Improve downscale #720
Merged
Improve downscale #720
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The floating point "size" used for nearest should be 0, not 0.5. Fix clamping so at least one source pixel is used sampled for each output pixel, but also so the floating point "size" applies to source size. I'm not too sure of this for scaling UP, but for scaling _down_ it seems more or less right.
Fast integer-only downscaling.
|
Looks good to me. |
|
Merging. |
|
Thanks! |
kaj
added a commit
to kaj/image
that referenced
this pull request
Apr 18, 2018
Use resize_dimensions from PR image-rs#667 in thumbnail from PR image-rs#720.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When scaling images down, the
supportpixels was counted before scaling, so only a small part of the pixels that should contribute actually did contribute, giving rough edges and moire aliasing problems. Here are some down-scaled sample images scaled before and after this change.Also, the Gaussian method used a standard deviation of 1.0, where I think it should use a standard deviation of 0.5, giving a diameter of 1.0 (rather than a radius of 1.0) for the main contributing area. Here are some up-scaled smaple images before and after, note the blurry images for gaussian scale before the change.
Also, I added an integer-only
thumbnailmethod for fast down-scaling.(Writing this message, I realize I should probably have made this three separate PRs. Sorry.)