[FFmpeg-devel] [PATCH 0/5] Properly handle CEA-708 caption data when transcoding

[FFmpeg-devel] [PATCH 0/5] Properly handle CEA-708 caption data when transcoding

[Devin Heitmueller]

This patch series is intended to address long-standing known issues where
CEA-708 caption data is either duplicated or corrupted (i.e. 50% of the data
is lost).  We employ an intermediate queue, which is available to filters that
change the frame rate of the video, which stashes the 608/708 data and
reconstructs properly formed 708 tuples on the output side (injecting
appropriate padding as needed).

The common functions are implemented in libavutil, but I am open to
suggestions.  While all the users in this patch series are within
libavfilter, there are use cases where this functionality would
be useful within libavformat (which is why libavutil was chosen).

This version of the patch series, compared to the RFC, has a couple
of fixes to catch missed code paths for yadif and tinterlace.

Comments/feedback are welcomed.

Thanks,

Devin

Devin Heitmueller (5):
  ccfifo: Properly handle CEA-708 captions through framerate conversion
  vf_fps: properly preserve CEA-708 captions
  yadif: Properly preserve CEA-708 closed captions
  tinterlace: Properly preserve CEA-708 closed captions
  vf_ccrepack: Add new filter to repack CEA-708 side data

 doc/filters.texi            |  10 ++
 libavfilter/Makefile        |   1 +
 libavfilter/allfilters.c    |   1 +
 libavfilter/tinterlace.h    |   2 +
 libavfilter/vf_bwdif.c      |   7 ++
 libavfilter/vf_ccrepack.c   |  95 ++++++++++++++++++
 libavfilter/vf_fps.c        |   9 +-
 libavfilter/vf_tinterlace.c |   8 ++
 libavfilter/vf_yadif.c      |   6 ++
 libavfilter/vf_yadif_cuda.c |   8 ++
 libavfilter/yadif.h         |   2 +
 libavfilter/yadif_common.c  |   5 +
 libavutil/Makefile          |   2 +
 libavutil/ccfifo.c          | 191 ++++++++++++++++++++++++++++++++++++
 libavutil/ccfifo.h          |  85 ++++++++++++++++
 15 files changed, 431 insertions(+), 1 deletion(-)
 create mode 100644 libavfilter/vf_ccrepack.c
 create mode 100644 libavutil/ccfifo.c
 create mode 100644 libavutil/ccfifo.h

-- 
2.35.1.655.ga68dfadae5

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

[FFmpeg-devel] [PATCH 1/5] ccfifo: Properly handle CEA-708 captions through framerate conversion

[Devin Heitmueller]

When transcoding video that contains 708 closed captions, the
caption data is tied to the frames as side data.  Simply dropping
or adding frames to change the framerate will result in loss of
data, so the caption data needs to be preserved and reformatted.

For example, without this patch converting 720p59 to 1080i59
would result in loss of 50% of the caption bytes, resulting in
garbled 608 captions and 708 probably wouldn't render at all.
Further, the frames that are there will have an illegal
cc_count for the target framerate, so some decoders may ignore
the packets entirely.

Extract the 608 and 708 tuples and insert them onto queues.  Then
after dropping/adding frames, re-write the tuples back into the
resulting frames at the appropriate rate given the target
framerate.  This includes both having the correct cc_count as
well as clocking out the 608 pairs at the appropriate rate.

Signed-off-by: Devin Heitmueller <dheitmueller@ltnglobal.com>
---
 libavutil/Makefile |   2 +
 libavutil/ccfifo.c | 191 +++++++++++++++++++++++++++++++++++++++++++++
 libavutil/ccfifo.h |  85 ++++++++++++++++++++
 3 files changed, 278 insertions(+)
 create mode 100644 libavutil/ccfifo.c
 create mode 100644 libavutil/ccfifo.h

diff --git a/libavutil/Makefile b/libavutil/Makefile
index dc9012f9a8..b6115859fa 100644
--- a/libavutil/Makefile
+++ b/libavutil/Makefile
@@ -17,6 +17,7 @@ HEADERS = adler32.h                                                     \
           buffer.h                                                      \
           cast5.h                                                       \
           camellia.h                                                    \
+          ccfifo.h                                                      \
           channel_layout.h                                              \
           common.h                                                      \
           cpu.h                                                         \
@@ -113,6 +114,7 @@ OBJS = adler32.o                                                        \
        buffer.o                                                         \
        cast5.o                                                          \
        camellia.o                                                       \
+       ccfifo.o                                                         \
        channel_layout.o                                                 \
        cpu.o                                                            \
        crc.o                                                            \
diff --git a/libavutil/ccfifo.c b/libavutil/ccfifo.c
new file mode 100644
index 0000000000..5db4149f3b
--- /dev/null
+++ b/libavutil/ccfifo.c
@@ -0,0 +1,191 @@
+/*
+ * CEA-708 Closed Captioning FIFO
+ * Copyright (c) 2023 LTN Global Communications
+ *
+ * Author: Devin Heitmueller <dheitmueller@ltnglobal.com>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "ccfifo.h"
+
+struct AVCCFifo {
+    AVFifo *cc_608_fifo;
+    AVFifo *cc_708_fifo;
+    int expected_cc_count;
+    int expected_608;
+    int cc_detected;
+    void *log_ctx;
+};
+
+#define MAX_CC_ELEMENTS 128
+#define CC_BYTES_PER_ENTRY 3
+
+struct cc_lookup {
+    int num;
+    int den;
+    int cc_count;
+    int num_608;
+};
+
+const static struct cc_lookup cc_lookup_vals[] = {
+    { 15, 1, 40, 4 },
+    { 24, 1, 25, 3 },
+    { 24000, 1001, 25, 3 },
+    { 30, 1, 20, 2 },
+    { 30000, 1001, 20, 2},
+    { 60, 1, 10, 1 },
+    { 60000, 1001, 10, 1},
+};
+
+void av_ccfifo_freep(AVCCFifo **ccf)
+{
+    if (ccf && *ccf) {
+        AVCCFifo *tmp = *ccf;
+        if (tmp->cc_608_fifo)
+            av_fifo_freep2(&tmp->cc_608_fifo);
+        if (tmp->cc_708_fifo)
+            av_fifo_freep2(&tmp->cc_708_fifo);
+        av_freep(*ccf);
+    }
+}
+
+AVCCFifo *av_ccfifo_alloc(AVRational *framerate, void *log_ctx)
+{
+    AVCCFifo *ccf;
+    int i;
+
+    ccf = av_mallocz(sizeof(*ccf));
+    if (!ccf)
+        return NULL;
+
+    if (!(ccf->cc_708_fifo = av_fifo_alloc2(MAX_CC_ELEMENTS, CC_BYTES_PER_ENTRY, 0)))
+        goto error;
+
+    if (!(ccf->cc_608_fifo = av_fifo_alloc2(MAX_CC_ELEMENTS, CC_BYTES_PER_ENTRY, 0)))
+        goto error;
+
+    /* Based on the target FPS, figure out the expected cc_count and number of
+       608 tuples per packet.  See ANSI/CTA-708-E Sec 4.3.6.1. */
+    for (i = 0; i < (sizeof(cc_lookup_vals) / sizeof(struct cc_lookup)); i++) {
+        if (framerate->num == cc_lookup_vals[i].num &&
+            framerate->den == cc_lookup_vals[i].den) {
+            ccf->expected_cc_count = cc_lookup_vals[i].cc_count;
+            ccf->expected_608 = cc_lookup_vals[i].num_608;
+            break;
+        }
+    }
+
+    if (ccf->expected_608 == 0) {
+        av_log(ccf->log_ctx, AV_LOG_WARNING, "cc_fifo cannot transcode captions fps=%d/%d\n",
+               framerate->num, framerate->den);
+        return NULL;
+    }
+
+    return ccf;
+
+error:
+    av_ccfifo_freep(&ccf);
+    return NULL;
+}
+
+int av_ccfifo_inject(AVCCFifo *ccf, AVFrame *frame)
+{
+    AVFrameSideData *sd;
+    int cc_filled = 0;
+    int i;
+
+    if (!ccf)
+        return 0;
+
+    if (ccf->cc_detected == 0 || ccf->expected_cc_count == 0)
+        return 0;
+
+    sd = av_frame_new_side_data(frame, AV_FRAME_DATA_A53_CC,
+                                ccf->expected_cc_count * CC_BYTES_PER_ENTRY);
+    if (!sd)
+        return 0;
+
+    for (i = 0; i < ccf->expected_608; i++) {
+        if (av_fifo_can_read(ccf->cc_608_fifo) >= CC_BYTES_PER_ENTRY) {
+            av_fifo_read(ccf->cc_608_fifo, &sd->data[cc_filled * CC_BYTES_PER_ENTRY],
+                         CC_BYTES_PER_ENTRY);
+            cc_filled++;
+        } else {
+            break;
+        }
+    }
+
+    /* Insert any available data from the 708 FIFO */
+    while (cc_filled < ccf->expected_cc_count) {
+        if (av_fifo_can_read(ccf->cc_708_fifo) >= CC_BYTES_PER_ENTRY) {
+            av_fifo_read(ccf->cc_708_fifo, &sd->data[cc_filled * CC_BYTES_PER_ENTRY],
+                         CC_BYTES_PER_ENTRY);
+            cc_filled++;
+        } else {
+            break;
+        }
+    }
+
+    /* Insert 708 padding into any remaining fields */
+    while (cc_filled < ccf->expected_cc_count) {
+        sd->data[cc_filled * CC_BYTES_PER_ENTRY]     = 0xfa;
+        sd->data[cc_filled * CC_BYTES_PER_ENTRY + 1] = 0x00;
+        sd->data[cc_filled * CC_BYTES_PER_ENTRY + 2] = 0x00;
+        cc_filled++;
+    }
+
+    return 0;
+}
+
+int av_ccfifo_extract(AVCCFifo *ccf, AVFrame *frame)
+{
+    int i;
+
+    if (!ccf)
+        return 0;
+
+    /* Read the A53 side data, discard padding, and put 608/708 into
+       queues so we can ensure they get into the output frames at
+       the correct rate... */
+    if (ccf->expected_cc_count > 0) {
+        AVFrameSideData *side_data = av_frame_get_side_data(frame, AV_FRAME_DATA_A53_CC);
+        if (side_data) {
+            uint8_t *cc_bytes = side_data->data;
+            int cc_count = side_data->size / CC_BYTES_PER_ENTRY;
+            ccf->cc_detected = 1;
+
+            for (i = 0; i < cc_count; i++) {
+                /* See ANSI/CTA-708-E Sec 4.3, Table 3 */
+                uint8_t cc_valid = (cc_bytes[CC_BYTES_PER_ENTRY*i] & 0x04) >> 2;
+                uint8_t cc_type = cc_bytes[CC_BYTES_PER_ENTRY*i] & 0x03;
+                if (cc_type == 0x00 || cc_type == 0x01) {
+                    av_fifo_write(ccf->cc_608_fifo, &cc_bytes[CC_BYTES_PER_ENTRY*i],
+                                  CC_BYTES_PER_ENTRY);
+                } else if (cc_valid && (cc_type == 0x02 || cc_type == 0x03)) {
+                    av_fifo_write(ccf->cc_708_fifo, &cc_bytes[CC_BYTES_PER_ENTRY*i],
+                                  CC_BYTES_PER_ENTRY);
+                }
+            }
+
+            /* Remove the side data, as we will re-create it on the
+               output as needed */
+            av_frame_remove_side_data(frame, AV_FRAME_DATA_A53_CC);
+        }
+    }
+    return 0;
+}
diff --git a/libavutil/ccfifo.h b/libavutil/ccfifo.h
new file mode 100644
index 0000000000..ad48c0a39c
--- /dev/null
+++ b/libavutil/ccfifo.h
@@ -0,0 +1,85 @@
+/*
+ * CEA-708 Closed Captioning FIFO
+ * Copyright (c) 2023 LTN Global Communications
+ *
+ * Author: Devin Heitmueller <dheitmueller@ltnglobal.com>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * CC FIFO Buffer
+ */
+
+#ifndef AVUTIL_CCFIFO_H
+#define AVUTIL_CCFIFO_H
+
+#include "avutil.h"
+#include "frame.h"
+#include "fifo.h"
+
+typedef struct AVCCFifo AVCCFifo;
+
+/**
+ * Allocate an AVCCFifo.
+ *
+ * @param sample_fmt  sample format
+ * @param channels    number of channels
+ * @param nb_samples  initial allocation size, in samples
+ * @return            newly allocated AVCCFifo, or NULL on error
+ */
+AVCCFifo *av_ccfifo_alloc(AVRational *framerate, void *log_ctx);
+
+/**
+ * Free an AVCCFifo
+ *
+ * @param ccf Pointer to the pointer to the AVCCFifo which should be freed
+ * @note `*ptr = NULL` is safe and leads to no action.
+ */
+void av_ccfifo_freep(AVCCFifo **ccf);
+
+
+/**
+ * Read a frame into a CC Fifo
+ *
+ * Extract CC bytes from the AVFrame, insert them into our queue, and
+ * remove the side data from the AVFrame.  The side data is removed
+ * as it will be re-inserted at the appropriate rate later in the
+ * filter.
+ *
+ * @param af          AVCCFifo to write to
+ * @param frame       AVFrame with the video frame to operate on
+ * @return            Zero on success, or negative AVERROR
+ *                    code on failure.
+ */
+int av_ccfifo_extract(AVCCFifo *af, AVFrame *frame);
+
+/**
+ * Insert CC data from the FIFO into an AVFrame (as side data)
+ *
+ * Dequeue the appropriate number of CC tuples based on the
+ * frame rate, and insert them into the AVFrame
+ *
+ * @param af          AVCCFifo to read from
+ * @param frame       AVFrame with the video frame to operate on
+ * @return            Zero on success, or negative AVERROR
+ *                    code on failure.
+ */
+int av_ccfifo_inject(AVCCFifo *af, AVFrame *frame);
+
+#endif /* AVUTIL_CCFIFO_H */
-- 
2.35.1.655.ga68dfadae5

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

[FFmpeg-devel] [PATCH 2/5] vf_fps: properly preserve CEA-708 captions

[Devin Heitmueller]

The existing implementation made an attempt to remove duplicate
captions if increasing the framerate, but made no attempt to
handle reducing the framerate, nor did it rewrite the caption
payloads to have the appropriate cc_count (e.g. the cc_count needs
to change from 20 to 10 when going from 1080i59 to 720p59 and
vice-versa).

Make use of the new ccfifo mechanism to ensure that caption data
is properly preserved.

Signed-off-by: Devin Heitmueller <dheitmueller@ltnglobal.com>
---
 libavfilter/vf_fps.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/libavfilter/vf_fps.c b/libavfilter/vf_fps.c
index 051d278f54..d2fab09c66 100644
--- a/libavfilter/vf_fps.c
+++ b/libavfilter/vf_fps.c
@@ -30,6 +30,7 @@
 #include <stdint.h>
 
 #include "libavutil/avassert.h"
+#include "libavutil/ccfifo.h"
 #include "libavutil/eval.h"
 #include "libavutil/mathematics.h"
 #include "libavutil/opt.h"
@@ -85,6 +86,7 @@ typedef struct FPSContext {
 
     AVFrame *frames[2];     ///< buffered frames
     int      frames_count;  ///< number of buffered frames
+    AVCCFifo *cc_fifo;      ///< closed captions
 
     int64_t  next_pts;      ///< pts of the next frame to output
 
@@ -165,6 +167,7 @@ static av_cold void uninit(AVFilterContext *ctx)
         frame = shift_frame(ctx, s);
         av_frame_free(&frame);
     }
+    av_ccfifo_freep(&s->cc_fifo);
 
     av_log(ctx, AV_LOG_VERBOSE, "%d frames in, %d frames out; %d frames dropped, "
            "%d frames duplicated.\n", s->frames_in, s->frames_out, s->drop, s->dup);
@@ -210,6 +213,9 @@ static int config_props(AVFilterLink* outlink)
                s->in_pts_off, s->out_pts_off, s->start_time);
     }
 
+    if (!(s->cc_fifo = av_ccfifo_alloc(&outlink->frame_rate, ctx)))
+        av_log(ctx, AV_LOG_VERBOSE, "Failure to setup CC FIFO queue.  Captions will be passed through\n");
+
     av_log(ctx, AV_LOG_VERBOSE, "fps=%d/%d\n", outlink->frame_rate.num, outlink->frame_rate.den);
 
     return 0;
@@ -242,6 +248,7 @@ static int read_frame(AVFilterContext *ctx, FPSContext *s, AVFilterLink *inlink,
     av_log(ctx, AV_LOG_DEBUG, "Read frame with in pts %"PRId64", out pts %"PRId64"\n",
            in_pts, frame->pts);
 
+    av_ccfifo_extract(s->cc_fifo, frame);
     s->frames[s->frames_count++] = frame;
     s->frames_in++;
 
@@ -289,7 +296,7 @@ static int write_frame(AVFilterContext *ctx, FPSContext *s, AVFilterLink *outlin
         if (!frame)
             return AVERROR(ENOMEM);
         // Make sure Closed Captions will not be duplicated
-        av_frame_remove_side_data(s->frames[0], AV_FRAME_DATA_A53_CC);
+        av_ccfifo_inject(s->cc_fifo, frame);
         frame->pts = s->next_pts++;
         frame->duration = 1;
 
-- 
2.35.1.655.ga68dfadae5

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

[FFmpeg-devel] [PATCH 3/5] yadif: Properly preserve CEA-708 closed captions

[Devin Heitmueller]

Various deinterlacing modes have the effect of doubling the
framerate, and we need to ensure that the caption data isn't
duplicated (or else you get double captions on-screen).

Use the new ccfifo mechanism for yadif (and yadif_cuda and bwdif
since they use the same yadif core) so that CEA-708 data is
properly preserved through this filter.

Signed-off-by: Devin Heitmueller <dheitmueller@ltnglobal.com>
---
 libavfilter/vf_bwdif.c      | 7 +++++++
 libavfilter/vf_yadif.c      | 6 ++++++
 libavfilter/vf_yadif_cuda.c | 8 ++++++++
 libavfilter/yadif.h         | 2 ++
 libavfilter/yadif_common.c  | 5 +++++
 5 files changed, 28 insertions(+)

diff --git a/libavfilter/vf_bwdif.c b/libavfilter/vf_bwdif.c
index 34e8c5e234..c625b68914 100644
--- a/libavfilter/vf_bwdif.c
+++ b/libavfilter/vf_bwdif.c
@@ -297,6 +297,7 @@ static av_cold void uninit(AVFilterContext *ctx)
     av_frame_free(&yadif->prev);
     av_frame_free(&yadif->cur );
     av_frame_free(&yadif->next);
+    av_ccfifo_freep(&yadif->cc_fifo);
 }
 
 static const enum AVPixelFormat pix_fmts[] = {
@@ -332,6 +333,12 @@ static int config_props(AVFilterLink *link)
 
     if(yadif->mode&1)
         link->frame_rate = av_mul_q(link->src->inputs[0]->frame_rate, (AVRational){2,1});
+    else
+        link->frame_rate = ctx->inputs[0]->frame_rate;
+
+    if (!(yadif->cc_fifo = av_ccfifo_alloc(&link->frame_rate, ctx)))
+        av_log(ctx, AV_LOG_VERBOSE, "Failure to setup CC FIFO queue.  Captions will be passed through\n");
+
 
     if (link->w < 3 || link->h < 4) {
         av_log(ctx, AV_LOG_ERROR, "Video of less than 3 columns or 4 lines is not supported\n");
diff --git a/libavfilter/vf_yadif.c b/libavfilter/vf_yadif.c
index 1be02de1a9..b51d21f6ff 100644
--- a/libavfilter/vf_yadif.c
+++ b/libavfilter/vf_yadif.c
@@ -261,6 +261,7 @@ static av_cold void uninit(AVFilterContext *ctx)
     av_frame_free(&yadif->prev);
     av_frame_free(&yadif->cur );
     av_frame_free(&yadif->next);
+    av_ccfifo_freep(&yadif->cc_fifo);
 }
 
 static const enum AVPixelFormat pix_fmts[] = {
@@ -293,6 +294,11 @@ static int config_output(AVFilterLink *outlink)
     if(s->mode & 1)
         outlink->frame_rate = av_mul_q(ctx->inputs[0]->frame_rate,
                                     (AVRational){2, 1});
+    else
+        outlink->frame_rate = ctx->inputs[0]->frame_rate;
+
+    if (!(s->cc_fifo = av_ccfifo_alloc(&outlink->frame_rate, ctx)))
+        av_log(ctx, AV_LOG_VERBOSE, "Failure to setup CC FIFO queue.  Captions will be passed through\n");
 
     if (outlink->w < 3 || outlink->h < 3) {
         av_log(ctx, AV_LOG_ERROR, "Video of less than 3 columns or lines is not supported\n");
diff --git a/libavfilter/vf_yadif_cuda.c b/libavfilter/vf_yadif_cuda.c
index 685b8a2035..d96ec391a0 100644
--- a/libavfilter/vf_yadif_cuda.c
+++ b/libavfilter/vf_yadif_cuda.c
@@ -206,6 +206,9 @@ static av_cold void deint_cuda_uninit(AVFilterContext *ctx)
     av_frame_free(&y->cur);
     av_frame_free(&y->next);
 
+    if (yadif->cc_fifo)
+        av_cc_fifo_free(yadif->cc_fifo);
+
     av_buffer_unref(&s->device_ref);
     s->hwctx = NULL;
     av_buffer_unref(&s->input_frames_ref);
@@ -291,6 +294,11 @@ static int config_output(AVFilterLink *link)
     if(y->mode & 1)
         link->frame_rate = av_mul_q(ctx->inputs[0]->frame_rate,
                                     (AVRational){2, 1});
+    else
+        outlink->frame_rate = ctx->inputs[0]->frame_rate;
+
+    if (!(s->cc_fifo = av_cc_fifo_alloc(&outlink->frame_rate, ctx)))
+        av_log(ctx, AV_LOG_VERBOSE, "Failure to setup CC FIFO queue.  Captions will be passed through\n");
 
     if (link->w < 3 || link->h < 3) {
         av_log(ctx, AV_LOG_ERROR, "Video of less than 3 columns or lines is not supported\n");
diff --git a/libavfilter/yadif.h b/libavfilter/yadif.h
index c928911b35..ccd4304860 100644
--- a/libavfilter/yadif.h
+++ b/libavfilter/yadif.h
@@ -21,6 +21,7 @@
 
 #include "libavutil/opt.h"
 #include "libavutil/pixdesc.h"
+#include "libavutil/ccfifo.h"
 #include "avfilter.h"
 
 enum YADIFMode {
@@ -76,6 +77,7 @@ typedef struct YADIFContext {
     int eof;
     uint8_t *temp_line;
     int temp_line_size;
+    AVCCFifo *cc_fifo;
 
     /*
      * An algorithm that treats first and/or last fields in a sequence
diff --git a/libavfilter/yadif_common.c b/libavfilter/yadif_common.c
index a10cf7a17f..24cc72ab5f 100644
--- a/libavfilter/yadif_common.c
+++ b/libavfilter/yadif_common.c
@@ -60,6 +60,8 @@ static int return_frame(AVFilterContext *ctx, int is_second)
             yadif->out->pts = AV_NOPTS_VALUE;
         }
     }
+
+    av_ccfifo_inject(yadif->cc_fifo, yadif->out);
     ret = ff_filter_frame(ctx->outputs[0], yadif->out);
 
     yadif->frame_pending = (yadif->mode&1) && !is_second;
@@ -96,6 +98,8 @@ int ff_yadif_filter_frame(AVFilterLink *link, AVFrame *frame)
 
     av_assert0(frame);
 
+    av_ccfifo_extract(yadif->cc_fifo, frame);
+
     if (yadif->frame_pending)
         return_frame(ctx, 1);
 
@@ -137,6 +141,7 @@ int ff_yadif_filter_frame(AVFilterLink *link, AVFrame *frame)
         if (!yadif->out)
             return AVERROR(ENOMEM);
 
+        av_ccfifo_inject(yadif->cc_fifo, yadif->out);
         av_frame_free(&yadif->prev);
         if (yadif->out->pts != AV_NOPTS_VALUE)
             yadif->out->pts *= 2;
-- 
2.35.1.655.ga68dfadae5

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

[FFmpeg-devel] [PATCH 4/5] tinterlace: Properly preserve CEA-708 closed captions

[Devin Heitmueller]

Because the interlacing filter halves the effective framerate, we
need to ensure that no CEA-708 data is lost as frames are merged.

Make use of the new ccfifo mechanism to ensure that caption data
is properly preserved as frames pass through the filter.

Thanks to Thomas Mundt for review and noticing a couple of
missed codepaths for injection on output.

Signed-off-by: Devin Heitmueller <dheitmueller@ltnglobal.com>
---
 libavfilter/tinterlace.h    | 2 ++
 libavfilter/vf_tinterlace.c | 8 ++++++++
 2 files changed, 10 insertions(+)

diff --git a/libavfilter/tinterlace.h b/libavfilter/tinterlace.h
index 37b6c10c08..30574c2ebf 100644
--- a/libavfilter/tinterlace.h
+++ b/libavfilter/tinterlace.h
@@ -30,6 +30,7 @@
 #include "libavutil/bswap.h"
 #include "libavutil/opt.h"
 #include "libavutil/pixdesc.h"
+#include "libavutil/ccfifo.h"
 #include "drawutils.h"
 #include "avfilter.h"
 
@@ -77,6 +78,7 @@ typedef struct TInterlaceContext {
     const AVPixFmtDescriptor *csp;
     void (*lowpass_line)(uint8_t *dstp, ptrdiff_t width, const uint8_t *srcp,
                          ptrdiff_t mref, ptrdiff_t pref, int clip_max);
+    AVCCFifo *cc_fifo;
 } TInterlaceContext;
 
 void ff_tinterlace_init_x86(TInterlaceContext *interlace);
diff --git a/libavfilter/vf_tinterlace.c b/libavfilter/vf_tinterlace.c
index 032629279a..226983c645 100644
--- a/libavfilter/vf_tinterlace.c
+++ b/libavfilter/vf_tinterlace.c
@@ -291,6 +291,9 @@ static int config_out_props(AVFilterLink *outlink)
 #endif
     }
 
+    if (!(tinterlace->cc_fifo = av_ccfifo_alloc(&outlink->frame_rate, ctx)))
+        av_log(ctx, AV_LOG_VERBOSE, "Failure to setup CC FIFO queue.  Captions will be passed through\n");
+
     av_log(ctx, AV_LOG_VERBOSE, "mode:%d filter:%s h:%d -> h:%d\n", tinterlace->mode,
            (tinterlace->flags & TINTERLACE_FLAG_CVLPF) ? "complex" :
            (tinterlace->flags & TINTERLACE_FLAG_VLPF) ? "linear" : "off",
@@ -375,6 +378,8 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *picref)
     tinterlace->cur  = tinterlace->next;
     tinterlace->next = picref;
 
+    av_ccfifo_extract(tinterlace->cc_fifo, picref);
+
     cur = tinterlace->cur;
     next = tinterlace->next;
     /* we need at least two frames */
@@ -451,6 +456,7 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *picref)
             if (!out)
                 return AVERROR(ENOMEM);
             out->pts /= 2;  // adjust pts to new framerate
+            av_ccfifo_inject(tinterlace->cc_fifo, out);
             ret = ff_filter_frame(outlink, out);
             return ret;
         }
@@ -486,6 +492,7 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *picref)
             out->pts = cur->pts*2;
 
         out->pts = av_rescale_q(out->pts, tinterlace->preout_time_base, outlink->time_base);
+        av_ccfifo_inject(tinterlace->cc_fifo, out);
         if ((ret = ff_filter_frame(outlink, out)) < 0)
             return ret;
 
@@ -521,6 +528,7 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *picref)
 
     out->pts = av_rescale_q(out->pts, tinterlace->preout_time_base, outlink->time_base);
     out->duration = av_rescale_q(1, av_inv_q(outlink->frame_rate), outlink->time_base);
+    av_ccfifo_inject(tinterlace->cc_fifo, out);
     ret = ff_filter_frame(outlink, out);
 
     return ret;
-- 
2.35.1.655.ga68dfadae5

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

[FFmpeg-devel] [PATCH 5/5] vf_ccrepack: Add new filter to repack CEA-708 side data

[Devin Heitmueller]

THis filter can correct certain issues seen from upstream sources
where the cc_count is not properly set or the CEA-608 tuples are
not at the start of the payload as expected.

Make use of the ccfifo to extract and immediately repack the CEA-708
side data, thereby removing any extra padding and ensuring the 608
tuples are at the front of the payload.

Signed-off-by: Devin Heitmueller <dheitmueller@ltnglobal.com>
---
 doc/filters.texi          | 10 +++++
 libavfilter/Makefile      |  1 +
 libavfilter/allfilters.c  |  1 +
 libavfilter/vf_ccrepack.c | 95 +++++++++++++++++++++++++++++++++++++++
 4 files changed, 107 insertions(+)
 create mode 100644 libavfilter/vf_ccrepack.c

diff --git a/doc/filters.texi b/doc/filters.texi
index 4449998ba4..d042b034f8 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -8901,6 +8901,16 @@ Only deinterlace frames marked as interlaced.
 The default value is @code{all}.
 @end table
 
+@section ccrepack
+
+Repack CEA-708 closed captioning side data
+
+This filter fixes various issues seen with commerical encoders
+related to upstream malformed CEA-708 payloads, specifically
+incorrect number of tuples (wrong cc_count for the target FPS),
+and incorrect ordering of tuples (i.e. the CEA-608 tuples are not at
+the first entries in the payload).
+
 @section cas
 
 Apply Contrast Adaptive Sharpen filter to video stream.
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 71e198bbf9..6b0da0f34f 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -211,6 +211,7 @@ OBJS-$(CONFIG_BOXBLUR_OPENCL_FILTER)         += vf_avgblur_opencl.o opencl.o \
                                                 opencl/avgblur.o boxblur.o
 OBJS-$(CONFIG_BWDIF_FILTER)                  += vf_bwdif.o yadif_common.o
 OBJS-$(CONFIG_CAS_FILTER)                    += vf_cas.o
+OBJS-$(CONFIG_CCREPACK_FILTER)               += vf_ccrepack.o
 OBJS-$(CONFIG_CHROMABER_VULKAN_FILTER)       += vf_chromaber_vulkan.o vulkan.o vulkan_filter.o
 OBJS-$(CONFIG_CHROMAHOLD_FILTER)             += vf_chromakey.o
 OBJS-$(CONFIG_CHROMAKEY_FILTER)              += vf_chromakey.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index d7db46c2af..b38550b358 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -196,6 +196,7 @@ extern const AVFilter ff_vf_boxblur;
 extern const AVFilter ff_vf_boxblur_opencl;
 extern const AVFilter ff_vf_bwdif;
 extern const AVFilter ff_vf_cas;
+extern const AVFilter ff_vf_ccrepack;
 extern const AVFilter ff_vf_chromaber_vulkan;
 extern const AVFilter ff_vf_chromahold;
 extern const AVFilter ff_vf_chromakey;
diff --git a/libavfilter/vf_ccrepack.c b/libavfilter/vf_ccrepack.c
new file mode 100644
index 0000000000..ab8473e7c1
--- /dev/null
+++ b/libavfilter/vf_ccrepack.c
@@ -0,0 +1,95 @@
+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/*
+ * Repackage CEA-708 arrays, which deals with incorrect cc_count for a given
+ * output framerate, and incorrect 708 padding.
+ *
+ * See CEA CEA-10-A "EIA-708-B Implementation Guidance", Section 26.5
+ * "Grouping DTVCC Data Within user_data() Structure"
+ */
+
+#include "avfilter.h"
+#include "internal.h"
+#include "libavutil/opt.h"
+#include "libavutil/ccfifo.h"
+
+typedef struct CCRepackContext
+{
+    const AVClass *class;
+    AVCCFifo *cc_fifo;
+} CCRepackContext;
+
+static const AVOption ccrepack_options[] = {
+    {  NULL }
+};
+
+AVFILTER_DEFINE_CLASS(ccrepack);
+
+static int config_input(AVFilterLink *link)
+{
+    CCRepackContext *ctx = link->dst->priv;
+
+    if (!(ctx->cc_fifo = av_ccfifo_alloc(&link->frame_rate, ctx)))
+        av_log(ctx, AV_LOG_VERBOSE, "Failure to setup CC FIFO queue.  Captions will be passed through\n");
+
+    return 0;
+}
+
+static int filter_frame(AVFilterLink *inlink, AVFrame *frame)
+{
+    CCRepackContext *ctx = inlink->dst->priv;
+    AVFilterLink *outlink = inlink->dst->outputs[0];
+
+    av_ccfifo_extract(ctx->cc_fifo, frame);
+    av_ccfifo_inject(ctx->cc_fifo, frame);
+
+    return ff_filter_frame(outlink, frame);
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+    CCRepackContext *s = ctx->priv;
+    av_ccfifo_freep(&s->cc_fifo);
+}
+
+static const AVFilterPad avfilter_vf_ccrepack_inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_VIDEO,
+        .filter_frame = filter_frame,
+	.config_props = config_input,
+    },
+};
+
+static const AVFilterPad avfilter_vf_ccrepack_outputs[] = {
+    {
+        .name = "default",
+        .type = AVMEDIA_TYPE_VIDEO,
+    },
+};
+
+AVFilter ff_vf_ccrepack = {
+    .name        = "ccrepack",
+    .description = NULL_IF_CONFIG_SMALL("Repack CEA-708 closed caption metadata"),
+    .uninit      = uninit,
+    .priv_size   = sizeof(CCRepackContext),
+    .priv_class  = &ccrepack_class,
+    FILTER_INPUTS(avfilter_vf_ccrepack_inputs),
+    FILTER_OUTPUTS(avfilter_vf_ccrepack_outputs),
+};
-- 
2.35.1.655.ga68dfadae5

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-devel] [PATCH 0/5] Properly handle CEA-708 caption data when transcoding

[Andreas Rheinhardt]

Devin Heitmueller:
> This patch series is intended to address long-standing known issues where
> CEA-708 caption data is either duplicated or corrupted (i.e. 50% of the data
> is lost).  We employ an intermediate queue, which is available to filters that
> change the frame rate of the video, which stashes the 608/708 data and
> reconstructs properly formed 708 tuples on the output side (injecting
> appropriate padding as needed).
> 
> The common functions are implemented in libavutil, but I am open to
> suggestions.  While all the users in this patch series are within
> libavfilter, there are use cases where this functionality would
> be useful within libavformat (which is why libavutil was chosen).
> 

If it is not used now outside of lavfi, it should be in lavfi to allow
to change it when it turns out that changes are needed.

- Andreas

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

Re: [FFmpeg-devel] [PATCH 0/5] Properly handle CEA-708 caption data when transcoding

[Devin Heitmueller]

On Wed, Mar 29, 2023 at 8:46 PM Andreas Rheinhardt
<andreas.rheinhardt@outlook.com> wrote:
>
> Devin Heitmueller:
> > This patch series is intended to address long-standing known issues where
> > CEA-708 caption data is either duplicated or corrupted (i.e. 50% of the data
> > is lost).  We employ an intermediate queue, which is available to filters that
> > change the frame rate of the video, which stashes the 608/708 data and
> > reconstructs properly formed 708 tuples on the output side (injecting
> > appropriate padding as needed).
> >
> > The common functions are implemented in libavutil, but I am open to
> > suggestions.  While all the users in this patch series are within
> > libavfilter, there are use cases where this functionality would
> > be useful within libavformat (which is why libavutil was chosen).
> >
>
> If it is not used now outside of lavfi, it should be in lavfi to allow
> to change it when it turns out that changes are needed.

So I was thinking along the same lines.  This would allow the API to
mature and change as needed without worrying about ABI breakage, which
would be useful since the API is relatively new and might need
tweaking.

Thanks for the suggestion.  I'll move it into libavfilter in the next
patches I submit (pending any other feedback/comments).

Devin

-- 
Devin Heitmueller, Senior Software Engineer
LTN Global Communications
o: +1 (301) 363-1001
w: https://ltnglobal.com  e: devin.heitmueller@ltnglobal.com
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".