[FFmpeg-devel] [PATCHv2] lavfi: add gblur_opencl filter

Dylan Fernando dylanf123 at gmail.com
Tue May 7 03:27:23 EEST 2019


Anyone have any comments/feedback?

On Sat, 27 Apr 2019 at 1:05 am, Dylan Fernando <dylanf123 at gmail.com> 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
>
> + at section gblur_opencl
> +
> +Apply Gaussian blur filter.
> +
> +The filter accepts the following options:
> +
> + at table @option
> + at item sigma
> +Set horizontal sigma, standard deviation of Gaussian blur. Default is
> @code{0.5}.
> +
> + at item planes
> +Set which planes to filter. By default all planes are filtered.
> +
> + at item sigmaV
> +Set vertical sigma, if negative it will be same as @code{sigma}.
> +Default is @code{-1}.
> + at end table
> +
> + at subsection Example
> +
> + at itemize
> + at item
> +Apply Gaussian blur filter with horizontal and vertical sigma of 0.5.
> + at example
> +-i INPUT -vf "hwupload, gblur_opencl=0.5, hwdownload" OUTPUT
> + at end example
> + at 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 0000000000..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 <= half_matrix_dim;
> conv_j++) {
> +        float4 px = read_imagef(src, sampler, loc + (int2)(conv_j, 0));
> +        convPix += px * coef_matrix[(conv_j + half_matrix_dim)];
> +    }
> +
> +    write_imagef(dst, loc, convPix);
> +}
> +
> +__kernel void gblur_conv_vert(__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 <= half_matrix_dim;
> conv_j++) {
> +        float4 px = read_imagef(src, sampler, loc + (int2)(0, conv_j));
> +        convPix += px * coef_matrix[(conv_j + half_matrix_dim)];
> +    }
> +
> +    write_imagef(dst, loc, convPix);
> +}
> diff --git a/libavfilter/opencl_source.h b/libavfilter/opencl_source.h
> index 4118138c30..be7e826c4c 100644
> --- a/libavfilter/opencl_source.h
> +++ b/libavfilter/opencl_source.h
> @@ -20,6 +20,7 @@
>  #define AVFILTER_OPENCL_SOURCE_H
>
>  extern const char *ff_opencl_source_avgblur;
> +extern const char *ff_opencl_source_gblur;
>  extern const char *ff_opencl_source_colorspace_common;
>  extern const char *ff_opencl_source_convolution;
>  extern const char *ff_opencl_source_neighbor;
> diff --git a/libavfilter/vf_gblur_opencl.c b/libavfilter/vf_gblur_opencl.c
> new file mode 100644
> index 0000000000..37c2b24918
> --- /dev/null
> +++ b/libavfilter/vf_gblur_opencl.c
> @@ -0,0 +1,368 @@
> +/*
> + * 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
> + */
> +
> +#include "libavutil/common.h"
> +#include "libavutil/imgutils.h"
> +#include "libavutil/opt.h"
> +
> +#include "avfilter.h"
> +#include "internal.h"
> +#include "opencl.h"
> +#include "opencl_source.h"
> +#include "video.h"
> +
> +typedef struct GBlurOpenCLContext {
> +    OpenCLFilterContext ocf;
> +
> +    int              initialised;
> +    cl_kernel        kernel_horz;
> +    cl_kernel        kernel_vert;
> +
> +    cl_command_queue command_queue;
> +
> +    int planes;
> +    float sigma;
> +    float sigmaV;
> +
> +    cl_mem matrix_horiz;
> +    cl_mem matrix_vert;
> +    int kernel_dimension_horiz;
> +    int kernel_dimension_vert;
> +
> +} GBlurOpenCLContext;
> +
> +
> +static int gblur_opencl_init(AVFilterContext *avctx)
> +{
> +    GBlurOpenCLContext *ctx = avctx->priv;
> +    cl_int cle;
> +    int err;
> +    err = ff_opencl_filter_load_program(avctx, &ff_opencl_source_gblur,
> 1);
> +    if (err < 0)
> +        goto fail;
> +
> +    ctx->command_queue = clCreateCommandQueue(ctx->ocf.hwctx->context,
> +                                              ctx->ocf.hwctx->device_id,
> +                                              0, &cle);
> +    CL_FAIL_ON_ERROR(AVERROR(EIO), "Failed to create OpenCL "
> +                     "command queue %d.\n", cle);
> +
> +    ctx->kernel_horz = clCreateKernel(ctx->ocf.program,
> "gblur_conv_horz", &cle);
> +    CL_FAIL_ON_ERROR(AVERROR(EIO), "Failed to create "
> +                     "kernel %d.\n", cle);
> +    ctx->kernel_vert = clCreateKernel(ctx->ocf.program,
> "gblur_conv_vert", &cle);
> +    CL_FAIL_ON_ERROR(AVERROR(EIO), "Failed to create "
> +                     "kernel %d.\n", cle);
> +
> +    ctx->initialised = 1;
> +    return 0;
> +
> +fail:
> +    if (ctx->command_queue)
> +        clReleaseCommandQueue(ctx->command_queue);
> +    if (ctx->kernel_horz)
> +        clReleaseKernel(ctx->kernel_horz);
> +    if (ctx->kernel_vert)
> +        clReleaseKernel(ctx->kernel_vert);
> +    return err;
> +}
> +
> +
> +static int gblur_opencl_make_filter_params(AVFilterLink *inlink)
> +{
> +    AVFilterContext    *ctx = inlink->dst;
> +    GBlurOpenCLContext *s = ctx->priv;
> +
> +    float *matrix_horiz = NULL;
> +    float *matrix_vert = NULL;
> +    size_t matrix_bytes_horiz;
> +    size_t matrix_bytes_vert;
> +    cl_int cle;
> +    double acc;
> +    int kernel_dimension_vert, kernel_dimension_horiz;
> +
> +    if (s->sigmaV < 0) {
> +        s->sigmaV = s->sigma;
> +    }
> +
> +
> +    kernel_dimension_horiz = (int)ceilf(6 * s->sigma);
> +    if (kernel_dimension_horiz % 2 == 0)
> +        kernel_dimension_horiz++;
> +    matrix_bytes_horiz = sizeof(float)*kernel_dimension_horiz;
> +
> +    matrix_horiz = av_malloc(matrix_bytes_horiz);
> +    if (!matrix_horiz) {
> +        return AVERROR(ENOMEM);
> +    }
> +
> +    acc = 0;
> +
> +    for (int i = 0; i < kernel_dimension_horiz; i++) {
> +        int x = i - (kernel_dimension_horiz / 2);
> +
> +        matrix_horiz[i] = (1 / sqrt(2 * 3.14159*pow(s->sigma,
> 2)))*exp(-(pow(x, 2) / (2 * pow(s->sigma, 2))));
> +        acc += matrix_horiz[i];
> +    }
> +
> +    kernel_dimension_vert = (int)ceilf(6 * s->sigmaV);
> +    if (kernel_dimension_vert % 2 == 0)
> +        kernel_dimension_vert++;
> +
> +    matrix_bytes_vert = sizeof(float)*kernel_dimension_vert;
> +    matrix_vert = av_malloc(matrix_bytes_vert);
> +    if (!matrix_vert) {
> +        return AVERROR(ENOMEM);
> +    }
> +
> +    acc = 0;
> +    for (int i = 0; i < kernel_dimension_vert; i++) {
> +        int x = i - (kernel_dimension_vert / 2);
> +
> +        matrix_vert[i] = (1 / sqrt(2 * 3.14159*pow(s->sigmaV,
> 2)))*exp(-(pow(x, 2) / (2 * pow(s->sigmaV, 2))));
> +        acc += matrix_vert[i];
> +    }
> +
> +    s->kernel_dimension_horiz = kernel_dimension_horiz;
> +    s->matrix_horiz = clCreateBuffer(s->ocf.hwctx->context,
> +                               CL_MEM_READ_ONLY |
> +                               CL_MEM_COPY_HOST_PTR |
> +                               CL_MEM_HOST_NO_ACCESS,
> +                               matrix_bytes_horiz, matrix_horiz, &cle);
> +
> +    s->kernel_dimension_vert = kernel_dimension_vert;
> +    s->matrix_vert = clCreateBuffer(s->ocf.hwctx->context,
> +                               CL_MEM_READ_ONLY |
> +                               CL_MEM_COPY_HOST_PTR |
> +                               CL_MEM_HOST_NO_ACCESS,
> +                               matrix_bytes_vert, matrix_vert, &cle);
> +
> +    av_freep(&matrix_horiz);
> +    av_freep(&matrix_vert);
> +
> +    return 0;
> +}
> +
> +static int gblur_opencl_filter_frame(AVFilterLink *inlink, AVFrame *input)
> +{
> +    AVFilterContext    *avctx = inlink->dst;
> +    AVFilterLink     *outlink = avctx->outputs[0];
> +    GBlurOpenCLContext *ctx = avctx->priv;
> +
> +    AVFrame *output = NULL;
> +    AVFrame *intermediate = NULL;
> +    cl_int cle;
> +    size_t global_work[2];
> +    cl_mem src, dst, inter;
> +    size_t origin[3] = {0, 0, 0};
> +    size_t region[3] = {0, 0, 1};
> +    int err, p;
> +
> +    av_log(ctx, AV_LOG_DEBUG, "Filter input: %s, %ux%u (%"PRId64").\n",
> +           av_get_pix_fmt_name(input->format),
> +           input->width, input->height, input->pts);
> +
> +    if (!input->hw_frames_ctx)
> +        return AVERROR(EINVAL);
> +
> +    if (!ctx->initialised) {
> +        err = gblur_opencl_init(avctx);
> +        if (err < 0)
> +            goto fail;
> +
> +        err = gblur_opencl_make_filter_params(inlink);
> +        if (err < 0)
> +            goto fail;
> +    }
> +
> +    output = ff_get_video_buffer(outlink, outlink->w, outlink->h);
> +    if (!output) {
> +        err = AVERROR(ENOMEM);
> +        goto fail;
> +    }
> +    intermediate = ff_get_video_buffer(outlink, outlink->w, outlink->h);
> +    if (!intermediate) {
> +        err = AVERROR(ENOMEM);
> +        goto fail;
> +    }
> +
> +    for (p = 0; p < FF_ARRAY_ELEMS(output->data); p++) {
> +        src = (cl_mem) input->data[p];
> +        dst = (cl_mem) output->data[p];
> +        inter = (cl_mem)intermediate->data[p];
> +
> +        if (!dst)
> +            break;
> +
> +        if (!(ctx->planes & (1 << p))) {
> +            err = ff_opencl_filter_work_size_from_image(avctx, region,
> output, p, 0);
> +            if (err < 0)
> +                goto fail;
> +
> +            cle = clEnqueueCopyImage(ctx->command_queue, src, dst,
> +                                        origin, origin, region, 0, NULL,
> NULL);
> +            CL_FAIL_ON_ERROR(AVERROR(EIO), "Failed to copy plane %d:
> %d.\n",
> +                                p, cle);
> +        }
> +        else {
> +            CL_SET_KERNEL_ARG(ctx->kernel_horz, 0, cl_mem,   &inter);
> +            CL_SET_KERNEL_ARG(ctx->kernel_horz, 1, cl_mem,   &src);
> +            CL_SET_KERNEL_ARG(ctx->kernel_horz, 2, cl_int,
>  &ctx->kernel_dimension_horiz);
> +            CL_SET_KERNEL_ARG(ctx->kernel_horz, 3, cl_mem,
>  &ctx->matrix_horiz);
> +
> +            err = ff_opencl_filter_work_size_from_image(avctx,
> global_work, intermediate, p, 0);
> +            if (err < 0)
> +                goto fail;
> +
> +            av_log(avctx, AV_LOG_DEBUG, "Run kernel on plane %d "
> +                    "(%"SIZE_SPECIFIER"x%"SIZE_SPECIFIER").\n",
> +                    p, global_work[0], global_work[1]);
> +
> +            cle = clEnqueueNDRangeKernel(ctx->command_queue,
> ctx->kernel_horz, 2, NULL,
> +                                            global_work, NULL,
> +                                            0, NULL, NULL);
> +            CL_FAIL_ON_ERROR(AVERROR(EIO), "Failed to enqueue horizontal "
> +                                "kernel: %d.\n", cle);
> +
> +
> +            CL_SET_KERNEL_ARG(ctx->kernel_vert, 0, cl_mem,   &dst);
> +            CL_SET_KERNEL_ARG(ctx->kernel_vert, 1, cl_mem,   &inter);
> +            CL_SET_KERNEL_ARG(ctx->kernel_vert, 2, cl_int,
>  &ctx->kernel_dimension_vert);
> +            CL_SET_KERNEL_ARG(ctx->kernel_vert, 3, cl_mem,
>  &ctx->matrix_vert);
> +
> +            err = ff_opencl_filter_work_size_from_image(avctx,
> global_work, output, p, 0);
> +            if (err < 0)
> +                goto fail;
> +
> +            av_log(avctx, AV_LOG_DEBUG, "Run kernel on plane %d "
> +                    "(%"SIZE_SPECIFIER"x%"SIZE_SPECIFIER").\n",
> +                    p, global_work[0], global_work[1]);
> +
> +            cle = clEnqueueNDRangeKernel(ctx->command_queue,
> ctx->kernel_vert, 2, NULL,
> +                                            global_work, NULL,
> +                                            0, NULL, NULL);
> +            CL_FAIL_ON_ERROR(AVERROR(EIO), "Failed to enqueue vertical "
> +                                "kernel: %d.\n", cle);
> +        }
> +    }
> +
> +    cle = clFinish(ctx->command_queue);
> +    CL_FAIL_ON_ERROR(AVERROR(EIO), "Failed to finish command queue:
> %d.\n", cle);
> +
> +    err = av_frame_copy_props(output, input);
> +    if (err < 0)
> +        goto fail;
> +
> +    av_frame_free(&input);
> +    av_frame_free(&intermediate);
> +
> +    av_log(ctx, AV_LOG_DEBUG, "Filter output: %s, %ux%u (%"PRId64").\n",
> +           av_get_pix_fmt_name(output->format),
> +           output->width, output->height, output->pts);
> +
> +    return ff_filter_frame(outlink, output);
> +
> +fail:
> +    clFinish(ctx->command_queue);
> +    av_frame_free(&input);
> +    av_frame_free(&output);
> +    av_frame_free(&intermediate);
> +    return err;
> +}
> +
> +
> +static av_cold void gblur_opencl_uninit(AVFilterContext *avctx)
> +{
> +    GBlurOpenCLContext *ctx = avctx->priv;
> +    cl_int cle;
> +
> +    if (ctx->kernel_horz) {
> +        cle = clReleaseKernel(ctx->kernel_horz);
> +        if (cle != CL_SUCCESS)
> +            av_log(avctx, AV_LOG_ERROR, "Failed to release "
> +                   "kernel: %d.\n", cle);
> +    }
> +
> +    if (ctx->kernel_vert) {
> +        cle = clReleaseKernel(ctx->kernel_vert);
> +        if (cle != CL_SUCCESS)
> +            av_log(avctx, AV_LOG_ERROR, "Failed to release "
> +                   "kernel: %d.\n", cle);
> +    }
> +
> +    if (ctx->command_queue) {
> +        cle = clReleaseCommandQueue(ctx->command_queue);
> +        if (cle != CL_SUCCESS)
> +            av_log(avctx, AV_LOG_ERROR, "Failed to release "
> +                   "command queue: %d.\n", cle);
> +    }
> +
> +    ff_opencl_filter_uninit(avctx);
> +}
> +
> +
> +static const AVFilterPad gblur_opencl_inputs[] = {
> +    {
> +        .name         = "default",
> +        .type         = AVMEDIA_TYPE_VIDEO,
> +        .filter_frame = &gblur_opencl_filter_frame,
> +        .config_props = &ff_opencl_filter_config_input,
> +    },
> +    { NULL }
> +};
> +
> +
> +static const AVFilterPad gblur_opencl_outputs[] = {
> +    {
> +        .name         = "default",
> +        .type         = AVMEDIA_TYPE_VIDEO,
> +        .config_props = &ff_opencl_filter_config_output,
> +    },
> +    { NULL }
> +};
> +
> +
> +#define OFFSET(x) offsetof(GBlurOpenCLContext, x)
> +#define FLAGS (AV_OPT_FLAG_FILTERING_PARAM | AV_OPT_FLAG_VIDEO_PARAM)
> +
> +#if CONFIG_GBLUR_OPENCL_FILTER
> +
> +static const AVOption gblur_opencl_options[] = {
> +    { "sigma",  "set horizontal sigma",  OFFSET(sigma),
> AV_OPT_TYPE_FLOAT, {.dbl=0.5},   0.0, 1024, FLAGS },
> +    { "planes", "set planes to filter", OFFSET(planes),  AV_OPT_TYPE_INT,
> {.i64=0xF}, 0,  0xF, FLAGS },
> +    { "sigmaV", "set vertical sigma",   OFFSET(sigmaV),
> AV_OPT_TYPE_FLOAT, {.dbl=-1},   -1, 1024, FLAGS },
> +    { NULL }
> +};
> +
> +AVFILTER_DEFINE_CLASS(gblur_opencl);
> +
> +
> +AVFilter ff_vf_gblur_opencl = {
> +    .name           = "gblur_opencl",
> +    .description    = NULL_IF_CONFIG_SMALL("Apply gaussian blur filter"),
> +    .priv_size      = sizeof(GBlurOpenCLContext),
> +    .priv_class     = &gblur_opencl_class,
> +    .init           = &ff_opencl_filter_init,
> +    .uninit         = &gblur_opencl_uninit,
> +    .query_formats  = &ff_opencl_filter_query_formats,
> +    .inputs         = gblur_opencl_inputs,
> +    .outputs        = gblur_opencl_outputs,
> +    .flags_internal = FF_FILTER_FLAG_HWFRAME_AWARE,
> +};
> +
> +#endif /* CONFIG_GBLUR_OPENCL_FILTER */
> --
> 2.15.2 (Apple Git-101.1)
>
>


More information about the ffmpeg-devel mailing list