Log in

View Full Version : Finding Nemo: How did you beat this beast?


brashquido
22nd November 2005, 01:49
I've got Finding Nemo which is a region 4 DVD, and I have long been trying to make as close to a full backup as I can of this disc but am having trouble in trying to preserve the video quality of the main feature. I like to listen to movies in DTS when it is available, so I'd rather not remove that if I can avoid it. This is what I have done to date to try and get more space for the main feature;

- Used Voblanker to remove all video (replaced with stills) and audio from menu
- Set DVD-Rebuilder to resize all 56 Vob-ID's to half size, and then ran prepare
- Used RB-Opt to remove half size setting for the AVS script for Vob-ID 25~30 which is the main feature.
- Used RB-Opt to group extra and main feature data seperately
- Used RB-Opt to set extras bitrate group to 750, with main feature group set to autosize
- Set each RP-Opt group to use VBR

All this has resulted in the average bitrate for the main feature increasing from 2500, to just under 4000 which is around 75% of the original source. Through all my backups I've done, I would consider this to be a pretty high bitrate value for the main feature. With Nemo though there is still very noticable image artifacts (halo and pixelisation) at certain points in the movie where there is a lot of movement on the screen. This must be one bandwidth hungry movie, as I have set the extras down to about as low as they will go. I must admin though that I have not experimented with CCE settings at all as yet.

Apart from stripping things out, what can I do get the quality better in the main feature (especially in fast moving sections)?

DK
22nd November 2005, 02:26
could you please specify any such scene or better provide a screenshot

i cannot remember my backup of finding nemo to look bad at all

brashquido
22nd November 2005, 02:53
I will do as soon as I get home...

Zeul
22nd November 2005, 08:28
try using the original matrix

FredThompson
22nd November 2005, 11:11
Use the HC encoder.

jdobbs
22nd November 2005, 12:20
Just MHO -- but DTS really doesn't give you much for all that extra bandwidth. You are trading video quality for it.

manono
22nd November 2005, 14:35
Not just your opinion, jdobbs. DTS is only minutely better than AC3. It's mostly just louder, and certainly not worth the space it consumes when you're trying to compress a difficult movie down to a DVD5.

OvERaCiD23
22nd November 2005, 17:54
Strip all the extras and you should be fine in terms of video quality. It's what, a 1.5 hour movie? You should be able to keep the DTS and DD tracks and still have a fairly high bitrate. It's a compromise either way: drop the extras, drop the DTS track, or have bad video quality.

CWR03
22nd November 2005, 20:24
With a disk containing lots of extras such as Finding Nemo, I burn only the main movie on one disk, all the extras and menus on a second. Many special edition DVD's come that way anyway, so I'm accustomed to switching disks or drives to view the extras.

brashquido
23rd November 2005, 02:42
I might try the matrix idea suggested by Zeul. I've looked at the matrix editor before, but have never actually used it. I know each movie is different, but I've just never had problems with artifacts in a movie with a 4000kbps bitrate before. Nemo also seems to have a different structure from anything I've seen before, and DVD-RB doesn't seem to pick up the extras. That is why I set DVD-RB to prepare everything at half size, and then used RB-Opt to remove the resize command in the AVS script for the main feature.

jdobbs
23rd November 2005, 13:55
That does seem unusual at such a high bitrate. What was the maximum bitrate setting detected? Also, it's possible the pixelation was there on the original (I've seen that before) -- have you checked that?

Rippraff
23rd November 2005, 23:43
I've never backuped Finding Nemo with DVD-RB but I remember this movie is really a bitch (http://www.doom9.org/index.html?/mpg/Eyes_Duplicate_PGC_Guide_to_Finding_Nemo.htm) . Tried it with BIG3 two years ago and finally gave up and made just a movie only backup with Shrink.
Would be interesting to try it again with RB.

Cu Rippraff

manono
24th November 2005, 01:40
Hi-

I have the R1 version and, like Rippraff, haven't tried it with DVD-RB. I messed with it a couple of years ago for AVI. It has an unusual structure. There's a 100 minute main movie, but it has angles. You have the option of playing the movie alone, or of playing it with video excerpts, which kick you out of the movie from time to time to these demonstrations of how the movie was created. They total 36 minutes. There are 3 PGCs for this, the 100 minute movie, the 36 minute video demo which you can play separately, or the 136 minute PGC for the whole thing. The 100 minute movie alone, minus the extra audio, fits on a DVD5.

I have a feeling, but no evidence since I haven't done it myself, that DVD-RB is choosing the longest PGC as the main movie, but giving only the video demo the good bitrate, while the movie itself, since it's on a shorter PGC, is getting the shaft. I could easily be wrong about that. The R1 DVD itself, except for banding in some of the ocean scenes, is of very good quality, and has no pixelation. I may have not read carefully enough, but if brashquido is using the free DVD-RB, I don't think there's any way to do this successfully.

brashquido
24th November 2005, 03:22
@jdobbs

I'm at work now, and can't remember what the maximum detected bitrate was. I will post back tonight tonight. I took a screen dump of a few sections of the original as well as the re-encoded video (using the original matricies), and there are several sections where the re-encoded video has some servere macroblocking going on. All of the macroblocking is visable on segments of video where there is movement in the entire screen, and is generally moving quickly. I'll see about posting those screenshots tonight. If I am using the original matrices for the re-encode and there is still some servere macroblocking, then wouldn't that indicate a bandwidth issue (or lack of it)? I'll run a few files through a bitrate analyser tonight to see if the video is actually being encoded at the rate it is set to be.

@Rippraff & manono
Duplicate PGC's? Never seen that before, no wonder the structure looked weird. I'm using DVD-RB Pro 1.02. It seems as if DVD-RB does not detect the extras at all as when I do the prepare there is nothing shown in the segment editor, and if I use the matrix editor tool for DVD-RB and untick the main feature box in the preview section there are no files left in the list. I was starting to think it might be a CCE issue, but something like this would back more sense as CCE has never missed a beat for me before.

manono
24th November 2005, 06:06
Hi-

Here's what the R1 Finding Nemo looks like in DVD Decrypter's IFO Mode:

http://img514.imageshack.us/img514/9310/unusualdvd4up.jpg

PGC 1 is the movie, PGC 3 the movie + the video demo, and PGC 4 just the video demo. Thinking about it some more, maybe even the Pro version might not be able to do it right, as the shorter PGC 1 is the movie that you want to keep the highest quality. But others would know more about it than I. And I don't know if your R4 version, brashquido, looks the same. It should, I think, as DVD Decrypter says it's region coded for both R1 and R4.

EDIT later: Looking around the forum, I see that SAPSTAR has a similar thread about being able to choose which PGC(s) is considered the main movie, and thus gets the better quality:

http://forum.doom9.org/showthread.php?t=103146

jdobbs
24th November 2005, 06:26
DVD-RB selects the longest PGC of the longest VTS... so you're right, it will take PGC_3. I'll see if I can figure out an easy way to redirect it when needed.

blutach
24th November 2005, 07:05
I did R4 Nemo at 100%, but killed all extras. No need for RB at all.

Just MHO, but how many times do you wanna watch the extras?

Regards

spyhawk
24th November 2005, 07:19
Without having the disc to look at, I think PGC 3 must reuse vobid in PGC 1 & 4, thus it has the longest time. This (vobid), besides length of time, could be the determining factor in selecting PGC 1 instead.

brashquido
24th November 2005, 11:31
Just looked at Nemo R4, and it has a different PGC structure as shown in manono's screendump earlier, but the same in that PGC1 is the main feature but isn't the longest PGC...

http://beta.it-hq.org/uploads/nemo.gif