ffmpeg/libavfilter/vf_lensfun.c

543 lines
26 KiB
C
Raw Normal View History

Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
/*
* Copyright (C) 2007 by Andrew Zabolotny (author of lensfun, from which this filter derives from)
* Copyright (C) 2018 Stephen Seo
*
* This file is part of FFmpeg.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program 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 General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <https://www.gnu.org/licenses/>.
*/
/**
* @file
* Lensfun filter, applies lens correction with parameters from the lensfun database
*
* @see https://lensfun.sourceforge.net/
*/
#include <float.h>
#include <math.h>
#include "libavutil/imgutils.h"
#include "libavutil/opt.h"
#include "libswscale/swscale.h"
#include "avfilter.h"
#include "formats.h"
#include "internal.h"
#include "video.h"
#include <lensfun.h>
#define LANCZOS_RESOLUTION 256
enum Mode {
VIGNETTING = 0x1,
GEOMETRY_DISTORTION = 0x2,
SUBPIXEL_DISTORTION = 0x4
};
enum InterpolationType {
NEAREST,
LINEAR,
LANCZOS
};
typedef struct VignettingThreadData {
int width, height;
uint8_t *data_in;
int linesize_in;
int pixel_composition;
lfModifier *modifier;
} VignettingThreadData;
typedef struct DistortionCorrectionThreadData {
int width, height;
const float *distortion_coords;
const uint8_t *data_in;
uint8_t *data_out;
int linesize_in, linesize_out;
const float *interpolation;
int mode;
int interpolation_type;
} DistortionCorrectionThreadData;
typedef struct LensfunContext {
const AVClass *class;
const char *make, *model, *lens_model;
int mode;
float focal_length;
float aperture;
float focus_distance;
float scale;
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
int target_geometry;
int reverse;
int interpolation_type;
float *distortion_coords;
float *interpolation;
lfLens *lens;
lfCamera *camera;
lfModifier *modifier;
} LensfunContext;
#define OFFSET(x) offsetof(LensfunContext, x)
#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
static const AVOption lensfun_options[] = {
{ "make", "set camera maker", OFFSET(make), AV_OPT_TYPE_STRING, {.str=NULL}, 0, 0, FLAGS },
{ "model", "set camera model", OFFSET(model), AV_OPT_TYPE_STRING, {.str=NULL}, 0, 0, FLAGS },
{ "lens_model", "set lens model", OFFSET(lens_model), AV_OPT_TYPE_STRING, {.str=NULL}, 0, 0, FLAGS },
{ "mode", "set mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64=GEOMETRY_DISTORTION}, 0, VIGNETTING | GEOMETRY_DISTORTION | SUBPIXEL_DISTORTION, FLAGS, "mode" },
{ "vignetting", "fix lens vignetting", 0, AV_OPT_TYPE_CONST, {.i64=VIGNETTING}, 0, 0, FLAGS, "mode" },
{ "geometry", "correct geometry distortion", 0, AV_OPT_TYPE_CONST, {.i64=GEOMETRY_DISTORTION}, 0, 0, FLAGS, "mode" },
{ "subpixel", "fix chromatic aberrations", 0, AV_OPT_TYPE_CONST, {.i64=SUBPIXEL_DISTORTION}, 0, 0, FLAGS, "mode" },
{ "vig_geo", "fix lens vignetting and correct geometry distortion", 0, AV_OPT_TYPE_CONST, {.i64=VIGNETTING | GEOMETRY_DISTORTION}, 0, 0, FLAGS, "mode" },
{ "vig_subpixel", "fix lens vignetting and chromatic aberrations", 0, AV_OPT_TYPE_CONST, {.i64=VIGNETTING | SUBPIXEL_DISTORTION}, 0, 0, FLAGS, "mode" },
{ "distortion", "correct geometry distortion and chromatic aberrations", 0, AV_OPT_TYPE_CONST, {.i64=GEOMETRY_DISTORTION | SUBPIXEL_DISTORTION}, 0, 0, FLAGS, "mode" },
{ "all", NULL, 0, AV_OPT_TYPE_CONST, {.i64=VIGNETTING | GEOMETRY_DISTORTION | SUBPIXEL_DISTORTION}, 0, 0, FLAGS, "mode" },
{ "focal_length", "focal length of video (zoom; constant for the duration of the use of this filter)", OFFSET(focal_length), AV_OPT_TYPE_FLOAT, {.dbl=18}, 0.0, DBL_MAX, FLAGS },
{ "aperture", "aperture (constant for the duration of the use of this filter)", OFFSET(aperture), AV_OPT_TYPE_FLOAT, {.dbl=3.5}, 0.0, DBL_MAX, FLAGS },
{ "focus_distance", "focus distance (constant for the duration of the use of this filter)", OFFSET(focus_distance), AV_OPT_TYPE_FLOAT, {.dbl=1000.0f}, 0.0, DBL_MAX, FLAGS },
{ "scale", "scale factor applied after corrections (0.0 means automatic scaling)", OFFSET(scale), AV_OPT_TYPE_FLOAT, {.dbl=0.0}, 0.0, DBL_MAX, FLAGS },
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
{ "target_geometry", "target geometry of the lens correction (only when geometry correction is enabled)", OFFSET(target_geometry), AV_OPT_TYPE_INT, {.i64=LF_RECTILINEAR}, 0, INT_MAX, FLAGS, "lens_geometry" },
{ "rectilinear", "rectilinear lens (default)", 0, AV_OPT_TYPE_CONST, {.i64=LF_RECTILINEAR}, 0, 0, FLAGS, "lens_geometry" },
{ "fisheye", "fisheye lens", 0, AV_OPT_TYPE_CONST, {.i64=LF_FISHEYE}, 0, 0, FLAGS, "lens_geometry" },
{ "panoramic", "panoramic (cylindrical)", 0, AV_OPT_TYPE_CONST, {.i64=LF_PANORAMIC}, 0, 0, FLAGS, "lens_geometry" },
{ "equirectangular", "equirectangular", 0, AV_OPT_TYPE_CONST, {.i64=LF_EQUIRECTANGULAR}, 0, 0, FLAGS, "lens_geometry" },
{ "fisheye_orthographic", "orthographic fisheye", 0, AV_OPT_TYPE_CONST, {.i64=LF_FISHEYE_ORTHOGRAPHIC}, 0, 0, FLAGS, "lens_geometry" },
{ "fisheye_stereographic", "stereographic fisheye", 0, AV_OPT_TYPE_CONST, {.i64=LF_FISHEYE_STEREOGRAPHIC}, 0, 0, FLAGS, "lens_geometry" },
{ "fisheye_equisolid", "equisolid fisheye", 0, AV_OPT_TYPE_CONST, {.i64=LF_FISHEYE_EQUISOLID}, 0, 0, FLAGS, "lens_geometry" },
{ "fisheye_thoby", "fisheye as measured by thoby", 0, AV_OPT_TYPE_CONST, {.i64=LF_FISHEYE_THOBY}, 0, 0, FLAGS, "lens_geometry" },
{ "reverse", "Does reverse correction (regular image to lens distorted)", OFFSET(reverse), AV_OPT_TYPE_BOOL, {.i64=0}, 0, 1, FLAGS },
{ "interpolation", "Type of interpolation", OFFSET(interpolation_type), AV_OPT_TYPE_INT, {.i64=LINEAR}, 0, LANCZOS, FLAGS, "interpolation" },
{ "nearest", NULL, 0, AV_OPT_TYPE_CONST, {.i64=NEAREST}, 0, 0, FLAGS, "interpolation" },
{ "linear", NULL, 0, AV_OPT_TYPE_CONST, {.i64=LINEAR}, 0, 0, FLAGS, "interpolation" },
{ "lanczos", NULL, 0, AV_OPT_TYPE_CONST, {.i64=LANCZOS}, 0, 0, FLAGS, "interpolation" },
{ NULL }
};
AVFILTER_DEFINE_CLASS(lensfun);
static av_cold int init(AVFilterContext *ctx)
{
LensfunContext *lensfun = ctx->priv;
lfDatabase *db;
const lfCamera **cameras;
const lfLens **lenses;
db = lf_db_create();
if (lf_db_load(db) != LF_NO_ERROR) {
lf_db_destroy(db);
av_log(ctx, AV_LOG_FATAL, "Failed to load lensfun database\n");
return AVERROR_INVALIDDATA;
}
if (!lensfun->make || !lensfun->model) {
const lfCamera *const *cameras = lf_db_get_cameras(db);
av_log(ctx, AV_LOG_FATAL, "Option \"make\" or option \"model\" not specified\n");
av_log(ctx, AV_LOG_INFO, "Available values for \"make\" and \"model\":\n");
for (int i = 0; cameras && cameras[i]; i++)
av_log(ctx, AV_LOG_INFO, "\t%s\t%s\n", cameras[i]->Maker, cameras[i]->Model);
lf_db_destroy(db);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
return AVERROR(EINVAL);
} else if (!lensfun->lens_model) {
const lfLens *const *lenses = lf_db_get_lenses(db);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
av_log(ctx, AV_LOG_FATAL, "Option \"lens_model\" not specified\n");
av_log(ctx, AV_LOG_INFO, "Available values for \"lens_model\":\n");
for (int i = 0; lenses && lenses[i]; i++)
av_log(ctx, AV_LOG_INFO, "\t%s\t(make %s)\n", lenses[i]->Model, lenses[i]->Maker);
lf_db_destroy(db);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
return AVERROR(EINVAL);
}
lensfun->lens = lf_lens_create();
lensfun->camera = lf_camera_create();
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
cameras = lf_db_find_cameras(db, lensfun->make, lensfun->model);
if (cameras && *cameras) {
lf_camera_copy(lensfun->camera, *cameras);
av_log(ctx, AV_LOG_INFO, "Using camera %s\n", lensfun->camera->Model);
} else {
lf_free(cameras);
lf_db_destroy(db);
av_log(ctx, AV_LOG_FATAL, "Failed to find camera in lensfun database\n");
return AVERROR_INVALIDDATA;
}
lf_free(cameras);
lenses = lf_db_find_lenses(db, lensfun->camera, NULL, lensfun->lens_model, 0);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
if (lenses && *lenses) {
lf_lens_copy(lensfun->lens, *lenses);
av_log(ctx, AV_LOG_INFO, "Using lens %s\n", lensfun->lens->Model);
} else {
lf_free(lenses);
lf_db_destroy(db);
av_log(ctx, AV_LOG_FATAL, "Failed to find lens in lensfun database\n");
return AVERROR_INVALIDDATA;
}
lf_free(lenses);
lf_db_destroy(db);
return 0;
}
static int query_formats(AVFilterContext *ctx)
{
// Some of the functions provided by lensfun require pixels in RGB format
static const enum AVPixelFormat fmts[] = {AV_PIX_FMT_RGB24, AV_PIX_FMT_NONE};
return ff_set_common_formats_from_list(ctx, fmts);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
}
static float lanczos_kernel(float x)
{
if (x == 0.0f) {
return 1.0f;
} else if (x > -2.0f && x < 2.0f) {
return (2.0f * sin(M_PI * x) * sin(M_PI / 2.0f * x)) / (M_PI * M_PI * x * x);
} else {
return 0.0f;
}
}
static int config_props(AVFilterLink *inlink)
{
AVFilterContext *ctx = inlink->dst;
LensfunContext *lensfun = ctx->priv;
int index;
float a;
if (!lensfun->modifier) {
if (lensfun->camera && lensfun->lens) {
lensfun->modifier = lf_modifier_create(lensfun->lens,
lensfun->focal_length,
lensfun->camera->CropFactor,
inlink->w,
inlink->h, LF_PF_U8, lensfun->reverse);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
if (lensfun->mode & VIGNETTING)
lf_modifier_enable_vignetting_correction(lensfun->modifier, lensfun->aperture, lensfun->focus_distance);
if (lensfun->mode & GEOMETRY_DISTORTION) {
lf_modifier_enable_distortion_correction(lensfun->modifier);
lf_modifier_enable_projection_transform(lensfun->modifier, lensfun->target_geometry);
lf_modifier_enable_scaling(lensfun->modifier, lensfun->scale);
}
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
if (lensfun->mode & SUBPIXEL_DISTORTION)
lf_modifier_enable_tca_correction(lensfun->modifier);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
} else {
// lensfun->camera and lensfun->lens should have been initialized
return AVERROR_BUG;
}
}
if (!lensfun->distortion_coords) {
if (lensfun->mode & SUBPIXEL_DISTORTION) {
lensfun->distortion_coords = av_malloc_array(inlink->w * inlink->h, sizeof(float) * 2 * 3);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
if (!lensfun->distortion_coords)
return AVERROR(ENOMEM);
if (lensfun->mode & GEOMETRY_DISTORTION) {
// apply both geometry and subpixel distortion
lf_modifier_apply_subpixel_geometry_distortion(lensfun->modifier,
0, 0,
inlink->w, inlink->h,
lensfun->distortion_coords);
} else {
// apply only subpixel distortion
lf_modifier_apply_subpixel_distortion(lensfun->modifier,
0, 0,
inlink->w, inlink->h,
lensfun->distortion_coords);
}
} else if (lensfun->mode & GEOMETRY_DISTORTION) {
lensfun->distortion_coords = av_malloc_array(inlink->w * inlink->h, sizeof(float) * 2);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
if (!lensfun->distortion_coords)
return AVERROR(ENOMEM);
// apply only geometry distortion
lf_modifier_apply_geometry_distortion(lensfun->modifier,
0, 0,
inlink->w, inlink->h,
lensfun->distortion_coords);
}
}
if (!lensfun->interpolation)
if (lensfun->interpolation_type == LANCZOS) {
lensfun->interpolation = av_malloc_array(LANCZOS_RESOLUTION, sizeof(float) * 4);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
if (!lensfun->interpolation)
return AVERROR(ENOMEM);
for (index = 0; index < 4 * LANCZOS_RESOLUTION; ++index) {
if (index == 0) {
lensfun->interpolation[index] = 1.0f;
} else {
a = sqrtf((float)index / LANCZOS_RESOLUTION);
lensfun->interpolation[index] = lanczos_kernel(a);
}
}
}
return 0;
}
static int vignetting_filter_slice(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
const VignettingThreadData *thread_data = arg;
const int slice_start = thread_data->height * jobnr / nb_jobs;
const int slice_end = thread_data->height * (jobnr + 1) / nb_jobs;
lf_modifier_apply_color_modification(thread_data->modifier,
thread_data->data_in + slice_start * thread_data->linesize_in,
0,
slice_start,
thread_data->width,
slice_end - slice_start,
thread_data->pixel_composition,
thread_data->linesize_in);
return 0;
}
static float square(float x)
{
return x * x;
}
static int distortion_correction_filter_slice(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
{
const DistortionCorrectionThreadData *thread_data = arg;
const int slice_start = thread_data->height * jobnr / nb_jobs;
const int slice_end = thread_data->height * (jobnr + 1) / nb_jobs;
int x, y, i, j, rgb_index;
float interpolated, new_x, new_y, d, norm;
int new_x_int, new_y_int;
for (y = slice_start; y < slice_end; ++y)
for (x = 0; x < thread_data->width; ++x)
for (rgb_index = 0; rgb_index < 3; ++rgb_index) {
if (thread_data->mode & SUBPIXEL_DISTORTION) {
// subpixel (and possibly geometry) distortion correction was applied, correct distortion
switch(thread_data->interpolation_type) {
case NEAREST:
new_x_int = thread_data->distortion_coords[x * 2 * 3 + y * thread_data->width * 2 * 3 + rgb_index * 2] + 0.5f;
new_y_int = thread_data->distortion_coords[x * 2 * 3 + y * thread_data->width * 2 * 3 + rgb_index * 2 + 1] + 0.5f;
if (new_x_int < 0 || new_x_int >= thread_data->width || new_y_int < 0 || new_y_int >= thread_data->height) {
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = 0;
} else {
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = thread_data->data_in[new_x_int * 3 + rgb_index + new_y_int * thread_data->linesize_in];
}
break;
case LINEAR:
interpolated = 0.0f;
new_x = thread_data->distortion_coords[x * 2 * 3 + y * thread_data->width * 2 * 3 + rgb_index * 2];
new_x_int = new_x;
new_y = thread_data->distortion_coords[x * 2 * 3 + y * thread_data->width * 2 * 3 + rgb_index * 2 + 1];
new_y_int = new_y;
if (new_x_int < 0 || new_x_int + 1 >= thread_data->width || new_y_int < 0 || new_y_int + 1 >= thread_data->height) {
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = 0;
} else {
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] =
thread_data->data_in[ new_x_int * 3 + rgb_index + new_y_int * thread_data->linesize_in] * (new_x_int + 1 - new_x) * (new_y_int + 1 - new_y)
+ thread_data->data_in[(new_x_int + 1) * 3 + rgb_index + new_y_int * thread_data->linesize_in] * (new_x - new_x_int) * (new_y_int + 1 - new_y)
+ thread_data->data_in[ new_x_int * 3 + rgb_index + (new_y_int + 1) * thread_data->linesize_in] * (new_x_int + 1 - new_x) * (new_y - new_y_int)
+ thread_data->data_in[(new_x_int + 1) * 3 + rgb_index + (new_y_int + 1) * thread_data->linesize_in] * (new_x - new_x_int) * (new_y - new_y_int);
}
break;
case LANCZOS:
interpolated = 0.0f;
norm = 0.0f;
new_x = thread_data->distortion_coords[x * 2 * 3 + y * thread_data->width * 2 * 3 + rgb_index * 2];
new_x_int = new_x;
new_y = thread_data->distortion_coords[x * 2 * 3 + y * thread_data->width * 2 * 3 + rgb_index * 2 + 1];
new_y_int = new_y;
for (j = 0; j < 4; ++j)
for (i = 0; i < 4; ++i) {
if (new_x_int + i - 2 < 0 || new_x_int + i - 2 >= thread_data->width || new_y_int + j - 2 < 0 || new_y_int + j - 2 >= thread_data->height)
continue;
d = square(new_x - (new_x_int + i - 2)) * square(new_y - (new_y_int + j - 2));
if (d >= 4.0f)
continue;
d = thread_data->interpolation[(int)(d * LANCZOS_RESOLUTION)];
norm += d;
interpolated += thread_data->data_in[(new_x_int + i - 2) * 3 + rgb_index + (new_y_int + j - 2) * thread_data->linesize_in] * d;
}
if (norm == 0.0f) {
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = 0;
} else {
interpolated /= norm;
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = interpolated < 0.0f ? 0.0f : interpolated > 255.0f ? 255.0f : interpolated;
}
break;
}
} else if (thread_data->mode & GEOMETRY_DISTORTION) {
// geometry distortion correction was applied, correct distortion
switch(thread_data->interpolation_type) {
case NEAREST:
new_x_int = thread_data->distortion_coords[x * 2 + y * thread_data->width * 2] + 0.5f;
new_y_int = thread_data->distortion_coords[x * 2 + y * thread_data->width * 2 + 1] + 0.5f;
if (new_x_int < 0 || new_x_int >= thread_data->width || new_y_int < 0 || new_y_int >= thread_data->height) {
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = 0;
} else {
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = thread_data->data_in[new_x_int * 3 + rgb_index + new_y_int * thread_data->linesize_in];
}
break;
case LINEAR:
interpolated = 0.0f;
new_x = thread_data->distortion_coords[x * 2 + y * thread_data->width * 2];
new_x_int = new_x;
new_y = thread_data->distortion_coords[x * 2 + y * thread_data->width * 2 + 1];
new_y_int = new_y;
if (new_x_int < 0 || new_x_int + 1 >= thread_data->width || new_y_int < 0 || new_y_int + 1 >= thread_data->height) {
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = 0;
} else {
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] =
thread_data->data_in[ new_x_int * 3 + rgb_index + new_y_int * thread_data->linesize_in] * (new_x_int + 1 - new_x) * (new_y_int + 1 - new_y)
+ thread_data->data_in[(new_x_int + 1) * 3 + rgb_index + new_y_int * thread_data->linesize_in] * (new_x - new_x_int) * (new_y_int + 1 - new_y)
+ thread_data->data_in[ new_x_int * 3 + rgb_index + (new_y_int + 1) * thread_data->linesize_in] * (new_x_int + 1 - new_x) * (new_y - new_y_int)
+ thread_data->data_in[(new_x_int + 1) * 3 + rgb_index + (new_y_int + 1) * thread_data->linesize_in] * (new_x - new_x_int) * (new_y - new_y_int);
}
break;
case LANCZOS:
interpolated = 0.0f;
norm = 0.0f;
new_x = thread_data->distortion_coords[x * 2 + y * thread_data->width * 2];
new_x_int = new_x;
new_y = thread_data->distortion_coords[x * 2 + 1 + y * thread_data->width * 2];
new_y_int = new_y;
for (j = 0; j < 4; ++j)
for (i = 0; i < 4; ++i) {
if (new_x_int + i - 2 < 0 || new_x_int + i - 2 >= thread_data->width || new_y_int + j - 2 < 0 || new_y_int + j - 2 >= thread_data->height)
continue;
d = square(new_x - (new_x_int + i - 2)) * square(new_y - (new_y_int + j - 2));
if (d >= 4.0f)
continue;
d = thread_data->interpolation[(int)(d * LANCZOS_RESOLUTION)];
norm += d;
interpolated += thread_data->data_in[(new_x_int + i - 2) * 3 + rgb_index + (new_y_int + j - 2) * thread_data->linesize_in] * d;
}
if (norm == 0.0f) {
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = 0;
} else {
interpolated /= norm;
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = interpolated < 0.0f ? 0.0f : interpolated > 255.0f ? 255.0f : interpolated;
}
break;
}
} else {
// no distortion correction was applied
thread_data->data_out[x * 3 + rgb_index + y * thread_data->linesize_out] = thread_data->data_in[x * 3 + rgb_index + y * thread_data->linesize_in];
}
}
return 0;
}
static int filter_frame(AVFilterLink *inlink, AVFrame *in)
{
AVFilterContext *ctx = inlink->dst;
LensfunContext *lensfun = ctx->priv;
AVFilterLink *outlink = ctx->outputs[0];
AVFrame *out;
VignettingThreadData vignetting_thread_data;
DistortionCorrectionThreadData distortion_correction_thread_data;
if (lensfun->mode & VIGNETTING) {
av_frame_make_writable(in);
vignetting_thread_data = (VignettingThreadData) {
.width = inlink->w,
.height = inlink->h,
.data_in = in->data[0],
.linesize_in = in->linesize[0],
.pixel_composition = LF_CR_3(RED, GREEN, BLUE),
.modifier = lensfun->modifier
};
ff_filter_execute(ctx, vignetting_filter_slice,
&vignetting_thread_data, NULL,
FFMIN(outlink->h, ff_filter_get_nb_threads(ctx)));
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
}
if (lensfun->mode & (GEOMETRY_DISTORTION | SUBPIXEL_DISTORTION)) {
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);
distortion_correction_thread_data = (DistortionCorrectionThreadData) {
.width = inlink->w,
.height = inlink->h,
.distortion_coords = lensfun->distortion_coords,
.data_in = in->data[0],
.data_out = out->data[0],
.linesize_in = in->linesize[0],
.linesize_out = out->linesize[0],
.interpolation = lensfun->interpolation,
.mode = lensfun->mode,
.interpolation_type = lensfun->interpolation_type
};
ff_filter_execute(ctx, distortion_correction_filter_slice,
&distortion_correction_thread_data, NULL,
FFMIN(outlink->h, ff_filter_get_nb_threads(ctx)));
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
av_frame_free(&in);
return ff_filter_frame(outlink, out);
} else {
return ff_filter_frame(outlink, in);
}
}
static av_cold void uninit(AVFilterContext *ctx)
{
LensfunContext *lensfun = ctx->priv;
if (lensfun->camera)
lf_camera_destroy(lensfun->camera);
if (lensfun->lens)
lf_lens_destroy(lensfun->lens);
if (lensfun->modifier)
lf_modifier_destroy(lensfun->modifier);
2018-10-18 11:30:11 +02:00
av_freep(&lensfun->distortion_coords);
av_freep(&lensfun->interpolation);
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
}
static const AVFilterPad lensfun_inputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_VIDEO,
.config_props = config_props,
.filter_frame = filter_frame,
},
};
static const AVFilterPad lensfun_outputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_VIDEO,
},
};
const AVFilter ff_vf_lensfun = {
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
.name = "lensfun",
.description = NULL_IF_CONFIG_SMALL("Apply correction to an image based on info derived from the lensfun database."),
.priv_size = sizeof(LensfunContext),
.init = init,
.uninit = uninit,
2021-08-12 13:05:31 +02:00
FILTER_INPUTS(lensfun_inputs),
FILTER_OUTPUTS(lensfun_outputs),
avfilter: Replace query_formats callback with union of list and callback If one looks at the many query_formats callbacks in existence, one will immediately recognize that there is one type of default callback for video and a slightly different default callback for audio: It is "return ff_set_common_formats_from_list(ctx, pix_fmts);" for video with a filter-specific pix_fmts list. For audio, it is the same with a filter-specific sample_fmts list together with ff_set_common_all_samplerates() and ff_set_common_all_channel_counts(). This commit allows to remove the boilerplate query_formats callbacks by replacing said callback with a union consisting the old callback and pointers for pixel and sample format arrays. For the not uncommon case in which these lists only contain a single entry (besides the sentinel) enum AVPixelFormat and enum AVSampleFormat fields are also added to the union to store them directly in the AVFilter, thereby avoiding a relocation. The state of said union will be contained in a new, dedicated AVFilter field (the nb_inputs and nb_outputs fields have been shrunk to uint8_t in order to create a hole for this new field; this is no problem, as the maximum of all the nb_inputs is four; for nb_outputs it is only two). The state's default value coincides with the earlier default of query_formats being unset, namely that the filter accepts all formats (and also sample rates and channel counts/layouts for audio) provided that these properties agree coincide for all inputs and outputs. By using different union members for audio and video filters the type-unsafety of using the same functions for audio and video lists will furthermore be more confined to formats.c than before. When the new fields are used, they will also avoid allocations: Currently something nearly equivalent to ff_default_query_formats() is called after every successful call to a query_formats callback; yet in the common case that the newly allocated AVFilterFormats are not used at all (namely if there are no free links) these newly allocated AVFilterFormats are freed again without ever being used. Filters no longer using the callback will not exhibit this any more. Reviewed-by: Paul B Mahol <onemda@gmail.com> Reviewed-by: Nicolas George <george@nsup.org> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2021-09-27 12:07:35 +02:00
FILTER_QUERY_FUNC(query_formats),
Add lensfun filter Lensfun is a library that applies lens correction to an image using a database of cameras/lenses (you provide the camera and lens models, and it uses the corresponding database entry's parameters to apply lens correction). It is licensed under LGPL3. The lensfun filter utilizes the lensfun library to apply lens correction to videos as well as images. This filter was created out of necessity since I wanted to apply lens correction to a video and the lenscorrection filter did not work for me. While this filter requires little info from the user to apply lens correction, the flaw is that lensfun is intended to be used on indvidual images. When used on a video, the parameters such as focal length is constant, so lens correction may fail on videos where the camera's focal length changes (zooming in or out via zoom lens). To use this filter correctly on videos where such parameters change, timeline editing may be used since this filter supports it. Note that valgrind shows a small memory leak which is not from this filter but from the lensfun library (memory is allocated when loading the lensfun database but it somehow isn't deallocated even during cleanup; it is briefly created in the init function of the filter, and destroyed before the init function returns). This may have been fixed by the latest commit in the lensfun repository; the current latest release of lensfun is almost 3 years ago. Bi-Linear interpolation is used by default as lanczos interpolation shows more artifacts in the corrected image in my tests. The lanczos interpolation is derived from lenstool's implementation of lanczos interpolation. Lenstool is an app within the lensfun repository which is licensed under GPL3. v2 of this patch fixes license notice in libavfilter/vf_lensfun.c v3 of this patch fixes code style and dependency to gplv3 (thanks to Paul B Mahol for pointing out the mentioned issues). v4 of this patch fixes more code style issues that were missed in v3. v5 of this patch adds line breaks to some of the documentation in doc/filters.texi (thanks to Gyan Doshi for pointing out the issue). v6 of this patch fixes more problems (thanks to Moritz Barsnick for pointing them out). v7 of this patch fixes use of sqrt() (changed to sqrtf(); thanks to Moritz Barsnick for pointing this out). Also should be rebased off of latest master branch commits at this point. Signed-off-by: Stephen Seo <seo.disparate@gmail.com>
2018-07-13 12:33:12 +02:00
.priv_class = &lensfun_class,
.flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC | AVFILTER_FLAG_SLICE_THREADS,
};