diff mbox series

[FFmpeg-devel] avcodec/ccaption_dec: rework non-real-time mode by delaying subtitles

Message ID 20200618194304.19742-1-onemda@gmail.com
State New
Headers show
Series [FFmpeg-devel] avcodec/ccaption_dec: rework non-real-time mode by delaying subtitles
Related show

Checks

Context Check Description
andriy/default pending
andriy/configure warning Failed to apply patch

Commit Message

Paul B Mahol June 18, 2020, 7:43 p.m. UTC
So it give similar output as visual output of real-time mode.

FATE results changes for non-realtime mode so it is updated.

Signed-off-by: Paul B Mahol <onemda@gmail.com>
---
 libavcodec/ccaption_dec.c    | 117 ++++++++++++++++++++---------------
 tests/ref/fate/sub-cc        |   4 +-
 tests/ref/fate/sub-cc-scte20 |   6 +-
 3 files changed, 73 insertions(+), 54 deletions(-)

Comments

Carl Eugen Hoyos June 18, 2020, 8:07 p.m. UTC | #1
Am Do., 18. Juni 2020 um 22:05 Uhr schrieb Paul B Mahol <onemda@gmail.com>:
>
> So it give similar output as visual output of real-time mode.
>
> FATE results changes for non-realtime mode so it is updated.


> diff --git a/tests/ref/fate/sub-cc b/tests/ref/fate/sub-cc
> index 2b30a35be0..0528c846d4 100644
> --- a/tests/ref/fate/sub-cc
> +++ b/tests/ref/fate/sub-cc
> @@ -11,5 +11,5 @@ Style: Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
>
>  [Events]
>  Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
> -Dialogue: 0,0:00:12.36,0:00:40.83,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1} inaudible radio chatter{\i0} )
> -Dialogue: 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1} inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains our number one
> +Dialogue: 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1} inaudible radio chatter{\i0} )
> +Dialogue: 0,0:00:59.07,0:01:17.31,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1} inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains our number one
> diff --git a/tests/ref/fate/sub-cc-scte20 b/tests/ref/fate/sub-cc-scte20
> index be28084887..47ad05db72 100644
> --- a/tests/ref/fate/sub-cc-scte20
> +++ b/tests/ref/fate/sub-cc-scte20
> @@ -11,6 +11,6 @@ Style: Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
>
>  [Events]
>  Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
> -Dialogue: 0,0:00:00.00,0:00:01.44,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE
> -Dialogue: 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES THE \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR
> -Dialogue: 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING AND THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS
> +Dialogue: 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE
> +Dialogue: 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES THE \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR
> +Dialogue: 0,0:00:06.31,0:00:08.68,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING AND THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS

Were both outputs completely wrong before?

Carl Eugen
Paul B Mahol June 18, 2020, 8:10 p.m. UTC | #2
On 6/18/20, Carl Eugen Hoyos <ceffmpeg@gmail.com> wrote:
> Am Do., 18. Juni 2020 um 22:05 Uhr schrieb Paul B Mahol <onemda@gmail.com>:
>>
>> So it give similar output as visual output of real-time mode.
>>
>> FATE results changes for non-realtime mode so it is updated.
>
>
>> diff --git a/tests/ref/fate/sub-cc b/tests/ref/fate/sub-cc
>> index 2b30a35be0..0528c846d4 100644
>> --- a/tests/ref/fate/sub-cc
>> +++ b/tests/ref/fate/sub-cc
>> @@ -11,5 +11,5 @@ Style:
>> Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
>>
>>  [Events]
>>  Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV,
>> Effect, Text
>> -Dialogue:
>> 0,0:00:12.36,0:00:40.83,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1}
>> inaudible radio chatter{\i0} )
>> -Dialogue:
>> 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1}
>> inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains our
>> number one
>> +Dialogue:
>> 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1}
>> inaudible radio chatter{\i0} )
>> +Dialogue:
>> 0,0:00:59.07,0:01:17.31,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1}
>> inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains our
>> number one
>> diff --git a/tests/ref/fate/sub-cc-scte20 b/tests/ref/fate/sub-cc-scte20
>> index be28084887..47ad05db72 100644
>> --- a/tests/ref/fate/sub-cc-scte20
>> +++ b/tests/ref/fate/sub-cc-scte20
>> @@ -11,6 +11,6 @@ Style:
>> Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
>>
>>  [Events]
>>  Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV,
>> Effect, Text
>> -Dialogue:
>> 0,0:00:00.00,0:00:01.44,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE
>> -Dialogue:
>> 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES THE
>> \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR
>> -Dialogue:
>> 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING AND
>> THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS
>> +Dialogue:
>> 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE
>> +Dialogue:
>> 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES THE
>> \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR
>> +Dialogue:
>> 0,0:00:06.31,0:00:08.68,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING AND
>> THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS
>
> Were both outputs completely wrong before?
>

Can't tell. samples are small and without audio so no way to check.

> Carl Eugen
> _______________________________________________
> 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".
Carl Eugen Hoyos June 18, 2020, 8:19 p.m. UTC | #3
Am Do., 18. Juni 2020 um 22:10 Uhr schrieb Paul B Mahol <onemda@gmail.com>:
>
> On 6/18/20, Carl Eugen Hoyos <ceffmpeg@gmail.com> wrote:
> > Am Do., 18. Juni 2020 um 22:05 Uhr schrieb Paul B Mahol <onemda@gmail.com>:
> >>
> >> So it give similar output as visual output of real-time mode.
> >>
> >> FATE results changes for non-realtime mode so it is updated.
> >
> >
> >> diff --git a/tests/ref/fate/sub-cc b/tests/ref/fate/sub-cc
> >> index 2b30a35be0..0528c846d4 100644
> >> --- a/tests/ref/fate/sub-cc
> >> +++ b/tests/ref/fate/sub-cc
> >> @@ -11,5 +11,5 @@ Style:
> >> Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
> >>
> >>  [Events]
> >>  Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV,
> >> Effect, Text
> >> -Dialogue:
> >> 0,0:00:12.36,0:00:40.83,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1}
> >> inaudible radio chatter{\i0} )
> >> -Dialogue:
> >> 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1}
> >> inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains our
> >> number one
> >> +Dialogue:
> >> 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1}
> >> inaudible radio chatter{\i0} )
> >> +Dialogue:
> >> 0,0:00:59.07,0:01:17.31,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1}
> >> inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains our
> >> number one
> >> diff --git a/tests/ref/fate/sub-cc-scte20 b/tests/ref/fate/sub-cc-scte20
> >> index be28084887..47ad05db72 100644
> >> --- a/tests/ref/fate/sub-cc-scte20
> >> +++ b/tests/ref/fate/sub-cc-scte20
> >> @@ -11,6 +11,6 @@ Style:
> >> Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
> >>
> >>  [Events]
> >>  Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV,
> >> Effect, Text
> >> -Dialogue:
> >> 0,0:00:00.00,0:00:01.44,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE
> >> -Dialogue:
> >> 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES THE
> >> \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR
> >> -Dialogue:
> >> 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING AND
> >> THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS
> >> +Dialogue:
> >> 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE
> >> +Dialogue:
> >> 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES THE
> >> \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR
> >> +Dialogue:
> >> 0,0:00:06.31,0:00:08.68,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING AND
> >> THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS
> >
> > Were both outputs completely wrong before?
> >
>
> Can't tell. samples are small and without audio so no way to check.

One of the videos used to be here:

http://gsocdev.ccextractor.org/~anshul/test_video/

Carl Eugen
Carl Eugen Hoyos June 18, 2020, 8:29 p.m. UTC | #4
Am Do., 18. Juni 2020 um 22:10 Uhr schrieb Paul B Mahol <onemda@gmail.com>:
>
> On 6/18/20, Carl Eugen Hoyos <ceffmpeg@gmail.com> wrote:
> > Am Do., 18. Juni 2020 um 22:05 Uhr schrieb Paul B Mahol <onemda@gmail.com>:
> >>
> >> So it give similar output as visual output of real-time mode.
> >>
> >> FATE results changes for non-realtime mode so it is updated.
> >
> >
> >> diff --git a/tests/ref/fate/sub-cc b/tests/ref/fate/sub-cc
> >> index 2b30a35be0..0528c846d4 100644
> >> --- a/tests/ref/fate/sub-cc
> >> +++ b/tests/ref/fate/sub-cc
> >> @@ -11,5 +11,5 @@ Style:
> >> Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
> >>
> >>  [Events]
> >>  Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV,
> >> Effect, Text
> >> -Dialogue:
> >> 0,0:00:12.36,0:00:40.83,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1}
> >> inaudible radio chatter{\i0} )
> >> -Dialogue:
> >> 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1}
> >> inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains our
> >> number one
> >> +Dialogue:
> >> 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1}
> >> inaudible radio chatter{\i0} )
> >> +Dialogue:
> >> 0,0:00:59.07,0:01:17.31,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1}
> >> inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains our
> >> number one
> >> diff --git a/tests/ref/fate/sub-cc-scte20 b/tests/ref/fate/sub-cc-scte20
> >> index be28084887..47ad05db72 100644
> >> --- a/tests/ref/fate/sub-cc-scte20
> >> +++ b/tests/ref/fate/sub-cc-scte20
> >> @@ -11,6 +11,6 @@ Style:
> >> Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
> >>
> >>  [Events]
> >>  Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV,
> >> Effect, Text
> >> -Dialogue:
> >> 0,0:00:00.00,0:00:01.44,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE
> >> -Dialogue:
> >> 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES THE
> >> \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR
> >> -Dialogue:
> >> 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING AND
> >> THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS
> >> +Dialogue:
> >> 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE
> >> +Dialogue:
> >> 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES THE
> >> \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR
> >> +Dialogue:
> >> 0,0:00:06.31,0:00:08.68,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING AND
> >> THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS
> >
> > Were both outputs completely wrong before?
> >
>
> Can't tell. samples are small and without audio so no way to check.

The longest variant of the first sample file is now in samples/MPEG2/subcc
File name is Closedcaption_rollup.ts

Carl Eugen
Paul B Mahol June 19, 2020, 10:15 a.m. UTC | #5
On 6/18/20, Carl Eugen Hoyos <ceffmpeg@gmail.com> wrote:
> Am Do., 18. Juni 2020 um 22:10 Uhr schrieb Paul B Mahol <onemda@gmail.com>:
>>
>> On 6/18/20, Carl Eugen Hoyos <ceffmpeg@gmail.com> wrote:
>> > Am Do., 18. Juni 2020 um 22:05 Uhr schrieb Paul B Mahol
>> > <onemda@gmail.com>:
>> >>
>> >> So it give similar output as visual output of real-time mode.
>> >>
>> >> FATE results changes for non-realtime mode so it is updated.
>> >
>> >
>> >> diff --git a/tests/ref/fate/sub-cc b/tests/ref/fate/sub-cc
>> >> index 2b30a35be0..0528c846d4 100644
>> >> --- a/tests/ref/fate/sub-cc
>> >> +++ b/tests/ref/fate/sub-cc
>> >> @@ -11,5 +11,5 @@ Style:
>> >> Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
>> >>
>> >>  [Events]
>> >>  Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV,
>> >> Effect, Text
>> >> -Dialogue:
>> >> 0,0:00:12.36,0:00:40.83,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1}
>> >> inaudible radio chatter{\i0} )
>> >> -Dialogue:
>> >> 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1}
>> >> inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains
>> >> our
>> >> number one
>> >> +Dialogue:
>> >> 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1}
>> >> inaudible radio chatter{\i0} )
>> >> +Dialogue:
>> >> 0,0:00:59.07,0:01:17.31,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1}
>> >> inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains
>> >> our
>> >> number one
>> >> diff --git a/tests/ref/fate/sub-cc-scte20
>> >> b/tests/ref/fate/sub-cc-scte20
>> >> index be28084887..47ad05db72 100644
>> >> --- a/tests/ref/fate/sub-cc-scte20
>> >> +++ b/tests/ref/fate/sub-cc-scte20
>> >> @@ -11,6 +11,6 @@ Style:
>> >> Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
>> >>
>> >>  [Events]
>> >>  Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV,
>> >> Effect, Text
>> >> -Dialogue:
>> >> 0,0:00:00.00,0:00:01.44,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE
>> >> -Dialogue:
>> >> 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES
>> >> THE
>> >> \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR
>> >> -Dialogue:
>> >> 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING
>> >> AND
>> >> THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS
>> >> +Dialogue:
>> >> 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE
>> >> +Dialogue:
>> >> 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES
>> >> THE
>> >> \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR
>> >> +Dialogue:
>> >> 0,0:00:06.31,0:00:08.68,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING
>> >> AND
>> >> THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS
>> >
>> > Were both outputs completely wrong before?
>> >
>>
>> Can't tell. samples are small and without audio so no way to check.
>
> The longest variant of the first sample file is now in samples/MPEG2/subcc
> File name is Closedcaption_rollup.ts
>

Thanks.

Fixed patch to only delay for  pop-on captions.
Thus fate does not need a change.
For others modes delaying is not needed.

> Carl Eugen
> _______________________________________________
> 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".
diff mbox series

Patch

diff --git a/libavcodec/ccaption_dec.c b/libavcodec/ccaption_dec.c
index 87ac23e0d1..5128f6d91f 100644
--- a/libavcodec/ccaption_dec.c
+++ b/libavcodec/ccaption_dec.c
@@ -231,27 +231,25 @@  typedef struct CCaptionSubContext {
     uint8_t cursor_color;
     uint8_t cursor_font;
     uint8_t cursor_charset;
-    AVBPrint buffer;
+    AVBPrint buffer[2];
+    int buffer_index;
     int buffer_changed;
     int rollup;
     enum cc_mode mode;
-    int64_t start_time;
-    /* visible screen time */
-    int64_t startv_time;
-    int64_t end_time;
+    int64_t buffer_time[2];
     int screen_touched;
     int64_t last_real_time;
     char prev_cmd[2];
     int readorder;
 } CCaptionSubContext;
 
-
 static av_cold int init_decoder(AVCodecContext *avctx)
 {
     int ret;
     CCaptionSubContext *ctx = avctx->priv_data;
 
-    av_bprint_init(&ctx->buffer, 0, AV_BPRINT_SIZE_UNLIMITED);
+    av_bprint_init(&ctx->buffer[0], 0, AV_BPRINT_SIZE_UNLIMITED);
+    av_bprint_init(&ctx->buffer[1], 0, AV_BPRINT_SIZE_UNLIMITED);
     /* taking by default roll up to 2 */
     ctx->mode = CCMODE_ROLLUP;
     ctx->rollup = 2;
@@ -275,7 +273,8 @@  static av_cold int init_decoder(AVCodecContext *avctx)
 static av_cold int close_decoder(AVCodecContext *avctx)
 {
     CCaptionSubContext *ctx = avctx->priv_data;
-    av_bprint_finalize(&ctx->buffer, NULL);
+    av_bprint_finalize(&ctx->buffer[0], NULL);
+    av_bprint_finalize(&ctx->buffer[1], NULL);
     return 0;
 }
 
@@ -299,7 +298,8 @@  static void flush_decoder(AVCodecContext *avctx)
     ctx->buffer_changed = 0;
     if (!(avctx->flags2 & AV_CODEC_FLAG2_RO_FLUSH_NOOP))
         ctx->readorder = 0;
-    av_bprint_clear(&ctx->buffer);
+    av_bprint_clear(&ctx->buffer[0]);
+    av_bprint_clear(&ctx->buffer[1]);
 }
 
 /**
@@ -427,7 +427,9 @@  static int capture_screen(CCaptionSubContext *ctx)
     struct Screen *screen = ctx->screen + ctx->active_screen;
     enum cc_font prev_font = CCFONT_REGULAR;
     enum cc_color_code prev_color = CCCOL_WHITE;
-    av_bprint_clear(&ctx->buffer);
+    const int bidx = ctx->buffer_index;
+
+    av_bprint_clear(&ctx->buffer[bidx]);
 
     for (i = 0; screen->row_used && i < SCREEN_ROWS; i++)
     {
@@ -459,7 +461,7 @@  static int capture_screen(CCaptionSubContext *ctx)
 
             x = ASS_DEFAULT_PLAYRESX * (0.1 + 0.0250 * j);
             y = ASS_DEFAULT_PLAYRESY * (0.1 + 0.0533 * i);
-            av_bprintf(&ctx->buffer, "{\\an7}{\\pos(%d,%d)}", x, y);
+            av_bprintf(&ctx->buffer[bidx], "{\\an7}{\\pos(%d,%d)}", x, y);
 
             for (; j < SCREEN_COLUMNS; j++) {
                 const char *e_tag = "", *s_tag = "", *c_tag = "";
@@ -521,35 +523,33 @@  static int capture_screen(CCaptionSubContext *ctx)
                 prev_color = color[j];
                 override = charset_overrides[(int)charset[j]][(int)row[j]];
                 if (override) {
-                    av_bprintf(&ctx->buffer, "%s%s%s%s", e_tag, s_tag, c_tag, override);
+                    av_bprintf(&ctx->buffer[bidx], "%s%s%s%s", e_tag, s_tag, c_tag, override);
                     seen_char = 1;
                 } else if (row[j] == ' ' && !seen_char) {
-                    av_bprintf(&ctx->buffer, "%s%s%s\\h", e_tag, s_tag, c_tag);
+                    av_bprintf(&ctx->buffer[bidx], "%s%s%s\\h", e_tag, s_tag, c_tag);
                 } else {
-                    av_bprintf(&ctx->buffer, "%s%s%s%c", e_tag, s_tag, c_tag, row[j]);
+                    av_bprintf(&ctx->buffer[bidx], "%s%s%s%c", e_tag, s_tag, c_tag, row[j]);
                     seen_char = 1;
                 }
 
             }
-            av_bprintf(&ctx->buffer, "\\N");
+            av_bprintf(&ctx->buffer[bidx], "\\N");
         }
     }
-    if (!av_bprint_is_complete(&ctx->buffer))
+    if (!av_bprint_is_complete(&ctx->buffer[bidx]))
         return AVERROR(ENOMEM);
-    if (screen->row_used && ctx->buffer.len >= 2) {
-        ctx->buffer.len -= 2;
-        ctx->buffer.str[ctx->buffer.len] = 0;
+    if (screen->row_used && ctx->buffer[bidx].len >= 2) {
+        ctx->buffer[bidx].len -= 2;
+        ctx->buffer[bidx].str[ctx->buffer[bidx].len] = 0;
     }
     ctx->buffer_changed = 1;
     return 0;
 }
 
-static int reap_screen(CCaptionSubContext *ctx, int64_t pts)
+static void update_time(CCaptionSubContext *ctx, int64_t pts)
 {
-    ctx->start_time = ctx->startv_time;
-    ctx->startv_time = pts;
-    ctx->end_time = pts;
-    return capture_screen(ctx);
+    ctx->buffer_time[0] = ctx->buffer_time[1];
+    ctx->buffer_time[1] = pts;
 }
 
 static void handle_textattr(CCaptionSubContext *ctx, uint8_t hi, uint8_t lo)
@@ -594,10 +594,7 @@  static void handle_pac(CCaptionSubContext *ctx, uint8_t hi, uint8_t lo)
     }
 }
 
-/**
- * @param pts it is required to set end time
- */
-static int handle_edm(CCaptionSubContext *ctx, int64_t pts)
+static int handle_edm(CCaptionSubContext *ctx)
 {
     struct Screen *screen = ctx->screen + ctx->active_screen;
     int ret;
@@ -605,35 +602,35 @@  static int handle_edm(CCaptionSubContext *ctx, int64_t pts)
     // In buffered mode, keep writing to screen until it is wiped.
     // Before wiping the display, capture contents to emit subtitle.
     if (!ctx->real_time)
-        ret = reap_screen(ctx, pts);
+        ret = capture_screen(ctx);
 
     screen->row_used = 0;
 
     // In realtime mode, emit an empty caption so the last one doesn't
     // stay on the screen.
     if (ctx->real_time)
-        ret = reap_screen(ctx, pts);
+        ret = capture_screen(ctx);
 
     return ret;
 }
 
-static int handle_eoc(CCaptionSubContext *ctx, int64_t pts)
+static int handle_eoc(CCaptionSubContext *ctx)
 {
     int ret;
 
     ctx->active_screen = !ctx->active_screen;
 
     // In buffered mode, we wait til the *next* EOC and
-    // reap what was already on the screen since the last EOC.
+    // capture what was already on the screen since the last EOC.
     if (!ctx->real_time)
-        ret = handle_edm(ctx, pts);
+        ret = handle_edm(ctx);
 
     ctx->cursor_column = 0;
 
     // In realtime mode, we display the buffered contents (after
     // flipping the buffer to active above) as soon as EOC arrives.
     if (ctx->real_time)
-        ret = reap_screen(ctx, pts);
+        ret = capture_screen(ctx);
 
     return ret;
 }
@@ -684,7 +681,7 @@  static void handle_char(CCaptionSubContext *ctx, char hi, char lo)
        ff_dlog(ctx, "(%c)\n", hi);
 }
 
-static int process_cc608(CCaptionSubContext *ctx, int64_t pts, uint8_t hi, uint8_t lo)
+static int process_cc608(CCaptionSubContext *ctx, uint8_t hi, uint8_t lo)
 {
     int ret = 0;
 
@@ -727,13 +724,13 @@  static int process_cc608(CCaptionSubContext *ctx, int64_t pts, uint8_t hi, uint8
             break;
         case 0x2c:
             /* erase display memory */
-            handle_edm(ctx, pts);
+            handle_edm(ctx);
             break;
         case 0x2d:
             /* carriage return */
             ff_dlog(ctx, "carriage return\n");
             if (!ctx->real_time)
-                ret = reap_screen(ctx, pts);
+                ret = capture_screen(ctx);
             roll_up(ctx);
             ctx->cursor_column = 0;
             break;
@@ -749,7 +746,7 @@  static int process_cc608(CCaptionSubContext *ctx, int64_t pts, uint8_t hi, uint8
         case 0x2f:
             /* end of caption */
             ff_dlog(ctx, "handle_eoc\n");
-            ret = handle_eoc(ctx, pts);
+            ret = handle_eoc(ctx);
             break;
         default:
             ff_dlog(ctx, "Unknown command 0x%hhx 0x%hhx\n", hi, lo);
@@ -780,7 +777,10 @@  static int decode(AVCodecContext *avctx, void *data, int *got_sub, AVPacket *avp
 {
     CCaptionSubContext *ctx = avctx->priv_data;
     AVSubtitle *sub = data;
-    const int64_t start_time = sub->pts;
+    int64_t in_time = sub->pts;
+    int64_t start_time;
+    int64_t end_time;
+    int bidx = ctx->buffer_index;
     uint8_t *bptr = NULL;
     int len = avpkt->size;
     int ret = 0;
@@ -796,7 +796,7 @@  static int decode(AVCodecContext *avctx, void *data, int *got_sub, AVPacket *avp
         if (cc_type == 1)
             continue;
 
-        ret = process_cc608(ctx, start_time, bptr[i + 1] & 0x7f, bptr[i + 2] & 0x7f);
+        ret = process_cc608(ctx, bptr[i + 1] & 0x7f, bptr[i + 2] & 0x7f);
         if (ret < 0)
             return ret;
 
@@ -804,23 +804,41 @@  static int decode(AVCodecContext *avctx, void *data, int *got_sub, AVPacket *avp
             continue;
         ctx->buffer_changed = 0;
 
-        if (ctx->buffer.str[0] || ctx->real_time) {
-            ff_dlog(ctx, "cdp writing data (%s)\n", ctx->buffer.str);
-            ret = ff_ass_add_rect(sub, ctx->buffer.str, ctx->readorder++, 0, NULL, NULL);
-            if (ret < 0)
-                return ret;
-            sub->pts = ctx->start_time;
+        if (!ctx->real_time)
+            ctx->buffer_index = bidx = !ctx->buffer_index;
+
+        update_time(ctx, in_time);
+
+        if (ctx->buffer[bidx].str[0] || ctx->real_time) {
+            ff_dlog(ctx, "cdp writing data (%s)\n", ctx->buffer[bidx].str);
+            start_time = ctx->buffer_time[0];
+            sub->pts = start_time;
+            end_time = ctx->buffer_time[1];
             if (!ctx->real_time)
-                sub->end_display_time = av_rescale_q(ctx->end_time - ctx->start_time,
+                sub->end_display_time = av_rescale_q(end_time - start_time,
                                                      AV_TIME_BASE_Q, ms_tb);
             else
                 sub->end_display_time = -1;
-            ctx->buffer_changed = 0;
+            ret = ff_ass_add_rect(sub, ctx->buffer[bidx].str, ctx->readorder++, 0, NULL, NULL);
+            if (ret < 0)
+                return ret;
             ctx->last_real_time = sub->pts;
             ctx->screen_touched = 0;
         }
     }
 
+    if (!bptr && !ctx->real_time && ctx->buffer[!ctx->buffer_index].str[0]) {
+        bidx = !ctx->buffer_index;
+        ret = ff_ass_add_rect(sub, ctx->buffer[bidx].str, ctx->readorder++, 0, NULL, NULL);
+        if (ret < 0)
+            return ret;
+        sub->pts = ctx->buffer_time[1];
+        sub->end_display_time = av_rescale_q(ctx->buffer_time[1] - ctx->buffer_time[0],
+                                             AV_TIME_BASE_Q, ms_tb);
+        if (sub->end_display_time == 0)
+            sub->end_display_time = ctx->buffer[bidx].len * 20;
+    }
+
     if (ctx->real_time && ctx->screen_touched &&
         sub->pts > ctx->last_real_time + av_rescale_q(200, ms_tb, AV_TIME_BASE_Q)) {
         ctx->last_real_time = sub->pts;
@@ -829,7 +847,7 @@  static int decode(AVCodecContext *avctx, void *data, int *got_sub, AVPacket *avp
         capture_screen(ctx);
         ctx->buffer_changed = 0;
 
-        ret = ff_ass_add_rect(sub, ctx->buffer.str, ctx->readorder++, 0, NULL, NULL);
+        ret = ff_ass_add_rect(sub, ctx->buffer[bidx].str, ctx->readorder++, 0, NULL, NULL);
         if (ret < 0)
             return ret;
         sub->end_display_time = -1;
@@ -864,4 +882,5 @@  AVCodec ff_ccaption_decoder = {
     .flush          = flush_decoder,
     .decode         = decode,
     .priv_class     = &ccaption_dec_class,
+    .capabilities   = AV_CODEC_CAP_DELAY,
 };
diff --git a/tests/ref/fate/sub-cc b/tests/ref/fate/sub-cc
index 2b30a35be0..0528c846d4 100644
--- a/tests/ref/fate/sub-cc
+++ b/tests/ref/fate/sub-cc
@@ -11,5 +11,5 @@  Style: Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
 
 [Events]
 Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
-Dialogue: 0,0:00:12.36,0:00:40.83,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1} inaudible radio chatter{\i0} )
-Dialogue: 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1} inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains our number one
+Dialogue: 0,0:00:40.83,0:00:59.07,Default,,0,0,0,,{\an7}{\pos(38,44)}({\i1} inaudible radio chatter{\i0} )
+Dialogue: 0,0:00:59.07,0:01:17.31,Default,,0,0,0,,{\an7}{\pos(38,28)}({\i1} inaudible radio chatter{\i0} )\N{\an7}{\pos(38,44)}>> Safety remains our number one
diff --git a/tests/ref/fate/sub-cc-scte20 b/tests/ref/fate/sub-cc-scte20
index be28084887..47ad05db72 100644
--- a/tests/ref/fate/sub-cc-scte20
+++ b/tests/ref/fate/sub-cc-scte20
@@ -11,6 +11,6 @@  Style: Default,Monospace,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,0,100,100,0,0,3,1,0,
 
 [Events]
 Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
-Dialogue: 0,0:00:00.00,0:00:01.44,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE 
-Dialogue: 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES THE \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR 
-Dialogue: 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING AND THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS 
+Dialogue: 0,0:00:01.43,0:00:03.93,Default,,0,0,0,,{\an7}{\pos(48,182)}BESIDES THE 
+Dialogue: 0,0:00:03.94,0:00:06.31,Default,,0,0,0,,{\an7}{\pos(38,166)}\hBESIDES THE \N{\an7}{\pos(38,197)}SPENDING AND THIS, IS THAT CAR 
+Dialogue: 0,0:00:06.31,0:00:08.68,Default,,0,0,0,,{\an7}{\pos(38,182)}SPENDING AND THIS, IS THAT CAR \N{\an7}{\pos(38,197)}MANUFACTURERS ARE ABOUT AS