Use noncustom kernel when possible #41
Open
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
sgt0
force-pushed
the
bicubic-bilinear-prefer-noncustom
branch
from
4acca74 to
57b31cb
Compare
sgt0
changed the title
`descale.Decustom` init is slow and AFAICT it's only being used for symmetry with `resize2.Custom`, which itself was introduced to enable setting `blur` (am I missing something else?). If that's the case, then everything is being made slower just for the (probably) few cases of blur usage. This hurts the native-res plugin especially. With this change, things work like so: 1. If descaling, always use `descale.Debicubic`/`descale.Debilinear`. They support `blur` so there should be no need for custom kernels. 2. If scaling/resampling, use `resize2.Bicubic`/`resize2.Bilinear` if `blur` is not set to the default of 1.0. This isn't here to speed anything up, rather it just seems like a good idea to use the plugin's available filters when possible instead of custom ones. 3. Otherwise, use the slower custom kernels since blur is needed and not provided by the plugin. With this, native-res execution (height [502, 999], step size 1, Catrom) goes from ~3 minutes to ~30 seconds.
sgt0
force-pushed
the
bicubic-bilinear-prefer-noncustom
branch
from
57b31cb to
c4b6ce8
Compare
|
Updated to cover lanczos and the splines as well. Moved the logic into |
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Outdated suggestions cannot be applied.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.
descale.Decustominit is slow and AFAICT it's only being used for symmetry withresize2.Custom, which itself was introduced to enable settingblur(am I missing something else?). If that's the case, then everything is being made slower just for the (probably) few cases of blur usage. This hurts the native-res plugin especially.With this change, things work like so:
descale.Debicubic/descale.Debilinear. They supportblurso there should be no need for custom kernels.resize2.Bicubic/resize2.Bilinearifbluris not set to the default of 1.0. This isn't here to speed anything up, rather it just seems like a good idea to use the plugin's available filters when possible instead of custom ones.With this, native-res execution (height [502, 999], step size 1, Catrom) goes from ~3 minutes to ~30 seconds. If we like this approach and it works well then we can bring it to the other kernels.