then exported as EXR. Both files were exported with Nuke to JPG (with the same display viewer curve RRT&ODT) The older files (IMG_8454) were converted with an older version of Affinity Photo (April 2017) I can provide the .CR2 files if necessary. Can someone explain why this happens? Best regards Daniel

Hi, I posted this feature request a while ago on Facebook and I was asked to post this here in the right place. I am very happy with the ability to retouch quickly 360° OpenEXR images in Affinity Photo now, for example to remove the tripod. For this I am using the Live Projection function. At times a stitched equirectangular image has not a leveled horizon or the most interesting part of the image is not inftont of the initial position. For video (plugin for FCPX) there is a plugin from dashwood3D.com that is called re-orient sphere, Autodesk flame has this function as well. Here it is called Map Convert. in Foundry´s Nuke its called Spherical Transform. And there is Skybox (Mettle) Convert-Re-orient-Camera.png http://www.dashwood3d.com/360vrtoolbox.php https://knowledge.autodesk.com/support/flame-products/learn-explore/caas/CloudHelp/cloudhelp/2017/ENU/Flame/files/GUID-26C43C08-4320-441F-83FE-201083E6A9E4-htm.html I guess with the Live Projection this functionality actually already exists in Affinity Photo. I could think of a wireframe overlay in the Live Projection, where I can rotate "texture" inside the sphere for example. Or there could be a filter like the Affine filter in the Distort menu in the 2:1 equirectangular view. There I could simply "turn" the 360 image in the desired direction. Best regards Daniel Brylka www.toodee.de

yes sure, but first of all it is only describing the "one way" from Photo to Resolve. And secondly it makes very little sense to activate this 3D-Lut for all the output of Resolve as shown in the tutorial. Unless you have a project that works in mostly the same set or "environment". You will rather use this per clip in the timeline of Resolve or for a whole timeline as a timeline post grade. What I am looking for is the other way around. You got a look created in Resolve and load that into Photo. And also... FCPX with "mLut" from MotionVFX reads the 3D-Lut from Photo, "Color Finale" from Denver Riddle can't read the .cube from Photo. (Error Message: Malformed DOMAIN_MIN tage Line:8" .But both plugins can read the .cube from Resolve.

Hi Kirk, I found this website too yesterday, but the launch app button is only telling me that the server is not reachable. I found also out that Resolve is writing out its own .cube format that is not the same as the Adobe Speedgrade .cube format. How clever of both of the companies :) Yes, there are LUT conversion tools, but this is not a very good workflow. I hope Affinity (and the other vendors) can improve their formats and make them more compatible. Greets Daniel

Hi there, I am testing 3D-Luts in Affinity Photo and how to work with them in other tools and vice versa. It would be very helpful if Photo can tell you (or name them in the help) which adjustment layers and how many can be used to bake them into a 3D-LUT. Please add a preference for the most used LUT type and size. Right now I am getting a .3dl from Photo into Nuke (result is not the same although color spaces are set correctly - maybe I did too many adjustments?) The .cube file is not working in Nuke. The .cube works in Resolve 12 (beta 3), a newly created 3D-LUT from Resolve (.cube) on the other hand I can't import into Photo. Nuke can read the .cube file from Resolve. We use in our job mostly .cube file between Resolve, Nuke and Flame. Is there anything to consider using .cube file from Photo? Best regards Daniel www.toodee.de