PL3 does not read lens sharpness correctly

Hi Tom.

Are your PL2 & PL3 screenshots taken from PL’s display - or from the images exported from PLx ?

If the former then that may explain the difference between PL3 vs Affinity (as result was exported) - - Also, did you apply any interpolation during the export ?

John

This would help explain the speed improvements in PhotoLab 3. The shortcut was always to lower the quality of previews. There’s a very nice technical explanation of sizes and preview quality somewhere in this forum by a DxO developer (one of the very few substantial contributions by an actual developer).

Here we go, from @wolf:

When magnification is >= 75%, HQ (Fast) noise reduction is indeed used in the Preview window, independently of whether PRIME or HQ (Fast) is selected.
When magnification is < 75% or while a slider is moved, we apply a different raw conversion (including a simpler noise reduction algorithm) to speed up the display. This aims at delivering results similar to HQ (Fast), but it’s not strictly equal.

I strongly recommend to visualize the image at >= 75% to have meaningful feedback on denoising sliders when using HQ (Fast). And you must use the magnifier in the Noise Reduction palette to have meaningful feedback on denoising sliders when using PRIME.

Some interesting notes about GPU acceleration which corresponds to my Mac experience:

About the GPU, I can give some insight on what we do today. But please note that these details are subject to change at any moment, as our processing evolves. Currently, the Mac version of PhotoLab 2 does not use the GPU much. The Windows version makes better use of the GPU, but mainly during image export (and it does not accelerate PRIME). So, as of today, to get fast slider updates, the best thing you can do is get a fast CPU.

3 Likes

Here is another comparison.
Let’s ignore, for the moment, that the exported TIFF from PL3 and opened in Affinity Photo restores the sharpening.
The image on the left is a screenshot for PL2. The Right PL3.

Both images were opened from a photo with no sidecar.
The view is 100% with only lens sharpening applied.
The screenshots are from within PL.

If this is how PL3 is operating I will not be able to use it. I can’t be guessing how the photos will look after export.

4 Likes

FWIW, I couldn’t reproduce this with a test image copied into new test folders on Windows.The results I get with PL2 and PL3 are identical. This was with a Panasonic GX85 camera and 12-60mm kit lens.

Have you tried your test with more than one lens module and made sure you’re using the same camera and lens modules between the two versions of PhotoLab?

Thanks for trying it.
Yes I have tried photos with different lenses. The comparison is always on one photo opened in PL2 and 3.

I can simulate the lens sharpness of PL2 in PL3 by boosting the Global slider to about +2.60. But then when I open the exported TIFF in Affinity it is over sharpened.
I suppose a workaround would be to modify the preset to set the Global slider to +2.60 just for viewing. Then reset it to 0 before export. But why?

I’m astonished by the image you’ve shown above. The difference in quality between PhotoLab 3 vs PhotoLab 2 previews is shocking. Can anyone else with access to PhotoLab 3 confirm Tom’s result?

Tom, I’m not sure if you understood my post above. We asked for better speed and faster previews. It appears we’ve been given faster previews but at the expense of quality of preview. This is the Aftershot solution (very fast but fairly awful previews).

3 Likes

I tried two different pictures, from two different cameras and lenses. I saw no difference in the previews between PL1 and PL3. I then even took screenshots and layered them in Pixelmator with Difference blending mode. To my surprise the screenshots were pixel to pixel identical. Since I couldn’t believe this I took two more screenshots and, after careful alignment, once again found them to be pixel wise identical.

I did set the “No Correction” preset and used only lens sharpness. I also tried different settings of lens sharpness. Also, to make sure I am not fooling myself, I tried pictures with different lens sharpness in the Difference layers. I am then able to see the difference, but also with my naked eye.

So it least I can say for my machine (iMac 5K, ATI Radeon Pro 580X) and the tested camera/lens combinations (Canon EOS 400D+Sigma 30mm 1.4 EX DC HSM; 77D+EF 70-300 IS USM) there is absolutely no difference in rendering the preview of lens sharpness between PhotoLab 1 and PhotoLab 3.

But your quote refers to viewing noise reduction. My test images are without noise reduction.

No, what I posted about is PhotoLab speed and quality of preview. It doesn’t matter whether the specific feature discussed is noise reduction. What matters is how the previews are calculated and in what resolution and quality.

Well unless someone comes along to validate what I am seeing, my complaints will be moot.

I’m looking forward to seeing some others’ examples. You may be onto something. It’s great that you posted visual examples which we can all see for ourselves.

I tried it now with PL2 vs PL3, Canon 40D, 24-105mm lens and I do not see any difference - but it was only one file I tried it.

2 Likes

Tom, if possible you could offer your RAW files for download. That would at least allow us confirm camera/lens combination, if I can see the difference on my machine with your RAW files.

If I can’t see the difference then it may be due to different hardware. What are the specs of your machine? Monitor, graphics card especially?

2 Likes

That’s very good news, Sigi. If a few more no difference reports come in, then Tom will need some help troubleshooting what’s different in his PL3 settings than in his PL2 settings.

Tom, if possible you could offer your RAW files for download. That would at least allow us confirm camera/lens combination, if I can see the difference on my machine with your RAW files.

Great idea, Christian.

Also - Tom, are your preferences settings the same in both PL2 and PL3?

Prefs are the same.
Download file here:

Mac mini (2018) 3.2 GHz Intel Core i7 64 GB Radeon Pro 580 8 GB

They are certainly not moot for you. Something is causing the significant difference we are seeing in your examples. It’s not subtle. Perhaps it is related to a specific camera and lens system which is why some other people are saying they don’t see an issue, or perhaps it’s related to your testing procedure. I suppose there may be a number of possible variables. In any case, I think we need a controlled, repeatable test that others can also try in order to come up with a consensus.

Mark

Thanks. I can confirm your finding. I have the same between PL1 and PL3. I’d never spotted it myself, though. I also downloaded a picture with 77D/18-135 from dpreview and it shows the same effect, but not as strong. Also, just like you said, in the exported picture the difference is not visible.

What I found with the two pictures by using the difference layers, though, is also a very subtle difference in the preview with all corrections off. I guess what we see is a difference in the RAW processing which is exaggerated through the lens sharpness, but doesn’t show on all pictures.

I then repeated my test with the picture from the 400D to make sure once more I am not fooling myself and still the results in the preview are identical.

So the effect is there, but not with every picture. It may be related to camera/lens combinations, but for this we would have to test many more pictures.

Just to demonstrate a picture where there is zero difference I attached the example below. It is the picture from the 400D/Sigma 30 1.4 EX. I used lens sharpness +2 because otherwise it has no visible effect on this picture. But also with lens sharpness 0 and off the results are identical:

2 Likes

PS: I do feel like a fool now. I tried to replicate once more to show a picture with Tom’s example. Now I get identical results again. What I changed meanwhile is going out of fullscreen mode (which I use PL usually in). But then, even back in fullscreen mode, I can’t replicate the previous finding. The difference is now again pixel wise identical between PL1 and PL3.

I still believe there is something fishy with the preview in PL3, but it doesn’t seem to be so easy to replicate for me.

Ok now this.
I thought I would open the image with PL2 and 3 at the same time to get a screen shot.
And discovered the following. Open the image, in PL3, that has been reset and only lens sharpness applied.
Expand to 100%. The photo looks soft (compared to PL2) as I have been saying.
Then if I grab the right side of the PL window and start dragging to the left as to shrink the window size, the photo snaps into clarity . It only takes a slight dragging movement.

2 Likes