Hi, Guido Fiala wrote: >>> I observed that marks.vdr seems not to be compatible between vdr 1.2.6 >>> and 1.3.44, in 1.3.44 the marks seem to be behind their correct position >>> about a few minutes. >> Do you really mean minutes? > > Yes, of course the error is cummulative, the first mark is off about 5 minutes > (full length of advertising), the next more and so on. > >>> (i'am using analogtv plugin for recordings, not sat) >>> The following applies: (ok= marks in correct position, not ok= marks >>> wrong) >>> >>> Recording made with 1.2.6, played with 1.2.6 = ok >>> Recording made with 1.3.44, played with 1.2.6 = ok >>> Recording made with 1.2.6, played with 1.3.44 = not ok >>> Recording made with 1.3.44, played with 1.3.44 = not ok >>> >>> So the error/change may not be in noad nor in analogtv - or is my logic >>> wrong? >>> >>> (used noad 0.71 but was same with older noad, with 1.2.6 i have a much >>> older analogtv plugin than with 1.3.44 but the recordings are >>> interoperable so i assume no relevant change and it might not a failure >>> of noad either) >> Would you please explain, how you did check the correctness? > > I made recordings and let noad run and the cutting marks did not fit under > 1.3.44, then i stopped vdr and started 1.2.6 and checked the same recording > without touching the marks.vdr file and now the marks are at the correct > position. The same procedure the other way around... > >> E. g. do you see a different image, when you position to the same >> cutting mark under 1.2.6 respectively 1.3.44? > > Yes, different point in video-stream while viewing the progressbar, it seems > the marks are moved to the right in the progressbar under 1.3.44 or the > "actual position point" is to the left. > > How can i further track down this? I've just recorded two DVB-S samples with one commercial break, one with VDR-1.2.6 and one with VDR-1.3.44 and let noad-0.6.0 create marks.vdr. Besides two false positives, the cutting marks for the commercial breaks were correct for both recordings in any of the above four test cases. Bye. -- Dipl.-Inform. (FH) Reinhard Nissl mailto:rnissl@xxxxxx
