githubmirror/FFmpeg

doc/scaler.texi
237 строк5 KB
Niklas Haas
swscale: add new SwsContext.backends option
04 июн 2026, 00:39
04 июн 2026, 00:39972c0cf
@anchor{scaler_options}
@chapter Scaler Options
@c man begin SCALER OPTIONS

The video scaler supports the following named options.

Options may be set by specifying -@var{option} @var{value} in the
FFmpeg tools, with a few API-only exceptions noted below.
For programmatic use, they can be set explicitly in the
@code{SwsContext} options or through the @file{libavutil/opt.h} API.

@table @option

@anchor{scaler}
@item scaler, scaler_sub
Choose the scaling algorithm to use. Default value is @samp{auto} for both.
It accepts the following values:

@table @samp
@item auto
Aumotic choice. For @samp{scaler_sub}, this means the same algorithm as
@samp{scaler}. For @samp{scaler}, this defaults to the scaler flag selected
by @samp{sws_flags}.

@item bilinear
Bilinear filter. (AKA triangle filter)

@item bicubic
2-tap cubic BC-spline (AKA Mitchell-Netravali spline). The B and C parameters
can be configured by setting @code{param0} and @code{param1}, defaulting to
0.0 and 0.6 respectively.

@item point, neighbor
Point sampling (AKA nearest neighbor).

@item area
Area averaging. Equivalent to @samp{bilinear} for upscaling.

@item gaussian
2-tap Gaussian filter approximation. The sharpness parameter can be configured
by setting @code{param0}, defaulting to 3.0.

@item sinc
Unwindowed sinc filter.

@item lanczos
Lanczos resampling (sinc windowed sinc). The number of filter taps can
be configured by setting @code{param0}, defaulting to 3.

@item spline
Unwindowed natural bicubic spline.
@end table

@anchor{sws_flags}
@item sws_flags
Set the scaler flags. This is also used to set the scaling
algorithm, though this usage is deprecated in favor of setting @samp{scaler}.
Only a single algorithm may be selected. Default value is @samp{bicubic}.

It accepts the following values:
@table @samp
@item fast_bilinear
Select fast bilinear scaling algorithm. (Deprecated)

@item bilinear
Select bilinear scaling algorithm. (Deprecated)

@item bicubic
Select bicubic scaling algorithm. (Deprecated)

@item experimental
Select experimental scaling algorithm. (Deprecated)

@item neighbor
Select nearest neighbor rescaling algorithm. (Deprecated)

@item area
Select averaging area rescaling algorithm. (Deprecated)

@item bicublin
Select bicubic scaling algorithm for the luma component, bilinear for
chroma components. (Deprecated)

@item gauss
Select Gaussian rescaling algorithm. (Deprecated)

@item sinc
Select sinc rescaling algorithm. (Deprecated)

@item lanczos
Select Lanczos rescaling algorithm. The default width (alpha) is 3 and can be
changed by setting @code{param0}. (Deprecated)

@item spline
Select natural bicubic spline rescaling algorithm. (Deprecated)

@item print_info
Enable printing/debug logging.

@item accurate_rnd
Enable accurate rounding.

@item full_chroma_int
Enable full chroma interpolation.

@item full_chroma_inp
Select full chroma input.

@item bitexact
Enable bitexact output.

@item unstable
Allow the use of experimental new code. May subtly affect the output or even
produce wrong results. For testing only.
@end table

@item srcw @var{(API only)}
Set source width.

@item srch @var{(API only)}
Set source height.

@item dstw @var{(API only)}
Set destination width.

@item dsth @var{(API only)}
Set destination height.

@item src_format @var{(API only)}
Set source pixel format (must be expressed as an integer).

@item dst_format @var{(API only)}
Set destination pixel format (must be expressed as an integer).

@item src_range @var{(boolean)}
If value is set to @code{1}, indicates source is full range. Default value is
@code{0}, which indicates source is limited range.

@item dst_range @var{(boolean)}
If value is set to @code{1}, enable full range for destination. Default value
is @code{0}, which enables limited range.

@item gamma @var{(boolean)}
If value is set to @code{1}, enable gamma correct scaling. Default value is @code{0}.

@anchor{sws_params}
@item param0, param1
Set scaling algorithm parameters. The specified values are specific of
some scaling algorithms and ignored by others. The specified values
are floating point number values.

@item sws_dither
Set the dithering algorithm. Accepts one of the following
values. Default value is @samp{auto}.

@table @samp
@item auto
automatic choice

@item none
no dithering

@item bayer
bayer dither

@item ed
error diffusion dither

@item a_dither
arithmetic dither, based using addition

@item x_dither
arithmetic dither, based using xor (more random/less apparent patterning that
a_dither).

@end table

@item alphablend
Set the alpha blending to use when the input has alpha but the output does not.
Default value is @samp{none}.

@table @samp
@item uniform_color
Blend onto a uniform background color

@item checkerboard
Blend onto a checkerboard

@item none
No blending

@end table

@item sws_backends
Set the allowed swscale backends. This is a flags option, so multiple backends
may be combined.

@table @samp
@item auto
Automatic selection. Equal to either @samp{stable} or @samp{all} depending on
whether or not the @samp{unstable} flag is set. This is the default value.

@item stable
All stable backends.

@item unstable
All unstable backends.

@item all
All available backends.

@item legacy
Legacy swscale code.

@item c
Template-based reference code.

@item memcpy
Fast path using libc @code{memcpy}.

@item x86
x86 SIMD kernels.

@item aarch64
AArch64 NEON kernels.

@item spirv
Vulkan SPIR-V backend.

@item glsl
Vulkan GLSL backend.

@end table

@end table

@c man end SCALER OPTIONS