vo: enable correct-downscaling, linear-downscaling, sigmoid-upscaling

This commit is contained in:
Kacper Michajłow 2023-09-16 21:42:03 +02:00 committed by Niklas Haas
parent 27a78276eb
commit 8121d41245
3 changed files with 8 additions and 3 deletions

View File

@ -63,6 +63,7 @@ Interface changes
- remove `bcspline` filter (`bicubic` is now the same as `bcspline`)
- rename `--cache-dir` and `--cache-unlink-files` to `--demuxer-cache-dir` and
`--demuxer-cache-unlink-files`
- enable `--correct-downscaling`, `--linear-downscaling`, `--sigmoid-upscaling`
--- mpv 0.36.0 ---
- add `--target-contrast`
- Target luminance value is now also applied when ICC profile is used.

View File

@ -5293,7 +5293,7 @@ them.
Catmull-Rom. A Cubic filter in the same vein as ``mitchell``, where
the ``B`` and ``C`` parameters are ``0.0`` and ``0.5`` respectively.
This filter is sharper than ``mitchell``, but it results in mild
ringing. Like ``mitchell``, this filter is good at downscaling (see
ringing. Like ``mitchell``, this filter is good at downscaling (see
``--dscale``).
``oversample``
@ -5438,6 +5438,7 @@ them.
``--correct-downscaling``
When using convolution based filters, extend the filter size when
downscaling. Increases quality, but reduces performance while downscaling.
Enabled by default.
This will perform slightly sub-optimally for anamorphic video (but still
better than without it) since it will extend the size to match only the
@ -5448,7 +5449,7 @@ them.
``--linear-downscaling``
Scale in linear light when downscaling. It should only be used with a
``--fbo-format`` that has at least 16 bit precision. This option
has no effect on HDR content.
has no effect on HDR content. Enabled by default.
``--linear-upscaling``
Scale in linear light when upscaling. Like ``--linear-downscaling``, it
@ -5459,7 +5460,7 @@ them.
``--sigmoid-upscaling``
When upscaling, use a sigmoidal color transform to avoid emphasizing
ringing artifacts. This is incompatible with and replaces
ringing artifacts. Enabled by default. This is incompatible with and replaces
``--linear-upscaling``. (Note that sigmoidization also requires
linearization, so the ``LINEAR`` rendering step fires in both cases)

View File

@ -313,6 +313,9 @@ static const struct gl_video_opts gl_video_opts_def = {
.clamp = 1, }, // tscale
},
.scaler_resizes_only = true,
.correct_downscaling = true,
.linear_downscaling = true,
.sigmoid_upscaling = true,
.scaler_lut_size = 6,
.interpolation_threshold = 0.01,
.alpha_mode = ALPHA_BLEND_TILES,