Anyone have any comments/feedback?
On Sat, 27 Apr 2019 at 1:05 am, Dylan Fernando wrote:
> ---
> configure | 1 +
> doc/filters.texi | 28
> libavfilter/Makefile | 2 +
> libavfilter/allfilters.c | 1 +
> libavfilter/opencl/gblur.cl | 62 +++
> libavfilter/opencl_source.h | 1 +
> libavfilter/vf_gblur_opencl.c | 368
> ++
> 7 files changed, 463 insertions(+)
> create mode 100644 libavfilter/opencl/gblur.cl
> create mode 100644 libavfilter/vf_gblur_opencl.c
>
> diff --git a/configure b/configure
> index bbeaf2fadc..8c1d3cdf92 100755
> --- a/configure
> +++ b/configure
> @@ -3451,6 +3451,7 @@ freezedetect_filter_select="scene_sad"
> frei0r_filter_deps="frei0r libdl"
> frei0r_src_filter_deps="frei0r libdl"
> fspp_filter_deps="gpl"
> +gblur_opencl_filter_deps="opencl"
> geq_filter_deps="gpl"
> histeq_filter_deps="gpl"
> hqdn3d_filter_deps="gpl"
> diff --git a/doc/filters.texi b/doc/filters.texi
> index 4bf96b6d90..80485fed83 100644
> --- a/doc/filters.texi
> +++ b/doc/filters.texi
> @@ -19010,6 +19010,34 @@ Apply erosion filter with threshold0 set to 30,
> threshold1 set 40, threshold2 se
> @end example
> @end itemize
>
> +@section gblur_opencl
> +
> +Apply Gaussian blur filter.
> +
> +The filter accepts the following options:
> +
> +@table @option
> +@item sigma
> +Set horizontal sigma, standard deviation of Gaussian blur. Default is
> @code{0.5}.
> +
> +@item planes
> +Set which planes to filter. By default all planes are filtered.
> +
> +@item sigmaV
> +Set vertical sigma, if negative it will be same as @code{sigma}.
> +Default is @code{-1}.
> +@end table
> +
> +@subsection Example
> +
> +@itemize
> +@item
> +Apply Gaussian blur filter with horizontal and vertical sigma of 0.5.
> +@example
> +-i INPUT -vf "hwupload, gblur_opencl=0.5, hwdownload" OUTPUT
> +@end example
> +@end itemize
> +
> @section overlay_opencl
>
> Overlay one video on top of another.
> diff --git a/libavfilter/Makefile b/libavfilter/Makefile
> index fef6ec5c55..230315ef39 100644
> --- a/libavfilter/Makefile
> +++ b/libavfilter/Makefile
> @@ -243,6 +243,8 @@ OBJS-$(CONFIG_FREEZEDETECT_FILTER) +=
> vf_freezedetect.o
> OBJS-$(CONFIG_FREI0R_FILTER) += vf_frei0r.o
> OBJS-$(CONFIG_FSPP_FILTER) += vf_fspp.o
> OBJS-$(CONFIG_GBLUR_FILTER) += vf_gblur.o
> +OBJS-$(CONFIG_GBLUR_OPENCL_FILTER) += vf_gblur_opencl.o
> opencl.o \
> +opencl/gblur.o
> OBJS-$(CONFIG_GEQ_FILTER)+= vf_geq.o
> OBJS-$(CONFIG_GRADFUN_FILTER)+= vf_gradfun.o
> OBJS-$(CONFIG_GRAPHMONITOR_FILTER) += f_graphmonitor.o
> diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
> index c51ae0f3c7..cb0fc051cc 100644
> --- a/libavfilter/allfilters.c
> +++ b/libavfilter/allfilters.c
> @@ -229,6 +229,7 @@ extern AVFilter ff_vf_freezedetect;
> extern AVFilter ff_vf_frei0r;
> extern AVFilter ff_vf_fspp;
> extern AVFilter ff_vf_gblur;
> +extern AVFilter ff_vf_gblur_opencl;
> extern AVFilter ff_vf_geq;
> extern AVFilter ff_vf_gradfun;
> extern AVFilter ff_vf_graphmonitor;
> diff --git a/libavfilter/opencl/gblur.cl b/libavfilter/opencl/gblur.cl
> new file mode 100644
> index 00..4fece30d4a
> --- /dev/null
> +++ b/libavfilter/opencl/gblur.cl
> @@ -0,0 +1,62 @@
> +/*
> + * Copyright (c) 2018 Dylan Fernando
> + *
> + * This file is part of FFmpeg.
> + *
> + * FFmpeg is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU Lesser General Public
> + * License as published by the Free Software Foundation; either
> + * version 2.1 of the License, or (at your option) any later version.
> + *
> + * FFmpeg is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + * Lesser General Public License for more details.
> + *
> + * You should have received a copy of the GNU Lesser General Public
> + * License along with FFmpeg; if not, write to the Free Software
> + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
> 02110-1301 USA
> + */
> +
> +
> +__kernel void gblur_conv_horz(__write_only image2d_t dst,
> + __read_only image2d_t src,
> + int coef_matrix_dim,
> + __constant float *coef_matrix)
> +{
> +const sampler_t sampler = (CLK_NORMALIZED_COORDS_FALSE |
> + CLK_ADDRESS_CLAMP_TO_EDGE |
> + CLK_FILTER_NEAREST);
> +
> +const int half_matrix_dim = (coef_matrix_dim / 2);
> +int2 loc = (int2)(get_global_id(0), get_global_id(1));
> +float4 convPix = (float4)(0.0f, 0.0f, 0.0f, 0.0f);
> +
> +for (int conv_j = -half_matrix_dim; conv_j <=
