Conversation
|
Thanks for working on this! According to Any thoughts on why? If not, I'd be glad to take a closer look. |
kleisauke
left a comment
There was a problem hiding this comment.
It looks like the support window of the VIPS_KERNEL_MKS2013 was wrong. For reference, here are the updated ResampleScope graphs:
https://github.com/kleisauke/libvips-pr-4237/tree/after-review#mks2013
|
Thank you @akimon658 for the PR. Do you think it's worth including both kernels, or would we be better to focus on the more recent 2021 "edition" only? (In addition, I note this kernel is rather useful for edge detection, so a future possible improvement there also.) |
Currently, my team is using the 2013 version. I think adding the 2013 edition would be beneficial as it allows us to migrate to libvips without altering application's behavior. |
|
Thanks for doing this work, @akimon658 ! Could you also:
|
|
In the ChangeLog file, which version should I put my change under? |
|
This is an API addition, so it'll be in the next stable version (8.17). Thanks! |
|
👍 |
|
Ah, sorry I missed this—my son Jack just flagged it to me. I would have been happy to provide advice, but it looks like you have it sorted. |
|
Hopefully! Thanks for checking @j-p-c |
For extreme clarity: I have not checked the correctness of your implementation; I assume that that has been covered off. |
Closes #4089
Added support for Magic Kernel Sharp (both 2013 and 2021 version).
Additions in reduceh.cpp are based on this comment: