diff mbox series

[FFmpeg-devel,v4,5/6] ffmpeg: Add display_matrix option

Message ID 20220919094604.4645-6-thilo.borgmann@mail.de
State New
Headers show
Series Add display_matrix option | expand

Checks

Context Check Description
yinshiyou/make_loongarch64 success Make finished
yinshiyou/make_fate_loongarch64 success Make fate finished
andriy/make_x86 success Make finished
andriy/make_fate_x86 success Make fate finished

Commit Message

Thilo Borgmann Sept. 19, 2022, 9:46 a.m. UTC
From: Jan Ekström <jeebjp@gmail.com>

This enables overriding the rotation as well as horizontal/vertical
flip state of a specific video stream on the input side.

Additionally, switch the singular test that was utilizing the rotation
metadata to instead override the input display rotation, thus leading
to the same result.
---
 doc/ffmpeg.texi             |  18 ++++++
 fftools/ffmpeg.h            |   2 +
 fftools/ffmpeg_filter.c     |  13 ++++
 fftools/ffmpeg_opt.c        | 123 ++++++++++++++++++++++++++++++++++++
 tests/fate/filter-video.mak |   2 +-
 5 files changed, 157 insertions(+), 1 deletion(-)
diff mbox series

Patch

diff --git a/doc/ffmpeg.texi b/doc/ffmpeg.texi
index 42440d93b4..ee86faedc8 100644
--- a/doc/ffmpeg.texi
+++ b/doc/ffmpeg.texi
@@ -912,6 +912,24 @@  If used together with @option{-vcodec copy}, it will affect the aspect ratio
 stored at container level, but not the aspect ratio stored in encoded
 frames, if it exists.
 
+@item -display_matrix[:@var{stream_specifier}] @var{opt1=val1[,opt2=val2]...} (@emph{input,per-stream})
+Set the video display matrix according to given options.
+We support scaling, flipping and rotation. The order of application to an Image is scale, flip, rotate.
+
+@table @option
+@item rotation=@var{number}
+Set the rotation using a floating point number that describes a pure
+counter-clockwise rotation in degrees.
+The @code{-autorotate} logic will be affected.
+@item hflip=@var{[0,1]}
+@item vflip=@var{[0,1]}
+Set a horizontal or vertical flip.
+@item hscale=@var{number}
+Set a horizontal scaling by factor of the given floating-point value.
+@item vscale=@var{number}
+Set a vertical scaling by factor of the given floating-point value.
+@end table
+
 @item -vn (@emph{input/output})
 As an input option, blocks all video streams of a file from being filtered or
 being automatically selected or mapped for any output. See @code{-discard}
diff --git a/fftools/ffmpeg.h b/fftools/ffmpeg.h
index ede0b2bd96..524a1d415a 100644
--- a/fftools/ffmpeg.h
+++ b/fftools/ffmpeg.h
@@ -193,6 +193,8 @@  typedef struct OptionsContext {
     int        nb_force_fps;
     SpecifierOpt *frame_aspect_ratios;
     int        nb_frame_aspect_ratios;
+    SpecifierOpt *display_matrixes;
+    int        nb_display_matrixes;
     SpecifierOpt *rc_overrides;
     int        nb_rc_overrides;
     SpecifierOpt *intra_matrices;
diff --git a/fftools/ffmpeg_filter.c b/fftools/ffmpeg_filter.c
index 7a5308425d..8bcba0d3a2 100644
--- a/fftools/ffmpeg_filter.c
+++ b/fftools/ffmpeg_filter.c
@@ -779,9 +779,22 @@  static int configure_input_video_filter(FilterGraph *fg, InputFilter *ifilter,
     if (ist->autorotate && !(desc->flags & AV_PIX_FMT_FLAG_HWACCEL)) {
         int32_t *displaymatrix = ifilter->displaymatrix;
         double theta;
+        double hscale = 1.0f;
+        double vscale = 1.0f;
 
         if (!displaymatrix)
             displaymatrix = (int32_t *)av_stream_get_side_data(ist->st, AV_PKT_DATA_DISPLAYMATRIX, NULL);
+
+        if (displaymatrix) {
+            hscale = av_display_hscale_get(displaymatrix);
+            vscale = av_display_vscale_get(displaymatrix);
+            if (hscale != 1.0f || vscale != 1.0f) {
+                char scale_buf[128];
+                snprintf(scale_buf, sizeof(scale_buf), "%f*iw:%f*ih", hscale, vscale);
+                ret = insert_filter(&last_filter, &pad_idx, "scale", scale_buf);
+            }
+        }
+
         theta = get_rotation(displaymatrix);
 
         if (fabs(theta - 90) < 1.0) {
diff --git a/fftools/ffmpeg_opt.c b/fftools/ffmpeg_opt.c
index be1cd673f6..ba56267460 100644
--- a/fftools/ffmpeg_opt.c
+++ b/fftools/ffmpeg_opt.c
@@ -20,6 +20,7 @@ 
 
 #include "config.h"
 
+#include <float.h>
 #include <stdint.h>
 
 #if HAVE_SYS_RESOURCE_H
@@ -45,6 +46,7 @@ 
 #include "libavutil/avutil.h"
 #include "libavutil/bprint.h"
 #include "libavutil/channel_layout.h"
+#include "libavutil/display.h"
 #include "libavutil/getenv_utf8.h"
 #include "libavutil/intreadwrite.h"
 #include "libavutil/fifo.h"
@@ -101,6 +103,7 @@  static const char *const opt_name_forced_key_frames[]         = {"forced_key_fra
 static const char *const opt_name_fps_mode[]                  = {"fps_mode", NULL};
 static const char *const opt_name_force_fps[]                 = {"force_fps", NULL};
 static const char *const opt_name_frame_aspect_ratios[]       = {"aspect", NULL};
+static const char *const opt_name_display_matrixes[]          = {"display_matrix", NULL};
 static const char *const opt_name_rc_overrides[]              = {"rc_override", NULL};
 static const char *const opt_name_intra_matrices[]            = {"intra_matrix", NULL};
 static const char *const opt_name_inter_matrices[]            = {"inter_matrix", NULL};
@@ -126,6 +129,38 @@  static const char *const opt_name_time_bases[]                = {"time_base", NU
 static const char *const opt_name_enc_time_bases[]            = {"enc_time_base", NULL};
 static const char *const opt_name_bits_per_raw_sample[]       = {"bits_per_raw_sample", NULL};
 
+// XXX this should probably go into a seperate file <name>_args.c and #included here
+    struct DisplayMatrix {
+        const AVClass *class;
+        double  rotation;
+        int     hflip;
+        int     vflip;
+        double  hscale;
+        double  vscale;
+    };
+#define OFFSET(x) offsetof(struct DisplayMatrix, x)
+    static const AVOption display_matrix_args[] = {
+        { "rotation", "set rotation", OFFSET(rotation), AV_OPT_TYPE_DOUBLE,
+            { .dbl = DBL_MAX }, -(DBL_MAX), DBL_MAX - 1.0f, AV_OPT_FLAG_ARGUMENT},
+        { "hflip",    "set hflip", OFFSET(hflip),    AV_OPT_TYPE_BOOL,
+            { .i64 = -1 }, 0, 1, AV_OPT_FLAG_ARGUMENT},
+        { "vflip",    "set vflip", OFFSET(vflip),    AV_OPT_TYPE_BOOL,
+            { .i64 = -1 }, 0, 1, AV_OPT_FLAG_ARGUMENT},
+        { "hscale", "set horizontal scale factor", OFFSET(hscale), AV_OPT_TYPE_DOUBLE,
+            { .dbl = DBL_MAX }, 0.0f, DBL_MAX - 1.0f, AV_OPT_FLAG_ARGUMENT},
+        { "vscale", "set vertical scale factor", OFFSET(vscale), AV_OPT_TYPE_DOUBLE,
+            { .dbl = DBL_MAX }, 0.0f, DBL_MAX - 1.0f, AV_OPT_FLAG_ARGUMENT},
+        { NULL },
+    };
+    static const AVClass class_display_matrix_args = {
+        .class_name = "display_matrix_args",
+        .item_name  = av_default_item_name,
+        .option     = display_matrix_args,
+        .version    = LIBAVUTIL_VERSION_INT,
+    };
+#undef OFFSET
+// XXX
+
 #define WARN_MULTIPLE_OPT_USAGE(name, type, so, st)\
 {\
     char namestr[128] = "";\
@@ -824,6 +859,85 @@  static int opt_recording_timestamp(void *optctx, const char *opt, const char *ar
     return 0;
 }
 
+static void add_display_matrix_to_stream(OptionsContext *o,
+                                         AVFormatContext *ctx, AVStream *st)
+{
+    int hflip_set = 0, vflip_set = 0, display_rotation_set = 0, hscale_set = 0, vscale_set = 0;
+    uint8_t *buf = NULL;
+
+    struct DisplayMatrix test_args = {
+        .class    = &class_display_matrix_args,
+        .rotation = DBL_MAX,
+        .hflip    = -1,
+        .vflip    = -1,
+        .hscale    = DBL_MAX,
+        .vscale    = DBL_MAX,
+    };
+
+    AVDictionary *global_args = NULL;
+    AVDictionary *local_args  = NULL;
+    AVDictionaryEntry *en = NULL;
+
+    MATCH_PER_STREAM_OPT(display_matrixes, dict, global_args, ctx, st);
+
+    if (!global_args)
+        return;
+
+    // make a copy of the dict so it doesn't get freed from underneath us
+    if (av_dict_copy(&local_args, global_args, 0) < 0) {
+        av_log(NULL, AV_LOG_FATAL,
+               "Failed to copy argument dict for display matrix!\n");
+        exit_program(1);
+    }
+
+    if (av_opt_set_dict2(&test_args, &local_args, 0) < 0) {
+        av_log(NULL, AV_LOG_FATAL,
+               "Failed to set options for a display matrix!\n");
+        exit_program(1);
+    }
+
+    while ((en = av_dict_get(local_args, "", en, AV_DICT_IGNORE_SUFFIX))) {
+        av_log(NULL, AV_LOG_FATAL,
+               "Unknown option=value pair for display matrix: "
+               "key: '%s', value: '%s'!\n",
+               en->key, en->value);
+    }
+
+    if (av_dict_count(local_args)) {
+        exit_program(1);
+    }
+
+    av_dict_free(&local_args);
+
+    display_rotation_set = test_args.rotation != DBL_MAX;
+    hflip_set            = test_args.hflip != -1;
+    vflip_set            = test_args.vflip != -1;
+    hscale_set           = test_args.hscale != DBL_MAX;
+    vscale_set           = test_args.vscale != DBL_MAX;
+
+    if (!display_rotation_set && !hflip_set && !vflip_set &&
+        !hscale_set && !vscale_set)
+        return;
+
+    if (!(buf = av_stream_new_side_data(st, AV_PKT_DATA_DISPLAYMATRIX,
+                                        sizeof(int32_t) * 9))) {
+        av_log(NULL, AV_LOG_FATAL, "Failed to generate a display matrix!\n");
+        exit_program(1);
+    }
+
+    av_display_rotation_set((int32_t *)buf,
+                            display_rotation_set ? -(test_args.rotation) :
+                                                   -0.0f);
+    av_display_matrix_flip((int32_t *)buf,
+                           hflip_set ? test_args.hflip : 0,
+                           vflip_set ? test_args.vflip : 0);
+
+    av_display_matrix_scale((int32_t *)buf,
+                            hscale_set ? test_args.hscale : 1.0f,
+                            vscale_set ? test_args.vscale : 1.0f);
+}
+
+
 static const AVCodec *find_codec_or_die(const char *name, enum AVMediaType type, int encoder)
 {
     const AVCodecDescriptor *desc;
@@ -958,6 +1072,8 @@  static void add_input_streams(OptionsContext *o, AVFormatContext *ic)
         }
 
         if (st->codecpar->codec_type == AVMEDIA_TYPE_VIDEO) {
+            add_display_matrix_to_stream(o, ic, st);
+
             MATCH_PER_STREAM_OPT(hwaccels, str, hwaccel, ic, st);
             MATCH_PER_STREAM_OPT(hwaccel_output_formats, str,
                                  hwaccel_output_format, ic, st);
@@ -1883,6 +1999,8 @@  static OutputStream *new_video_stream(OptionsContext *o, AVFormatContext *oc, in
         ost->frame_aspect_ratio = q;
     }
 
+    add_display_matrix_to_stream(o, oc, st);
+
     MATCH_PER_STREAM_OPT(filter_scripts, str, ost->filters_script, oc, st);
     MATCH_PER_STREAM_OPT(filters,        str, ost->filters,        oc, st);
 
@@ -4044,6 +4162,11 @@  const OptionDef options[] = {
     { "aspect",       OPT_VIDEO | HAS_ARG  | OPT_STRING | OPT_SPEC |
                       OPT_OUTPUT,                                                { .off = OFFSET(frame_aspect_ratios) },
         "set aspect ratio (4:3, 16:9 or 1.3333, 1.7777)", "aspect" },
+    { "display_matrix", OPT_VIDEO | HAS_ARG | OPT_DICT | OPT_SPEC |
+                        OPT_INPUT,                                              { .off = OFFSET(display_matrixes) },
+        "define a display matrix with rotation, and/or horizontal/vertical "
+        "flip/scale for stream(s)",
+        "arguments", &class_display_matrix_args },
     { "pix_fmt",      OPT_VIDEO | HAS_ARG | OPT_EXPERT  | OPT_STRING | OPT_SPEC |
                       OPT_INPUT | OPT_OUTPUT,                                    { .off = OFFSET(frame_pix_fmts) },
         "set pixel format", "format" },
diff --git a/tests/fate/filter-video.mak b/tests/fate/filter-video.mak
index 372c70bba7..763390ea51 100644
--- a/tests/fate/filter-video.mak
+++ b/tests/fate/filter-video.mak
@@ -691,7 +691,7 @@  fate-filter-metadata-avf-aphase-meter-out-of-phase: SRC = $(TARGET_SAMPLES)/filt
 fate-filter-metadata-avf-aphase-meter-out-of-phase: CMD = run $(FILTER_METADATA_COMMAND) "amovie='$(SRC)',aphasemeter=video=0"
 
 FATE_FILTER_SAMPLES-$(call TRANSCODE, RAWVIDEO H264, MOV, ARESAMPLE_FILTER  AAC_FIXED_DECODER) += fate-filter-meta-4560-rotate0
-fate-filter-meta-4560-rotate0: CMD = transcode mov $(TARGET_SAMPLES)/filter/sample-in-issue-505.mov mov "-c copy -metadata:s:v:0 rotate=0" "-af aresample" "" "" "-flags +bitexact -c:a aac_fixed"
+fate-filter-meta-4560-rotate0: CMD = transcode "mov -display_matrix:v:0 rotation=0" $(TARGET_SAMPLES)/filter/sample-in-issue-505.mov mov "-c copy" "-af aresample" "" "" "-flags +bitexact -c:a aac_fixed"
 
 FATE_FILTER_CMP_METADATA-$(CONFIG_BLOCKDETECT_FILTER) += fate-filter-refcmp-blockdetect-yuv
 fate-filter-refcmp-blockdetect-yuv: CMD = cmp_metadata blockdetect yuv420p 0.015