[Bug 787300] basesink: Add property to handle seamless-seek by sink
GStreamer (GNOME Bugzilla)
bugzilla at gnome.org
Sat Sep 23 07:59:44 UTC 2017
https://bugzilla.gnome.org/show_bug.cgi?id=787300
--- Comment #9 from HoonHee Lee <hoonh83.lee at gmail.com> ---
Dear All.
I want to explain more.
And I want to assume that new segment is delivered directly to sink and it is
not forwarding to upstream (demux layer).
For example,
1) play 10s with 1x.
2) trick with 2x at 10s.
3) trick with 0.25 at 20s.
4) trick with 1.x at 25s.
5) paused while 10s.
6) play
Assume clock's absolute time at starting 0s.
For 1)
* clock(absolute-time): 0s -> 10s
* c.start-time: 0s
* c.base-time: 0s
* running-time: 0s -> 10s
For 2)
* clock: 11s -> 20s
* c.running-time vs b.running-time
11s -> 12s
12s -> 14s
13s -> 16s
14s -> 18s
15s -> 20s
16s -> 22s
17s -> 24s
18s -> 26s
19s -> 28s
20s -> 30s
For 3)
* clock: 21s -> 25s
* c.running-time vs b.running-time
21s -> 30.25s
22s -> 30.50s
23s -> 30.75s
24s -> 31.00s
25s -> 31.25s
For 4)
* clock: 26s -> 30s
* c.running-time vs b.running-time
26s -> 32.25s
27s -> 33.25s
28s -> 34.25s
29s -> 35.25s
30s -> 36.25s
For 5)
* clock: 31 -> 40s
* c.start-time: 30s
* c.base-time: 0s
* c.running-time: 30s
For 6)
* clock: 41s -> ...
* c.start-time: 30s
* c.base-time: 10s
* c.running-time: 30s
* c.running-time vs b.running-time
31s -> 37.25s
My analysis is correct??
As I know, c.running-time is 31s and sink may would expect a buffer with
timestamp 31s after paused to playing. I am guessing there will be a jitter.
I am not sure how to synchronize correctly between clock, buffer and segment.
It would be helpful if you give some advices.
Thanks.
--
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
More information about the gstreamer-bugs
mailing list