Why is the picture "shaky"?

[dazzle]

I've tried to encode a PAL movie consisting of 2 VTS-sets. The movie is in VTS01, which hold 9 PGC:s (lots of logos and stuff). Extras are in VTS02 and also consists of 9 PGC:s.

The main movie is encoded perfectly; progressive and top field are checked.

But there are something wrong when I encode the extras.
When I check the files they are interlaced (stripes in Dvd2Avi or in Win Mediaplayer).

The first time around I checked Top Field and Alternate Scan as described in the Guide. But when I later checked on my standalone (Philips 963SA) all the extras had that "shaky" and unstable look.

I decided to try to encode once more, and discovered that Bitrate Viewer said "interlaced and zig zag".

Great I thought, that must have been the problem, and checked only Top Field (leaving the rest blank).

But the same problem re-occured...:(

So what am I doing wrong?

[jel]

hi dazzle and welcome to the forums!
the problem you are having is of incorrect field order. if you run a search in this forum, cce forum, dvd burning forum etc, you will find lots of questions regarding top field first (tff). so have a bit of a read to gain a clearer (?!?!?) understanding - it is also discussed in the cce faq that RB has written.
to solve your problem though find a programe called 'restream' - just google it - and invert field order on your .mpv stream. this will save you the hassle of re-encoding your .m2v streams again.
let me know if that solves your 'shakiness' :D
cheers
j
oh and i am also in PAL land and as a general rule (and others will disagree) i always encode interlaced material with 'alternate scan' selected in DoCCE4U only. i also up the max. bitrate to 1950 - 2100 as interlaced sources do not compress as well as progressive.

[dazzle]

Thanks for the reply Jel!
I already deleted the files though...:(

But if I decide to recode the film again, I should NOT check the Top Field First I guess?

Good thing there are programs like Restream though, for future mistakes :)

About the scanning order; yes usually I do Alternate Scan with Interlaced sources. But since it didn't work I tried the Zig Zag.
...but of course the problem was the last check box...Top Field :)

Oh, and by the way: I use CCE 2.50.

[idbirch2]

I hope that solution works for you but I wouldn't go applying those settings to all your films. I only encode PAL movies and have found that extras or episodic DVDs are always interlaced. Every movie I have done, without excepetion, has used the settings you mentioned in your first post (Progressive unchecked and TFF checked)for interacerd material and its always worked. Either something is wrong with your setup or this is a one-off movie with a different field order.

[dazzle]

Yeah, hopefully this is a "one-of-a-kind-movie".
I have the "Invert file names..." checked in DIF4U.

So the only thing I haven't tried is TFF unchecked.
So my guess is the setting would be:

Everything unchecked (or maybe check Alt. Scan, but since Bitrate Viewer reported Zig Zag maybe it actually is).

Is this right you think?

[idbirch2]

Yep, from what you've said already it sounds like they are the right settings to use but in 99% of PAL extras you'll just want TFF checked and everything else unchecked.

[dazzle]

Thx mate!
I'll try this now, and se how I end up :)

And if it does't work, I could use the Restream program to fix this later on?

Oh, and by the way:
I've seen this exact problem on a dvd that a friend encoded:
Bruce Almighty PAL.

[jel]

@ idbirch
????:confused: ????
hmmm...now i am confused.... i have always NOT selected 'TFF' in DoCCE4U and have never noticed wrong field order issues.
from now on i will try the methods discussed for detecting TFF/BFF (cant find the post right now but i know its somewhere ;) ) prior to adjusting any settings.
cheers
j

[dazzle]

from now on i will try the methods discussed for detecting TFF/BFF (cant find the post right now but i know its somewhere ) prior to adjusting any settings.

If you find, please post the link here as well, since I'm very intrested in how you can detect that. :)

[idbirch2]

@jel

If you've left TFF unchecked and the output looks OK on a TV then you've obviously used the right settings. All I know is that from all the UK PAL DVDs I've done these settings have worked every time. I set DoItFast4U to use frame diagnosis to decide whether material is interlaced. Once BatchCCEws opens, anything with an "-I" in the file name indicates interlaced and for these I uncheck progressive and check TFF.

[jel]

ahhhhh the wonders of multi-region dvds eh ? :D
so far with all of my interlaced dvds - r4 PAL - i have found that the settings i used have given acceptable to excellent results (depending on the bitrate/passes used).
i actually dont own any uk dvds (i did import the father ted series but it ended up being cheaper getting them from america - go figure?!?!?!-usually the majority of uk tv series i like get released here in oz anyway) so i dont have anything to compare it to.
.....makes it confusing though when you are trying to help somebody with encoding issues ... :rolleyes:

[dazzle]

Just want to let you know, that it worked flawless this time around.
The TopFieldFirst was the problem.

I didn't check anything for the extras this time around (since Bitrate Viewer reported Zig Zag scanning order).

Thanx for all your help!

If now I could only find an easy way to see if the source is TTF or BFF _before_ I start encoding...:D

[idbirch2]

Frame Diagnosis should detect field order correcly, seems to work fine for me. Make sure you have it enabled in DoItFast4U > Global Options > Other Options. If I remember right Frame Diagnosis is best for PAL material and Header Flags is best for NTSC.

[dazzle]

Aha, I didn't have them checked.
Will do that next time then.

And if you're not remembering right...let me know, okay. ;)

[idbirch2]

Sorry about the uncertainty there, I checked the DoItFast4U help file:

Frame Diagnosis: Uses more complex diagnosis routines developed by Hard Code (the same author of newest version of pulldown.exe) to determine the true field order of each video file

The help file also describes the Header Flags option as "Generally good enough for PAL" but I can remember using that and it mis-reports interlaced material as progressive, especially with PAL extras.

Stick to Frame Diagnosis with PAL material.

[dazzle]

Thanx m8!
I will try that the next time around. And for ntsc, you don't need that at all if you use "Auto Detect NTSC" or..?

[idbirch2]

Sorry I don't understand what you're getting at.

And for ntsc, you don't need that at all if you use "Auto Detect NTSC" or..?

Don't need what? Frame Disgnosis? If thats what you mean I don't know I have never encoeded any NTSC sources so have never touched the "Auto Detect NTSC" option.

[dazzle]

Yepp, that was what I meant. But I've done a few without trouble, so my guess is I don't need it for NTSC.
:)

[dazzle]

Okay Top-Field Masters...I have one more question :D

I've now selected Frame Diagnosis (and the Invert Field Order in Filenames) in DIF4U.

In DoCCE4U it says:
Added job VTS__01_P02.I-BFF.16~9_1

Does this mean that it's TFF, but because of the Invert-option it's now named BFF?

Or is it actually BFF?

And finally, should or should I not check the TTF box in a situation like this?

All the Inverts and stuff made me confused :D

[Wensleydale]

Hey you guys,

I've also come to the conclusion that episodiscs and extras (PAL) are interlaced, hence the checking of Alternate scan in docc4u. But what I would like to know is if I also need to check 2:3 pulldown(?) I've only made one disc after my discovery and that came out ok but perhaps I was lucky?

[dazzle]

Pulldown is only needed for NTSC films.

[Wensleydale]

Ahhh guess I should stop checking that box then! :D

Thnx!

[dazzle]

Okay Top-Field Masters...I have one more question:

I've now selected Frame Diagnosis (and the Invert Field Order in Filenames) in DIF4U.

In DoCCE4U it says:
Added job VTS__01_P02.I-BFF.16~9_1

Does this mean that it's TFF, but because of the Invert-option it's now named BFF?

Or is it actually BFF?

And finally, should or should I not check the TTF box in a situation like this?

All the Inverts and stuff made me confused. :D

Okay, I've given this question some thought...
I previosly have used another CCE-based encoding program, and there the solution to the TTF or BFF issue was like this (here transfered to CCE 2.50 and CCE 2.67):

IF you use CCE 2.50 the TTF check box should NEVER be checked.
IF you use CCE 2.67 the TTF cehck box should be checked.

This since CCE operate this different in the 2 versions.
Am I right in stating this or...?

For progressive sources, it doesn't matter if you check it or not.

[idbirch2]

@dazzle

Yes, in this case you would uncheck Progressive (the 'I' in the filename tells you its interlaced) and check TFF for exactly the reason you mentioned - the filename is inverted. I think it inverts so that BatchCCEws sets the field for you but I've found this never works and you have to do it manually. You've got me on the CCE version thing though - I know they changed the syntax required to enable TFF between 2.50 and 2.67 but I don't know if that meant you had to reverse the TFF setting.

@Wensleydale

Can't remember the exact layout of DoCCE4U because I've been using BatchCCEws since it was released. For all the PAL DVDs I've done whether its extras or episodic, 'Alternate Scan' is not required. I've found that as long as 'TFF' is checked and everything else in unchecked it works every time (so far). I have done it with 'Alternate Scan' checked before but it made absolutely no difference to the output. Looked fine both ways.

[dazzle]

Originally posted by idbirch2
[B]@dazzle

Yes, in this case you would uncheck Progressive (the 'I' in the filename tells you its interlaced) and check TFF for exactly the reason you mentioned - the filename is inverted. I think it inverts so that BatchCCEws sets the field for you but I've found this never works and you have to do it manually. You've got me on the CCE version thing though - I know they changed the syntax required to enable TFF between 2.50 and 2.67 but I don't know if that meant you had to reverse the TFF setting.


Thanks idbirch2.
The only problem is that the videostream is progressive...but I guess I always have to check that manually anyway :)

But just to sort things out:

If the file is VTS__01_P02.I-BFF.16~9_1 (With the Inverse Field Order-box checked), would that mean that its actually TopFieldFirst in the original?

And if the file is VTS__01_P02.I-TFF.16~9_1 (With the Inverse Field Order-box checked), would that mean that the original is actually BottomFieldFirst?

[idbirch2]

The only problem is that the videostream is progressive...but I guess I always have to check that manually anyway

You mean there's an 'I' in the file name but the video stream is progressive? If so I may be full of crap and the I just means the file name has been inverted - not sure on that one now!



But just to sort things out:

If the file is VTS__01_P02.I-BFF.16~9_1 (With the Inverse Field Order-box checked), would that mean that its actually TopFieldFirst in the original?

And if the file is VTS__01_P02.I-TFF.16~9_1 (With the Inverse Field Order-box checked), would that mean that the original is actually BottomFieldFirst?

Yes, spot on.

[dazzle]

So the sum of all this should be:

If it says BFF in DoCCE, then Top Field First should be unchecked.
If it says TFF in DoCCE, then Top Field First should be checked.

Right? :D

[Trahald]

For use with cce 2.66 it works like this...
if your file is Bottom Frame First then you should have top frame first checked

and if its top frame first you should NOT check top frame first

keep in mind if you have invert field order option ON in dif4u then the file name will be the opposite of what it is.. ie I-BFF means you have a Top field first video.. so you should NOT check top field first. heres a little test avs script for before encoding


LoadPlugin("C:\Program Files\DoItFast4U\new.d2v\mpeg2dec3dg.dll")
Mpeg2Source("test.d2v",idct=0)
AssumeTFF().SeparateFields()
ConvertTOYUY2(interlaced=true)


(taken from a thread by RB and modified a tad)

put your d2v file name in there and run it through windows media player. if it studders then its Bottom Frame First (so you should Check Top Field First in batchccews) if its smooth then it truely is Top Field First and you should NOT check top field first. the cce faq explains why we have to work backwards.

its also a good test for your output before burning. since CCE's out put will ALWAYS be flagged as TFF, you can just make a .d2v for the video you made with cce (using dvd2avi , force film OFF) and plug that in the test script. if its smooth, then you did fine. if it jitters, you screwed up.. fortunately you can just run restream on your video to change it to BFF flagged.

With 2.67 its the opposite, you will set the flag as the name appears in the filename ie I-TFF means set Top Frame First in batchccews. I-BFF then you uncheck Top Frame First.

[dazzle]

Hehe...old thread waking to life :)

Yepp, that's true.

That is TFF should never be checked w 2.50 (unless it's actually bottom field first, which should be TFF in filename). With CCE 2.67 I think it's the opposite, since the whole matter is reworked.

This is only interesting for interlaced material anyway, since it doesn't really matter for progressive material (nothing gets screwed up that is), it just shifts the picture up/down 1 line.

[Trahald]

yeah.. Fmalibu pointed that out to me.. i was editing my post as you typed yours.. lol

anywho.. what a mess

i didnt test 2.50 but i'll trust you on that one (im getting a headache.. hehe)

[dazzle]

Hehe.
At least we're getting close to the goal now :)