QTGMC best quality settings for PAL SD Video Capture

[Johnnysh]

Hello Guys

ok a few questions can someone please advise the best script for QTGMC deinterlacing

first of all, there is NO other deinterlacer other than QTGMC that I have seen that can create great looking video play back on digital equipment, the playback from the original interlaced capture and newly created progressive QTGMC is like night and day

QTGMC converted video is NOT JERKRY / NO INTERLACE ARTIFACTS, plays back buttery smooth

when I convert interlaced to progressive on adobe premiere and every other transcoding software the result are rubbish compared to QTGMC ?

Ok so yes we know the original (CAMCORDER) capture at 10bit I have made has all the information stored and ok for archive, but sadly playback of these prores 422 HQ .mov files and any other INTERLACED video files on PC is rubbish and awful looking, because its interlaced and jerky, it looks horrible unless playing back out through my capture card via my pro JVC CRT monitor (analogue Svideo / Composite)


So now for some magic converting analogue interlaced captured video to playback and look correct on pc/web thus QTGMC, the result are far better than adobe premier or anything else I have seen or used converting interlaced to progressive (deinterlacing).

It may have something to do with QTGMC creating 50 frames, thus creating a really smooth playback ? I don't know ? but what I do know is the results are amazing and the type of results I would have liked to have seen from the original capture (interlaced pro res) when playing back on the pc!

A few questions please, QTGMC is very processor hungry taking several hours of conversion for my DV capture to deinterlace

I have a high end video card is there a CUDA version that works ?? I have seen KTGMC but sure how to create a script to work??

QTGMC is so SLOOOW even with the high end hardware I have, also I have a powerful video that is not being used by QTGMC when converting/rendering


Working currently with QTGMC I am trying to create a script file to output the very best acceptable quality for pc/web playback in H.264 from my original DV captures at 10 bit

Below is the source capture file, so we can see its

10bit Pro Res 422HQ
4:3 Ratio
constant Frame Rate
Frame rate : 25.000 FPS
Standard : PAL
Color space : YUV
Chroma subsampling : 4:2:2

720x576.... PAL SIZE

Below is the captured camcorder file

General
Complete name                            : C:\JERSEY\JERSEY TAPE 1 Final Interlaced.mov
Format                                   : MPEG-4
Format profile                           : QuickTime
Codec ID                                 : qt   2005.03 (qt  )
File size                                : 45.5 GiB
Duration                                 : 1 h 32 min
Overall bit rate mode                    : Variable
Overall bit rate                         : 70.3 Mb/s
Encoded date                             : UTC 2021-08-29 19:29:24
Tagged date                              : UTC 2021-08-29 19:33:53
Writing library                          : Apple QuickTime
TIM                                      : 00:00:00:00
TSC                                      : 25
TSZ                                      : 1

Video
ID                                       : 1
Format                                   : ProRes
Format version                           : Version 0
Format profile                           : 422 HQ
Codec ID                                 : apch
Duration                                 : 1 h 32 min
Bit rate mode                            : Variable
Bit rate                                 : 68.7 Mb/s
Width                                    : 720 pixels
Clean aperture width                     : 703 pixels
Height                                   : 576 pixels
Clean aperture height                    : 576 pixels
Display aspect ratio                     : 4:3
Clean aperture display aspect ratio      : 4:3
Frame rate mode                          : Constant
Frame rate                               : 25.000 FPS
Standard                                 : PAL
Color space                              : YUV
Chroma subsampling                       : 4:2:2
Scan type                                : Interlaced
Scan type, store method                  : Interleaved fields
Scan order                               : Top Field First
Bits/(Pixel*Frame)                       : 6.629
Stream size                              : 44.5 GiB (98%)
Writing library                          : adb0
Language                                 : English
Encoded date                             : UTC 2021-08-29 19:29:24
Tagged date                              : UTC 2021-08-29 19:29:24
Color primaries                          : BT.709
Transfer characteristics                 : BT.709
Matrix coefficients                      : BT.709

Audio
ID                                       : 2
Format                                   : PCM
Format settings                          : Little / Signed
Codec ID                                 : sowt
Duration                                 : 1 h 32 min
Bit rate mode                            : Constant
Bit rate                                 : 1 536 kb/s
Channel(s)                               : 2 channels
Channel layout                           : L R
Sampling rate                            : 48.0 kHz
Bit depth                                : 16 bits
Stream size                              : 1 018 MiB (2%)
Language                                 : English
Encoded date                             : UTC 2021-08-29 19:29:24
Tagged date                              : UTC 2021-08-29 19:29:24

Now when I create my QTGMC script and convert, the quality is good, but size ad aspect ratio seem to change ??

when I add various cropping and resizing I seem to loose video at the edge

I am working with PAL and PAL only so please do not confuse me about NTSC

Can some one please suggest a good ultra High quality QTGMC script to do what I want



This is my current AVIsynth script file with some items disabled for testing that I use in Virtualdub2

I am working with 10bit prores 422 files

SetFilterMTMode("QTGMC",2)

FFMPEGSource2("JERSEY TAPE 1 Final Interlaced.mov",atrack=1)

# ConverttoYV12(interlaced=true)    ONLY GOOD FOR 8bit Video !!

ConvertToYUV422(interlaced=true)  # MY VIDEO CAPTURE IS 10bit     DO I NEED THIS ?? as already 10bit

AssumeTFF() # My Interlacing is TOP FRAME FIRST !

# crop(0, 2, 0, -10)   ??
#addborders(0,6,0,6)  ??
#crop(5, 2, -5, -2)   ??
#Crop(0,0,704,576)  ??

QTGMC(preset="Medium", EdiThreads=3, Sharpness=1)

# BilinearResize(720,540) ??
#LanczosResize(720,540) ??
#Spline64Resize(720,540) ??

#LanczosResize(696,520) ??

Prefetch(10) # used for multi thread processing

I don't understand resizing and cropping etc, please keep it very very simple for me, I just need to understand how to create the right script

Please also suggest anything that is missing in my script or anything that is wrong, i'm very new to this and trying to learn.



also when rendering the final script, what is the best and maximum 10bit H.264 MP4 setting i should have ? my current ones are in virtual dub 2 are


[Attachment 60568 - Click to enlarge]

are these acceptable or should I go high bit rate / slow / slower etc. with the settings

I also not sure about the colour space, I think I am correct by keeping everything 422 YUV


With my current QTGMC script file with the cropping and resizing disabled the output files gives a different aspect ratio 5:4 why ?
the original is 4:3

yet the screen size is still 720 576 ?

General
Complete name                            : C:\JERSEY\New AviSynth+ Script2.mp4
Format                                   : MPEG-4
Format profile                           : Base Media
Codec ID                                 : isom (isom/iso2/avc1/mp41)
File size                                : 7.23 MiB
Duration                                 : 17 s 220 ms
Overall bit rate mode                    : Variable
Overall bit rate                         : 3 521 kb/s
Writing application                      : Lavf57.79.100

Video
ID                                       : 1
Format                                   : AVC
Format/Info                              : Advanced Video Codec
Format profile                           : High 4:2:2@L3.1
Format settings                          : CABAC / 4 Ref Frames
Format settings, CABAC                   : Yes
Format settings, Reference frames        : 4 frames
Codec ID                                 : avc1
Codec ID/Info                            : Advanced Video Coding
Duration                                 : 17 s 220 ms
Bit rate                                 : 3 177 kb/s
Width                                    : 720 pixels
Height                                   : 576 pixels
Display aspect ratio                     : 5:4
Frame rate mode                          : Constant
Frame rate                               : 50.000 FPS
Color space                              : YUV
Chroma subsampling                       : 4:2:2
Bit depth                                : 10 bits
Scan type                                : Progressive
Bits/(Pixel*Frame)                       : 0.153
Stream size                              : 6.52 MiB (90%)
Writing library                          : x264 core 157 r2935M 545de2f
Encoding settings                        : cabac=1 / ref=3 / deblock=1:0:0 / analyse=0x3:0x113 / me=hex / subme=7 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=-2 / threads=18 / lookahead_threads=3 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=3 / b_pyramid=2 / b_adapt=1 / b_bias=0 / direct=1 / weightb=1 / open_gop=0 / weightp=2 / keyint=250 / keyint_min=25 / scenecut=40 / intra_refresh=0 / rc_lookahead=40 / rc=crf / mbtree=1 / crf=23.0 / qcomp=0.60 / qpmin=0 / qpmax=81 / qpstep=4 / ip_ratio=1.40 / aq=1:1.00
Color range                              : Limited
Matrix coefficients                      : BT.470 System B/G
Codec configuration box                  : avcC

I really would appreciate some help from an expert who knows about converting high quality captured PAL interlaced video and deinterlacing with QTGMC at the very highest quality in h.264 with perfect crop and resize

I feel I am well on my way but would like a little more help please


Thank you in advance



as you can see below there is a VERY SMALL amount of video that could be cropped around the edges and I guess then resized ?

but surly once cropped it should be left, to preserve original quality ?

Won't Resizing (Upscaling slightly) hinder quality ?

But then cropping only may mess up the aspect ratio oh I'm confused ??? Why is everything such hard work ?

original Interlaced

[Attachment 60569 - Click to enlarge]

QTGMC

[Attachment 60570 - Click to enlarge]


Regards John

[DB83]

I'll make a comment or two about the cropping and aspect ratio.


When it IMO comes down to cropping it all depends if the borders annoy you. And even if you cropped them away watch the vid on a screen wider than your image ie 4:3 on 16:9 the player will add borders regardless. The mediainfo stated that the true image was 703 pixels so there is a total of 13 pixels to crop. Normally one would see 704 pixels so 703 is rather 'odd'.


Aspect ratio is quite easy to fix. You full image was 720*576 which is 5:4. It was only displayed at 4:3 due to a flag. When you convert you can add a flag but will reply on your player to recognise the flag and not all will do so. You set the PAR at 1:1 which ignores any flag and kept the 720*576. So the fix is to resize to 720*540. This is a good choice if you plan to watch on a 1920*1080 tv since the height is exactly 50%.


I think the colorspace at 4.2.2 is ok. But what I would find annoying from your sample is that pink edge. Is that on the camera original or due to the capture ?

[jagabo]

Set SAR Width to 12, SAR height to 11.

Regarding speed of QTGMC, are you using a multithreaded build of AviSynth?

[Johnnysh]

I'll make a comment or two about the cropping and aspect ratio.


When it IMO comes down to cropping it all depends if the borders annoy you. And even if you cropped them away watch the vid on a screen wider than your image ie 4:3 on 16:9 the player will add borders regardless. The mediainfo stated that the true image was 703 pixels so there is a total of 13 pixels to crop. Normally one would see 704 pixels so 703 is rather 'odd'.


Aspect ratio is quite easy to fix. You full image was 720*576 which is 5:4. It was only displayed at 4:3 due to a flag. When you convert you can add a flag but will reply on your player to recognise the flag and not all will do so. You set the PAR at 1:1 which ignores any flag and kept the 720*576. So the fix is to resize to 720*540. This is a good choice if you plan to watch on a 1920*1080 tv since the height is exactly 50%.


I think the colorspace at 4.2.2 is ok. But what I would find annoying from your sample is that pink edge. Is that on the camera original or due to the capture ?

Thanks for the info, appreciated,

Aspect ratio is quite easy to fix. You full image was 720*576 which is 5:4. It was only displayed at 4:3 due to a flag. When you convert you can add a flag but will reply on your player to recognise the flag and not all will do so. You set the PAR at 1:1 which ignores any flag and kept the 720*576. So the fix is to resize to 720*540. This is a good choice if you plan to watch on a 1920*1080 tv since the height is exactly 50%.

so where did I set the PAR 1:1 should I not have done this ?

So let me get this straight, the original captured file is at 720*576 which is 5:4 but file info is stating this has an aspect ratio of 4:3 because of a flag setting
correct ?

for true 4:3 aspect ratio I need to resize to 720*540 and one would crop first and then resize
correct ?


so i'm guessing my script would look like so

Crop(0,0,720,540)

QTGMC(preset="Medium", EdiThreads=3, Sharpness=1)
BilinearResize(720,540)

would this be correct ?

Is BilinearResize the best option ? as there are others I have seen?

BilinearResize(640,480)
LanczosResize(720,540)
Spline64Resize(720,540)

so what you are saying is, it maybe better to live with the original non cropped video ??

Yes the pink edge is I believe due to not having a TBC and feeding the SVIDEO in to my very high end AJA kona video card with no timebase corrector ??

I don't get this from my SONY SVO-5800 VHS machine when capturing VHS, I guess I either require an Time base corrector in-between the Sony Hi8 camcorder and capture card or invest in a similar grade of machine as the svo-5800 for Hi8 & mini dv that has onboard TBC


Thanks for any feed back in advance


Regards

[Johnnysh]

Set SAR Width to 12, SAR height to 11.


Where do you set the SAR ?? in the script ? or in virtual dub2

can you please explain thanks



Regarding speed of QTGMC, are you using a multithreaded build of AviSynth?

hello thanks for your reply


how would I know if I'm using multi thread build ?

[OS/Hardware info]
Operating system:           Windows 10 (x64) (Build 19043)

CPU:                        Intel(R) Xeon(R) CPU E5-2643 v3 @ 3.40GHz / Xeon (Woodcrest)
                            MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, FMA3, AVX, AVX2
                            6 physical cores / 12 logical cores


[Avisynth info]
VersionString:              AviSynth+ 3.7.0 (r3382, 3.7, i386)
VersionNumber:              2.60
File / Product version:     3.7.0.0 / 3.7.0.0
Interface Version:          8
Multi-threading support:    Yes
Avisynth.dll location:      C:\WINDOWS\SysWOW64\avisynth.dll
Avisynth.dll time stamp:    2021-01-11, 15:46:40 (UTC)
PluginDir2_5 (HKLM, x86):   C:\Users\HP-Z640\AviSynth+\plugins
PluginDir+   (HKLM, x86):   C:\Users\HP-Z640\AviSynth+\plugins+


[C++ 2.6 Plugins (32 Bit)]
C:\Users\HP-Z640\AviSynth+\plugins+\ConvertStacked.dll  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins+\DirectShowSource.dll  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins+\ImageSeq.dll  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins+\Shibatch.dll  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins+\TimeStretch.dll  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins+\VDubFilter.dll  [2021-01-11]

[Scripts (AVSI)]
C:\Users\HP-Z640\AviSynth+\plugins+\colors_rgb.avsi  [2021-01-11]

[Uncategorized files]
C:\Users\HP-Z640\AviSynth+\plugins+\colors_rgb.txt  [2021-01-11]

[OS/Hardware info]
Operating system:           Windows 10 (x64) (Build 19043)

CPU:                        Intel(R) Xeon(R) CPU E5-2643 v3 @ 3.40GHz / Xeon (Woodcrest)
                            MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, FMA3, AVX, AVX2
                            6 physical cores / 12 logical cores


[Avisynth info]
VersionString:              AviSynth+ 3.7.0 (r3382, 3.7, x86_64)
VersionNumber:              2.60
File / Product version:     3.7.0.0 / 3.7.0.0
Interface Version:          8
Multi-threading support:    Yes
Avisynth.dll location:      C:\WINDOWS\SYSTEM32\avisynth.dll
Avisynth.dll time stamp:    2021-01-11, 15:47:06 (UTC)
PluginDir2_5 (HKLM, x64):   C:\Users\HP-Z640\AviSynth+\plugins64
PluginDir+   (HKLM, x64):   C:\Users\HP-Z640\AviSynth+\plugins64+


[C++ 2.6 Plugins (64 Bit)]
C:\Users\HP-Z640\AviSynth+\plugins64+\ConvertStacked.dll  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins64+\DePan.dll  [2.13.1.6]
C:\Users\HP-Z640\AviSynth+\plugins64+\DePanEstimate.dll  [2.10.0.4]
C:\Users\HP-Z640\AviSynth+\plugins64+\DirectShowSource.dll  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins64+\ffms2.dll  [2020-08-22]
C:\Users\HP-Z640\AviSynth+\plugins64+\ImageSeq.dll  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins64+\KNNEDI3.dll  [0.9.4.49]
C:\Users\HP-Z640\AviSynth+\plugins64+\KTGMC.dll  [2018-11-23]
C:\Users\HP-Z640\AviSynth+\plugins64+\masktools2.dll  [2.2.26.0]
C:\Users\HP-Z640\AviSynth+\plugins64+\mvtools2.dll  [2.7.45.0]
C:\Users\HP-Z640\AviSynth+\plugins64+\nnedi3.dll  [0.9.4.59]
C:\Users\HP-Z640\AviSynth+\plugins64+\RgTools.dll  [1.2.0.0]
C:\Users\HP-Z640\AviSynth+\plugins64+\Shibatch.dll  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins64+\TimeStretch.dll  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins64+\VDubFilter.dll  [2021-01-11]

[Scripts (AVSI)]
C:\Users\HP-Z640\AviSynth+\plugins64+\colors_rgb.avsi  [2021-01-11]
C:\Users\HP-Z640\AviSynth+\plugins64+\KSMDegrain.avsi  [2018-02-12]
C:\Users\HP-Z640\AviSynth+\plugins64+\KTGMC.avsi  [2018-07-08]
C:\Users\HP-Z640\AviSynth+\plugins64+\QTGMC.avsi  [2021-09-03]
C:\Users\HP-Z640\AviSynth+\plugins64+\Zs_RF_Shared.avsi  [2021-09-03]

[Uncategorized files]
C:\Users\HP-Z640\AviSynth+\plugins64+\colors_rgb.txt  [2021-01-11]



[Plugin errors/warnings]
_______________________________________________________________________________________________________

Function duplicates:

"undefined" : "[InternalFunction]"
"Undefined" : "C:\Users\HP-Z640\AviSynth+\plugins64+\Zs_RF_Shared.avsi"

_______________________________________________________________________________________________________

"C:\Users\HP-Z640\AviSynth+\plugins64+\DePanEstimate.dll" requires the FFTW library for some functions.

Note: 'libfftw3f-3.dll' can be downloaded here:
      http://www.fftw.org/install/windows.html

      'libfftw3f-3.dll' should be placed in "System32" and/or "SysWoW64"
_______________________________________________________________________________________________________

"C:\Users\HP-Z640\AviSynth+\plugins64+\mvtools2.dll" requires the FFTW library for some functions.

Note: 'libfftw3f-3.dll' can be downloaded here:
      http://www.fftw.org/install/windows.html

      'libfftw3f-3.dll' should be placed in "System32" and/or "SysWoW64"
_______________________________________________________________________________________________________

thank you in advance for any hep

[_Al_]

-crop 8 from left and 8 right or total 16
-QTGMC
-change YUV if you want
-resize to 4:3, possibly good idea is to keep lines (576) and just stretch them to 768
-if you still at this point insist to crop some faulty parts, you can do that, but it is not possible to resize it again without getting ar error (if someone upsizes in the future), because your current aspect ratio would be lost (even if shown correctly on screen after this border glitch crop), so good idea is just to mask those border glitches

vapoursynth script(you do Avisynth, follow pattern):

import vapoursynth as vs
from vapoursynth import core
import havsfunc
clip = core.lsmas.LibavSMASHSource("my_422P10_captured_as_720x576_from_VHS.mp4")
clip = core.std.Crop(clip, left=8, right=8, top=0, bottom=0)
clip = havsfunc.QTGMC(clip, Preset='Slower', TFF=True)
clip = core.resize.Spline36(clip, width=768, height=576) 
#clip = core.resize.Spline36(clip, width=768, height=576, format=vs.YUV420P8) #if YUV420P8 needed
clip.set_output()

[Johnnysh]

-crop 8 from left and 8 right or total 16
-QTGMC
-change YUV if you want
-resize to 4:3, possibly good idea is to keep lines (576) and just stretch them to 768
-if you still at this point insist to crop some faulty parts, you can do that, but it is not possible to resize it again without getting ar error, because your current aspect ratio would be lost (even if shown correctly on screen), so good idea is just to mask those border glitches

vapoursynth script(you do Avisynth, follow pattern):

import vapoursynth as vs
from vapoursynth import core
import havsfunc
clip = core.lsmas.LibavSMASHSource("my_422P10_captured_as_720x576_from_VHS.mp4")
clip = core.std.Crop(clip, left=8, right=8, top=0, left=0)
clip = havsfunc.QTGMC(clip, Preset='Slower', TFF=True)
clip = core.resize.Spline36(clip, width=768, height=576) 
#clip = core.resize.Spline36(clip, width=768, height=576, format=vs.YUV420P8) #if YUV420P8 needed
clip.set_output()

thanks for the info

Sorry I don't really understand ??

do I resize ?

don't I resize ?

should I do nothing ?

by resizing will i lose aspect ratio or degrade image quality ?

should I just crop and not resize ?


should I just capture and do nothing and live with the artefacts around the edge ??



I'm so confused

thanks

[DB83]

Well as I said, the cropping is down to personal choice. Borders have never troubled me.


Resize is more critical. As it stands the 5:4 display will be slightly distorted. The choice comes down to two camps. I believe that jagabo's method, which will also result in 768*576 - the same that your original video displays as - relies on the player. Al's method forces the frame to 768*576 so will display correctly with all players. The only 'advantage' with 720*540 is the scaling. But if you watch a 4:3 dvd on a 1080p display with no obvious visual distraction it is perfectly good, and native, choice.


As for the pink I really do not think this would be solved by a TBC. Possibly your high end capture card is 'too good' for the source.

[Johnnysh]

OK by following DB83 advice and cropping and resizing with the following

Crop(0,0,720,540)
QTGMC(preset="Medium", EdiThreads=3, Sharpness=1)
BilinearResize(720,540)

I get this


[Attachment 60571 - Click to enlarge]


the original was this


[Attachment 60572 - Click to enlarge]



Please note the screen grabs of the two images are not as the same frame time,


and the media info for the newly cropped and resized file looks good, I'm thinking 4:3 aspect

General
Complete name : C:\JERSEY\New AviSynth+ Script2.mp4
Format : MPEG-4
Format profile : Base Media
Codec ID : isom (isom/iso2/avc1/mp41)
File size : 23.1 MiB
Duration : 47 s 480 ms
Overall bit rate mode : Variable
Overall bit rate : 4 087 kb/s
Writing application : Lavf57.79.100

Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High 4:2:2@L3.1
Format settings : CABAC / 4 Ref Frames
Format settings, CABAC : Yes
Format settings, Reference frames : 4 frames
Codec ID : avc1
Codec ID/Info : Advanced Video Coding
Duration : 47 s 480 ms
Bit rate : 3 766 kb/s
Width : 720 pixels
Height : 540 pixels
Display aspect ratio : 4:3
Frame rate mode : Constant
Frame rate : 50.000 FPS
Color space : YUV
Chroma subsampling : 4:2:2
Bit depth : 10 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 0.194
Stream size : 21.3 MiB (92%)
Writing library : x264 core 157 r2935M 545de2f
Encoding settings : cabac=1 / ref=3 / deblock=1:0:0 / analyse=0x3:0x113 / me=hex / subme=7 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=-2 / threads=17 / lookahead_threads=2 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=3 / b_pyramid=2 / b_adapt=1 / b_bias=0 / direct=1 / weightb=1 / open_gop=0 / weightp=2 / keyint=250 / keyint_min=25 / scenecut=40 / intra_refresh=0 / rc_lookahead=40 / rc=crf / mbtree=1 / crf=23.0 / qcomp=0.60 / qpmin=0 / qpmax=81 / qpstep=4 / ip_ratio=1.40 / aq=1:1.00
Color range : Limited
Matrix coefficients : BT.470 System B/G
Codec configuration box : avcC

Audio
ID : 2
Format : AAC LC
Format/Info : Advanced Audio Codec Low Complexity
Codec ID : mp4a-40-2
Duration : 47 s 467 ms
Bit rate mode : Variable
Bit rate : 311 kb/s
Maximum bit rate : 576 kb/s
Channel(s) : 2 channels
Channel layout : L R
Sampling rate : 48.0 kHz
Frame rate : 46.875 FPS (1024 SPF)
Compression mode : Lossy
Stream size : 1.76 MiB (8%)
Default : Yes
Alternate group : 1

so do we all agree there is nothing wrong with the script settings


thanks in advance

Regards

[Johnnysh]

Is BilinearResize the best option ? as there are others I have seen?


BilinearResize(640,480)
LanczosResize(720,540)
Spline64Resize(720,540)

[DB83]

Not so sure about the crop line. My interpretation of that is you have removed 36 lines which then renders the resize pointless.


Better to post the same frame or sample before and after video.

[_Al_]

I'm so confused

Because there is a ton of scripts on web for VHS that confuses you.
Principle is:
- crop left and right to 704, cutting 16 pixels, whatever side needs more or less, use multiples by two pixels
IF your video was captured as 704x576, you do not do that
-use QTGMC
-resize to 4:3 (if you encode with SAR in mp4 file you do not do that)
-can crop more to get rid of border glitches, but I'd simply mask those to keep perfect 4:3 aspect ratio if someone later decides ti upscale. You skip this part for now, focus on getting it up to here

SetFilterMTMode("QTGMC",2) #not using these so just copying
FFMPEGSource2("JERSEY TAPE 1 Final Interlaced.mov",atrack=1)
AssumeTFF()
Crop(left=8, top=0,width=704, height=576)
QTGMC(preset="Medium", EdiThreads=3, Sharpness=1)
Spline64Resize(768,576)  #768/576 is 4/3 , I would not use Lanczos, but that is a preference
#now here you can crop border glitches but you should know the consequences as I explained
#ConverttoYV12() #if your input was 422
Prefetch(10) #again, not using this so just copying your code

[Johnnysh]

Thank you for your reply

will try your script


Kind Regards

[jagabo]

You're using AviSynth+ so you have a multi threaded build. And the "prefetch(10)" (I missed that earlier) indicates you are enabling mutithreading. So you should be getting good use of your CPU. Without the prefetch the script would be much slower.

With an ITU standard 720x576 video capture the 4:3 image is in a 704x576 (it's actually closer to 702x576 but most people like to stick with mod 16 frame sizes) portion of that frame. With a 704x576 frame the sample aspect ratio (SAR) is 12:11. (The SAR doesn't change when you crop the frame. Like when you cut the edges off a piece of graph paper the shape of the boxes on the remaining paper is still the same.) If you encode with that frame size and that SAR the player should play the video with a 4:3 frames size like 768x576, 1440x1080, etc. But not all players respect the SAR setting so it's safer to resize the frame yourself and set the SAR to 1:1. So crop to 704x576, QTGMC, then resize to a 4:3 frame size like 704x528, 768x576, etc. Of those two frame sizes the first loses a little vertical resolution. The second keeps the vertical resolution but requires a slightly higher bitrate.

Regarding which algorithm to use for resizing: Bilinear is the least sharp. Spline64() is sharper. Lanczos is the sharpest. If your source is very sharp you may want to avoid the sharper resizers as they will produce oversharpening halos and aliasing artifacts. Bilinear will reduce oversharpening halos a bit. Use whatever is appropriate for your video. And it's possible to use one resizer for one axis, another for the other axis. So for example you could use a combination like BilinearResize(640,576) followed by Spline64Resize(640x480) to produce a final 640x480 (4:3) frame.

[Johnnysh]

Thank you very much for your detailed reply

I will follow what you say and suggest and report back

Thank you once again


Regards John