JpegXL opening bugs

[Toastman3000]

When opening JPEG XL files the following bugs appear

• PS can't determine the difference between 8bit, 16bit - PS automatically opens the files as 32bit regardless
• PS displays JPEG XL files in completely the wrong gamma range, Affinity Photo doesn't do this
• When converting a 16bit JPEG XL to an 8Bit, it just goes pure black

[CShubert]

Hi @Toastman3000 testing here and not seeing an issue.  Could you share some screenshots of your workspace and the errors that you are seeing?

1. Try resetting Prefs for the Beta:
/Users/yourname/Library/Preferences/ Photoshop Beta Settings (Mac Ventura)
/Users/yourname/Library/Preferences/Adobe\ Photoshop\ \(Beta\)\ Settings (Mac)


C:\Users\user name\AppData\Roaming\Adobe\Adobe Photoshop (Beta)\Adobe Photoshop (Beta) Settings

Launch BETA Photoshop While Pressing The Keyboard Shortcut
With Photoshop closed, press and hold Shift+Ctrl+Alt (Win) / Shift+Command+Option (Mac) on your keyboard and then Launch Photoshop the way you normally would.
https://helpx.adobe.com/photoshop/using/preferences.html#Manually 

2. It may help if we could see your Photoshop beta System Info. Launch Photoshop beta, and select Help >System Info...and copy/paste the text in a reply.  

Thanks,

Cory - Photoshop Product Manager

[Toastman3000]

Thanks for the reply. This is a fresh install of PS Beta on macos.

Is there a way I can send someone the JPEG xl files I'm working with? They're being created by Film Lab App, software which just converts negatives into positives. It can also save tiffs which is obviously fine.

When it saves jpegxl I'm only seeing these bugs I've described with Photoshop beta only, affinity photo reads them just fine as does macos itself.

[CShubert]

Yes @Toastman3000 send them to me, shubert@adobe.com

Thanks,

Cory - Photoshop Product Manager

[gregbenz]

Are you opening the images as HDR through Adobe Camera RAW? That would make everything 32-bit, regardless of whether the source is 32-bit. Be sure to open directly in PS without ACR or at least don't use HDR mode in ACR if you don't want a 32-bit image.

[CShubert]

Hi @Toastman3000 chatting with the team, this is working as expected.  The issue, or difference between Ps and your app is that:

"The files shared were encoded as 16b half float. Half float is always assumed linear and opened in 32b (Ps do not have a 16 float format, only 32f).

 If you open a floating point file in Ps, we read it as floating point, and it would not be desirable to do otherwise. The issue here appears to be that the app exporting the JXL is accidentally writing 16b floats regardless of what you choose in the export settings. We downloaded FilmLab and regardless if we choose 8/16b (it doesn't specify if it's supposed to be integer or float), we end up with 16b float files."

Hope that clears this up for you.

Thanks,

Cory - Photoshop Product Manager

[Toastman3000]

It doesn't really make sense to me that macOS preview, Affinity Photo read the file with the correct gamma but PS is the odd one out. It reminds me of the same situation openEXR we were in with Photoshop 15 years ago where the gamma was read wrong. That was changed later on,

[gregbenz]

@Toastman3000I don't disagree with you that support should be as comprehensive as possible, but there's a cost to supporting every edge case too. 32-bit generally means linear (not that there's some universal spec around it).

Do you see a valid use case for using the very wide range of 16-bit float encoding with a gamma other than linear? If there isn't, it seems like the encoder should be using something which will be more accurate (less quantization error) or more efficient (fewer bits).