View Full Version : PAL to NTSC, 25i to 29.976i (originally NTSC)
zerowalker
7th October 2015, 03:44
Okay this has been asked hundreds of times and i just can't find what i am supposed to do on this.
The video is PAL 25fps and it's origin is (i think) 29.976i NTSC.
So, i just want to convert PAL to NTSC as good as possible,
i am sure there will be lost frames from when it was converted in the first place and stuff like that which is not much that can be done about.
(As i think this video is one of the few where they didn't slow/speed up do get to PAL, but instead cut off the frames), meaning the video is "identical" in terms of speed.
So now there is this stutter effect that can happen as well so i am guessing it's ConvertFPS that needs to be used, but i just can't figure out if i can use it just like that or if i have to deinterlace first or something and then interlace it again etc?
The process is probably very simply and it's just me that is confused:P
Thanks:)
Boulder
7th October 2015, 04:11
What is your destination format? If it's DVD, you can re-encode at an NTSC resolution while retaining the framerate and then use DGPulldown to get to 29.97fps.
zerowalker
7th October 2015, 09:42
Yes it's DVD, i am not the one making it but merely helping with the clip.
Will have to ask what DVD Software he uses then.
What does DGPulldown do, isn't that something like telling the player to repeat certain frames to do a "virtual changefps()" thingy?
Music Fan
7th October 2015, 10:46
Yes. But it's generally used with progressive content, I wonder how it does handle interlaced videos. The problem being that you might see 2 fields repeated in a wrong order and it would look like this (A, B, C are frames divided in top and bottom fields) : Atf, Abf, Btf, Bbf, Btf, Bbf, Cbf, Ctf ...
Btf is not supposed to be displayed again after Bbf (if the video is TFF). This would be worse than the classical 3:2 pulldown stuttering.
Boulder
7th October 2015, 11:04
If you don't need to have the result as interlaced, it might be a good idea to test deinterlacing with QTGMC using the single-rate output mode and then running that through DGPulldown (after encoding to an NTSC resolution). IMO the single-rate output looks quite good and DGPulldown will produce less jittery motion compared to field blending etc.
Music Fan
7th October 2015, 13:37
One can also de-interlace to 50p, resize in 720.480, change fps to 60p, then interlace it.
It will probably be more fluid.
zerowalker
7th October 2015, 17:49
So i should use QTGMC(), then change fps to 60, then interlace that?
manolito
7th October 2015, 19:01
So i should use QTGMC(), then change fps to 60, then interlace that?
No, I don't think so...
If you deinterlace your source first then you should use DGPulldown to repeat fields as necessary to get NTSC frame rate.
If you want to keep things interlaced then the standard method is still this one:
PALtoNTSC_Interlaced.avs:
LoadCPlugin("...PATH...\Yadif.dll")
AssumeTFF() # Specify source field order
PointResize(width,height+8,0,-4,width,height+8) # Padding
Yadif(mode=1)
Crop(0,4,-0,-4)
Spline36Resize(720,480)
ChangeFPS(59.94)
AssumeTFF() # Specify target field order
SeparateFields()
SelectEvery(4, 0, 3)
Weave()
Change your bob deinterlacer and resizer as you prefer. The encoder must be set to "Interlaced" for this method.
Cheers
manolito
Music Fan
7th October 2015, 20:59
So i should use QTGMC(), then change fps to 60, then interlace that?
Yes, with the kind of script showed by manolito (replace Yadif by QTGMC).
But I don't know what PointResize is supposed to do in this script.
manolito
7th October 2015, 22:24
But I don't know what PointResize is supposed to do in this script.
This applies padding to the top and bottom as a workaround against the bobbing artifacts (thanks Didée). It is necessary for Yadif, no idea about QTGMC.
Cheers
manolito
creaothceann
7th October 2015, 22:34
>padding
Then why not AddBorders?
manolito
7th October 2015, 22:57
See here:
http://forum.doom9.org/showthread.php?p=1325630#post1325630
Music Fan
8th October 2015, 10:45
This applies padding to the top and bottom as a workaround against the bobbing artifacts (thanks Didée). It is necessary for Yadif, no idea about QTGMC.
It seems useful too with QTGMC ;
See here:
http://forum.doom9.org/showthread.php?p=1325630#post1325630
calming of the blinking scanlines "improved" Yadif's compressibility by 0.005%. { (14396-14320)/14396 }
Whereas TGMC improved compressibility by 48%. { (14320-7452)/14320 }
By the way, you removed the .001 included in Didée's script ;
mpeg2source(...)
pointresize(width,height+8,0,-4,width,height+8.001)
Yadif(mode=1).crop(0,4,-0,-4)
Is it useful for all de-interlacings or only in some cases ?
As creaothceann, I don't really understand the difference between this padding and addborders.
Boulder
8th October 2015, 12:07
To the OP: can you upload a short sample for us to test various options?
manolito
8th October 2015, 19:08
By the way, you removed the .001 included in Didée's script ;
You have to read further down in this thread. There was a discussion between Didée and Gavino with the result that this ".001" was not needed.
Cheers
manolito
Music Fan
8th October 2015, 21:02
Thanks, and what about the difference between this padding and addborders ?
manolito
8th October 2015, 21:59
Sorry, no idea...
Cheers
manolito
Music Fan
9th October 2015, 10:09
Ok, thanks.
zerowalker
9th October 2015, 17:12
Thanks indeed for the tips;)
vBulletin® v3.8.11, Copyright ©2000-2026, vBulletin Solutions Inc.