QTGMC in VapourSynth without 2x Frame Rate

[expanseFan]

Hello,
I want to deinterlace a video with QTGMC but I want to keep original framerate. QTGMC automatically doubles the framerate. In Avisynth, all I had to do was write SelectEven() after QTGMC.
How do I do that in VapourSynth? I tried clip = havsfunc.SelectEven() and other combinations but nothing works.
Could you help me please?

[Myrsloik]

clip = clip[::2]

[expanseFan]

Originally Posted by Myrsloik

clip = clip[::2]

Thank you VERY much!!!!! I can't believe it was this simple.

[Jukus]

Damn, avisynth.nl/index.php/QTGMC does not currently work.
Alright, you only need to use 'FPSDivisor=2'.

[Cary Knoop]

Originally Posted by expanseFan

I want to deinterlace a video with QTGMC but I want to keep original framerate.

If you want the original framerate you do nothing. Interlaced material is based on a field-based time resolution of half the frame's vertical resolution.

Originally Posted by expanseFan

QTGMC automatically doubles the framerate.

It doesn't!
It's you who wants to cut the framerate in half.

[expanseFan]

I'm using VapourSynth with StaxRip and when I enable the QTGMC filter, the framerate of the result video is shown as having double the frame-rate. I want to keep original framerate. How am I supposed to do nothing? I see it with my own eyes that the end result is double frame-rate!
Also, when using QTGMC in VapourSynth I noticed that CPU Utilization is ~55%. How can I make encoding faster? I was originally using Avisynth but it was also slow with low cpu utilization and though that VapourSynth was faster but as it turns out it isn't. Is QTGMC multi-threaded? My CPU is AMD Ryzen 6-core/12-threads.

[lansing]

Originally Posted by expanseFan

I'm using VapourSynth with StaxRip and when I enable the QTGMC filter, the framerate of the result video is shown as having double the frame-rate. I want to keep original framerate. How am I supposed to do nothing? I see it with my own eyes that the end result is double frame-rate!
Also, when using QTGMC in VapourSynth I noticed that CPU Utilization is ~55%. How can I make encoding faster? I was originally using Avisynth but it was also slow with low cpu utilization and though that VapourSynth was faster but as it turns out it isn't. Is QTGMC multi-threaded? My CPU is AMD Ryzen 6-core/12-threads.

If your interlaced video is 30i, then its original frame rate is almost certainly 60fps, so by reducing it to 30fps, you're throwing away half of the information.

[expanseFan]

Originally Posted by lansing

If your interlaced video is 30i, then its original frame rate is almost certainly 60fps, so by reducing it to 30fps, you're throwing away half of the information.

I didn't know that, thanks. One question though, why other deinterlacers like yadifmod if video was 30i, they deinterlace it and result in 30p instead of 60p?
If 30i is actually 60p, does that mean that videos were really filmed at 60fps and simply interlaced and reduced to 30fps?

[poisondeathray]

Originally Posted by expanseFan

I didn't know that, thanks. One question though, why other deinterlacers like yadifmod if video was 30i, they deinterlace it and result in 30p instead of 60p?
If 30i is actually 60p, does that mean that videos were really filmed at 60fps and simply interlaced and reduced to 30fps?

No, just the settings used for mode

####	mode -

		Controls double rate vs same rate output, and whether or not the spatial interlacing check is performed.

		0(default) = same rate, do spatial check
		1 = double rate, do spatial check
		2 = same rate, no spatial check
		3 = double rate, no spatial check

[expanseFan]

I want to convert an interlaced DVD to progressive. Should I convert it to progressive with double the frame rate or keep it at original frame rate?
If for example the dvd has a frame rate of ~29.9fps, should I let QTGMC give me a progressive video with ~59fps? If the DVD wasn't filmed at 59fps then, 59 fps wouldn't be true 59fps.

[poisondeathray]

Originally Posted by expanseFan

I want to convert an interlaced DVD to progressive. Should I convert it to progressive with double the frame rate or keep it at original frame rate?
If for example the dvd has a frame rate of ~29.9fps, should I let QTGMC give me a progressive video with ~59fps? If the DVD wasn't filmed at 59fps then, 59 fps wouldn't be true 59fps.

It depends on what the actual content is. Is it film "24p" content telecined, such as a theatrical "movie" ?

All DVD's are "interlaced", but they should not necessaily all be deinterlaced . In general, you should not deinterlace progressive content or you degrade it

[expanseFan]

Originally Posted by poisondeathray

It depends on what the actual content is. Is it film "24p" content telecined, such as a theatrical "movie" ?

All DVD's are "interlaced", but they should not necessaily all be deinterlaced . In general, you should not deinterlace progressive content or you degrade it

How do I learn to recognize if a video needs deinterlacing or not, if it also needs detelecine, etc? Any good tutorial, guide?

Video
ID : 1
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings : BVOP
Format settings, BVOP : Yes
Format settings, Matrix : Default
Format settings, GOP : M=3, N=13
Format settings, picture st : Field
Codec ID : V_MPEG2
Codec ID/Info : MPEG 1 or 2 Video
Bit rate mode : Constant
Bit rate : 5 000 kb/s
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 4:3
Original display aspect rat : 4:3
Frame rate mode : Constant
Frame rate : 29.970 (30000/1001) FPS
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan order : Top Field First
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.483
Time code source : Group of pictures header
GOP, Open/Closed : Open
Color primaries : BT.601 NTSC
Transfer characteristics : BT.601
Matrix coefficients : BT.601

[poisondeathray]

If you double rate deinterlace it to 59.94, then step through frame by frame , look at the pattern during a scene with motion (such as a camera pan)

1) If there is motion every frame, then it's interlaced content (and you would want to use 59.94p when deinterlacing)
2) If you get duplicates , then it's 29.97p content (then should not deinterlace - you want to field match to 29.97p)
3) If you get 3:2 pattern of repeats, then it's 23.976p content (then you should not deinterlace - it needs to be IVTC'ed)

There are exceptions, and weird cadences, changing patterns, but those 3 should cover the majority of common scenarios

[Cary Knoop]

Originally Posted by expanseFan

I didn't know that, thanks. One question though, why other deinterlacers like yadifmod if video was 30i, they deinterlace it and result in 30p instead of 60p?
If 30i is actually 60p, does that mean that videos were really filmed at 60fps and simply interlaced and reduced to 30fps?

A video camera (not a film camera) that records interlaced takes a capture each 1/50th (PAL) or 1/60th (NTSC) of a second.

The time resolution is 50 and 60, not 25 or 30!

If you deinterlace correctly you should reconstruct the missing scan lines instead of trying to get scanlines from one time capture to consolidate with one from a different time capture.

And yes many "deinterlacers" do it completely wrong!

[feisty2]

if the content is not native progressive (fake interlace like telecine), I suggest that you do not waste your time trying to process it, it is already garbage.

[expanseFan]

Thank you everyone for your help, I think I now understand a lot better about when to use deinterlace, detelecine and such. I'll try to watch frame by frame to determine what needs to be done.

[feisty2]

before the time that modern digital cameras like Arri Alexa were introduced, the image quality of any non-film based material is basically garbage. back then you gotta use film cameras like Arriflex 435 to shoot high quality videos. since footages shot on film are apparently progressive, that leaves all native interlaced content nothing but abysmal quality, which is not worth saving anyways.

[Cary Knoop]

Originally Posted by feisty2

before the time that modern digital cameras like Arri Alexa were introduced, the image quality of any non-film based material is basically garbage. back then you gotta use film cameras like Arriflex 435 to shoot high quality videos. since footages shot on film are apparently progressive, that leaves all native interlaced content nothing but abysmal quality, which is not worth saving anyways.

That's where I have to disagree with you. The quality may be dismal but quality is just one component. What is recorded may be interesting and valuable despite the bad quality.

So no, I disagree that some things are not worthy to be saved because of a bad quality.

[Jukus]

Originally Posted by expanseFan

Also, when using QTGMC in VapourSynth I noticed that CPU Utilization is ~55%. How can I make encoding faster? I was originally using Avisynth but it was also slow with low cpu utilization and though that VapourSynth was faster but as it turns out it isn't. Is QTGMC multi-threaded? My CPU is AMD Ryzen 6-core/12-threads.

Maybe ~ 55% is used for filtering, others ~ 45% for encoding?

[logicom]

check field order of the video and accordingly try

 haf.QTGMC(clip, Preset='Slower', FPSDivisor=2, TFF=False)

Keyword is FPSDivisor

Cheers!