vidstabdetect and vidstabtransform common functions for interfacing
vid.stab are in libavfilter/vidstabutils.c
Signed-off-by: Georg Martius <martius@mis.mpg.de>
... | ... |
@@ -226,6 +226,7 @@ External library support: |
226 | 226 |
--enable-libtwolame enable MP2 encoding via libtwolame [no] |
227 | 227 |
--enable-libutvideo enable Ut Video encoding and decoding via libutvideo [no] |
228 | 228 |
--enable-libv4l2 enable libv4l2/v4l-utils [no] |
229 |
+ --enable-libvidstab enable video stabilization using vid.stab [no] |
|
229 | 230 |
--enable-libvo-aacenc enable AAC encoding via libvo-aacenc [no] |
230 | 231 |
--enable-libvo-amrwbenc enable AMR-WB encoding via libvo-amrwbenc [no] |
231 | 232 |
--enable-libvorbis enable Vorbis en/decoding via libvorbis, |
... | ... |
@@ -1181,6 +1182,7 @@ EXTERNAL_LIBRARY_LIST=" |
1181 | 1181 |
libtwolame |
1182 | 1182 |
libutvideo |
1183 | 1183 |
libv4l2 |
1184 |
+ libvidstab |
|
1184 | 1185 |
libvo_aacenc |
1185 | 1186 |
libvo_amrwbenc |
1186 | 1187 |
libvorbis |
... | ... |
@@ -2152,6 +2154,8 @@ stereo3d_filter_deps="gpl" |
2152 | 2152 |
subtitles_filter_deps="avformat avcodec libass" |
2153 | 2153 |
super2xsai_filter_deps="gpl" |
2154 | 2154 |
tinterlace_filter_deps="gpl" |
2155 |
+vidstabdetect_filter_deps="libvidstab" |
|
2156 |
+vidstabtransform_filter_deps="libvidstab" |
|
2155 | 2157 |
yadif_filter_deps="gpl" |
2156 | 2158 |
pixfmts_super2xsai_test_deps="super2xsai_filter" |
2157 | 2159 |
tinterlace_merge_test_deps="tinterlace_filter" |
... | ... |
@@ -3573,6 +3577,7 @@ die_license_disabled_gpl() { |
3573 | 3573 |
|
3574 | 3574 |
die_license_disabled gpl libcdio |
3575 | 3575 |
die_license_disabled gpl libutvideo |
3576 |
+die_license_disabled gpl libvidstab |
|
3576 | 3577 |
die_license_disabled gpl libx264 |
3577 | 3578 |
die_license_disabled gpl libxavs |
3578 | 3579 |
die_license_disabled gpl libxvid |
... | ... |
@@ -4015,6 +4020,7 @@ enabled libtwolame && require libtwolame twolame.h twolame_init -ltwolame && |
4015 | 4015 |
die "ERROR: libtwolame must be installed and version must be >= 0.3.10"; } |
4016 | 4016 |
enabled libutvideo && require_cpp utvideo "stdint.h stdlib.h utvideo/utvideo.h utvideo/Codec.h" 'CCodec*' -lutvideo -lstdc++ |
4017 | 4017 |
enabled libv4l2 && require_pkg_config libv4l2 libv4l2.h v4l2_ioctl |
4018 |
+enabled libvidstab && require_pkg_config vidstab vid.stab/libvidstab.h vsMotionDetectInit |
|
4018 | 4019 |
enabled libvo_aacenc && require libvo_aacenc vo-aacenc/voAAC.h voGetAACEncAPI -lvo-aacenc |
4019 | 4020 |
enabled libvo_amrwbenc && require libvo_amrwbenc vo-amrwbenc/enc_if.h E_IF_init -lvo-amrwbenc |
4020 | 4021 |
enabled libvorbis && require libvorbis vorbis/vorbisenc.h vorbis_info_init -lvorbisenc -lvorbis -logg |
... | ... |
@@ -5308,6 +5308,141 @@ much, but it will increase the amount of blurring needed to cover over |
5308 | 5308 |
the image and will destroy more information than necessary, and extra |
5309 | 5309 |
pixels will slow things down on a large logo. |
5310 | 5310 |
|
5311 |
+@anchor{vidstabtransform} |
|
5312 |
+@section vidstabtransform |
|
5313 |
+ |
|
5314 |
+Video stabilization/deshaking: pass 2 of 2, |
|
5315 |
+see @ref{vidstabdetect} for pass 1. |
|
5316 |
+ |
|
5317 |
+Read a file with transform information for each frame and |
|
5318 |
+apply/compensate them. Together with the @ref{vidstabdetect} |
|
5319 |
+filter this can be used to deshake videos. See also |
|
5320 |
+@url{http://public.hronopik.de/vid.stab}. It is important to also use |
|
5321 |
+the unsharp filter, see below. |
|
5322 |
+ |
|
5323 |
+To enable compilation of this filter you need to configure FFmpeg with |
|
5324 |
+@code{--enable-libvidstab}. |
|
5325 |
+ |
|
5326 |
+This filter accepts the following named options, expressed as a |
|
5327 |
+sequence of @var{key}=@var{value} pairs, separated by ":". |
|
5328 |
+ |
|
5329 |
+@table @option |
|
5330 |
+ |
|
5331 |
+@item input |
|
5332 |
+path to the file used to read the transforms (default: @file{transforms.trf}) |
|
5333 |
+ |
|
5334 |
+@item smoothing |
|
5335 |
+number of frames (value*2 + 1) used for lowpass filtering the camera movements |
|
5336 |
+(default: 10). For example a number of 10 means that 21 frames are used |
|
5337 |
+(10 in the past and 10 in the future) to smoothen the motion in the |
|
5338 |
+video. A larger values leads to a smoother video, but limits the |
|
5339 |
+acceleration of the camera (pan/tilt movements). |
|
5340 |
+ |
|
5341 |
+@item maxshift |
|
5342 |
+maximal number of pixels to translate frames (default: -1 no limit) |
|
5343 |
+ |
|
5344 |
+@item maxangle |
|
5345 |
+maximal angle in radians (degree*PI/180) to rotate frames (default: -1 |
|
5346 |
+no limit) |
|
5347 |
+ |
|
5348 |
+@item crop |
|
5349 |
+How to deal with borders that may be visible due to movement |
|
5350 |
+compensation. Available values are: |
|
5351 |
+ |
|
5352 |
+@table @samp |
|
5353 |
+@item keep |
|
5354 |
+keep image information from previous frame (default) |
|
5355 |
+@item black |
|
5356 |
+fill the border black |
|
5357 |
+@end table |
|
5358 |
+ |
|
5359 |
+@item invert |
|
5360 |
+@table @samp |
|
5361 |
+@item 0 |
|
5362 |
+ keep transforms normal (default) |
|
5363 |
+@item 1 |
|
5364 |
+ invert transforms |
|
5365 |
+@end table |
|
5366 |
+ |
|
5367 |
+ |
|
5368 |
+@item relative |
|
5369 |
+consider transforms as |
|
5370 |
+@table @samp |
|
5371 |
+@item 0 |
|
5372 |
+ absolute |
|
5373 |
+@item 1 |
|
5374 |
+ relative to previous frame (default) |
|
5375 |
+@end table |
|
5376 |
+ |
|
5377 |
+ |
|
5378 |
+@item zoom |
|
5379 |
+percentage to zoom (default: 0) |
|
5380 |
+@table @samp |
|
5381 |
+@item >0 |
|
5382 |
+ zoom in |
|
5383 |
+@item <0 |
|
5384 |
+ zoom out |
|
5385 |
+@end table |
|
5386 |
+ |
|
5387 |
+@item optzoom |
|
5388 |
+if 1 then optimal zoom value is determined (default). |
|
5389 |
+Optimal zoom means no (or only little) border should be visible. |
|
5390 |
+Note that the value given at zoom is added to the one calculated |
|
5391 |
+here. |
|
5392 |
+ |
|
5393 |
+@item interpol |
|
5394 |
+type of interpolation |
|
5395 |
+ |
|
5396 |
+Available values are: |
|
5397 |
+@table @samp |
|
5398 |
+@item no |
|
5399 |
+no interpolation |
|
5400 |
+@item linear |
|
5401 |
+linear only horizontal |
|
5402 |
+@item bilinear |
|
5403 |
+linear in both directions (default) |
|
5404 |
+@item bicubic |
|
5405 |
+cubic in both directions (slow) |
|
5406 |
+@end table |
|
5407 |
+ |
|
5408 |
+@item tripod |
|
5409 |
+virtual tripod mode means that the video is stabilized such that the |
|
5410 |
+camera stays stationary. Use also @code{tripod} option of |
|
5411 |
+@ref{vidstabdetect}. |
|
5412 |
+@table @samp |
|
5413 |
+@item 0 |
|
5414 |
+off (default) |
|
5415 |
+@item 1 |
|
5416 |
+virtual tripod mode: equivalent to @code{relative=0:smoothing=0} |
|
5417 |
+@end table |
|
5418 |
+ |
|
5419 |
+@end table |
|
5420 |
+ |
|
5421 |
+@subsection Examples |
|
5422 |
+ |
|
5423 |
+@itemize |
|
5424 |
+@item |
|
5425 |
+typical call with default default values: |
|
5426 |
+ (note the unsharp filter which is always recommended) |
|
5427 |
+@example |
|
5428 |
+ffmpeg -i inp.mpeg -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 inp_stabilized.mpeg |
|
5429 |
+@end example |
|
5430 |
+ |
|
5431 |
+@item |
|
5432 |
+zoom in a bit more and load transform data from a given file |
|
5433 |
+@example |
|
5434 |
+vidstabtransform=zoom=5:input="mytransforms.trf" |
|
5435 |
+@end example |
|
5436 |
+ |
|
5437 |
+@item |
|
5438 |
+smoothen the video even more |
|
5439 |
+@example |
|
5440 |
+vidstabtransform=smoothing=30 |
|
5441 |
+@end example |
|
5442 |
+ |
|
5443 |
+@end itemize |
|
5444 |
+ |
|
5445 |
+ |
|
5311 | 5446 |
@section scale |
5312 | 5447 |
|
5313 | 5448 |
Scale (resize) the input video, using the libswscale library. |
... | ... |
@@ -5706,6 +5841,93 @@ in [-30,0] will filter edges. Default value is 0. |
5706 | 5706 |
If a chroma option is not explicitly set, the corresponding luma value |
5707 | 5707 |
is set. |
5708 | 5708 |
|
5709 |
+@anchor{vidstabdetect} |
|
5710 |
+@section vidstabdetect |
|
5711 |
+ |
|
5712 |
+Video stabilization/deshaking: pass 1 of 2, see @ref{vidstabtransform} |
|
5713 |
+for pass 2. |
|
5714 |
+ |
|
5715 |
+Generates a file with relative transform information translation, |
|
5716 |
+rotation about subsequent frames. |
|
5717 |
+ |
|
5718 |
+To enable compilation of this filter you need to configure FFmpeg with |
|
5719 |
+@code{--enable-libvidstab}. |
|
5720 |
+ |
|
5721 |
+This filter accepts the following named options, expressed as a |
|
5722 |
+sequence of @var{key}=@var{value} pairs, separated by ":". |
|
5723 |
+ |
|
5724 |
+@table @option |
|
5725 |
+@item result |
|
5726 |
+path to the file used to write the transforms (default:@file{transforms.trf}) |
|
5727 |
+ |
|
5728 |
+@item shakiness |
|
5729 |
+how shaky is the video and how quick is the camera? (default: 5) |
|
5730 |
+@table @samp |
|
5731 |
+@item 1 |
|
5732 |
+ little (fast) |
|
5733 |
+@item ... |
|
5734 |
+@item 10 |
|
5735 |
+ very strong/quick (slow) |
|
5736 |
+@end table |
|
5737 |
+ |
|
5738 |
+@item accuracy |
|
5739 |
+accuracy of detection process (>=shakiness) (default: 9) |
|
5740 |
+@table @samp |
|
5741 |
+@item 1 |
|
5742 |
+ low (fast) |
|
5743 |
+@item 15 |
|
5744 |
+ high (slow) |
|
5745 |
+@end table |
|
5746 |
+ |
|
5747 |
+@item stepsize |
|
5748 |
+stepsize of search process, region around minimum is scanned with 1 pixel |
|
5749 |
+resolution (default: 6) |
|
5750 |
+ |
|
5751 |
+@item mincontrast |
|
5752 |
+below this contrast a local measurement field is discarded (0-1) (default: 0.3) |
|
5753 |
+ |
|
5754 |
+@item tripod |
|
5755 |
+virtual tripod mode: @code{tripod=framenum} if framenum>0 otherwise disabled. |
|
5756 |
+The motion of the frames is compared to a reference frame (framenum). |
|
5757 |
+The idea is to compensate all movements in a more-or-less static scene |
|
5758 |
+ and keep the camera view absolutely still. |
|
5759 |
+(default: 0 (disabled)) |
|
5760 |
+ |
|
5761 |
+@item show |
|
5762 |
+draw nothing (default); 1,2: show fields and transforms in the resulting frames |
|
5763 |
+ |
|
5764 |
+@end table |
|
5765 |
+ |
|
5766 |
+@subsection Examples |
|
5767 |
+ |
|
5768 |
+@itemize |
|
5769 |
+@item |
|
5770 |
+use default values: |
|
5771 |
+@example |
|
5772 |
+vidstabdetect |
|
5773 |
+@end example |
|
5774 |
+ |
|
5775 |
+@item |
|
5776 |
+strongly shaky movie and put the results in @code{mytransforms.trf} |
|
5777 |
+@example |
|
5778 |
+vidstabdetect=shakiness=10:accuracy=15:result="mytransforms.trf" |
|
5779 |
+@end example |
|
5780 |
+ |
|
5781 |
+@item |
|
5782 |
+visualize some internals in the resulting video |
|
5783 |
+@example |
|
5784 |
+vidstabdetect=show=1 |
|
5785 |
+@end example |
|
5786 |
+ |
|
5787 |
+ |
|
5788 |
+@item |
|
5789 |
+Typical call with visualization |
|
5790 |
+@example |
|
5791 |
+ffmpeg -i input -vf vidstabdetect=shakiness=5:show=1 dummy.avi |
|
5792 |
+@end example |
|
5793 |
+@end itemize |
|
5794 |
+ |
|
5795 |
+ |
|
5709 | 5796 |
@section stereo3d |
5710 | 5797 |
|
5711 | 5798 |
Convert between different stereoscopic image formats. |
... | ... |
@@ -179,6 +179,8 @@ OBJS-$(CONFIG_TINTERLACE_FILTER) += vf_tinterlace.o |
179 | 179 |
OBJS-$(CONFIG_TRANSPOSE_FILTER) += vf_transpose.o |
180 | 180 |
OBJS-$(CONFIG_UNSHARP_FILTER) += vf_unsharp.o |
181 | 181 |
OBJS-$(CONFIG_VFLIP_FILTER) += vf_vflip.o |
182 |
+OBJS-$(CONFIG_VIDSTABDETECT_FILTER) += vidstabutils.o vf_vidstabdetect.o |
|
183 |
+OBJS-$(CONFIG_VIDSTABTRANSFORM_FILTER) += vidstabutils.o vf_vidstabtransform.o |
|
182 | 184 |
OBJS-$(CONFIG_YADIF_FILTER) += vf_yadif.o |
183 | 185 |
|
184 | 186 |
OBJS-$(CONFIG_CELLAUTO_FILTER) += vsrc_cellauto.o |
... | ... |
@@ -176,6 +176,8 @@ void avfilter_register_all(void) |
176 | 176 |
REGISTER_FILTER(TRANSPOSE, transpose, vf); |
177 | 177 |
REGISTER_FILTER(UNSHARP, unsharp, vf); |
178 | 178 |
REGISTER_FILTER(VFLIP, vflip, vf); |
179 |
+ REGISTER_FILTER(VIDSTABDETECT, vidstabdetect, vf); |
|
180 |
+ REGISTER_FILTER(VIDSTABTRANSFORM, vidstabtransform, vf); |
|
179 | 181 |
REGISTER_FILTER(YADIF, yadif, vf); |
180 | 182 |
|
181 | 183 |
REGISTER_FILTER(CELLAUTO, cellauto, vsrc); |
182 | 184 |
new file mode 100644 |
... | ... |
@@ -0,0 +1,239 @@ |
0 |
+/* |
|
1 |
+ * Copyright (c) 2013 Georg Martius <georg dot martius at web dot de> |
|
2 |
+ * |
|
3 |
+ * This file is part of FFmpeg. |
|
4 |
+ * |
|
5 |
+ * FFmpeg is free software; you can redistribute it and/or |
|
6 |
+ * modify it under the terms of the GNU Lesser General Public |
|
7 |
+ * License as published by the Free Software Foundation; either |
|
8 |
+ * version 2.1 of the License, or (at your option) any later version. |
|
9 |
+ * |
|
10 |
+ * FFmpeg is distributed in the hope that it will be useful, |
|
11 |
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
12 |
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
|
13 |
+ * Lesser General Public License for more details. |
|
14 |
+ * |
|
15 |
+ * You should have received a copy of the GNU Lesser General Public |
|
16 |
+ * License along with FFmpeg; if not, write to the Free Software |
|
17 |
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
18 |
+ */ |
|
19 |
+ |
|
20 |
+#define DEFAULT_RESULT_NAME "transforms.trf" |
|
21 |
+ |
|
22 |
+#include <vid.stab/libvidstab.h> |
|
23 |
+ |
|
24 |
+#include "libavutil/common.h" |
|
25 |
+#include "libavutil/opt.h" |
|
26 |
+#include "libavutil/imgutils.h" |
|
27 |
+#include "avfilter.h" |
|
28 |
+#include "internal.h" |
|
29 |
+ |
|
30 |
+#include "vidstabutils.h" |
|
31 |
+ |
|
32 |
+typedef struct { |
|
33 |
+ const AVClass* class; |
|
34 |
+ |
|
35 |
+ VSMotionDetect md; |
|
36 |
+ VSMotionDetectConfig conf; |
|
37 |
+ |
|
38 |
+ char* result; |
|
39 |
+ FILE* f; |
|
40 |
+} StabData; |
|
41 |
+ |
|
42 |
+ |
|
43 |
+#define OFFSET(x) offsetof(StabData, x) |
|
44 |
+#define OFFSETC(x) (offsetof(StabData, conf)+offsetof(VSMotionDetectConfig, x)) |
|
45 |
+#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM |
|
46 |
+ |
|
47 |
+static const AVOption vidstabdetect_options[]= { |
|
48 |
+ {"result", "path to the file used to write the transforms (def:transforms.trf)", OFFSET(result), AV_OPT_TYPE_STRING, {.str = DEFAULT_RESULT_NAME}}, |
|
49 |
+ {"shakiness", "how shaky is the video and how quick is the camera?" |
|
50 |
+ " 1: little (fast) 10: very strong/quick (slow) (def: 5)", OFFSETC(shakiness), AV_OPT_TYPE_INT, {.i64 = 5}, 1, 10, FLAGS}, |
|
51 |
+ {"accuracy", "(>=shakiness) 1: low 15: high (slow) (def: 9)", OFFSETC(accuracy), AV_OPT_TYPE_INT, {.i64 = 9 }, 1, 15, FLAGS}, |
|
52 |
+ {"stepsize", "region around minimum is scanned with 1 pixel resolution (def: 6)", OFFSETC(stepSize), AV_OPT_TYPE_INT, {.i64 = 6}, 1, 32, FLAGS}, |
|
53 |
+ {"mincontrast", "below this contrast a field is discarded (0-1) (def: 0.3)", OFFSETC(contrastThreshold), AV_OPT_TYPE_DOUBLE, {.dbl = 0.25}, 0.0, 1.0, FLAGS}, |
|
54 |
+ {"show", "0: draw nothing (def); 1,2: show fields and transforms", OFFSETC(show), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 2, FLAGS}, |
|
55 |
+ {"tripod", "virtual tripod mode (if >0): motion is compared to a reference" |
|
56 |
+ " reference frame (frame # is the value) (def: 0)", OFFSETC(virtualTripod), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, FLAGS}, |
|
57 |
+ {NULL}, |
|
58 |
+}; |
|
59 |
+ |
|
60 |
+AVFILTER_DEFINE_CLASS(vidstabdetect); |
|
61 |
+ |
|
62 |
+static av_cold int init(AVFilterContext *ctx) |
|
63 |
+{ |
|
64 |
+ StabData* sd = ctx->priv; |
|
65 |
+ vs_set_mem_and_log_functions(); |
|
66 |
+ sd->class = &vidstabdetect_class; |
|
67 |
+ av_log(ctx, AV_LOG_VERBOSE, "vidstabdetect filter: init %s\n", LIBVIDSTAB_VERSION); |
|
68 |
+ return 0; |
|
69 |
+} |
|
70 |
+ |
|
71 |
+static av_cold void uninit(AVFilterContext *ctx) |
|
72 |
+{ |
|
73 |
+ StabData *sd = ctx->priv; |
|
74 |
+ VSMotionDetect* md = &(sd->md); |
|
75 |
+ |
|
76 |
+ if (sd->f) { |
|
77 |
+ fclose(sd->f); |
|
78 |
+ sd->f = NULL; |
|
79 |
+ } |
|
80 |
+ |
|
81 |
+ vsMotionDetectionCleanup(md); |
|
82 |
+ |
|
83 |
+} |
|
84 |
+ |
|
85 |
+static int query_formats(AVFilterContext *ctx) |
|
86 |
+{ |
|
87 |
+ // If you add something here also add it in vidstabutils.c |
|
88 |
+ static const enum AVPixelFormat pix_fmts[] = { |
|
89 |
+ AV_PIX_FMT_YUV444P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV420P, |
|
90 |
+ AV_PIX_FMT_YUV411P, AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUVA420P, |
|
91 |
+ AV_PIX_FMT_YUV440P, AV_PIX_FMT_GRAY8, |
|
92 |
+ AV_PIX_FMT_RGB24, AV_PIX_FMT_BGR24, AV_PIX_FMT_RGBA, |
|
93 |
+ AV_PIX_FMT_NONE |
|
94 |
+ }; |
|
95 |
+ |
|
96 |
+ ff_set_common_formats(ctx, ff_make_format_list(pix_fmts)); |
|
97 |
+ return 0; |
|
98 |
+} |
|
99 |
+ |
|
100 |
+ |
|
101 |
+static int config_input(AVFilterLink *inlink) |
|
102 |
+{ |
|
103 |
+ AVFilterContext *ctx = inlink->dst; |
|
104 |
+ StabData *sd = ctx->priv; |
|
105 |
+ |
|
106 |
+ VSMotionDetect* md = &(sd->md); |
|
107 |
+ VSFrameInfo fi; |
|
108 |
+ const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format); |
|
109 |
+ |
|
110 |
+ vsFrameInfoInit(&fi,inlink->w, inlink->h, av_2_vs_pixel_format(ctx, inlink->format)); |
|
111 |
+ if(fi.bytesPerPixel != av_get_bits_per_pixel(desc)/8){ |
|
112 |
+ av_log(ctx, AV_LOG_ERROR, "pixel-format error: wrong bits/per/pixel, please report a BUG"); |
|
113 |
+ return AVERROR(EINVAL); |
|
114 |
+ } |
|
115 |
+ if(fi.log2ChromaW != desc->log2_chroma_w){ |
|
116 |
+ av_log(ctx, AV_LOG_ERROR, "pixel-format error: log2_chroma_w, please report a BUG"); |
|
117 |
+ return AVERROR(EINVAL); |
|
118 |
+ } |
|
119 |
+ |
|
120 |
+ if(fi.log2ChromaH != desc->log2_chroma_h){ |
|
121 |
+ av_log(ctx, AV_LOG_ERROR, "pixel-format error: log2_chroma_h, please report a BUG"); |
|
122 |
+ return AVERROR(EINVAL); |
|
123 |
+ } |
|
124 |
+ |
|
125 |
+ // set values that are not initializes by the options |
|
126 |
+ sd->conf.algo = 1; |
|
127 |
+ sd->conf.modName = "vidstabdetect"; |
|
128 |
+ if(vsMotionDetectInit(md, &sd->conf, &fi) != VS_OK){ |
|
129 |
+ av_log(ctx, AV_LOG_ERROR, "initialization of Motion Detection failed, please report a BUG"); |
|
130 |
+ return AVERROR(EINVAL); |
|
131 |
+ } |
|
132 |
+ |
|
133 |
+ vsMotionDetectGetConfig(&sd->conf, md); |
|
134 |
+ av_log(ctx, AV_LOG_INFO, "Video stabilization settings (pass 1/2):\n"); |
|
135 |
+ av_log(ctx, AV_LOG_INFO, " shakiness = %d\n", sd->conf.shakiness); |
|
136 |
+ av_log(ctx, AV_LOG_INFO, " accuracy = %d\n", sd->conf.accuracy); |
|
137 |
+ av_log(ctx, AV_LOG_INFO, " stepsize = %d\n", sd->conf.stepSize); |
|
138 |
+ av_log(ctx, AV_LOG_INFO, " mincontrast = %f\n", sd->conf.contrastThreshold); |
|
139 |
+ av_log(ctx, AV_LOG_INFO, " show = %d\n", sd->conf.show); |
|
140 |
+ av_log(ctx, AV_LOG_INFO, " result = %s\n", sd->result); |
|
141 |
+ |
|
142 |
+ sd->f = fopen(sd->result, "w"); |
|
143 |
+ if (sd->f == NULL) { |
|
144 |
+ av_log(ctx, AV_LOG_ERROR, "cannot open transform file %s\n", sd->result); |
|
145 |
+ return AVERROR(EINVAL); |
|
146 |
+ }else{ |
|
147 |
+ if(vsPrepareFile(md, sd->f) != VS_OK){ |
|
148 |
+ av_log(ctx, AV_LOG_ERROR, "cannot write to transform file %s\n", sd->result); |
|
149 |
+ return AVERROR(EINVAL); |
|
150 |
+ } |
|
151 |
+ } |
|
152 |
+ return 0; |
|
153 |
+} |
|
154 |
+ |
|
155 |
+ |
|
156 |
+static int filter_frame(AVFilterLink *inlink, AVFrame *in) |
|
157 |
+{ |
|
158 |
+ AVFilterContext *ctx = inlink->dst; |
|
159 |
+ StabData *sd = ctx->priv; |
|
160 |
+ VSMotionDetect* md = &(sd->md); |
|
161 |
+ LocalMotions localmotions; |
|
162 |
+ |
|
163 |
+ AVFilterLink *outlink = inlink->dst->outputs[0]; |
|
164 |
+ int direct = 0; |
|
165 |
+ AVFrame *out; |
|
166 |
+ VSFrame frame; |
|
167 |
+ int plane; |
|
168 |
+ |
|
169 |
+ if (av_frame_is_writable(in)) { |
|
170 |
+ direct = 1; |
|
171 |
+ out = in; |
|
172 |
+ } else { |
|
173 |
+ out = ff_get_video_buffer(outlink, outlink->w, outlink->h); |
|
174 |
+ if (!out) { |
|
175 |
+ av_frame_free(&in); |
|
176 |
+ return AVERROR(ENOMEM); |
|
177 |
+ } |
|
178 |
+ av_frame_copy_props(out, in); |
|
179 |
+ } |
|
180 |
+ |
|
181 |
+ for(plane=0; plane < md->fi.planes; plane++){ |
|
182 |
+ frame.data[plane] = in->data[plane]; |
|
183 |
+ frame.linesize[plane] = in->linesize[plane]; |
|
184 |
+ } |
|
185 |
+ if(vsMotionDetection(md, &localmotions, &frame) != VS_OK){ |
|
186 |
+ av_log(ctx, AV_LOG_ERROR, "motion detection failed"); |
|
187 |
+ return AVERROR(AVERROR_EXTERNAL); |
|
188 |
+ } else { |
|
189 |
+ if(vsWriteToFile(md, sd->f, &localmotions) != VS_OK){ |
|
190 |
+ av_log(ctx, AV_LOG_ERROR, "cannot write to transform file"); |
|
191 |
+ return AVERROR(errno); |
|
192 |
+ } |
|
193 |
+ vs_vector_del(&localmotions); |
|
194 |
+ } |
|
195 |
+ if(sd->conf.show>0 && !direct){ |
|
196 |
+ av_image_copy(out->data, out->linesize, |
|
197 |
+ (void*)in->data, in->linesize, |
|
198 |
+ in->format, in->width, in->height); |
|
199 |
+ } |
|
200 |
+ |
|
201 |
+ if (!direct) |
|
202 |
+ av_frame_free(&in); |
|
203 |
+ |
|
204 |
+ return ff_filter_frame(outlink, out); |
|
205 |
+} |
|
206 |
+ |
|
207 |
+static const AVFilterPad avfilter_vf_vidstabdetect_inputs[] = { |
|
208 |
+ { |
|
209 |
+ .name = "default", |
|
210 |
+ .type = AVMEDIA_TYPE_VIDEO, |
|
211 |
+ .filter_frame = filter_frame, |
|
212 |
+ .config_props = config_input, |
|
213 |
+ }, |
|
214 |
+ { NULL } |
|
215 |
+}; |
|
216 |
+ |
|
217 |
+static const AVFilterPad avfilter_vf_vidstabdetect_outputs[] = { |
|
218 |
+ { |
|
219 |
+ .name = "default", |
|
220 |
+ .type = AVMEDIA_TYPE_VIDEO, |
|
221 |
+ }, |
|
222 |
+ { NULL } |
|
223 |
+}; |
|
224 |
+ |
|
225 |
+AVFilter avfilter_vf_vidstabdetect = { |
|
226 |
+ .name = "vidstabdetect", |
|
227 |
+ .description = NULL_IF_CONFIG_SMALL("pass 1 of 2 for stabilization" |
|
228 |
+ "extracts relative transformations" |
|
229 |
+ "(pass 2 see vidstabtransform)"), |
|
230 |
+ .priv_size = sizeof(StabData), |
|
231 |
+ .init = init, |
|
232 |
+ .uninit = uninit, |
|
233 |
+ .query_formats = query_formats, |
|
234 |
+ |
|
235 |
+ .inputs = avfilter_vf_vidstabdetect_inputs, |
|
236 |
+ .outputs = avfilter_vf_vidstabdetect_outputs, |
|
237 |
+ .priv_class = &vidstabdetect_class, |
|
238 |
+}; |
0 | 239 |
new file mode 100644 |
... | ... |
@@ -0,0 +1,294 @@ |
0 |
+/* |
|
1 |
+ * Copyright (c) 2013 Georg Martius <georg dot martius at web dot de> |
|
2 |
+ * |
|
3 |
+ * This file is part of FFmpeg. |
|
4 |
+ * |
|
5 |
+ * FFmpeg is free software; you can redistribute it and/or |
|
6 |
+ * modify it under the terms of the GNU Lesser General Public |
|
7 |
+ * License as published by the Free Software Foundation; either |
|
8 |
+ * version 2.1 of the License, or (at your option) any later version. |
|
9 |
+ * |
|
10 |
+ * FFmpeg is distributed in the hope that it will be useful, |
|
11 |
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
12 |
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
|
13 |
+ * Lesser General Public License for more details. |
|
14 |
+ * |
|
15 |
+ * You should have received a copy of the GNU Lesser General Public |
|
16 |
+ * License along with FFmpeg; if not, write to the Free Software |
|
17 |
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
18 |
+ */ |
|
19 |
+ |
|
20 |
+#define DEFAULT_INPUT_NAME "transforms.trf" |
|
21 |
+ |
|
22 |
+#include <vid.stab/libvidstab.h> |
|
23 |
+ |
|
24 |
+#include "libavutil/common.h" |
|
25 |
+#include "libavutil/opt.h" |
|
26 |
+#include "libavutil/imgutils.h" |
|
27 |
+#include "avfilter.h" |
|
28 |
+#include "internal.h" |
|
29 |
+ |
|
30 |
+#include "vidstabutils.h" |
|
31 |
+ |
|
32 |
+typedef struct { |
|
33 |
+ const AVClass* class; |
|
34 |
+ |
|
35 |
+ VSTransformData td; |
|
36 |
+ VSTransformConfig conf; |
|
37 |
+ |
|
38 |
+ VSTransformations trans; // transformations |
|
39 |
+ char* input; // name of transform file |
|
40 |
+ int tripod; |
|
41 |
+} TransformContext; |
|
42 |
+ |
|
43 |
+#define OFFSET(x) offsetof(TransformContext, x) |
|
44 |
+#define OFFSETC(x) (offsetof(TransformContext, conf)+offsetof(VSTransformConfig, x)) |
|
45 |
+#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM |
|
46 |
+ |
|
47 |
+static const AVOption vidstabtransform_options[]= { |
|
48 |
+ {"input", "path to the file storing the transforms (def:transforms.trf)", OFFSET(input), |
|
49 |
+ AV_OPT_TYPE_STRING, {.str = DEFAULT_INPUT_NAME} }, |
|
50 |
+ {"smoothing", "number of frames*2 + 1 used for lowpass filtering (def: 10)", OFFSETC(smoothing), |
|
51 |
+ AV_OPT_TYPE_INT, {.i64 = 10}, 1, 1000, FLAGS}, |
|
52 |
+ {"maxshift", "maximal number of pixels to translate image (def: -1 no limit)", OFFSETC(maxShift), |
|
53 |
+ AV_OPT_TYPE_INT, {.i64 = -1}, -1, 500, FLAGS}, |
|
54 |
+ {"maxangle", "maximal angle in rad to rotate image (def: -1 no limit)", OFFSETC(maxAngle), |
|
55 |
+ AV_OPT_TYPE_DOUBLE, {.dbl = -1.0}, -1.0, 3.14, FLAGS}, |
|
56 |
+ {"crop", "keep: (def), black", OFFSETC(crop), |
|
57 |
+ AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1, FLAGS, "crop"}, |
|
58 |
+ { "keep", "keep border", 0, |
|
59 |
+ AV_OPT_TYPE_CONST, {.i64 = VSKeepBorder }, 0, 0, FLAGS, "crop"}, |
|
60 |
+ { "black", "black border", 0, |
|
61 |
+ AV_OPT_TYPE_CONST, {.i64 = VSCropBorder }, 0, 0, FLAGS, "crop"}, |
|
62 |
+ {"invert", "1: invert transforms (def: 0)", OFFSETC(invert), |
|
63 |
+ AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1, FLAGS}, |
|
64 |
+ {"relative", "consider transforms as 0: abslute, 1: relative (def)", OFFSETC(relative), |
|
65 |
+ AV_OPT_TYPE_INT, {.i64 = 1}, 0, 1, FLAGS}, |
|
66 |
+ {"zoom", "percentage to zoom >0: zoom in, <0 zoom out (def: 0)", OFFSETC(zoom), |
|
67 |
+ AV_OPT_TYPE_DOUBLE, {.dbl = 0}, 0, 100, FLAGS}, |
|
68 |
+ {"optzoom", "0: nothing, 1: determine optimal zoom (def) (added to 'zoom')", OFFSETC(optZoom), |
|
69 |
+ AV_OPT_TYPE_INT, {.i64 = 1}, 0, 1, FLAGS}, |
|
70 |
+ {"interpol", "type of interpolation, no, linear, bilinear (def) , bicubic", OFFSETC(interpolType), |
|
71 |
+ AV_OPT_TYPE_INT, {.i64 = 2}, 0, 3, FLAGS, "interpol"}, |
|
72 |
+ { "no", "no interpolation", 0, |
|
73 |
+ AV_OPT_TYPE_CONST, {.i64 = VS_Zero }, 0, 0, FLAGS, "interpol"}, |
|
74 |
+ { "linear", "linear (horizontal)", 0, |
|
75 |
+ AV_OPT_TYPE_CONST, {.i64 = VS_Linear }, 0, 0, FLAGS, "interpol"}, |
|
76 |
+ { "bilinear","bi-linear", 0, |
|
77 |
+ AV_OPT_TYPE_CONST, {.i64 = VS_BiLinear},0, 0, FLAGS, "interpol"}, |
|
78 |
+ { "bicubic", "bi-cubic", 0, |
|
79 |
+ AV_OPT_TYPE_CONST, {.i64 = VS_BiCubic },0, 0, FLAGS, "interpol"}, |
|
80 |
+ {"tripod", "if 1: virtual tripod mode (equiv. to relative=0:smoothing=0)", OFFSET(tripod), |
|
81 |
+ AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1, FLAGS}, |
|
82 |
+ {NULL}, |
|
83 |
+}; |
|
84 |
+ |
|
85 |
+AVFILTER_DEFINE_CLASS(vidstabtransform); |
|
86 |
+ |
|
87 |
+static av_cold int init(AVFilterContext *ctx) |
|
88 |
+{ |
|
89 |
+ TransformContext* tc = ctx->priv; |
|
90 |
+ vs_set_mem_and_log_functions(); |
|
91 |
+ tc->class = &vidstabtransform_class; |
|
92 |
+ av_log(ctx, AV_LOG_VERBOSE, "vidstabtransform filter: init %s\n", LIBVIDSTAB_VERSION); |
|
93 |
+ return 0; |
|
94 |
+} |
|
95 |
+ |
|
96 |
+static av_cold void uninit(AVFilterContext *ctx) |
|
97 |
+{ |
|
98 |
+ TransformContext *tc = ctx->priv; |
|
99 |
+ |
|
100 |
+ vsTransformDataCleanup(&tc->td); |
|
101 |
+ vsTransformationsCleanup(&tc->trans); |
|
102 |
+} |
|
103 |
+ |
|
104 |
+static int query_formats(AVFilterContext *ctx) |
|
105 |
+{ |
|
106 |
+ // If you add something here also add it in vidstabutils.c |
|
107 |
+ static const enum AVPixelFormat pix_fmts[] = { |
|
108 |
+ AV_PIX_FMT_YUV444P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV420P, |
|
109 |
+ AV_PIX_FMT_YUV411P, AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUVA420P, |
|
110 |
+ AV_PIX_FMT_YUV440P, AV_PIX_FMT_GRAY8, |
|
111 |
+ AV_PIX_FMT_RGB24, AV_PIX_FMT_BGR24, AV_PIX_FMT_RGBA, |
|
112 |
+ AV_PIX_FMT_NONE |
|
113 |
+ }; |
|
114 |
+ |
|
115 |
+ ff_set_common_formats(ctx, ff_make_format_list(pix_fmts)); |
|
116 |
+ return 0; |
|
117 |
+} |
|
118 |
+ |
|
119 |
+ |
|
120 |
+static int config_input(AVFilterLink *inlink) |
|
121 |
+{ |
|
122 |
+ AVFilterContext *ctx = inlink->dst; |
|
123 |
+ TransformContext *tc = ctx->priv; |
|
124 |
+ FILE* f; |
|
125 |
+ |
|
126 |
+ const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format); |
|
127 |
+ |
|
128 |
+ VSTransformData* td = &(tc->td); |
|
129 |
+ |
|
130 |
+ VSFrameInfo fi_src; |
|
131 |
+ VSFrameInfo fi_dest; |
|
132 |
+ |
|
133 |
+ if(!vsFrameInfoInit(&fi_src, inlink->w, inlink->h, |
|
134 |
+ av_2_vs_pixel_format(ctx,inlink->format)) || |
|
135 |
+ !vsFrameInfoInit(&fi_dest, inlink->w, inlink->h, |
|
136 |
+ av_2_vs_pixel_format(ctx, inlink->format))){ |
|
137 |
+ av_log(ctx, AV_LOG_ERROR, "unknown pixel format: %i (%s)", |
|
138 |
+ inlink->format, desc->name); |
|
139 |
+ return AVERROR(EINVAL); |
|
140 |
+ } |
|
141 |
+ |
|
142 |
+ if(fi_src.bytesPerPixel != av_get_bits_per_pixel(desc)/8 || |
|
143 |
+ fi_src.log2ChromaW != desc->log2_chroma_w || |
|
144 |
+ fi_src.log2ChromaH != desc->log2_chroma_h){ |
|
145 |
+ av_log(ctx, AV_LOG_ERROR, "pixel-format error: bpp %i<>%i ", |
|
146 |
+ fi_src.bytesPerPixel, av_get_bits_per_pixel(desc)/8); |
|
147 |
+ av_log(ctx, AV_LOG_ERROR, "chroma_subsampl: w: %i<>%i h: %i<>%i\n", |
|
148 |
+ fi_src.log2ChromaW, desc->log2_chroma_w, |
|
149 |
+ fi_src.log2ChromaH, desc->log2_chroma_h); |
|
150 |
+ return AVERROR(EINVAL); |
|
151 |
+ } |
|
152 |
+ |
|
153 |
+ // set values that are not initializes by the options |
|
154 |
+ tc->conf.modName = "vidstabtransform"; |
|
155 |
+ tc->conf.verbose =1; |
|
156 |
+ if(tc->tripod){ |
|
157 |
+ av_log(ctx, AV_LOG_INFO, "Virtual tripod mode: relative=0, smoothing=0"); |
|
158 |
+ tc->conf.relative=0; |
|
159 |
+ tc->conf.smoothing=0; |
|
160 |
+ } |
|
161 |
+ |
|
162 |
+ if(vsTransformDataInit(td, &tc->conf, &fi_src, &fi_dest) != VS_OK){ |
|
163 |
+ av_log(ctx, AV_LOG_ERROR, "initialization of vid.stab transform failed, please report a BUG\n"); |
|
164 |
+ return AVERROR(EINVAL); |
|
165 |
+ } |
|
166 |
+ |
|
167 |
+ vsTransformGetConfig(&tc->conf,td); |
|
168 |
+ av_log(ctx, AV_LOG_INFO, "Video transformation/stabilization settings (pass 2/2):\n"); |
|
169 |
+ av_log(ctx, AV_LOG_INFO, " input = %s\n", tc->input); |
|
170 |
+ av_log(ctx, AV_LOG_INFO, " smoothing = %d\n", tc->conf.smoothing); |
|
171 |
+ av_log(ctx, AV_LOG_INFO, " maxshift = %d\n", tc->conf.maxShift); |
|
172 |
+ av_log(ctx, AV_LOG_INFO, " maxangle = %f\n", tc->conf.maxAngle); |
|
173 |
+ av_log(ctx, AV_LOG_INFO, " crop = %s\n", tc->conf.crop ? "Black" : "Keep"); |
|
174 |
+ av_log(ctx, AV_LOG_INFO, " relative = %s\n", tc->conf.relative ? "True": "False"); |
|
175 |
+ av_log(ctx, AV_LOG_INFO, " invert = %s\n", tc->conf.invert ? "True" : "False"); |
|
176 |
+ av_log(ctx, AV_LOG_INFO, " zoom = %f\n", tc->conf.zoom); |
|
177 |
+ av_log(ctx, AV_LOG_INFO, " optzoom = %s\n", tc->conf.optZoom ? "On" : "Off"); |
|
178 |
+ av_log(ctx, AV_LOG_INFO, " interpol = %s\n", getInterpolationTypeName(tc->conf.interpolType)); |
|
179 |
+ |
|
180 |
+ f = fopen(tc->input, "r"); |
|
181 |
+ if (f == NULL) { |
|
182 |
+ av_log(ctx, AV_LOG_ERROR, "cannot open input file %s\n", tc->input); |
|
183 |
+ return AVERROR(errno); |
|
184 |
+ } else { |
|
185 |
+ VSManyLocalMotions mlms; |
|
186 |
+ if(vsReadLocalMotionsFile(f,&mlms)==VS_OK){ |
|
187 |
+ // calculate the actual transforms from the localmotions |
|
188 |
+ if(vsLocalmotions2TransformsSimple(td, &mlms,&tc->trans)!=VS_OK){ |
|
189 |
+ av_log(ctx, AV_LOG_ERROR, "calculating transformations failed\n"); |
|
190 |
+ return AVERROR(EINVAL); |
|
191 |
+ } |
|
192 |
+ }else{ // try to read old format |
|
193 |
+ if (!vsReadOldTransforms(td, f, &tc->trans)) { /* read input file */ |
|
194 |
+ av_log(ctx, AV_LOG_ERROR, "error parsing input file %s\n", tc->input); |
|
195 |
+ return AVERROR(EINVAL); |
|
196 |
+ } |
|
197 |
+ } |
|
198 |
+ } |
|
199 |
+ fclose(f); |
|
200 |
+ |
|
201 |
+ if (vsPreprocessTransforms(td, &tc->trans)!= VS_OK ) { |
|
202 |
+ av_log(ctx, AV_LOG_ERROR, "error while preprocessing transforms\n"); |
|
203 |
+ return AVERROR(EINVAL); |
|
204 |
+ } |
|
205 |
+ |
|
206 |
+ // TODO: add sharpening, so far the user needs to call the unsharp filter manually |
|
207 |
+ return 0; |
|
208 |
+} |
|
209 |
+ |
|
210 |
+ |
|
211 |
+static int filter_frame(AVFilterLink *inlink, AVFrame *in) |
|
212 |
+{ |
|
213 |
+ AVFilterContext *ctx = inlink->dst; |
|
214 |
+ TransformContext *tc = ctx->priv; |
|
215 |
+ VSTransformData* td = &(tc->td); |
|
216 |
+ |
|
217 |
+ AVFilterLink *outlink = inlink->dst->outputs[0]; |
|
218 |
+ int direct = 0; |
|
219 |
+ AVFrame *out; |
|
220 |
+ VSFrame inframe; |
|
221 |
+ int plane; |
|
222 |
+ |
|
223 |
+ if (av_frame_is_writable(in)) { |
|
224 |
+ direct = 1; |
|
225 |
+ out = in; |
|
226 |
+ } else { |
|
227 |
+ out = ff_get_video_buffer(outlink, outlink->w, outlink->h); |
|
228 |
+ if (!out) { |
|
229 |
+ av_frame_free(&in); |
|
230 |
+ return AVERROR(ENOMEM); |
|
231 |
+ } |
|
232 |
+ av_frame_copy_props(out, in); |
|
233 |
+ } |
|
234 |
+ |
|
235 |
+ for(plane=0; plane < vsTransformGetSrcFrameInfo(td)->planes; plane++){ |
|
236 |
+ inframe.data[plane] = in->data[plane]; |
|
237 |
+ inframe.linesize[plane] = in->linesize[plane]; |
|
238 |
+ } |
|
239 |
+ if(out == in){ // inplace |
|
240 |
+ vsTransformPrepare(td, &inframe, &inframe); |
|
241 |
+ }else{ // seperate frames |
|
242 |
+ VSFrame outframe; |
|
243 |
+ for(plane=0; plane < vsTransformGetDestFrameInfo(td)->planes; plane++){ |
|
244 |
+ outframe.data[plane] = out->data[plane]; |
|
245 |
+ outframe.linesize[plane] = out->linesize[plane]; |
|
246 |
+ } |
|
247 |
+ vsTransformPrepare(td, &inframe, &outframe); |
|
248 |
+ } |
|
249 |
+ |
|
250 |
+ vsDoTransform(td, vsGetNextTransform(td, &tc->trans)); |
|
251 |
+ |
|
252 |
+ vsTransformFinish(td); |
|
253 |
+ |
|
254 |
+ if (!direct) |
|
255 |
+ av_frame_free(&in); |
|
256 |
+ |
|
257 |
+ return ff_filter_frame(outlink, out); |
|
258 |
+} |
|
259 |
+ |
|
260 |
+static const AVFilterPad avfilter_vf_vidstabtransform_inputs[] = { |
|
261 |
+ { |
|
262 |
+ .name = "default", |
|
263 |
+ .type = AVMEDIA_TYPE_VIDEO, |
|
264 |
+ .filter_frame = filter_frame, |
|
265 |
+ .config_props = config_input, |
|
266 |
+ }, |
|
267 |
+ { NULL } |
|
268 |
+}; |
|
269 |
+ |
|
270 |
+static const AVFilterPad avfilter_vf_vidstabtransform_outputs[] = { |
|
271 |
+ { |
|
272 |
+ .name = "default", |
|
273 |
+ .type = AVMEDIA_TYPE_VIDEO, |
|
274 |
+ }, |
|
275 |
+ { NULL } |
|
276 |
+}; |
|
277 |
+ |
|
278 |
+AVFilter avfilter_vf_vidstabtransform = { |
|
279 |
+ .name = "vidstabtransform", |
|
280 |
+ .description = NULL_IF_CONFIG_SMALL("pass 2 of stabilization" |
|
281 |
+ "transforms the frames" |
|
282 |
+ "(see vidstabdetect for pass 1)"), |
|
283 |
+ .priv_size = sizeof(TransformContext), |
|
284 |
+ .init = init, |
|
285 |
+ .uninit = uninit, |
|
286 |
+ .query_formats = query_formats, |
|
287 |
+ |
|
288 |
+ .inputs = avfilter_vf_vidstabtransform_inputs, |
|
289 |
+ .outputs = avfilter_vf_vidstabtransform_outputs, |
|
290 |
+ .priv_class = &vidstabtransform_class, |
|
291 |
+ |
|
292 |
+}; |
|
293 |
+ |
0 | 294 |
new file mode 100644 |
... | ... |
@@ -0,0 +1,84 @@ |
0 |
+/* |
|
1 |
+ * Copyright (c) 2013 Georg Martius <georg dot martius at web dot de> |
|
2 |
+ * |
|
3 |
+ * This file is part of FFmpeg. |
|
4 |
+ * |
|
5 |
+ * FFmpeg is free software; you can redistribute it and/or |
|
6 |
+ * modify it under the terms of the GNU Lesser General Public |
|
7 |
+ * License as published by the Free Software Foundation; either |
|
8 |
+ * version 2.1 of the License, or (at your option) any later version. |
|
9 |
+ * |
|
10 |
+ * FFmpeg is distributed in the hope that it will be useful, |
|
11 |
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
12 |
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
|
13 |
+ * Lesser General Public License for more details. |
|
14 |
+ * |
|
15 |
+ * You should have received a copy of the GNU Lesser General Public |
|
16 |
+ * License along with FFmpeg; if not, write to the Free Software |
|
17 |
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
18 |
+ */ |
|
19 |
+ |
|
20 |
+#include "vidstabutils.h" |
|
21 |
+ |
|
22 |
+ |
|
23 |
+/** convert AV's pixelformat to vid.stab pixelformat */ |
|
24 |
+VSPixelFormat av_2_vs_pixel_format(AVFilterContext *ctx, enum AVPixelFormat pf){ |
|
25 |
+ switch(pf){ |
|
26 |
+ case AV_PIX_FMT_YUV420P: return PF_YUV420P; |
|
27 |
+ case AV_PIX_FMT_YUV422P: return PF_YUV422P; |
|
28 |
+ case AV_PIX_FMT_YUV444P: return PF_YUV444P; |
|
29 |
+ case AV_PIX_FMT_YUV410P: return PF_YUV410P; |
|
30 |
+ case AV_PIX_FMT_YUV411P: return PF_YUV411P; |
|
31 |
+ case AV_PIX_FMT_YUV440P: return PF_YUV440P; |
|
32 |
+ case AV_PIX_FMT_YUVA420P: return PF_YUVA420P; |
|
33 |
+ case AV_PIX_FMT_GRAY8: return PF_GRAY8; |
|
34 |
+ case AV_PIX_FMT_RGB24: return PF_RGB24; |
|
35 |
+ case AV_PIX_FMT_BGR24: return PF_BGR24; |
|
36 |
+ case AV_PIX_FMT_RGBA: return PF_RGBA; |
|
37 |
+ default: |
|
38 |
+ av_log(ctx, AV_LOG_ERROR, "cannot deal with pixel format %i\n", pf); |
|
39 |
+ return PF_NONE; |
|
40 |
+ } |
|
41 |
+} |
|
42 |
+ |
|
43 |
+ |
|
44 |
+/** struct to hold a valid context for logging from within vid.stab lib */ |
|
45 |
+typedef struct { |
|
46 |
+ const AVClass* class; |
|
47 |
+} VS2AVLogCtx; |
|
48 |
+ |
|
49 |
+/** wrapper to log vs_log into av_log */ |
|
50 |
+static int vs_2_av_log_wrapper(int type, const char* tag, const char* format, ...){ |
|
51 |
+ va_list ap; |
|
52 |
+ VS2AVLogCtx ctx; |
|
53 |
+ AVClass class = { |
|
54 |
+ .class_name = tag, |
|
55 |
+ .item_name = av_default_item_name, |
|
56 |
+ .option = 0, |
|
57 |
+ .version = LIBAVUTIL_VERSION_INT, |
|
58 |
+ .category = AV_CLASS_CATEGORY_FILTER, |
|
59 |
+ }; |
|
60 |
+ ctx.class = &class; |
|
61 |
+ va_start (ap, format); |
|
62 |
+ av_vlog(&ctx, type, format, ap); |
|
63 |
+ va_end (ap); |
|
64 |
+ return VS_OK; |
|
65 |
+} |
|
66 |
+ |
|
67 |
+/** sets the memory allocation function and logging constants to av versions */ |
|
68 |
+void vs_set_mem_and_log_functions(void){ |
|
69 |
+ vs_malloc = av_malloc; |
|
70 |
+ vs_zalloc = av_mallocz; |
|
71 |
+ vs_realloc = av_realloc; |
|
72 |
+ vs_free = av_free; |
|
73 |
+ |
|
74 |
+ VS_ERROR_TYPE = AV_LOG_ERROR; |
|
75 |
+ VS_WARN_TYPE = AV_LOG_WARNING; |
|
76 |
+ VS_INFO_TYPE = AV_LOG_INFO; |
|
77 |
+ VS_MSG_TYPE = AV_LOG_VERBOSE; |
|
78 |
+ |
|
79 |
+ vs_log = vs_2_av_log_wrapper; |
|
80 |
+ |
|
81 |
+ VS_ERROR = 0; |
|
82 |
+ VS_OK = 1; |
|
83 |
+} |
0 | 84 |
new file mode 100644 |
... | ... |
@@ -0,0 +1,36 @@ |
0 |
+/* |
|
1 |
+ * Copyright (c) 2013 Georg Martius <georg dot martius at web dot de> |
|
2 |
+ * |
|
3 |
+ * This file is part of FFmpeg. |
|
4 |
+ * |
|
5 |
+ * FFmpeg is free software; you can redistribute it and/or |
|
6 |
+ * modify it under the terms of the GNU Lesser General Public |
|
7 |
+ * License as published by the Free Software Foundation; either |
|
8 |
+ * version 2.1 of the License, or (at your option) any later version. |
|
9 |
+ * |
|
10 |
+ * FFmpeg is distributed in the hope that it will be useful, |
|
11 |
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
12 |
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
|
13 |
+ * Lesser General Public License for more details. |
|
14 |
+ * |
|
15 |
+ * You should have received a copy of the GNU Lesser General Public |
|
16 |
+ * License along with FFmpeg; if not, write to the Free Software |
|
17 |
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
18 |
+ */ |
|
19 |
+ |
|
20 |
+#ifndef AVFILTER_VIDSTABUTILS_H |
|
21 |
+#define AVFILTER_VIDSTABUTILS_H |
|
22 |
+ |
|
23 |
+#include <vid.stab/libvidstab.h> |
|
24 |
+ |
|
25 |
+#include "avfilter.h" |
|
26 |
+ |
|
27 |
+/* ** some conversions from avlib to vid.stab constants and functions *** */ |
|
28 |
+ |
|
29 |
+/** converts the pixelformat of avlib into the one of the vid.stab library */ |
|
30 |
+VSPixelFormat av_2_vs_pixel_format(AVFilterContext *ctx, enum AVPixelFormat pf); |
|
31 |
+ |
|
32 |
+/** sets the memory allocation function and logging constants to av versions */ |
|
33 |
+void vs_set_mem_and_log_functions(void); |
|
34 |
+ |
|
35 |
+#endif |