-
Notifications
You must be signed in to change notification settings - Fork 15.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Bug]: unexpected 4 and 16 bit image results using electron 21.4.1 and newer #37152
Comments
Hello @robertmassman. Thanks for reporting this and helping to make Electron better! Would it be possible for you to make a standalone testcase with only the code necessary to reproduce the issue? For example, Electron Fiddle is a great tool for making small test cases and makes it easy to publish your test case to a gist that Electron maintainers can use. Stand-alone test cases make fixing issues go more smoothly: it ensure everyone's looking at the same issue, it removes all unnecessary variables from the equation, and it can also provide the basis for automated regression tests. Now adding the |
Hey @robertmassman, adding to the bot's response above - it's a bit hard for us to tell exactly what might be failing without seeing a code repro or any error logs. If you could share a repo with us, or some kind of error logs so we can see what GPU errors are corresponding with the images you linked above, happy to try and take a look! |
Thank you for the response Keeley.
I'd like to help chase down what's going wrong.
So you know, I write code in my free time so I don't have a lot of formal
education. Just to clarify, would you like to see the code I've written as
is or would you like me to create a version that just deals with the GPU
code?
Best,
Rob
…On Mon, Feb 6, 2023 at 11:51 PM Keeley Hammond ***@***.***> wrote:
Hey @robertmassman <https://github.com/robertmassman>, adding to the
bot's response above - it's a bit hard for us to tell exactly what might be
failing without seeing a code repro or any error logs. If you could share a
repo with us, or some kind of error logs so we can see what GPU errors are
corresponding with the images you linked above, happy to try and take a
look!
|
@robertmassman Either would be great! If you're able to share the code written as is, that would be very helpful - basically, what helps us most is some code that allows us to quickly reproduce the issue. That will let us use it as a test case to track down exactly what's different per version and where in Electron is failing. 🙂 Thanks for the fast reply! |
@VerteDinde |
@VerteDinde |
@codebytere |
@robertmassman Thanks so much! Just confirming that I have access to the project, taking a look 🙇♀️ Would you mind if I added a few more maintainers to the project, so they can also take a look? Thanks again for making this and sharing it with us! |
@VerteDinde No problem at all. Feel free to add whoever can help look into the issue with you. |
Hi @robertmassman, I'm so sorry - it looks like I don't have access to the bit depth repo anymore. Would you mind re-adding me? If you're not seeing the issue anymore, no worries! |
@VerteDinde Yes I’m still seeing the problem.
Just added you.
…On Mon, Feb 27, 2023 at 2:46 PM Keeley Hammond ***@***.***> wrote:
Hi @robertmassman <https://github.com/robertmassman>, I'm so sorry - it
looks like I don't have access to the bit depth repo anymore. Would you
mind re-adding me? If you're not seeing the issue anymore, no worries!
|
@robertmassman is there any way you could just provide a gist or public repo? It'd be much easier and better in case we need to return to this later. |
@VerteDinde will do.
…On Tue, Feb 28, 2023 at 9:03 AM Shelley Vohr ***@***.***> wrote:
@robertmassman <https://github.com/robertmassman> is there any way you
could just provide a gist or public repo? It'd be much easier and better in
case we need to return to this later.
|
@VerteDinde says it’s public now.
On Tue, Feb 28, 2023 at 9:09 AM Robert Massman ***@***.***>
wrote:
…
@VerteDinde will do.
On Tue, Feb 28, 2023 at 9:03 AM Shelley Vohr ***@***.***>
wrote:
> @robertmassman <https://github.com/robertmassman> is there any way you
> could just provide a gist or public repo? It'd be much easier and better in
> case we need to return to this later.
|
Any updates on the bit depth investigation? |
hey @robertmassman - are there steps missing in your repo? It doesn't start as is -
|
@codebytere After I use terminal to locate the folder on my computer I type. "npm run start" |
@codebytere PS. Once you are able to run the demo and you decide to see what happens with the most current version of electron by updating "package.json" you will need to re-compile the demo "npm run compile" before running the demo again. Or you could type this into terminal "npm run compile && npm run start" to compile and start the demo. |
@codebytere Hope you are good. Any updates on the bit depth issue? |
Hi @codebytere. Any updates on the above issue? |
Below is a link to a gist I just created. Please use electron 17.4.11 to see the expected behavior and any other version above to see the unexpected results. I hope this helps. https://gist.github.com/robertmassman/80f3379d0d7333ff546681cedf17eac4 To hopefully help your investigation. Please see the comments I left in “gist” at line 10538. It’s in all caps so that I can find it more easily not because I’m shouting. Please don’t take offense. Note that if you enable the above, write to screen code and then comment out the code below that line. Even though the demo is using the 4-bit buffer, the results will be as if you are not working with 8-bit data and display as if it's 8-bit. Similar to what is observed when updating from electron 17.4.11 to electron 21.4.1 or greater and leaving the code as it was before. |
@robertmassman and you've confirmed this does not happen in Chrome with the same code? |
@codebytere If I open index.html in the version of google chrome listed below. I do not see the expected results. The image looks smooth. If I open index.html in the version of Firefox listed below. I see the expected results. The image looks blocky. If I open index.html in the version of Safari listed below. I do not see the expected results. The image looks smooth. |
Unfortunately then there's not a lot we can do - that means this is an image processing issue in Chromium/Webkit. I recommend filing a bug at https://crbug.com and explaining what should be the case vs what you're seeing. When they fix it, we will inherit that! |
Preflight Checklist
Electron Version
21.4.1 and newer
What operating system are you using?
macOS
Operating System Version
macOS 13.1
What arch are you using?
x64
Last Known Working Electron version
17.4.11
Expected Behavior
At the moment I’m using macOS with an Intel processor. I’m running the program with npm and electron and I’m currently using "electron": "^17.4.11" which has been telling me for some time that I need to update it because of the vulnerabilities that are stacking up. Every time I try to update electron I received errors when running the program. As of "electron": "^21.4.1" the program is able to work but I’m noticing some GPU oddities in the results. Please see below.
Is there something that needs to be configured for it to act like it was in previous versions?
Right now I have the program set up with the ability to switch between bit-depths “4, 8 and 16”.
When I investigated further I can see that the starting image texture looks like it has the appropriate bit depth. But when I use the gaussian shader it looks like the gaussian filter is using 8 bit data, Not really sure. I’m including 2 videos that I hope show the differences between using electron 17.4.11 and 21.4.1.
In the first video shows a 4 bit image being blurred using electron 17.4.11. You can see that the image data looks like it's 4 bits during the blur.
https://user-images.githubusercontent.com/12559691/217069777-169fa018-ed0b-432f-9add-5f86d42cdf21.mov
In the second video shows a 4 bit image being blurred using electron 21.4.1. You can see it starts with an image that looks like the one in the first video but as the image is blurred it becomes smooth and soft losing the 4 bit quality that's preserved in the first video.
https://user-images.githubusercontent.com/12559691/217069821-68d7cd44-d95b-4c2c-80e1-d2c60804a7b5.mov
Actual Behavior
Please see above
Testcase Gist URL
No response
Additional Information
No response
The text was updated successfully, but these errors were encountered: