Very strange PAL cadence...Please Help

[nfv]

Hello everyone,

What script would you recommend to bring this clip back to 23.976p?

http://www.mediafire.com/file/wcdepknwl7wp0ho/palclip.avi/file

The pattern appears to be 7 progressive and 7 interlaced.

Any help would be much appreciated.

Thanks!

[poisondeathray]

AVISource("palclip.avi")
AssumeTFF()
TFM(pp=0)
TDecimate(cycler=1, cycle=25)

[nfv]

Thanks just ran it on the whole episode and it worked great!

After the script it is now 24fps...not 23.976.

[poisondeathray]

Yes, because 25-1=24 . You're field matching , then deleting the 1 duplicate frame , out every 25

Did you need 23.976 for a reason (maybe DVD) ? If so you, can slow it down, along with the audio (or use audio editor to do the audio part if you want)

AssumeFPS(24000,1001, sync_audio=true)
SSRC(48000)

[Sharc]

Hello everyone,

http://www.mediafire.com/file/wcdepknwl7wp0ho/palclip.avi/file

The pattern appears to be 7 progressive and 7 interlaced.

I am seeing a pattern of 13 progressive and 12 interlaced, which is standard 'Euro-pulldown' for converting 24p film to 25i for PAL DVD. It's just a bit less common than speedup but preserves the original film length and audio pitch.
The script for inverse telecining is what poisondeathray has suggested.

[bruno321]

To remove the residual combing you may want to use Vinverse.

@poisondeathray: two questions. Does setting pp=0 result in any visible change? Or does it just save encoding time? Also, is it really necessary to AssumeTFF(), and if so, why, and how did you tell you had to use it?

[poisondeathray]

@poisondeathray: two questions. Does setting pp=0 result in any visible change? Or does it just save encoding time?


In his sample clip default default pp makes no difference. You'd be ok to leave it default. There are various tests for differences in avisynth if you wanted to double check. pp only make a difference if it detects combing. You can adjust the thresholds if you think it's not sensitive enough, or too sensitive

You'd have to check if some areas are combed on the rest of the movie, and whether or not pp makes a difference. If this is a representative sample, I suspect it makes no difference for entire DVD. People make suggestions on the forum as a starting point, but it's up to the user to check and customize to their needs

I never decide to use PP( or not) based on speed. It's a negligible difference unless you're using some slow clip2 mode and some comb is detected. PP is nice for "auto" handling combing sometimes, but it can also degrade the image if it mistakenly detects something as combing. eg. Maybe some curtain pattern ,or clothing pattern




Also, is it really necessary to AssumeTFF(), and if so, why, and how did you tell you had to use it?


You bet it does. Try removing it and you will see problems, combing, in some sections . If you then decide to deinterlace or pp or vinverse or anything you degrade it. This is sample clean in terms of fields, you just have to get the field order correct

You always start with basics.


Use AssumeBFF().SeparateFields() and step through. If you go foward/backwards jerky then it's the wrong field order.

Then Try AssumeTFF().SeparateFields() and step through, and you will see it goes forward through the fields. That's the correct field order

Each progressive frame has 2 fields (even /odd scan lines) taken in the same point in time, so you expect 2 duplicate fields in a row when you step through (just a slight up/down motion from the even/odd scan line offset), but the progression is forward, not back/forth. (There is the duplicate frame very 25 frames , but ignore that for now)

Sometimes there are changes to the pattern, tricky things like switching field orders, but in terms of the overall field order, you start with the basics . There is nothing like that on this one.

[bruno321]

Thanks for the clear and informative explanation.

[nfv]

Yes, because 25-1=24 . You're field matching , then deleting the 1 duplicate frame , out every 25

Did you need 23.976 for a reason (maybe DVD) ? If so you, can slow it down, along with the audio (or use audio editor to do the audio part if you want)

AssumeFPS(24000,1001, sync_audio=true)
SSRC(48000)

Correct, for a DVD. Usually I use the following for that:

AssumeFPS(23.976)

Then use a program to slow the audio to match.

[manono]

Or just keep it at 24fps and soft telecine off of that using DGPulldown. It saves reencoding and slightly degrading the audio.