[FFmpeg-devel] [PATCH] video stabilization plugins using vid.stab library
Clément Bœsch
ubitux at gmail.com
Tue Apr 9 10:27:35 CEST 2013
On Tue, Apr 02, 2013 at 12:36:00AM +0200, Georg Martius wrote:
> Hi,
>
> attached is my new patch with new names for the filters:
> vidstabdetect and vidstabtransform
>
> with configure script and makefile adaptations
> with documentation in filters.texi
> without tabs ;-)
>
Sorry for the delay.
The following review is mostly cosmetics now; if you honor the comments
below, it should be OK at the next iteration.
> Regards,
> Georg
>
[...]
> From d1474504e130f170bd235f69969e67a1a32b6fb2 Mon Sep 17 00:00:00 2001
> From: Georg Martius <martius at mis.mpg.de>
> Date: Thu, 28 Mar 2013 22:59:16 +0100
> Subject: [PATCH] video stabilization plugins using vid.stab library:
> vidstabdetect and vidstabtransform
>
>
> Signed-off-by: Georg Martius <martius at mis.mpg.de>
> ---
> Changelog | 1 +
> configure | 6 +
> doc/filters.texi | 184 +++++++++++++++++
> libavfilter/Makefile | 2 +
> libavfilter/allfilters.c | 2 +
> libavfilter/vf_vidstabdetect.c | 331 +++++++++++++++++++++++++++++++
> libavfilter/vf_vidstabtransform.c | 396 +++++++++++++++++++++++++++++++++++++
> 7 files changed, 922 insertions(+)
> create mode 100644 libavfilter/vf_vidstabdetect.c
> create mode 100644 libavfilter/vf_vidstabtransform.c
>
> diff --git a/Changelog b/Changelog
> index 72a51ea..537deff 100644
> --- a/Changelog
> +++ b/Changelog
> @@ -14,6 +14,7 @@ version <next>:
> - WebP demuxing and decoding support
> - new ffmpeg options -filter_script and -filter_complex_script, which allow a
> filtergraph description to be read from a file
> +- stabilize and transform filter: video stabilization using vid.stab library
>
>
> version 1.2:
> diff --git a/configure b/configure
> index 8dc2500..d6284cd 100755
> --- a/configure
> +++ b/configure
> @@ -224,6 +224,7 @@ External library support:
> --enable-libtwolame enable MP2 encoding via libtwolame [no]
> --enable-libutvideo enable Ut Video encoding and decoding via libutvideo [no]
> --enable-libv4l2 enable libv4l2/v4l-utils [no]
> + --enable-libvidstab enable video stabilization using vid.stab [no]
> --enable-libvo-aacenc enable AAC encoding via libvo-aacenc [no]
> --enable-libvo-amrwbenc enable AMR-WB encoding via libvo-amrwbenc [no]
> --enable-libvorbis enable Vorbis en/decoding via libvorbis,
> @@ -1171,6 +1172,7 @@ EXTERNAL_LIBRARY_LIST="
> libtwolame
> libutvideo
> libv4l2
> + libvidstab
> libvo_aacenc
> libvo_amrwbenc
> libvorbis
> @@ -2115,6 +2117,8 @@ stereo3d_filter_deps="gpl"
> subtitles_filter_deps="avformat avcodec libass"
> super2xsai_filter_deps="gpl"
> tinterlace_filter_deps="gpl"
> +vidstabdetect_filter_deps="libvidstab"
> +vidstabtransform_filter_deps="libvidstab"
> yadif_filter_deps="gpl"
> pixfmts_super2xsai_test_deps="super2xsai_filter"
> tinterlace_merge_test_deps="tinterlace_filter"
> @@ -3532,6 +3536,7 @@ die_license_disabled_gpl() {
>
> die_license_disabled gpl libcdio
> die_license_disabled gpl libutvideo
> +die_license_disabled gpl libvidstab
Please add just an entry in "compatible libraries" in the section of the
LICENSE file.
> die_license_disabled gpl libx264
> die_license_disabled gpl libxavs
> die_license_disabled gpl libxvid
> @@ -3967,6 +3972,7 @@ enabled libtwolame && require libtwolame twolame.h twolame_init -ltwolame &&
> die "ERROR: libtwolame must be installed and version must be >= 0.3.10"; }
> enabled libutvideo && require_cpp utvideo "stdint.h stdlib.h utvideo/utvideo.h utvideo/Codec.h" 'CCodec*' -lutvideo -lstdc++
> enabled libv4l2 && require_pkg_config libv4l2 libv4l2.h v4l2_ioctl
> +enabled libvidstab && require_pkg_config vidstab vid.stab/libvidstab.h vsMotionDetectInit
> enabled libvo_aacenc && require libvo_aacenc vo-aacenc/voAAC.h voGetAACEncAPI -lvo-aacenc
> enabled libvo_amrwbenc && require libvo_amrwbenc vo-amrwbenc/enc_if.h E_IF_init -lvo-amrwbenc
> enabled libvorbis && require libvorbis vorbis/vorbisenc.h vorbis_info_init -lvorbisenc -lvorbis -logg
> diff --git a/doc/filters.texi b/doc/filters.texi
> index 4190cca..7bed8f7 100644
> --- a/doc/filters.texi
> +++ b/doc/filters.texi
> @@ -4706,6 +4706,114 @@ much, but it will increase the amount of blurring needed to cover over
> the image and will destroy more information than necessary, and extra
> pixels will slow things down on a large logo.
>
> + at anchor{vidstabtransform}
> + at section vidstabtransform
> +
> +Video stabilization/deshaking: pass 2 of 2,
> +see @ref{vidstabdetect} for pass 1
> +
Period missing.
> +Reads a file with transform information for each frame and
> +applies/compensates them. Together with the @ref{vidstabdetect}
Read, apply, compensate; we don't use the third person for some reason.
> +filter this can be used to deshake videos. See also
> + at url{http://public.hronopik.de/vid.stab}. It is important to also use
> +the unsharp filter, see below.
> +
> +To enable compilation of this filter you need to configure FFmpeg with
> + at code{--enable-libvidstab}.
> +
> +This filter accepts the following named options, expressed as a
> +sequence of @var{key}=@var{value} pairs, separated by ":".
> +
> + at table @option
> +
> + at item input
> +path to the file used to read the transforms (default: transforms.trf)
> +
@file{transforms.trf}
> + at item smoothing
> +number of frames*2 + 1 used for lowpass filtering used for stabilizing
> +(default: 10) For example a number of 10 means that 21 frames are used
Please add some little punctuation here.
> +(10 in the past and 10 in the future) to smoothen the motion in the
> +video. A larger values leads to a smoother video but limits the
> +acceleration of the camera (pan/tilt movements)
> +
> + at item maxshift
> +maximal number of pixels to translate frames (default: -1 no limit)
> +
> + at item maxangle
> +maximal angle in radians (degree*PI/180) to rotate frames (default: -1
> +no limit)
> +
> + at item crop
> +How to deal with borders that may be visible due to movement
> +compensation
"Available values are:"
> + at table @option
@table @samp
> + at item keep
> +keep image information from previous frame (default)
> + at item black
> +fill the border black
> + at end table
> +
> + at item invert
> + 0: keep normal (default), 1: invert transforms
> +
A table with the two entries would be better.
> + at item relative
> +consider transforms as 0: absolute, 1: relative to previous frame
> +(default)
> +
Ditto
> + at item zoom
> +percentage to zoom >0: zoom in, <0 zoom out (default: 0)
> +
> + at item optzoom
> +0: nothing, 1: determine optimal zoom (default).
> +Optimal zoom means no (or only little) border should be visible.
> +Note that the value given at zoom is added to the one calculated
> +here.
> +
> + at item interpol
> +type of interpolation
> +
"Available values are:"
> + at table @option
@table @samp
> + at item no
> +no interpolation
> + at item linear
> +linear only horizontal
> + at item bilinear
> +linear in both directions (default)
> + at item bicubic
> +cubic in both directions (slow)
> + at end table
> +
> + at item tripod
> +0: off (default),
> +1: virtual tripod mode: equivalent to @code{relative=0:smoothing=0}
> +
Again a table would be welcome
> + at end table
> +
> + at subsection Examples
> +
> + at itemize
> + at item
> +typical call with default default values:
> + (note the unsharp filter which is always recommended)
> + at example
> + ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg
> + at end example
> +
Please don't indent the examples; since it's in a <pre> it will be shifted
in comparison with the others.
> + at item
> +zoom in a bit more and load transform data from a given file
> + at example
> + vidstabtransform=zoom=5:input="mytransforms.trf"
> + at end example
> +
> + at item
> +smoothen the video even more
> + at example
> + vidstabtransform=smoothing=30
> + at end example
> +
> + at end itemize
> +
> +
> @section scale
>
> Scale (resize) the input video, using the libswscale library.
> @@ -5084,6 +5192,82 @@ in [-30,0] will filter edges. Default value is 0.
> If a chroma option is not explicitly set, the corresponding luma value
> is set.
>
> + at anchor{vidstabdetect}
> + at section vidstabdetect
> +
> +Video stabilization/deshaking: pass 1 of 2 see @ref{vidstabtransform}
> +for pass 2
> +
> +Generates a file with relative transform information translation,
> +rotation about subsequent frames.
> +
> +To enable compilation of this filter you need to configure FFmpeg with
> + at code{--enable-libvidstab}.
> +
> +This filter accepts the following named options, expressed as a
> +sequence of @var{key}=@var{value} pairs, separated by ":".
> +
> + at table @option
> + at item result
> +path to the file used to write the transforms (default:inputfile.stab)
> +
@file
> + at item shakiness
> +how shaky is the video and how quick is the camera?
> +1: little (fast) 10: very strong/quick (slow) (defaultault: 5)
> +
@table ...
> + at item accuracy
> +accuracy of detection process (>=shakiness)
> +1: low (fast) 15: high (slow) (default: 9)
> +
@table ...
[...]
Same comments as previously apply here as well.
> @section stereo3d
>
> Convert between different stereoscopic image formats.
> diff --git a/libavfilter/Makefile b/libavfilter/Makefile
> index 690b1cb..6be4b31 100644
> --- a/libavfilter/Makefile
> +++ b/libavfilter/Makefile
> @@ -166,6 +166,8 @@ OBJS-$(CONFIG_TINTERLACE_FILTER) += vf_tinterlace.o
> OBJS-$(CONFIG_TRANSPOSE_FILTER) += vf_transpose.o
> OBJS-$(CONFIG_UNSHARP_FILTER) += vf_unsharp.o
> OBJS-$(CONFIG_VFLIP_FILTER) += vf_vflip.o
> +OBJS-$(CONFIG_VIDSTABDETECT_FILTER) += vf_vidstabdetect.o
> +OBJS-$(CONFIG_VIDSTABTRANSFORM_FILTER) += vf_vidstabtransform.o
> OBJS-$(CONFIG_YADIF_FILTER) += vf_yadif.o
>
> OBJS-$(CONFIG_CELLAUTO_FILTER) += vsrc_cellauto.o
> diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
> index 45a67e5..d921272 100644
> --- a/libavfilter/allfilters.c
> +++ b/libavfilter/allfilters.c
> @@ -162,6 +162,8 @@ void avfilter_register_all(void)
> REGISTER_FILTER(TRANSPOSE, transpose, vf);
> REGISTER_FILTER(UNSHARP, unsharp, vf);
> REGISTER_FILTER(VFLIP, vflip, vf);
> + REGISTER_FILTER(VIDSTABDETECT, vidstabdetect, vf);
> + REGISTER_FILTER(VIDSTABTRANSFORM, vidstabtransform, vf);
> REGISTER_FILTER(YADIF, yadif, vf);
>
> REGISTER_FILTER(CELLAUTO, cellauto, vsrc);
> diff --git a/libavfilter/vf_vidstabdetect.c b/libavfilter/vf_vidstabdetect.c
> new file mode 100644
> index 0000000..d7c73f3
> --- /dev/null
> +++ b/libavfilter/vf_vidstabdetect.c
> @@ -0,0 +1,331 @@
> +/*
> + * vf_vidstabdetect.c
> + *
This is is useless and will likely be forgotten in case of rename (let's
say if we want to remove all the vf_ etc prefixes at some point)
> + * Copyright (C) Georg Martius - Jan 2012
> + * georg dot martius at web dot de
> + *
This is a bit inconsistent with the other copyright forms of the codebase.
> + * 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
> + */
> +
> +#define DEFAULT_RESULT_NAME "transforms.trf"
> +
> +#include <vid.stab/libvidstab.h>
> +
> +#include "libavutil/common.h"
> +#include "libavutil/opt.h"
> +#include "libavutil/imgutils.h"
> +// #include "libavcodec/dsputil.h"
If you don't need it, please remove
> +#include "avfilter.h"
> +#include "internal.h"
> +
> +/* private date structure of this filter*/
Useless comment
> +typedef struct _stab_data {
Please remove the '_', or the _stab_data completely.
> + const AVClass* class;
> +
> + VSMotionDetect md;
> +
> + char* args;
> + char* result;
> + FILE* f;
> +} StabData;
> +
> +
> +/* ** Commandline options *** */
> +
Useless comment
> +#define OFFSET(x) offsetof(StabData, x)
> +#define OFFSETMD(x) (offsetof(StabData, md)+offsetof(VSMotionDetect, x))
> +#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
> +
> +static const AVOption vidstabdetect_options[]= {
> + {"result", "path to the file used to write the transforms (def:transforms.trf)", OFFSET(result), AV_OPT_TYPE_STRING, {.str = DEFAULT_RESULT_NAME}},
> + {"shakiness", "how shaky is the video and how quick is the camera?"
> + " 1: little (fast) 10: very strong/quick (slow) (def: 5)", OFFSETMD(shakiness), AV_OPT_TYPE_INT, {.i64 = 5}, 1, 10, FLAGS},
> + {"accuracy", "(>=shakiness) 1: low 15: high (slow) (def: 9)", OFFSETMD(accuracy), AV_OPT_TYPE_INT, {.i64 = 9 }, 1, 15, FLAGS},
> + {"stepsize", "region around minimum is scanned with 1 pixel resolution (def: 6)", OFFSETMD(stepSize), AV_OPT_TYPE_INT, {.i64 = 6}, 1, 32, FLAGS},
> + {"mincontrast", "below this contrast a field is discarded (0-1) (def: 0.3)", OFFSETMD(contrastThreshold), AV_OPT_TYPE_DOUBLE, {.dbl = 0.25}, 0.0, 1.0, FLAGS},
> + {"show", "0: draw nothing (def); 1,2: show fields and transforms", OFFSETMD(show), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 2, FLAGS},
> + {"tripod", "virtual tripod mode (if >0): motion is compared to a reference"
> + " reference frame (frame # is the value) (def: 0)", OFFSETMD(virtualTripod), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, FLAGS},
> + {NULL},
> +};
> +
> +AVFILTER_DEFINE_CLASS(vidstabdetect);
> +
> +/* ** some conversions from avlib to vid.stab constants and functions *** */
> +
> +/** convert AV's pixelformat to vid.stab pixelformat */
> +static VSPixelFormat av_2_vs_pixel_format(AVFilterContext *ctx, enum AVPixelFormat pf){
> + switch(pf){
> + case AV_PIX_FMT_YUV420P: return PF_YUV420P;
> + case AV_PIX_FMT_YUV422P: return PF_YUV422P;
> + case AV_PIX_FMT_YUV444P: return PF_YUV444P;
> + case AV_PIX_FMT_YUV410P: return PF_YUV410P;
> + case AV_PIX_FMT_YUV411P: return PF_YUV411P;
> + case AV_PIX_FMT_YUV440P: return PF_YUV440P;
> + case AV_PIX_FMT_YUVA420P: return PF_YUVA420P;
> + case AV_PIX_FMT_GRAY8: return PF_GRAY8;
> + case AV_PIX_FMT_RGB24: return PF_RGB24;
> + case AV_PIX_FMT_BGR24: return PF_BGR24;
> + case AV_PIX_FMT_RGBA: return PF_RGBA;
> + default:
> + av_log(ctx, AV_LOG_ERROR, "cannot deal with pixel format %i\n", pf);
> + return PF_NONE;
> + }
> +}
> +
> +/// struct to hold a valid context for logging from within vid.stab lib
> +typedef struct {
> + const AVClass* class;
> +} StabLogCtx;
> +
> +/** wrapper to log vs_log into av_log */
> +static int av_log_wrapper(int type, const char* tag, const char* format, ...){
> + va_list ap;
> + StabLogCtx ctx;
> + ctx.class = &vidstabdetect_class;
> + av_log(&ctx, type, "%s: ", tag);
> + va_start (ap, format);
> + av_vlog(&ctx, type, format, ap);
> + va_end (ap);
> + return VS_OK;
> +}
> +
> +/** sets the memory allocation function and logging constants to av versions */
> +static void set_mem_and_log_functions(void){
> + vs_malloc = av_malloc;
> + vs_zalloc = av_mallocz;
> + vs_realloc = av_realloc;
> + vs_free = av_free;
> +
> + VS_ERROR_TYPE = AV_LOG_ERROR;
> + VS_WARN_TYPE = AV_LOG_WARNING;
> + VS_INFO_TYPE = AV_LOG_INFO;
> + VS_MSG_TYPE = AV_LOG_VERBOSE;
> +
> + vs_log = av_log_wrapper;
> +
> + VS_ERROR = 0;
> + VS_OK = 1;
> +}
> +
> +
> +/*************************************************************************/
> +
> +/* Module interface routines and data. */
> +
> +/*************************************************************************/
> +
I believe this huge comment is not really necessary.
> +static av_cold int init(AVFilterContext *ctx, const char *args)
> +{
> + StabData* sd = ctx->priv;
> + set_mem_and_log_functions();
> +
> + sd->class = &vidstabdetect_class;
> + av_opt_set_defaults(sd); // the default values are overwritten by initMotiondetect later
> +
> + av_log(ctx, AV_LOG_VERBOSE, "vidstabdetect filter: init %s\n", LIBVIDSTAB_VERSION);
> +
> + // save args for later, because the initialization of the vid.stab library requires
> + // knowledge about the input.
> + sd->args=av_strdup(args);
> +
> + return 0;
> +}
> +
> +static av_cold void uninit(AVFilterContext *ctx)
> +{
> + StabData *sd = ctx->priv;
> + VSMotionDetect* md = &(sd->md);
> +
> + av_opt_free(sd);
> + if (sd->f) {
> + fclose(sd->f);
> + sd->f = NULL;
> + }
> +
> + vsMotionDetectionCleanup(md);
> + av_free(sd->args);
> +}
> +
> +static int query_formats(AVFilterContext *ctx)
> +{
> + // If you add something here also add it to the above mapping function
> + static const enum AVPixelFormat pix_fmts[] = {
> + AV_PIX_FMT_YUV444P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV420P,
> + AV_PIX_FMT_YUV411P, AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUVA420P,
> + AV_PIX_FMT_YUV440P, AV_PIX_FMT_GRAY8,
> + AV_PIX_FMT_RGB24, AV_PIX_FMT_BGR24, AV_PIX_FMT_RGBA,
> + AV_PIX_FMT_NONE
> + };
> +
> + ff_set_common_formats(ctx, ff_make_format_list(pix_fmts));
> + return 0;
> +}
> +
> +
> +static int config_input(AVFilterLink *inlink)
> +{
> + AVFilterContext *ctx = inlink->dst;
> + StabData *sd = ctx->priv;
> + int returnval;
> +
> + VSMotionDetect* md = &(sd->md);
> + VSFrameInfo fi;
> + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format);
> +
> + vsFrameInfoInit(&fi,inlink->w, inlink->h, av_2_vs_pixel_format(ctx, inlink->format));
> + // check
useless comment
> + if(fi.bytesPerPixel != av_get_bits_per_pixel(desc)/8){
> + av_log(ctx, AV_LOG_ERROR, "pixel-format error: wrong bits/per/pixel");
> + return AVERROR(EINVAL);
> + }
> + if(fi.log2ChromaW != desc->log2_chroma_w){
> + av_log(ctx, AV_LOG_ERROR, "pixel-format error: log2_chroma_w");
> + return AVERROR(EINVAL);
> + }
> +
> + if(fi.log2ChromaH != desc->log2_chroma_h){
> + av_log(ctx, AV_LOG_ERROR, "pixel-format error: log2_chroma_h");
> + return AVERROR(EINVAL);
> + }
> +
These error messages will let users desperate.
> + if(vsMotionDetectInit(md, &fi, "vidstabdetect") != VS_OK){
> + av_log(ctx, AV_LOG_ERROR, "initialization of Motion Detection failed");
> + return AVERROR(EINVAL);
> + }
> +
> + // we need to do it after vsMotionDetectInit because otherwise the values are overwritten
> + if ((returnval = (av_set_options_string(sd, sd->args, "=", ":"))) < 0)
> + return returnval;
> +
> + // display help
> + /* if(optstr_lookup(sd->options, "help")) { */
> + /* av_log(ctx, AV_LOG_INFO, vs_motiondetect_help); */
> + /* return AVERROR(EINVAL); */
> + /* } */
> +
Could you remove such dead code?
> + if(vsMotionDetectConfigure(md)!= VS_OK){
> + av_log(ctx, AV_LOG_ERROR, "configuration of Motion Detection failed\n");
> + return AVERROR(EINVAL);
> + }
> +
> + av_log(ctx, AV_LOG_INFO, "Image Stabilization Settings:\n");
> + av_log(ctx, AV_LOG_INFO, " shakiness = %d\n", md->shakiness);
> + av_log(ctx, AV_LOG_INFO, " accuracy = %d\n", md->accuracy);
> + av_log(ctx, AV_LOG_INFO, " stepsize = %d\n", md->stepSize);
> + av_log(ctx, AV_LOG_INFO, " mincontrast = %f\n", md->contrastThreshold);
> + av_log(ctx, AV_LOG_INFO, " show = %d\n", md->show);
> + av_log(ctx, AV_LOG_INFO, " result = %s\n", sd->result);
> +
> + sd->f = fopen(sd->result, "w");
> + if (sd->f == NULL) {
> + av_log(ctx, AV_LOG_ERROR, "cannot open transform file %s\n", sd->result);
> + return AVERROR(EINVAL);
> + }else{
> + if(vsPrepareFile(md, sd->f) != VS_OK){
> + av_log(ctx, AV_LOG_ERROR, "cannot write to transform file %s\n", sd->result);
> + return AVERROR(EINVAL);
> + }
> + }
> + return 0;
> +}
> +
> +
> +static int filter_frame(AVFilterLink *inlink, AVFrame *in)
> +{
> + AVFilterContext *ctx = inlink->dst;
> + StabData *sd = ctx->priv;
> + VSMotionDetect* md = &(sd->md);
> + LocalMotions localmotions;
> +
> + AVFilterLink *outlink = inlink->dst->outputs[0];
> + int direct = 0;
> + AVFrame *out;
> + VSFrame frame;
> + int plane;
> +
> + if (av_frame_is_writable(in)) {
> + direct = 1;
> + out = in;
> + } else {
> + out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
> + if (!out) {
> + av_frame_free(&in);
> + return AVERROR(ENOMEM);
> + }
> + av_frame_copy_props(out, in);
> + }
> +
> + for(plane=0; plane < md->fi.planes; plane++){
> + frame.data[plane] = in->data[plane];
> + frame.linesize[plane] = in->linesize[plane];
> + }
> + if(vsMotionDetection(md, &localmotions, &frame) != VS_OK){
> + av_log(ctx, AV_LOG_ERROR, "motion detection failed");
> + return AVERROR(AVERROR_EXTERNAL);
> + } else {
> + if(vsWriteToFile(md, sd->f, &localmotions) != VS_OK){
> + av_log(ctx, AV_LOG_ERROR, "cannot write to transform file!");
The exclamation mark is too violent.
> + return AVERROR(errno);
> + }
> + vs_vector_del(&localmotions);
> + }
> + if(md->show>0 && !direct){
> + // copy
We can guess from the next line that "av_image_copy()" is a copy.
> + av_image_copy(out->data, out->linesize,
> + (void*)in->data, in->linesize,
> + in->format, in->width, in->height);
> + }
> +
> + if (!direct)
> + av_frame_free(&in);
> +
> + return ff_filter_frame(outlink, out);
> +}
> +
> +static const AVFilterPad avfilter_vf_vidstabdetect_inputs[] = {
> + {
> + .name = "default",
> + .type = AVMEDIA_TYPE_VIDEO,
> + .filter_frame = filter_frame,
> + .config_props = config_input,
> + },
> + { NULL }
> +};
> +
> +static const AVFilterPad avfilter_vf_vidstabdetect_outputs[] = {
> + {
> + .name = "default",
> + .type = AVMEDIA_TYPE_VIDEO,
> + },
> + { NULL }
> +};
> +
> +AVFilter avfilter_vf_vidstabdetect = {
> + .name = "vidstabdetect",
> + .description = NULL_IF_CONFIG_SMALL("pass 1 of 2 for stabilization"
> + "extracts relative transformations"
> + "(pass 2 see vidstabtransform)"),
> + .priv_size = sizeof(StabData),
> + .init = init,
> + .uninit = uninit,
> + .query_formats = query_formats,
> +
> + .inputs = avfilter_vf_vidstabdetect_inputs,
> + .outputs = avfilter_vf_vidstabdetect_outputs,
> + .priv_class = &vidstabdetect_class,
> +};
Most of the comments above apply to vf_vidstabtransform.c as well.
> diff --git a/libavfilter/vf_vidstabtransform.c b/libavfilter/vf_vidstabtransform.c
[...]
> +
> +/** convert AV's pixelformat to vid.stab pixelformat */
> +static VSPixelFormat av_2_vs_pixel_format(AVFilterContext *ctx, enum AVPixelFormat pf){
> + switch(pf){
> + case AV_PIX_FMT_YUV420P: return PF_YUV420P;
> + case AV_PIX_FMT_YUV422P: return PF_YUV422P;
> + case AV_PIX_FMT_YUV444P: return PF_YUV444P;
> + case AV_PIX_FMT_YUV410P: return PF_YUV410P;
> + case AV_PIX_FMT_YUV411P: return PF_YUV411P;
> + case AV_PIX_FMT_YUV440P: return PF_YUV440P;
> + case AV_PIX_FMT_YUVA420P: return PF_YUVA420P;
> + case AV_PIX_FMT_GRAY8: return PF_GRAY8;
> + case AV_PIX_FMT_RGB24: return PF_RGB24;
> + case AV_PIX_FMT_BGR24: return PF_BGR24;
> + case AV_PIX_FMT_RGBA: return PF_RGBA;
> + default:
> + av_log(ctx, AV_LOG_ERROR, "cannot deal with pixel format %i\n", pf);
> + return PF_NONE;
> + }
> +}
> +
> +
> +/// struct to hold a valid context for logging from within vid.stab lib
> +typedef struct {
> + const AVClass* class;
> +} TransformLogCtx;
> +
> +/** wrapper to log vs_log into av_log */
> +static int av_log_wrapper(int type, const char* tag, const char* format, ...){
> + va_list ap;
> + TransformLogCtx ctx;
> + ctx.class = &vidstabtransform_class;
> + av_log(&ctx, type, "%s: ", tag);
> + va_start (ap, format);
> + av_vlog(&ctx, type, format, ap);
> + va_end (ap);
> + return VS_OK;
> +}
> +
> +/** sets the memory allocation function and logging constants to av versions */
> +static void set_mem_and_log_functions(void){
> + vs_malloc = av_malloc;
> + vs_zalloc = av_mallocz;
> + vs_realloc = av_realloc;
> + vs_free = av_free;
> +
> + VS_ERROR_TYPE = AV_LOG_ERROR;
> + VS_WARN_TYPE = AV_LOG_WARNING;
> + VS_INFO_TYPE = AV_LOG_INFO;
> + VS_MSG_TYPE = AV_LOG_VERBOSE;
> +
> + vs_log = av_log_wrapper;
> +
> + VS_ERROR = 0;
> + VS_OK = 1;
> +}
> +
Could it make sense to factorize all these 3 functions in a
libavfilter/vidstabutils.c or similar?
[...]
--
Clément B.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 490 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20130409/76c4b49d/attachment.asc>
More information about the ffmpeg-devel
mailing list