audio files wurden eingeschraenkt am ende von demux

[Oscar]

Ich habe 2 Recordings bei welchem die audio file zu klein ist nach demuxen. Man kann das sehen wenn man der mpg-file nach muxen wieder abspielt in powerDVD. Es gibt nur 13 minuten Sound und danach nicht mehr weil die video file 31 Minuten dauert. Ich habe das getestet in dsjar 0.81.4 Beim demuxen mit dsJar koentte Ich in Explorer sehen das am Anfang die verschiedene Audio Files anwachsten bis 86 MB. Am Ende beim demuxen wird angezeigt 'checking and synchronising audio file' und die originale audio filegroesse van 86 MB wird eingeschraenkt zu 25 MB. In die log wird gezeigt das die Video file 31 minuten dauert und dabei die audio files nur 13 minuten dauern. Was koentte das Problem sein ?

Oscar

[TheHorse]

Du müsstest unbedingt den Logauszug aus dem X-Fenster (d*.jar gibt's nicht mehr :) ) kopieren und posten. Darin sieht man bestimmt., warum was faul war.

PS: Frage: Bist du der Oscar aus Norwegen mit dem Topfield?

[Oscar]

Ich ben der Oscar aus die Niederlaende...bitte verzeiche mir fuer mein Deutsches Schreiben...English ist fuer Mich besser aber das ist nicht passend fuer dieses Board.

Ich habe bis letztes Woche viel mit DVR-topA fuer den Topfield gearbeitet aber leider gibt es zu viel Probleme mit das demuxen damit. Ich war sehr gluecklich zu erfahren dass die demux crash und audio delay Probleme bei DVR-topA verschwunden waren mit X . Nur die preview screen fuers erstellen von Schnitten koentte meiner Meinung nach verbessert werden durch ein klares (es gibt Welle/Streife in image) und auf schnellere weise das Bild zu erstellen. Auch wurde es besser sein wenn es ein kleineres unit time step wure geben damit mann ein bischen genauer die Schnittpunkte kann eingeben. Ich bin aber sehr begeistert von die Stabilitaet und viele moeglichkeiten von X.

Ich habe hierunter der Log aus X kopiert. Ich hoffe Du siehst wo das Problem ist.

Oscar

zondag 7 september 2003 19:08:55 uur CEST

X V0.81.4 (29.08.2003)



=== session infos ===



=> working with collection 0

-> write output files to :

E:\Topfield\



=> File 0: E:\Topfield\Testmovie.tts (1610239552 bytes)

=> File is DVB/MPEG-2 TS (Video/Audio/TTX PES)

=> demux

-> SID 0x474

-> PMT 0x474 refer to these Video/Audio/Text related PIDs :

< 0x475 0x47B 0x476 0x477 0x478 0x479 >

ok> PID 0x478 has PES-ID 0xC9 (MPEG Audio) (0x3948 #49)

--> PID 0x3E8 (payload: 00 88 70 23 01 12 5A E4 ..) (0x5820 #91) -> ignored

ok> PID 0x47B has PES-ID 0xBD (private stream 1) (TTX) (0x81FC #148)

ok> PID 0x479 has PES-ID 0xCA (MPEG Audio) (0x9224 #170)

--> PID 0x1A2C (ECM) (0xA24C #192) -> ignored

ok> PID 0x475 has PES-ID 0xE0 (MPEG Video) (0xB564 #218)

--> PID 0x474 (PMT) (0xD208 #257) -> ignored

--> PID 0x1 (CAT) (0xE2EC #280) -> ignored

ok> PID 0x476 has PES-ID 0xC7 (MPEG Audio) (0x13F74 #406)

ok> PID 0x477 has PES-ID 0xC8 (MPEG Audio) (0x18F80 #515)

-> video basics: 720*576 @ 25fps @ 0.6735 (4:3) @ 15000000bps, vbvBuffer 112

dropping useless B-Frames @ GOP#0 / new Timecode 00:00:00.000

dropping gop(sequence) @ orig.PTS 11:59:33.533 / new Timecode 00:02:19.720

GOP#290, Pics expected/count 12/16 ; PTS difference 960ms (should be 0)

dropping useless B-Frames @ GOP#291 / new Timecode 00:02:19.720

dropping gop(sequence) @ orig.PTS 12:00:20.573 / new Timecode 00:03:05.240

GOP#386, Pics expected/count 12/16 ; PTS difference 480ms (should be 0)

dropping useless B-Frames @ GOP#387 / new Timecode 00:03:05.240

dropping gop(sequence) @ orig.PTS 12:10:38.213 / new Timecode 00:13:21.840

GOP#1669, Pics expected/count 12/5 ; PTS difference 47721858ms (should be 0)

dropping useless B-Frames @ GOP#1670 / new Timecode 00:13:21.840

dropping gop(sequence) @ orig.PTS 01:27:07.871 / new Timecode 00:14:29.080

GOP#1810, Pics expected/count 12/13 ; PTS difference 480ms (should be 0)

dropping useless B-Frames @ GOP#1811 / new Timecode 00:14:29.080

--> PID 0x475 -> packet 4108302 @ pos. 772366228 out of sequence (7/11) (shifting..)

--> PID 0x477 -> packet 4108303 @ pos. 772366416 out of sequence (3/6) (shifting..)

--> PID 0x47B -> packet 4108305 @ pos. 772366792 out of sequence (2/10) (shifting..)

--> PID 0x478 -> packet 4108312 @ pos. 772368108 out of sequence (15/2) (shifting..)

--> PID 0x479 -> packet 4108317 @ pos. 772369048 out of sequence (11/14) (shifting..)

--> PID 0x476 -> packet 4108325 @ pos. 772370552 out of sequence (8/10) (shifting..)

--> PID 0x475 -> packet 4108785 @ pos. 772457032 out of sequence (6/2) (shifting..)

--> PID 0x476 -> packet 4108787 @ pos. 772457408 out of sequence (2/9) (shifting..)

--> PID 0x477 -> packet 4108792 @ pos. 772458348 out of sequence (14/5) (shifting..)

--> PID 0x478 -> packet 4108796 @ pos. 772459100 out of sequence (10/1) (shifting..)

--> PID 0x479 -> packet 4108801 @ pos. 772460040 out of sequence (6/13) (shifting..)

--> PID 0x47B -> packet 4108819 @ pos. 772463424 out of sequence (10/7) (shifting..)

dropping gop(sequence) @ orig.PTS 12:28:04.973 / new Timecode 00:14:58.280

GOP#1872, Pics expected/count 12/10 ; PTS difference 0ms (should be 0)

dropping useless B-Frames @ GOP#1873 / new Timecode 00:14:58.280

dropping gop(sequence) @ orig.PTS 12:31:28.253 / new Timecode 00:18:21.000

GOP#2294, Pics expected/count 12/15 ; PTS difference 480ms (should be 0)

dropping useless B-Frames @ GOP#2295 / new Timecode 00:18:21.000

dropping gop(sequence) @ orig.PTS 12:41:44.813 / new Timecode 00:28:36.520

GOP#3574, Pics expected/count 12/21 ; PTS difference 960ms (should be 0)

dropping useless B-Frames @ GOP#3575 / new Timecode 00:28:36.520

dropping gop(sequence) @ orig.PTS 12:42:21.773 / new Timecode 00:29:11.960

GOP#3649, Pics expected/count 12/10 ; PTS difference 0ms (should be 0)

dropping useless B-Frames @ GOP#3650 / new Timecode 00:29:11.960

packs: 8043567 100% 1610239552

video: fr/ct/1p/cg/og/dg 46791/10/0/3891/0/8

videolength: 46791f @ 00:31:11.640

avg. nom. bitrate 5179336bps (min/max: 2852800/7188800)

set first sequence_header bitrate to 7188800bps

===> new File: E:\Topfield\Testmovie.mpv

--> MPEG Audio (0xC9) on PID 0x478

2 PTS's recalculated in stream

Audio PTS: first packet 11:57:13.126, last packet 12:44:22.246

Video PTS: start 1.GOP 11:57:13.813, end last GOP 12:44:22.493

-> adjusting audio at video-timeline

=> src_audio: MPEG-1,Layer2,48000Hz,stereo,256kbps,noCRC @ 00:00:00.000

-> skipped sourceframe(s) @ 00:02:19.704

-> skipped sourceframe(s) @ 00:02:19.704

-> skipped sourceframe(s) @ 00:03:05.232

-> skipped sourceframe(s) @ 00:03:05.232

-> 6 frame(s) (144ms) inserted @ 00:13:21.504

-> 8 frame(s) (192ms) inserted @ 00:13:21.648

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

audio frames: wri/pre/skip/ins/add 33410/0/11/14/0 @ 00:13:21.840 done..

===> new File: E:\Topfield\Testmovie.mp2

--> Teletext on PID 0x47B

--> MPEG Audio (0xCA) on PID 0x479

1 PTS's recalculated in stream

Audio PTS: first packet 11:57:13.150, last packet 12:44:22.150

Video PTS: start 1.GOP 11:57:13.813, end last GOP 12:44:22.493

-> adjusting audio at video-timeline

=> src_audio: MPEG-1,Layer2,48000Hz,stereo,256kbps,noCRC @ 00:00:00.000

-> skipped sourceframe(s) @ 00:02:19.704

-> skipped sourceframe(s) @ 00:13:21.504

-> skipped sourceframe(s) @ 00:13:21.504

-> 14 frame(s) (336ms) inserted @ 00:13:21.504

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

audio frames: wri/pre/skip/ins/add 33410/0/8/14/0 @ 00:13:21.840 done..

===> new File: E:\Topfield\Testmovie_1.mp2

--> MPEG Audio (0xC7) on PID 0x476

1 PTS's recalculated in stream

Audio PTS: first packet 11:57:13.198, last packet 12:44:22.198

Video PTS: start 1.GOP 11:57:13.813, end last GOP 12:44:22.493

-> adjusting audio at video-timeline

=> src_audio: MPEG-1,Layer2,48000Hz,stereo,256kbps,noCRC @ 00:00:00.000

-> skipped sourceframe(s) @ 00:02:19.704

-> skipped sourceframe(s) @ 00:02:19.704

-> skipped sourceframe(s) @ 00:02:19.704

-> skipped sourceframe(s) @ 00:03:05.232

-> skipped sourceframe(s) @ 00:03:05.232

-> skipped sourceframe(s) @ 00:13:21.504

-> skipped sourceframe(s) @ 00:13:21.504

-> 14 frame(s) (336ms) inserted @ 00:13:21.504

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

audio frames: wri/pre/skip/ins/add 33410/0/10/14/0 @ 00:13:21.840 done..

===> new File: E:\Topfield\Testmovie_2.mp2

--> MPEG Audio (0xC8) on PID 0x477

1 PTS's recalculated in stream

Audio PTS: first packet 11:57:13.222, last packet 12:44:22.222

Video PTS: start 1.GOP 11:57:13.813, end last GOP 12:44:22.493

-> adjusting audio at video-timeline

=> src_audio: MPEG-1,Layer2,48000Hz,stereo,256kbps,noCRC @ 00:00:00.000

-> skipped sourceframe(s) @ 00:02:19.704

-> skipped sourceframe(s) @ 00:13:21.504

-> 14 frame(s) (336ms) inserted @ 00:13:21.504

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

-> skipped sourceframe(s) @ 00:13:21.840

audio frames: wri/pre/skip/ins/add 33410/0/7/14/0 @ 00:13:21.840 done..

===> new File: E:\Topfield\Testmovie_3.mp2



summary of created media files:

.Video: 46791 Frames 00:31:11.640 E:\Topfield\Testmovie.mpv

Audio 0: 33410 Frames 00:13:21.840 0/11/14/0 E:\Topfield\Testmovie.mp2

Audio 1: 33410 Frames 00:13:21.840 0/8/14/0 E:\Topfield\Testmovie_1.mp2

Audio 2: 33410 Frames 00:13:21.840 0/10/14/0 E:\Topfield\Testmovie_2.mp2

Audio 3: 33410 Frames 00:13:21.840 0/7/14/0 E:\Topfield\Testmovie_3.mp2

=> 1314367241 bytes written...

[Oscar]

Ich habe jezt probiert die Recording zu demuxen nachdem ich es in vob-Format mit TFget exportiert hatte statt das raw Format. Jetzt hat das demuxen gut gelaufen. Der audio und video file habben dieselben Laenge. Dieses Problem bezieht also nur auf dem raw Format.

Oscar

[dvb.matt]

@Oscar
you may write in english/nederlands, if that's better for you...

es gibt Welle/Streife in image
will be reduced with an available enhancement.

auf schnellere weise das Bild zu erstellen
we're still awaiting an answer of a developer from the old X forums

ein kleineres unit time step wure geben damit mann ein bischen genauer die Schnittpunkte
using X, we can only cut at GOP boundaries, thats the limit.

--------
probably there's a problem with a big jump in PTS,
maybe caused by a manual/automatic pre-cut on TF ? (on a bad 'place')

cause:
'big' PTS difference in one or more packets:
47721858ms *90 = 4294967220 PTS cycles = 0xFFFFFFB4
(PTS resetted by broadcaster?)

dropping gop(sequence) @ orig.PTS 12:10:38.213 / new Timecode 00:13:21.840
GOP#1669, Pics expected/count 12/5 ; PTS difference 47721858ms (should be 0)
dropping useless B-Frames @ GOP#1670 / new Timecode 00:13:21.840
...
dropping gop(sequence) @ orig.PTS 01:27:07.871 / new Timecode 00:14:29.080
timeline PTS must be later, not earlier, so Audio stops here at 00:13:21.840
...
some packets are out of sequence (shifting..) (pre-cut)
...
dropping gop(sequence) @ orig.PTS 12:28:04.973 / new Timecode 00:14:58.280
GOP#1872, Pics expected/count 12/10 ; PTS difference 0ms (should be 0)
seems to be the original (continued) timeline PTS again, but cannot handled (Audio already stopped).

also for each Audio:
1 PTS's recalculated in stream
confirm one 'bad' PTS in stream data.

conclusion:
(1) known problem, see "Problemfälle, BR/WDR..."
(2) bad cluster copy from TF?
(3) workaround, using global PTS shift (spec), probably does not help
(4) workaround, removing the bad packets, seems to be difficult, manually or automatically
(5) filetype is irrelevant

using X, I guess there is no handy way to solve this prob,
except in binary modification of input stream data.


bye

[Oscar]

@dvb.matt

I hope the member who adds to a fast processing of the preview image during scrolling through the recording will respond to the thread. In fact it's the main issue that I would wish (besides a larger preview window) at the moment. I've searched upon this subject within the old forum but couldn't find any thread concerned.

Concerning the too small audio file problem I can say that I have imported the raw-file both by TFGet and DVR-topA and it both cases the audio file length remained the same small size (13 minutes) . There were also no PRE-cuts performed at the TF. I can't agree to the point that the format is irrelevant because as I said when the file is exported in vob format instead of the raw-format, X demuxes the file properly and the audio file(s) have the same length at the end. So, X must have another approach to such a particular situation as this one when the original file is in vob format rather than raw format. Anyway I am glad that X managed to demux it at the end.

Oscar

[dvb.matt]

hope the member who adds to a fast processing of the preview
hmm, seems, that thread is not transferred from the old forum (http://www.nv-systems.net/dvbforum/viewtopic.php?p=235#235)location..

AFAIK he currently works on other things.. (not X)
I don't know of any success, because he didn't published the code.

I can't agree to the point that the format is irrelevant because as I said when the file is exported in vob format instead of the raw-format, X demuxes the file properly and the audio file(s) have the same length at the end
well, but the difference in TF.exe is the modification of PTS when creating .vob (maybe in that special case better than X and DVR)
that's why it works in that way.. :wink:

rgds

[Oscar]

Well from the link I saw that java.lang is the member who had ideas to accomplish this task. However he doesn't seem to be a member at the moment since his name is not present in the member list.

I understand that Java could be the problem for the quick image refresh and that it probably therefore could not produce the speed of DVR-topA but any optimalisation should be in place with respect to this subject since one needs to scroll through the entire recording to find the environment first to place a cut. In some cases one needs to add more than 10 cuts which takes too long. Anyway java.lang had a nice approach to this subject and I hope that he will work on it.

Oscar

[Daywalker]

Well from the link I saw that java.lang is the member who had ideas to accomplish this task. However he doesn't seem to be a member at the moment since his name is not present in the member list.


Java.lang is not working on X at the moment! He works on another project based on an MPEG editor. He will mod the editor to read .ts files from WinSTB recordings, set cutpoints and demux them. Don't know 100% but i think this editor will support MPEG-1 and MPEG-2 streams including AC-3, MPEG and LPCM audio-substreams too. In one month he will release a version [beta testing in the moment..].

Sorry for my bad english.....and the bad news....