POTC: Dead Man's Chest - R2 Pixelated Edition

[desta]

I just bought meself the new Pirates of The Caribbean dvd.. Dead Man's Chest (region 2). Sat down to watch it, and literally about half way through (right near the layer break) it pixelated for a couple frames. Thinking that I might have just bought a bad disc, I borrowed the same movie off someone, who had also just bought it. The exact same thing.

Frame #115453
http://img95.imageshack.us/img95/7383/115453si5.png

Frame #115454
http://img291.imageshack.us/img291/209/115454ej3.png

Frame #115455
http://img95.imageshack.us/img95/867/115455ci6.png

Frame #115456
http://img291.imageshack.us/img291/4392/115456ol8.png


Thought it was pretty bad that they would put out a retail with such an obvious flaw, to be honest. I'm wondering (if anyone has bought it yet), is this something that only us lucky Region 2 folks have got, or is it on the R1 (etc) versions too?

[Mug Funky]

i've only seen the r4 rental, and it didn't do it as far as i know (i didn't watch it as critically as i could, but i'm sure i'd have seen that).

could you try find out if the title is open GOP? i have a theory... there may have been an edit done in authoring. if they were unlucky, the title would have been watched and signed off (i'm not sure of other authoring houses, but mine makes sure every encode is seen by someone, usually a few times), before the edit was made.

[setarip_old]

or is it on the R1The R1 version doesn't come out until December 08.

Quite honestly, I'm surprised you were able to spot something that lasts for only approximately 1/15th of a second (2 frames)...

[Chainmax]

Is is a pretty glaring artifact...

[Mug Funky]

actually 2/25 of a second... r2 = PAL (japan excepted).

hehe.. the AV testers here are expected to spot smaller glitches that last 1 field (beta SP dropouts do this). of course they often miss them which is why things are watched several times (once for every track, including subtitles).

i think this is a splice on an open GOP - being 2 frames long and on a scene change seems to imply this.

for any australians: SBS seem to encode their station IDs and program information with an open first GOP, and they'll produce a similar artefact when they're put immediately after an ad. you can see this in their analog broadcast too.

[mpucoder]

That certainly looks like what happens when an edit is made on an open GOP. Setting the "broken" bit in the GOP header should allow the decoder to skip the two frames (and freeze slightly). Or maybe the broken bit is already set and your decoder ignored it.

[desta]

I decrypted this to my hdd, as I was going to do an encode from it. Had to run it through FixVTS, because it uses Ripguard protection.

Anyway, as people were asking if it was open GOP or not, I started up PgcDemux and loaded the ifo. When I decrypted it earlier, there were 8 vobs. When I loaded the ifo in PgcDM, and chose 'by VOB id', it only detected 4. [On a side note, I actually doublechecked this by loading the ifo in pgcdemux, straight from the dvd - this time it found 8, but the first 4 all started before the first second of the movie. The remaning 4 had the same starting times as the decrypted version.]

I demuxed each separately, and going by the screens I posted above, VID03 finishes on frame 115453, and VID04 starts on 115456. The two pixelated frames are nowhere to be found.

And yes, after checking, they are open GOPs.