View Full Version : HC Encoding with autogop on FILM source
scraper
20th January 2006, 23:52
Hi all,
I just encoded a FILM-source with HC encoder (which is absolutely great by the way, did i ever mention that?) with autogop turned on.
It gave me a max goplength op 15, which i found weird, being used to a max goplength of 12 for film-sources. But checking this forum I found some interesting discussions about it being possible to pulldown a film-source with goplength 15, provided the encoder is aware of some limitations.
So I tried applying pulldown. Got no messages there, but after multiplexing with muxman i recieved this message:
Largest GOP of 38 fields exceeds DVD specification of 36.
Resulting DVD is non-standard.
So here are my questions:
Should I expect problems when playing this dvd?
Am i doing something wrong with pulldown?
Or is HCencoder (i'm using the latest version available from http://hcencoder.bitburners.com/) not able to produce streams from a film-source which can "legally" be pulled down when auto-gop is on?
If not, is there a way to use max-goplength of 12 frames when using autogop?
Thanks in advance,
Jan Peter
hank315
21st January 2006, 00:47
Hm... strange, will PM you.
is there a way to use max-goplength of 12 frames when using autogop?No, that's not possible...
Trahald
23rd January 2006, 19:22
i usually set *gop 12 2 when encoding film source
@hank315
is it possible on ntsc / 23.976 source to use 12 as max? as stated once rffs are added for pulldown with 15 frame gop you get 38 fields.. 12 brings you to 30 fields which is legal.
hank315
23rd January 2006, 23:06
There's still one thing which puzzles me and that's why it should be set to 12.
The maximum GOP length is 15 for PAL and 18 for NTSC, counting the actual frames in a GOP, pulldown doesn't change that.
Quote by Brent Beyeler:
The GOP size is a count of physical frames in the bitstream and is independent of the display duration of the frames in the GOP.
DVD limits the GOP sizes to 18 for NTSC and 15 for PAL, but the count is actual encoded frames, not the display duration of those frames.
I'm absolutely not an expert on authoring, maybe there's someone whith more insight in this matter??
For the complete post by Brent Beyeler: http://forum.mainconcept.com/viewtopic.php?p=5891#5920
dragongodz
24th January 2006, 01:46
just my humble opinion.
The maximum GOP length is 15 for PAL and 18 for NTSC, counting the actual frames in a GOP, pulldown doesn't change that.
the problem is after pulldown you have to think of fields and not just frames. pulldown is repeating fields so thats what its seen as, even when the footage is really progressive. now the number of real fields of course doesnt change but it seems the repeated fields are counted as new fields. so if you have an 18 frame GOP, which would equate to 36 fields and then repeat/reuse some of those fields as pulldown does you go over the maximum limit for NTSC. so the original number of frames/fields needs to be real fields + repeated fields =< max fields.
as i said thats just my opinion. :)
Trahald
24th January 2006, 07:08
its supposed to equate to time based on 29.97/ 25/23.976 fps.. must be under .6 seconds display time. i'll round for ease of math
15 frames for pal is (16/25=).6 seconds
18 frames for ntsc is (18/30=).6 seconds
12 frames for film is (12/24 =).5 seconds
all fine
15 frames for film would be (15/24=)0.625 which goes over
here is a quote from mpucoder (from a different doom9.org thread)
There is a reason for the gop length restriction, and it has nothing to do with picture quality (arithmetic error accummulation) or any decoder related issues. The gop length restriction sets the maximum size of the most basic unit used to compose a VOBU. And VOBU's in turn have length restrictions as well of 0.4 to 1.0 second except the last of a cell, which may be as long as 1.2 second. Having gop's longer than 0.6 seconds (the difference between the shortest and longest allowable VOBU) would create situations where it is impossible to compose a compliant VOBU.
scraper
24th January 2006, 12:48
but wouldn't the max gopsize be 14 then?
14 / 23.976 = 0.58 secs
:confused:
And reading Brent's post would suggest 15 as a perfectly healthy amount of frames in a GOP, as long as there would not be a 9 frame gop following it, which itself was followed by a 15 frame gop. That would make it impossible to fit into a vobu of legal length.
I've check the structure of my mpeg-stream and such is not the case...
This would suggest that my DVD is still compliant and seeing there were no muxing problems I can just disregard the muxman message... Still haven't tried playing it on my player. Will have a go at that tonight.
Jan Peter
scraper
24th January 2006, 13:59
well... it would be, were it devidable by 3, right??
that's what a nice walk to store does for your mind... ;)
scraper
24th January 2006, 19:54
Well... just burned it and had a quick view of it on my pioneer-player. Looks like any other pulled down disc to me. I didn't have time to view the whole movie, but it didn't skip, stutter or lose av-sync, so for me it works dandy ;)
but... i'm still gonna leave it on a dvd-rw 'till i can watch the whole thing just to be sure of course :D
Hank, I noticed the avarage bitrate being below the specified wanted averege... i also noticed the movie is black&white with not a lot of movement so in my opinion it's just an easily compressible source (avg quant 1.050 seems to point in the same direction... or am i wrong?)
Anyway thanks for all your insights... will post back if i do happen to run into trouble on playback...
See you all later,
Jan Peter
:thanks:
hank315
24th January 2006, 21:49
I noticed the avarage bitrate being below the specified wanted averege... Your movie is 98 min, B&W, not a lot of movement.
That's indeed a candidate for saturation.
And an average quant of 1.05 surely points to that.
Mug Funky
27th January 2006, 06:44
on the too-many-fields thing, i'll just throw in that the spruce cards at work (yeah, the weird-closed-gop ones) will allow GOPs of 15 frame while auto-IVTC is enabled. this in theory can produce too-long GOPs, but Maestro doesn't seem to have a problem.
however, probably best to stick to 12 frame GOPs for FILM just in case. pehaps a "mild" warning can be added to the HC info box?
also, while we're on the info box, it can be good to use colours on the text so we know how severe the warnings are. for example, Maestro gives 2 kinds of warning for buffer underflow - a "green" one where it still allows you to compile the project (and there's no apparent playback difficulties), and a "red" one where it stops compiling and requires you to fix the problem before you can re-compile (nicely enough it doesn't tell you exactly where the spike is.. you have to go hunting for it. big time-waster there.).
so anyway, you could use green warning text if autogop is on and the source is 23.976, and you could have red warning text for any clear-cut DVD noncompliance issues (like DC prec = 11, or max bitrate > 9800, or even no max bitrate CQ...).
Largest GOP of 37 fields exceeds DVD specification of 36.
Resulting DVD is non-standard.
I just got the above error from muxman but the DVD played fine throughout. :confused:
Encoding was done using CCE SP with a GOP size of 15.
vBulletin® v3.8.11, Copyright ©2000-2026, vBulletin Solutions Inc.