Skip to content

Suggestion: provide shader for each p5js filter #4820

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

Closed
2 of 13 tasks
dhowe opened this issue Oct 1, 2020 · 9 comments · Fixed by #6324
Closed
2 of 13 tasks

Suggestion: provide shader for each p5js filter #4820

dhowe opened this issue Oct 1, 2020 · 9 comments · Fixed by #6324

Comments

@dhowe
Copy link
Contributor

dhowe commented Oct 1, 2020

Suggestion: make available the set of filters in p5js (currently THRESHOLD, GRAY, OPAQUE, INVERT, POSTERIZE, BLUR, ERODE, DILATE) as shaders (with the same names), so that users can more easily port sketches between default and WEBGL renderers.

USE-CASE: Student has performance problems in sketch using filters. Instructor would like to suggest WEBGL, but the filter doesn't exist as a shader (or is not easily found).

Another option would be to have the filter call actually work in WEBGL by loading/invoking the correct shader automatically.

Most appropriate sub-area of p5.js?

  • Accessibility (Web Accessibility)
  • Color
  • Core/Environment/Rendering
  • Data
  • Dom
  • Events
  • Image
  • IO
  • Math
  • Typography
  • Utilities
  • WebGL
  • Other (specify if possible)
@stalgiag
Copy link
Contributor

stalgiag commented Oct 2, 2020

Interesting idea! @aferriss has already done most (all?) of the hard work for this in this repo.

Typically, I would direct a student to this repo when I was suggesting this route. But perhaps it is worth the addition to library size (likely small) to have these as p5.Shader constants.

Performance is really important to consider right now and definitely falls under the umbrella of accessibility. Working with students through Zoom, I am painfully aware of the hard limits that students are hitting with performance. This is exacerbated by the fact that many students are using home computers for the first time when they have used school lab computers (which often have superior hardware) in the past.

For sketches that rely heavily on images with filters, the performance gains with WEBGL and shaders will be very significant.

@dhowe
Copy link
Contributor Author

dhowe commented Oct 3, 2020

Interesting idea! @aferriss has already done most (all?) of the hard work for this in this repo.

Exactly -- I think it would largely be a question of syncing up the names. And seems a good thing to do whether or not we want to have them invoked automatically when using the GL renderer.

@outofambit
Copy link
Contributor

This sounds like a great idea to me! 💜

@JalokimFreelance
Copy link

Hey, I would like to help with that issue, would just the .vert and .frag files be enough? Or do I have to convert them somehow into Shader object?

@outofambit
Copy link
Contributor

@JalokimFreelance i think that would be enough! i'm 90% sure we already have some build process for inlining shader files into the p5.js lib.

@JetStarBlues
Copy link
Contributor

JetStarBlues commented Jun 3, 2021

I spent some time looking at how this might work. What do you think of the approach here?

The shaders for each filter would be files that are handled the same way as other default shaders.

The method applyFilter(filterType, filterParam) would be provided and it would behave the same as filter() in a 2D context.

(applyFilter could also just be called filter, and then the appropriate method called behind-the-scenes following a check of whether the current canvas is 2D or 3D).

Update:

Here is a more robust version that creates a new canvas/buffer via createGraphics() to perform the filter.

@JetStarBlues
Copy link
Contributor

@stalgiag What do you think?

@stalgiag
Copy link
Contributor

Thanks for the draft @JetStarBlues !

I think using filter is a good idea. Your second implementation looks great to me but I am very curious whether get() is going to be a bottleneck. I am getting a low framerate on the example but it is unclear how much of this is caused by creating the shader in applyFilter. I think this is worth pursuing for now and we can see how performance shakes out once a draft PR is made.

@Qianqianye Qianqianye added this to the 1.5.0 milestone Jul 27, 2022
@wong-justin
Copy link
Contributor

Hey all, just a heads up that I'll be tackling this as part of my GSoC 2023 project in the coming weeks. I'll link to a PR when I make one. Looking forward to working on it!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
Status: DONE! 🎉
Development

Successfully merging a pull request may close this issue.

7 participants