GOP standards for PAL and NTSC players?
 

according to DVD standard:
max gop for ntsc is 18
max gop for pal is 15
In reality, how long gop standalone players usually do accept?

Re: GOP player
 

For (S)VCD, you can go way over 100, and it will still play on most players :!:
But results are crap, because of the extremely long GOP.
For DVDs, I'm not sure, because most DVD authoring programs always stick to 15 or 18, and they just won't let you import the mpeg if it's not a standard GOP.

-kwag

Correct me if I am wrong.
Processing of GOP is happening with decoder buffer, and you actually specify the size of this buffer in multiplexing options which are 46K for VCD and 230 for SVCD.
In reality, DVD players memory buffer is much larger than these outdated specifications.
Maybe that's why standalone decoders chuck on long GOPs and the only thing is they want to tell them that allowed buffer in multiplex stream is larger?

Yes, it's all (mostly) dependant on the VBV size.
I'm pretty sure most DVD players will handle long GOPs. It's just the authoring programs that are too strick, and don't let us do it :twisted:
Eventually, when more DVD authoring programs are available, maybe some will be more "relaxed", and allow non-standard stuff, and then we'll be able to try all sorts of weird things :cool:

-kwag

:encore: ahha!!
so did anybody tested different video fuffer sizes in multiplexing and then on standalone players?

@kwag: I thought TMPGEnc's DVD Authoring Software allows to import Streams with KVCD GOP :?: :roll:

I don't remember that :!:
Have you tried it :?:

-kwag

Here's the thread...

http://www.kvcd.net/forum/viewtopic.php?t=3416

I forgot all about that :lol:
But I would follow the recommendations, and stick to a standard GOP size, so the KDVDs will play in every standalone player.
This way, the only thing different is the KVCD notch matrix, and even though it's a customized matrix, there was never a "fixed" matrix constraint for DVD compliancy. So KDVDs are 100% DVD compliant, if we stick with the max GOP size ;)

-kwag

Technical notes in B-Book specs for DVD Updated: July 21, 1996. It is written even there: "MP@ML has no GOP size restriction. In fact, the GOP() is considered to be an insignificant layer in MPEG-2. Instead the sequence() layer serves as the most important boundary in the generic MPEG sense. "
Maybe restrictions on video buffer size are more important?
What are they?

GOP size max 36 fields or 18 frames (NTSC)
max 30 fields or 15 frames (PAL)

http://www.mpeg.org/MPEG/DVD/Book_B/Video.html

-kwag

Lucky you :!:
You can do that because you are in NTSC area and you do not suffer too much from the limitation (18 insteed of 24 that we use generally for KVCD).
But in PAL, that's a complete different story ! Using 15 insteed of 25 lead to a big difference in file size ! :-(

What about using 18 for PAL ? As players handle this value for NTSC, 18 seems to be "not so out of the standard" even in PAL. What do you think ?

Someone should try that ASAP... I'd like to know too! 8)

It speaks for itself :arrow: its out of standard and so you have to decide for yourself what to choose.

My recommendation:
- IF breaking the standards, go and encode using 24max / KVCD specs. ... But this will result in a less range of player which will accept this. Cause 18 on PAL already breaks the Std. and so if you don't care about Calories at least choose someting deliciuos to eat ;-)
But maybe 18 on PAL gives a smaller risk?! And thats what I think you want to assume with your question.
Its the same thing If we use 480x480 on PAL cause its a std. NTSC resolution to get more compression.

- IF keeping Std. ... well you know the answer ... 100% Player compatibility but less compression.

So choose your weapon...

@ JellyGoose
This would end up in a new Compatibility chart where almost everybody should provide his result which would give us all an average ;-)

For sure it is ! But the player handle a GOP of 18 for NTSC resolutions. So they have the capabilities to cope with such GOP. 24 is other the standard in both NTSC and PAL. That's why it seems to me that 18 can be acceptable also for PAL. It's just an idea :-)

EDIT: you edited your post while I was typing mine :-).
You added :

Yes, tht's exactly what I wanted to suggest :-)

I don't mind converting all my movies in NTSC. I never do it for KVCD because the gain is not so important. But in the case of a KDVD...

Well it really gave me an advantage in cases of Movies which defenitely didn't fit one one CD by archiving a CQ at least more than 60.

What's your challange DialHot? :wink:
Trying to get as most movies fitting on one DVD-R as you can ... also to break "in that case of DVD-R" the Moviecount value.

Well, if you get a DVD Burner (I don't know if you already got one) the first time you'll keep trying the "burn on that baby as much as it supports" to keep even there pushing the limits.
And suddenly you think this DVD-R burning and encoding way will be my way to give a main focus to quality.
(Ok thats my little story, and also many of my encodings do base on captures)

In fact I do not plan to put more than one movie per DVD (except perhaps for a movie and it's sequel like Blair wich project I and II for instance).

But what I do currently is TV shows. I use to put one complete season of 24 ep (22 min each) on 4 KVCD. That make 4x810 = 3240 Mo and the quality is widely enought for that kind of program.

On a KDVD I will have 1000 Mo more. So I have two choices :
- use the standard GOP (15) and the 1000 Mo will be used by the loose in compression. The quality won't change a lot (even if shorter GOP = better quality)
- use the KVCD GOP (25) and my CQ_VBR will raise (actually I use 18, I'll can surely go up to 20). So the picture will be better.

By using a GOP of 18, I can find a solution in the middle of the two above.

Soon, very soon :-)

That is also why I do not want to put more than one movie per DVD. My idea in making KDVD is to go from [480*576 - MIN 64 - MAX 2300] I use for KVCD to [704*576 - MIN 300 - MAX 6000].

As we use the NOTCH on KDVD encodings its not (really) nedded to go above 5000kbits at max. I very often do lower the max in case of a letterboxed KDVD to 4000-5000kbit and high bitrate peaks still come out nice. But sometimes KDVD is a good choice where we can also focus on anamorph encodings and there 5000kbit is perfect.
300min, .... well in case of KDVD 704x576 I every time set the min to 800kbit, I just figured out last week when I was re-encoding LOTR2 that at the beginning when the picture is almost dark ... the very little Parts within the almost black screen where you can see fire do come out a little blocky .. a jump to min=800 fixed that.

In case of buying a DVD Burner choose one of a brand where Firmwareupdates are well and fast provided... this lets you enjoy using cheaper DVD Medias which will be supported without problems.

[quote="incredible"]That's true ! I didn't think about that.

Why in this case ? Anormorphic target means more picture to encode, so greater need in bandwith. So why to limit in this case is more benefit than when encoding in 4:3 ?

Oh no ! My long time computer experience (15 years) tells me to trust known, reputed brands. I will buy a Pioneer DVR-106.

I meant these values in case of my choosen bitrates: 4000-5000 in case of Letterboxed streams and not lower than 5000 in case of anamorph streams, ... based on the reason you just explained well above;-)

And that's what I meant! Get this one its awesome.
I got an "old" *lol* DVR-105 on my own ... it "eats" and handles almost every cheap media, well provided Firmware updates on PioneerModels BUT ... (in case of mine 105) not able to overburn in CD-R Mode maybe they fixed that on the 106 model now.
These Pioneer Models are very well supported by most apps (and updates) internal Burning capabilities, .... except VCDeasy.
And you can find them all over the world ... I do not know if they still implementate it in the newer G5's, but all Apple G4's came with Pioneer DVD Writing devices.

For me kvcd works at 704x576 mpeg1, gop 25 on a dvd. Authored with tmpgenc dvd author. I've just made the test. I've tried also 480x576 with dvd lab and it works too. 8) I can put easily 5 movies on a dvd. That's great :D

What standalone do you have Aragorn ?

I've got a philips dvd 955. I've bought it 3 years ago.

I think that GOP is related to processing power of decoder
(hardware or software) and also to video buffer size,
and this one is more important.
As it mentioned already, nowdays standalone players are capabale to porocess GOPs over 25 (of course the quality depends also on sequience IPB type frames), but players have certain limitations on video buffer size.
So the question is - how many frames buffer holds?
The reason I am asking this because I can specify video buffer size in encoding and also it is specified in multiplexing.
Reading the thread while getting buffer underruns:
http://thread.gmane.org/gmane.comp.v...yer.user/22839

I just want to say that I did my first try on DVD-RW yesterday and my pioneer DV454 accepts a KDVD done with my "old" videos I did for KVCD.

In other words : a MPEG1 352*240 GOP24 video authored with TMPGENC DVD Author is read without any problem !

I do not try yet with higher resolutions.

Note: TMPGENC Author took 40 minutes to make a DVD image of 583 Mb !!! Is that normal ?
(96% of the CPU was use doruing all the process - I have a P4 1.4Ghz).

Returning to the question: how large video buffer can be?

Current windoze encoders suggested 40KB for VCD, 112KB for SVCD, 224KB for DVD and mentioned that these values are minimum. How these arbitrary figures sneaked into encoding procedures and source code?

Video buffer size is closely connected to bitrate control and maximum GOP.

In regular TmpgEnc DVD Author as any other Authoring programm doesn't need that much cpu power, its more a HD power thing.
But as your KVCDs in regular come with a 44.1khz sound, TmpgEnc DVD Author got an automatic resample engine, this option is viewable at the sound input field ... so your sound will be resampled to 48khz.
And thats another issue: If the Quality by doing this could suffer. Cause if you go in with a 44.1khz 112kbit Audio this will be resamled at 48khz 112kbit (as I know) = Quality loss in audio. ?! We should figure that out.

Did you only choose one movie to burn on cd for testing or did you put several KVCD based tracks in one project??
EDIT: Yes you did as I re-read your post again ..

That's also what I was thinking. But with my 120GB/7200rpm/8Mo memory HD, that shouldn't be a problem :-)

Okay ! I didn't think about that. That's sure that for at least one of the track the sound was in 44.1 because it was a regular VCD (I made a DVD with 2 tracks : one regular VCD and a KVCD).

You can't resample a compressed audio. You have to decode/resample/reencode. And as mp2 isn't a lossless format, for sure the quality suffer a little.

I just answered to that above :-)

I know, but there are differences in resampling engines and so lets see if it would be better to let do besweet the decode/resample/reencode way or if it also will come out in a smae quality if TmpgEncDVD Author reencodes it.
But anyhow "decode/resample/reencode" a 112kbit or even a 128kbit, .... oh oh ... and I think 95% of all encoded KVCDs around the globe do got 128kbit or below that. And so I think in case of re-authoring already made KVCDs to DVD we should use Bewseet to "reencode" that existing mpeg2 to at least 160kbit and do a comparison. Thats what I meant. :)

As I already do not trust tmpgenc audio engine, I won't either trust the one in tmgenc author (I guess it's the same !).
So I think about using besweet for that. And you're surely right : it's a good idea to do the new file in 160. This way the loses will be lesser than if we reencode in 128.