Differences

This shows you the differences between two versions of the page.

--- c3tracker:repair [2023/03/15 17:01] – [Manually repairing a Recording (TS)] jwacalex
+++ c3tracker:repair [2025/09/14 12:38] (current) – v0tti
@@ Line 43: / Line 43: @@
 ffmpeg -i repair_rl.ts -filter_complex "channelsplit=channel_layout=stereo[l][r]" -c:v copy -c:a mp2 -b:a 192k -ar:a 48000 -map 0:0 -map "[l]" -map "[r]" -ac 2 repair_2chan.ts
 </code>
+=== Hints for simple .ts repairs with ffmpeg
+For basic repairs (for example, removing sections from the middle of a recording), you can use ''ffmpeg'' directly. This approach not only offers excellent performance but also preserves secondary video and audio tracks with minimal effort.
+First, determine the start and end timestamps of the sections you want to keep, relative to ''uncut.ts''. Then, export these sections as separate ''.ts'' files to your local machine (or the ''tmp'' mount on local event storage), for example:
+<code>
+ffmpeg -i uncut.ts -ss 00:02:46.20 -to 00:17:25.18 -map 0 -c copy part1.ts
+ffmpeg -i uncut.ts -ss 00:19:12.12 -to 00:52:27.23 -map 0 -c copy part2.ts
+</code>
+After that, you can easily combine the parts with ''cat'':
+<code>
+cat part1.ts part2.ts > combined.ts
+</code>
+Finally, repair the timestamps with ''ffmpeg'':
+<code>
+ffmpeg -i combined.ts -c copy -map 0 repair.ts
+</code>
+Place this file in the repair directory of your current event and follow the steps as described above. With this method, there is no need to use KDEnlive or Shotcut for cutting anymore.
 == Hints for repairing / manual cutting with KDEnlive
@@ Line 77: / Line 103: @@
     -map 0:a -c:a mp2 -b:a 192k -ac:a 1 -ar:a 48000 \
     -f mpegts output.ts
+</code>
+=== Convert intro without audio to ts
+Input can be any file that can be read by ffmpeg. The intro needs a audio track in order to be compatible with the tracker pipeline.
+<code>
+ffmpeg -loop 1 -i 123.mp4 -ar 48000 -ac 1 -f s16le -i /dev/zero -ar 48000 -ac 1 -f s16le -i /dev/zero -ar 48000 -ac 1 -f s16le -i /dev/zero -ar 48000 -ac 1 -f s16le -i /dev/zero  -map 0:v -c:v mpeg2video -q:v 0 -aspect 16:9 -map 1:0 -map 2:0 -map 3:0 -map 4:0 -shortest -f mpegts 123.ts
 </code>
@@ Line 86: / Line 119: @@
     -f mpegts /video/tmp/$EVENT/repair/${TALKID}_repair.ts
 </code>können
+=== source replacement with remapped audio streams + av-delay
+<code>
+ffmpeg -i /video/fuse/38c3/stage_huff/838/uncut.ts -itsoffset 0.2 -i /video/fuse/38c3/stage_huff/838/uncut.ts -map 0:v -map 1:a:1 -map 1:a:2 -c copy -ss 2:23 -to 45:00 383-remuxed.ts
+</code>
+==Audio Stream Reihenfolge reparieren via Jobfile.xml
+Issue: Audio Spuren vertauscht, Translation + Main in falscher Reihenfolge
+Ausgangslagen: Tracker steht auf to cut, AudioSpuren sind vertauscht
+  - Sich selbst als encoding worker zuweisen
+  - Track nach Wiki Anleitung schneiden
+  - Warten bis Recording auf finalized steht
+  - TS- HD master mp4 aufrufen
+  - Job file runterladen
+  - Audio spuren in encoding teil vom Jobfile richtig einsortieren
+  - z.B. Audiospuren 1 und 2 sind vertauscht <code>-map '[audio1_mix]' -c:a:0 aac -b:a:0 128k -ar:a:0 48000 \
+  -map '[audio0_mix]' -c:a:1 aac -b:a:1 128k -ar:a:1 48000 \
+  -map '[audio2_mix]' -c:a:2 aac -b:a:2 128k -ar:a:2 48000 </code>
+  - Schnittmarken + AV delay macht der Tracker automatisch
+  - auf ein einem worker anmelden und auf voc user wechseln
+  - Angepasste encoding ffmpeg Zeile aus dem Jobfile ausführen
+  -
 == Audiospur bearbeiten