[Edit] this Topic was split from viewtopic.php?f=17&t=10254 to give more emphasis to HD TS.
I have been transcoding broadcast-captured HD TS(1920x1080,mpeg2-aac,23Mbps,implying RFF/TFFflags,2-3hours) into .m2ts(H.264-ac3) with MC. It takes much time, about 10-20 hours to convert each movies on my computer system.
Apart from Merobs's evaluation in his Test Report (viewtopic.php?f=17&t=9611) where the target is aimed to ac3 input and short one, my target has aac audio and has long duration. Then evaluation result differs from his result.
Most important qualities of transcoding software for me are following basical two.
(1) can convert whole input, not to be aborted neither ends on the way to give shorter output. can be trusted for completion during I am asleep.
(2) can keep A/V sync, or at least correctable afterward
If possible, two items are expected.
(3) can treat dual audio
(4) fast conversion speed
From this view point, my rough evaluation of MC versions is shown below.
4399, 4672-4742 is good as to (1) and (2), however, 2011-50xx versions are still invalid.
4399 : best choice now, can treat dual audio (some A/V sync issue, TSmuxer is not included)
4450-4668: A/V sync is out of order(some are good for not-implied RFF/TFF)
4672-4742: good choice, but cannot treat dual audio
4762-5025: crashes or terminate on the way with high probability(by unknown cause) in either 32bit or 64bit.
**A/V sync issue at 4399**
When input is TS(5.1ch), video become slower(dulation expanded), whereas video speed kept normally when forced output into stereo.
Sync kept output can be got from muxing .ac3 or .aac(audio only conversion) and .264(conversion into stereo) with Yamb or TSmuxeR.
**.m2ts by 4399**
By muxing audio(.ac3) and video(.264) afterwards with TSmuxer.
Test of MediaCoder Builds for HD TS input
Moderator: HuggiL
-
- Advanced
- Posts: 480
- Joined: Sun Dec 14, 2008 4:52 pm
- Location: JAPAN
Test of MediaCoder Builds for HD TS input
Akira Kawamoto
-
- Advanced
- Posts: 480
- Joined: Sun Dec 14, 2008 4:52 pm
- Location: JAPAN
A/V speed sync, and dual audio
Build 5100 is the first version that can keep A/V speed sync for soft-telecined(with RFF/TFF) input TS, and also can treat dual audio tracks simultaneously properly.
Assume four TS(mpeg2-aac) type,
(1)nonRFFst--having stereo audio, video without RFF/TFF, 3,2-pulldowned 29.97 interlaced
(2)nonRFF51------having 5.1ch audio, video without RFF/TFF, 3,2-pulldowned 29.97 interlaced
(3)RFFst-----having stereo audio, video with RFF/TFF, 23.976 progressive
(4)RFF51---------having 5.1ch audio, video with RFF/TFF, 23.976 progressive
History of MC's A/V speed sync and can treat dual audio simultaneously for TS is as follows:
MCversion----------------nonRFFst--nonRFF51---RFFst----RFF51---(dual audio)
0.7.0.4399----------------good------????--------good------bad----good--
0.7.1.4450-0.7.3.4668----absurd-----????--------absurd---absurd--absurd--
0.7.3.4672-0.7.3.4677-----good------????--------absurd---absurd--bad--
0.7.3.4685-0.7.5.4730-----good------????--------good------bad----bad--
0.7.5.4740-0.7.5.4797-----good------good-------good------good----bad--
0.7.5.4798-0.7.5.4799-----good------good-------good------bad-----good--
build5000-build5075--------good------good-------bad-------bad----good--
build5100------------------good------????------good------good----good--
Assume four TS(mpeg2-aac) type,
(1)nonRFFst--having stereo audio, video without RFF/TFF, 3,2-pulldowned 29.97 interlaced
(2)nonRFF51------having 5.1ch audio, video without RFF/TFF, 3,2-pulldowned 29.97 interlaced
(3)RFFst-----having stereo audio, video with RFF/TFF, 23.976 progressive
(4)RFF51---------having 5.1ch audio, video with RFF/TFF, 23.976 progressive
History of MC's A/V speed sync and can treat dual audio simultaneously for TS is as follows:
MCversion----------------nonRFFst--nonRFF51---RFFst----RFF51---(dual audio)
0.7.0.4399----------------good------????--------good------bad----good--
0.7.1.4450-0.7.3.4668----absurd-----????--------absurd---absurd--absurd--
0.7.3.4672-0.7.3.4677-----good------????--------absurd---absurd--bad--
0.7.3.4685-0.7.5.4730-----good------????--------good------bad----bad--
0.7.5.4740-0.7.5.4797-----good------good-------good------good----bad--
0.7.5.4798-0.7.5.4799-----good------good-------good------bad-----good--
build5000-build5075--------good------good-------bad-------bad----good--
build5100------------------good------????------good------good----good--
Akira Kawamoto