kirkt
-
Posts
440 -
Joined
-
Last visited
Posts posted by kirkt
-
-
Here is the AP version, with some work to make it pleasing ... :)
-
EDIT - I take my previous statement back. You can kludge Affinity Photo to do the combination.
EDIT 2 - I reinvented the wheel so succinctly described in post #12 by Mbd. I'm totally spacing out here. Anyway, Mbd is correct, but the images need to be linear for this to work accurately, and in 32bit.
I will post in a second, but you basically create a new document in 16bit RGB color and bring each of your R, G and B grayscale images into the document in their own layers. In the channels palette you clear the channels of the two color channels that are not the current grayscale image's channel. For example, in the grayscale image that represents your RED channel, you clear the GREEN and BLUE channel from the layer. Etc.
Set the Blend Mode of each layer to Add (or experiment with Screen or Lighten). To White Balance, you can add an exposure adjustment to each layer and adjust the exposure of each of the three layers to get the color correct. Because the grayscale images likely are not linear, the exposure adjustment will cause color shifts in shadows and highlights, so you will have to tweak the color in shadows and highlights to get neutrals as well (in using exposure you are scaling gamma-adjusted images, where the exposure relationship no longer is linear).
kirk
-
Ok great, channels and color space conversion are still the two weakest points in my knowledge so please bare with me.
When I opened your images they were opened in a greyscale color space and for whatever reason neither assign nor convert to color space let me choose something like sRGB. So I had to create a new document using sRGB and your files pixel dimensions. I then pasted the 4 files in (of course you could use Adobe or ProPhoto or whatever as well)
If anyone can tell me why I can´t assign/ convert to a RGB profile, I´d be very happy :)
You probably have to change the document mode to Color 16bit and then you can assign a color profile.
kirk
-
Unfortunately, if you want to do this the "correct" way, you need to do it in 32bit mode, using linear data. You add the linear images with color filters applied to each of the three grayscale color images and adjust the exposure of each of the three grayscale color images to white balance the image color. Once you get the color correct, you can apply that to the luma channel and get your composite. Affinity Photo does not operate in 32bit as far as I can tell - it will open a 32bit image in 16bit, which clips all of the information that is not within the 0 to 1 range. If you try to do this in 16bit mode, you may have some success, but you will quickly clip your image.
Attached is the composite made in Photoshop, 32bit mode for the math, converted to 16bit via Exposure and Gamma tone mapping and exported as an 8bit JPEG for posting here.
kirk
-
Take a look at a couple of other threads here on the forum that suggest that Affinity has a long-standing problem with dual-monitor color management. You may be having issues with your external display and how the plug-in honors the second display profile versus how AP may (not) honor the second display profile.
For example:
https://forum.affinity.serif.com/index.php?/topic/13363-colour-management-on-second-monitor/
kirk
-
Here is a screen shot where I did not use soft proofing. The AP image and the image ready to print are not the same.
AP is more saturated and brighter
But, to be clear, are the AP image and the "image ready to print" the same file - i.e., the same exact file, just opened in two different applications?
What happens if you open the exact same file in AP and in Preview? Do they look identical and both look different than PS? If so, it may be related to the way OS X handles black point in the tone curve of the display profile. See this thread on the luminous landscape:
http://forum.luminous-landscape.com/index.php?topic=108348.0
This issue has plagued OS X for a while now. Color Sync displays crushed blacks while Photoshop displays blacks properly (it has its own internal color management). AP may be using Color Sync for display, and your images may be affected by this. Just trying to think of anything I've experienced in the past that causes similar display differences across applications.
kirk
kirk
-
-
I have watched the sp video, over and over again. And again, just know. I did exactly as the voice told me to. No match. According to video, and please confirm, he suggests the sp is to see what the printed image will look like.
Sorry. Not on my display.
And would someone explain what is meant and the process to do it, when he says 'merge'? What are the motions I do to merge like he says?
As R C-R and others have described, soft proofing is just that - it is a software simulation of a proof print. Instead of opening an image, doing your edits and then printing a proof to see how the edits you made get rendered by the printer, a soft proof is the software's attempt, using the ICC profile for the output device, to simulate the proof on your display. You probably already know this, but there are people following this thread that may benefit from the explanation - it is not my intent to insult your intelligence or imply that you do not know this.
Soft proofing that you may be used to in Photoshop is implemented at the display level. It takes your layer stack and renders it through the ICC profile of the device you want to simulate, using the rendering intent, black point compensation and paper white simulation to give you an idea of how the image will look on your output device (in your case, your printer). This saves us the hassle of printing a proof, inspecting it in controlled viewing conditions, making judgements about various edits that are still necessary and then making the edits and the printing another proof, etc. In Photoshop, the RGB numbers of the document in the working space are converted through the soft proof profile so that they can be displayed as a simulation of the output device; however, the RGB numbers of the actual document are not changed in the document itself - it is just a simulation for viewing purposes.
In AP, soft proofing is not implemented in a purely display-based way - it is implemented as an adjustment layer. While this is interesting, it may cause a little confusion. When the Soft Proof adjustment layer is placed on the layer stack, it actually changes the RGB numbers of the document so that the rendering of the document takes on the appearance of the composite stack below the Soft Proof adjustment, with the transformation of the RGB numbers from the working space to the output device profile selected in the Soft Proof adjustment layer. This is an actual adjustment to the image, just like a curves adjustment or any other adjustment layer. As the video states, you need to make your compensatory tweaks to the image for print in AP with layers placed UNDER the Soft Proof adjustment layer to see how those compensatory tweaks will modify the look of the simulated print.
According to what has been stated in this thread, the Soft Proof adjustment layer IS NOT supposed to get exported or printed, regardless of whether or not it is visible/active in the layer stack. However, you can apply its effect to the layer stack by merging the Soft Proof adjustment layer into the stack - using the flatten command or the "Merge" button on the Soft Proof adjustment layer dialog panel. As the video states, you might want to do this for creative reasons - this is not something that I have seen in other applications, but it is analogous to using the ICC profile in soft proof mode as a LUT (Look Up Table).
I have also tried to understand how this compares to Assign Profile or Convert to Profile. The problem is, the ICC profiles available in the Assign and Convert dialogs are not the same as the Soft Proof ICC profiles. I do not know why this is the case.
Please understand - I know you are frustrated. But, if you have the time, can you please do the exercises that I proposed: just open the same file in AP and PS and apply the soft proof adjustment layer with the ICC and settings in AP and apply those same settings for soft proofing in PS and report the results. Do not flatten anything, do not export anything. Just take a screen shot of the AP and PS rendering on the same screen next to each other. This will at least give you an idea if the soft proofing in each application is comparable, without any other things to confound the comparison.
I am trying to isolate your problem in a methodical way, so we need to start with some simple experiments with well-known conditions. I say "your" problem, because I cannot replicate it on either of my Macs.
Kirk
-
... but i think this is only one of the problems here.
I agree.
kirk
-
That's interesting. Had to test it myself as the help menu reads different. Nothing in there about the need to disable the adjustment.
Screen Shot 2016-03-16 at 14.29.21.pngAllan
It may get excluded from Export and Print automatically (according to the screen shot in the post by mystrawberrymonkey), but if you flatten the layer stack while it is active, its effects are burned into your composite. Woah!
Thanks MBd and R C-R for the heads up about the "[M]" - got it!
kirk
-
What still concerns me is that the screenshot in Post #20 is apparently just comparing the image, as converted and edited (and eventually exported) in AP (with the ".CR2" extension in the document tab) to the exported TIFF opened in PS. Even though the history panel shows edits and flattening of the soft proof layer, etc., the flattened image at the end of the processing chain should appear the same in both applications. That is, because the history stack shows that the document was flattened, presumably any edits, adjustment layers, etc. have been committed destructively to the document and, presumably, exported as a16bit ProPhotoRGB TIFF that we see in the PS window. These two images should look identical. The PS window does not indicate that soft proofing is enabled in PS in that screen shot.
I also noticed in the Export > TIFF tab of the Export dialog that there is a checkbox to "Don't export layers hidden by Export Persona." This is probably an important box to check if you use the Export Persona to disable something like a soft proof adjustment layer prior to exporting for final output.
I would like to know what the "[M]" (capital M in square brackets) means at the end of the filename in the document tab of the image in AP in Post #20.
Ultimately, I am trying to narrow down the possible suspects in the editing workflow here, but the comparisons so far are a little scattered and need some tighter control with explicitly stated "knowns" - hence my requests for the two exercises above.
kirk
-
Elaborating a little on the soft-proof adjustment layer and the flatten operation in the history of the screen shot in post #20:
In AP, soft-proofing as an adjustment layer apparently has the power to change the actual RGB pixel values in your document. That is, the appearance of the image when viewed with a soft proof adjustment layer changes because the adjustment layer changes the pixel values of the composite image (layer stack). This is distinctly different than soft proofing in PS, where the appearance of the image on screen changes, but the RGB values of the actual document do not change. The critical thing to realize here is that if you flatten your layer stack with the Soft Proof adjustment layer enabled you are burning that soft proof change INTO THE RGB values of your document! It now becomes a permanent change. THIS IS NOT GOOD! Soft Proofing should be a simulation of the output device, not a permanent edit to your RGB data. Soft proofing with an adjustment layer should be done with the understanding that the soft proofing adjustment simulates the output device so that you can add, say, a curves adjustment to compensate for a slight loss of contrast due to the printer. So, you add the soft proof layer to simulate the output for your printer-paper combination, then you tweak the image to compensate for the output device. Then you DISABLE the soft proof layer and save your adjusted file for output to the specific device you simulated with the soft proof.
In PS, you enable soft-proofing and do the same thing - you choose your output profile to simulate the output device, add a curve, or whatever, to compensate for specifics of the printer-paper combination and save your file for output. Because soft-proofing in PS occurs at the display level, it does not alter the actual RGB values of your document, so there is no need to disable soft proofing to save your tweaked file.
This is really important and I was not really aware that the AP soft-proofing adjustment layer could actually be applied as a destructive change to your layer stack and the RGB values in the document.
So, if you use the soft-proofing adjustment layer to simulate your printer for final tweaking to output, make sure you disable the soft-proof layer when you save your final output - otherwise the soft-proof "adjustment" will get burned into the document (EDIT: Maybe?).
Someone who is more familiar with soft-proofing in AP, please correct me if my observations are off the mark.
Kirk
- anon1, A_B_C, My Strawberry Monkey and 1 other
-
4
-
What is soooo frustrating to me, as someone who looks to this forum for information and understanding of unfamiliar concepts, techniques and programs, is when posters use single letter abbreviations assuming everyone already knows what they mean, instead of spelling out the words, at least once, so those who don't know what RPP is, for example, might learn something from the post.
RPP is an abbreviation for Raw Photo Processor. Here is a link to the Raw Photo Processor website if you are interested in learning more about this raw converter:
https://www.raw-photo-processor.com/RPP/Overview.html
If you Google RPP (www.google.com) it shows up a couple of times in the top ten links. Given we are talking about raw conversion into AP (Affinity Photo) it is pretty safe to assume that RPP (Raw Photo Processor) has to do with photo processing.
In the above discussions, AP is an abbreviation for Affinity Photo and PS is an abbreviation for Photoshop. These are image processing applications that often are used for post-production of images converted from raw files into RGB files. ICC is an abbreviation for "International Color Consortium" which is an organization that defines the protocol and standardization for color management tools like device and color space profiles.
Hope that helps you get oriented with the use of abbreviations in this thread.
kirk
-
Kirk, please show me where I use absolute colorimetric rendering intent. Or are you messaging someone else?
Today, I have been using RP, LR, and PS CS5 for my job. It works for me.
Sure do wish someone could guide me in AP, should it be I don't know how to use the application. I thought the videos and other help were guiding me to use PA to it's fullest.
If AP is perfect as is, I need one superior tutorial, in slow motion, to guide me through the process of color management.
My work with RPP > LR > and PS CS5, is what I have been using and works. Again, my dream is to do this AP.
I was looking at your first screenshot in post #14 - that shows the soft proof adjustment layer dialog with Abs Col. selected.
Because you are using a custom printer-paper profile (and display profile), I am interested to know:
1) Is it ICC version 2 or version 4. Some applications do not support version 4.
2) If you use an Epson printer-paper profile, do you have this problem.
And I also want to make sure that I am fully understanding the issue - the comparisons that you have demonstrated here (for example, showing the rendering of the soft-proof image in AP versus PS) are with both images rendered in soft-proofing mode, with the same rendering intent selected, etc. in each application.
In the above screen shot (post #20) you appear to be comparing PS to AP. The HISTORY panel in AP appears to show that you opened the document, added a soft proof layer with some selected adjustment to the soft proof layer, then you FLATTENED the document, burning the soft proof appearance into the document. Then you added a curves layer, etc.
At this point, I am not sure we are comparing apples to apples.
Try this:
Exercise A
1) Open the RPP TIFF in AP and PS. Make no adjustments, etc. just open the image. Make sure AP and PS are both set up to use BetaRGB as the working space, and confirm that the opened image is in BetaRGB.
2) In AP, add a soft proof adjustment layer and select a canned Epson printer-paper profile (i.e., not one of your custom profiles) to display as the soft proof, with your preferred rendering intent. Enable BPC (black point compensation).
3) In PS, choose "View > Proof Setup > Custom..." and chose to soft proof with the exact same canned Epson profile you applied in (2) with the same rendering intent as in (2) and with BPC and "Simulate Paper Color" enabled.
Now compare the two images.
Exercise B
Repeat Exercise A, but with your custom printer-paper profile applied as the soft proofing profile in both AP and PS. You may have to re-apply/activate the specific rendering intent, BPC and Simulate Paper Color.
Compare the two images.
Questions:
1) In Exercise A, do the AP and PS images appear different?
2) In Exercise B, do the AP and PS images appear different? Do the Exercise B images appear different than the exercise A images?
Let us know how it goes.
kirk
-
Your screen shot shows your soft proof using absolute colorimetric rendering intent. Are your custom printer profiles v2 or v4?
Kirk
PS - RPP is my go-to raw converter - highest quality conversions.
-
Here is what I get when I use RPP in BetaRGB, open the resulting TIFF in AP, APbeta and PS and turn soft-proofing on in all three applications, relative colorimetric, Epson 3880 Photo Glossy profile. it appears that AP applies a "paper white" simulation to the soft proof adjustment.
kirk
-
Am I interpreting your screenshots correctly - you are using a beta version of AP? What version are you using?
Also - I can see in some of your screenshots three different windows of the same image (the color checker image) - the foreground image appears to be AP with a soft-proof adjustment, the middle appears to be Photoshop (with the icon in the title bar) and the rear image appears to be RPP (with a custom ICC camera profile applied). Is this correct?
The AP image definitely looks different and more saturated. Is the Photoshop rendering in the screenshot in Soft Proof mode, with the same ICC (Epson) profile applied as in AP? If soft proofing is disabled in both AP and Photoshop, do all of the images look identical? How about not soft-proofing in Absolute Colorimetric rendering intent, but in a more reasonable intent like Perceptual or Rel Colorimetric?
What happens if you do not use a beta version of AP?
kirk
-
Sometimes these kinds of graphics are easier to develop as 3D models with a non-photorealistic render (NPR) as the output. It is much more efficient to change the point of view, section the solid surfaces, generate textures/materials etc. in 3D than it is to redo the illustration that is purely 2D.
That said, if you chose to perform the 2D illustration, you should be aware of basic rules of perspective or isometric projection. You can set up grids and layout lines on layers that will not be visible in the final composite that can help you maintain the view and proportion of objects, etc.
If you have never tackled a 3D modeler/renderer before, you are in for a learning curve. But even a proxy 3D model that you render and then import into AP or AD to trace for the illustration in 2D might be helpful (something simple, like Google SketchUp: http://www.sketchup.com is easy to learn quickly).
kirk
-
A couple of questions for the frustrated OP:
1) How is your color management policy set up in the "Preferences > Colour Profiles" section? What is your RGB Color space (your working color space) and do you have any automatic conversion to the working space (with or without warning) enabled?
2) RPP cannot convert into ProPhotoRGB (it is not a choice in the RPP application) - assuming you do your raw conversion in RPP into color space "X" (BetaRGB, sRGB, etc.) and then have RPP automatically open the resulting TIFF in AP, does the color space into which the RPP conversion was performed persist in the AP interface (does the color profile for the image info read correctly in the tool bar area under the Persona icons)? Or has it changed to something else (through an unexpected conversion or assignment)?
- As an example, I do not have AP set up to convert incoming files automatically into the working space. I have set my AP working space as BetaRGB and I convert my raw files in RPP into BetaRGB when I export them into AP.
3) What does your image (soft-proof or not) look like compared to what you expect? Is it desaturated compared to what you expect? Are some types of images more strongly affected/erroneously displayed than others (more saturated, more shadow tones)?
Can you take a single screenshot of the image, as displayed incorrectly in AP, side-by-side with the same image displayed correctly in another application? Doing it this way (with both images displayed at the same time, side-by-side) eliminates problems with the potential mismatch between the color spaces of separate screen grabs. What is the "reference" application that you are using to make the determination that things in AP are not right?
It sounds like you have a color management issue somewhere in your AP workflow. Even if you export from RPP in a color space that is not your AP working space, color management should display and handle everything correctly. Same goes with soft-proofing, within the ability of your display to show you an accurate representation of the colors in your image.
kirk
-
I second the desire to see "Apply Image" and "Calculations" implemented in AP, especially, as Davide has pointed out, with both operations accessing the on-the-fly color space conversions that are seen the in Curves dialog, for example.
These are really important features for manipulating channels and creating masks in a repeatable, scriptable (Macro persona!!!) way.
thanks for your consideration,
kirk thibault
-
The Develop persona in AP does not make destructive adjustments to your raw file. It pretty much operates the same way that the Adobe Camera Raw plug-in in Photoshop works (a bunch of parametric sliders with local adjustment tools). I cannot imagine it would be that difficult to implement an XML or JSON sidecar file export in the AP developer interface to preserve a set of adjustments. That sidecar could sit next to your raw file and would not increase the file size, just like ACR.
There are a couple of open raw decoding libraries that are available and can be tailored to your specific needs and include several demosaic algorithms, etc. - dcraw and libraw come to mind - I don't think the whole process of raw development has to be reinvented, just implemented in a high-quality, fast and efficient way. I would think there is a population of folks out there that would not agree that Adobe offers the best image quality in terms of their raw conversion.
kirk
-
You might find this utility handy for plotting and converting LUTs.
http://www.vision-color.com/lutconverter/
Kirk
-
In case you need it, here is an online set of utilities to plot and convert LUTs from one format to another:
http://www.vision-color.com/lutconverter/
kirk
-
So, I still don't agree with you. And, this thread is about the quality of AP's RAW conversion. Not about that blog post, specifically. I posted that because it showed clearly the results I was getting. (meaning it wasn't limited to my exact setup) I didn't post any images because it's annoying to do that in a forum. That's why I submitted an original photo so you could see the results for yourself. Or more importantly, so the developers could see them.
I don't think this whole thing has much to do with noise reduction at all. And your statement that if I opened my RAF file in Photos and boosted some stuff it would look like a mess that no noise reduction would fix, is 100% false. The photo I adjusted in Photos (and ACR) looked remarkably good, after boosting things. (mostly though, for best results I adjusted the black point) Have a look for yourself. But let's keep this on topic to what is happening in AP- what's coming from the Develop RAW conversion process is producing remarkably different (and I would say undesired) results compared to the majority of popular RAW processors out there.
Of course you had to change the black point in Photos - that was my point. Each raw converter requires its own individual adjustments to achieve an acceptable image. I do not disagree with your assessment that there is room for improvement with the AP conversion. There are much better raw converters for the Mac than the Develop persona in AP - Raw Photo Processor and Iridient Developer to name two of them. Raw Therapee permits experimentation with a variety of demosaicing algorithms and a very detailed noise reduction control that gives the user plenty of control over how and where NR is applied. Fortunately for Mac users, we have plenty of viable raw converter alternatives, especially if one plans to incorporate AP into their image processing workflow to completely eschew Adobe products from their suite of tools.
kirk
combining L,R, G,B monochrome images
in Pre-V2 Archive of Desktop Questions (macOS and Windows)
Posted
But, once you open your grayscale image, you have to make it a COLOR image (Document > Color Format RGB (16bit)) THEN you can assign a color profile. I was not sure you were in color mode.
kirk