DVD Inverse Telecine ?

[dolpsdw]

Hi, i am starting on restoring and enhance a dvd (that comes from old TV VHS series)

I am From Europe, the DVD is on PAL, but the scene seems to be Telecined ?

Find attached a sample of the Original Stream.

Here is my current Script, wich fails at IVTelecine it.

#Indexing
LoadPlugin("C:\PORT\dgdecnv2053\DGDecodeNV.dll")
DGSource("DVD1_1.dgi", rw=1350, rh=1080)  

#Detelecine "ScanType": "Interlaced", "ScanOrder": "TFF" -> Input Type 0
QTGMC(Preset="Very Slow", MatchPreset="Very Slow", MatchPreset2="Very Slow", NoisePreset="Slower", InputType=0, FPSDivisor=2, SourceMatch=3, Lossless=1, Denoiser="dfttest", Edithreads=3, showsettings=true)

Prefetch(4)

Cheers.

[jagabo]

It's a little hard to tell because there's so little motion. But it appears to be a field blended NTSC/PAL conversion:

QTGMC()
SRestore()

Next time include a medium speed panning shot.

[dolpsdw]

Thanks you for your fast response, will dig it more today and report back.

Tested

QTGMC()
SRestore()
---------------AND---------
QTGMC()
SRestore(frate=25)

Still interlaced, will attach a better sample.

After Reading more docs i activate pure video deinterlacer and it do the job

DGSource("DVD1_1.dgi", deinterlace =1, use_pf=true)

[Ennio]

My apologies to dolpsdw for hopping in; hope you don't mind.

...appears to be a field blended NTSC/PAL conversion

Can you explain a bit more about "field blended NTSC-->PAL"?
Reason I'm asking is I always wondered how they'd convert "real" 29.97fps (59.94i) NTSC to 25fps (50i) PAL

[Sharc]

Simple explanation:
Instead of skipping certain frames or fields for framerate reduction one can merge 2 successive frames or fields into 1 by blending (overlay) successive fields (or frames) into 1 'ghost' field (or frame). Blending gives a smoother playback impression than just skipping which produces jerkiness and may break the field cadence.
There exist other methods though - if NTSC->PAL conversion is required at all as PAL DVD players can usually play NTSC discs without issues.

Edit:
You may find the discussion here helpful:
https://forum.videohelp.com/threads/378461-Query-about-identifying-field-blending

[dolpsdw]

I hope this is a better sample, there is no much action scenes on this one.

[Sharc]

This is different from the sample1.
Sample2 is regular PAL 720x576i25.
No ghosts or blends, just blocking artefacts. Has it been digitized with a VHS/DVD recorder?

I would simplify the deinterlacing like

QTGMC(preset="fast",EZDenoise=3.0)  #50fps bob-deinterlace
#selecteven()   #uncomment this line for single-rate (25fps) deinterlacing
derainbow()  #reduces the rainbows

[dolpsdw]

History:

The samples are part of a spanish "muppets like" show.
The Original signal probably has scenes recorded in NTCS (some songs like sample1 and the first part of sample3) and scenes recorded in spain (probably in pal?, the ones that showcase a giant pink hedgehog)

The actual samples come from a digitalization made by spanish television that produce a nostalgic DVD series.

I want to end having a clean progressive footage.

My current plan is using the DGIndexNV deinterlacer only in frames detected interlaced by NVCUVID pure video, instead of deinterlace everithing.

DGSource("DVD1_1.dgi", deinterlace =1, use_pf=true)

And then QTGMC() with inputtype=2 to clean a bit, may derainbow can help there to.

Not sure if this is the best way to go.

Any help is welcomed.

Thanks.

[jagabo]

The second clip is live interlaced PAL video. You want to just use QTGMC. Don't use FPSDivisor=2. That will lose half the motion and make the video flickery.

You can use deblock_qed() to reduce the blockiness -- at the cost of some blurring. With interlaced sources use:

function Deblock_QED_i ( clip clp, int "quant1", int "quant2", int "aOff1", int "bOff1", int "aOff2", int "bOff2", int "uv" )
{
    quant1 = default( quant1, 24 ) # Strength of block edge deblocking
    quant2 = default( quant2, 26 ) # Strength of block internal deblocking

    aOff1 = default( aOff1, 1 ) # halfway "sensitivity" and halfway a strength modifier for borders
    aOff2 = default( aOff2, 1 ) # halfway "sensitivity" and halfway a strength modifier for block interiors
    bOff1 = default( bOff1, 2 ) # "sensitivity to detect blocking" for borders
    bOff2 = default( bOff2, 2 ) # "sensitivity to detect blocking" for block interiors

    uv    = default( uv, 3 )    # u=3 -> use proposed method for chroma deblocking
                                # u=2 -> no chroma deblocking at all (fastest method)
                                # u=1|-1 -> directly use chroma debl. from the normal|strong deblock()

    last=clp
    par=getparity()
    SeparateFields().PointResize(width,height)
    Deblock_QED(last, quant1, quant2, aOff1, aOff2, bOff1, bOff2, uv)
    AssumeFrameBased()
    SeparateFields()
    Merge(SelectEven(),SelectOdd())
    par ? AssumeTFF() : AssumeBFF()
    Weave() 
}

[dolpsdw]

Quick question,
I am looking at this sample4 file.

And after bobing it with QTGMC(preset="fast",EZDenoise=3.0) #50fps bob-deinterlace
*i like the 50 fps output

I see like some ghosting issues, (like what happen with sample 1)
Is this normal ?, can i do something with it?
Could this result be considered progresive ?

Thanks in advice