trying to make a dvd with this file

[smartel]

Mediainfo gives me this :

General

Format : MPEG-PS
File size : 29.0 MiB
Duration : 25s 272ms
Overall bit rate : 9 614 Kbps

Video
ID : 224 (0xE0)
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings, BVOP : Yes
Format settings, Matrix : Default
Duration : 25s 125ms
Bit rate mode : Variable
Bit rate : 9 586 Kbps
Width : 704 pixels
Height : 480 pixels
Display aspect ratio : 4:3
Frame rate : 29.970 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) : 0.947
Stream size : 27.5 MiB (95%)

Audio
ID : 192 (0xC0)
Format : MPEG Audio
Format version : Version 1
Format profile : Layer 2
Duration : 25s 272ms
Bit rate mode : Constant
Bit rate : 256 Kbps
Channel(s) : 2 channels
Sampling rate : 48.0 KHz
Compression mode : Lossy
Delay relative to video : -67ms
Stream size : 790 KiB (3%)

was mini-dv (i think) to panasonic dvd recorder ( model # ?)
iso file extracted with isobuster , edited with mpeg-vcr (giveaway version)

not much going on on the left side figured i'd crop 64 from left to make it 640x480
and let the encoder resize it ( tmpge ?? or better)

just want to make it more acceptable and make a dvd out of it (for this part it's ok)

i know a little about megui's avisynth script creator , i won't deinterlace

tried lot's of things in virtualdub nothing really satisfied me

tested some convolution3d but i'm not really shure

what would YOU do ??

thanks a lot for your time have a great day , sheers

[16mmJunkie]

(what would YOU do ?? ) A tripod would be needed for starters. if it means something salvage the audio and toss the video ...it's at a angle not worth anything to me. You asked.

[smartel]

i won't redo this! i got to work with what i got there

and i love the angle for my taste

we lost the drummer to cancer that year . we all knew at the time

i want to make a good dvd for his daughter

[16mmJunkie]

Okay, I'll buy the beauty is in the eye of the beholder. Anyone that has a bonding relationship will enjoy the video of their past loved one. So if you burn this to disc right now the daughter will still enjoy it.(hence, my comment "if it means something) She won't care if it is raw or creative. Any corrections will be a "redo" so I don't understand your asking if you've already made your mind up.

[smartel]

very helpfull ....

[jagabo]

I'd start with ColorYUV(off_y=-12, gain_y=-10). That will get you within valid IRE levels and prevent any futher blowout of the bright spots. Unfortunately, the bright spots are already blown out so you can't recover any details there.

[smartel]

so i add
ColorYUV(off_y=-12, gain_y=-10)
somewhere at the beginning of my MEgui script

[smartel]

something like this ?

# Set DAR in encoder to 46 : 37. The following line is for automatic signalling
global MeGUI_darx = 46
global MeGUI_dary = 37
LoadPlugin("C:\Documents and Settings\Steve\Bureau\MeGUI_1911_x86_WithoutInstal ler\tools\dgindex\DGDecode.dll")
DGDecode_mpeg2source("C:\Documents and Settings\Steve\Bureau\FNM\Title_1.d2v", info=3)
LoadPlugin("C:\Documents and Settings\Steve\Bureau\MeGUI_1911_x86_WithoutInstal ler\tools\avisynth_plugin\ColorMatrix.dll")
ColorMatrix(hints=true, threads=0)
#deinterlace
ColorYUV(off_y=-12, gain_y=-10)
crop( 64, 0, 0, 0)

[jagabo]

It will be fine there. You might want to check with your entire source to verify the adjustments are ok with the rest of the file. Use Histogram() and make sure the levels stay out of the brown bars.

[johns0]

smartel-in the future please use a more descriptive subject title in your posts to allow others to search for similar topics. I will change yours this time.

[smartel]

resulting file using this script :

# Set DAR in encoder to 46 : 37. The following line is for automatic signalling
global MeGUI_darx = 46
global MeGUI_dary = 37
LoadPlugin("C:\Documents and Settings\Steve\Bureau\MeGUI_1911_x86_WithoutInstal ler\tools\dgindex\DGDecode.dll")
DGDecode_mpeg2source("C:\Documents and Settings\Steve\Bureau\FNM\sample.d2v")
#deinterlace
ColorYUV(off_y=-12, gain_y=-10)
crop( 64, 0, 0, 0)
LanczosResize(640,480) # Lanczos (Sharp)
#denoise

for the histogram thing ..
the white line didn't touch the brown side bars ,
stayed more the the left ?!?

and johns0 sorry i'm kinda new here , trying to learn a little

[smartel]

sample hystogram from vob

Without ColorYUV(off_y=-12, gain_y=-10)

[jagabo]

Yes, sample2.mpg has the bad black level and is way over IRE100 at the top end, like the original file.

I'm not sure what the colors should be like but a little more saturation looks better. Try adding Tweak(sat=1.5) or thereabouts. See what you think.

[smartel]

one with the tweak only

the other with this :

Tweak(sat=1.5)
ColorYUV(off_y=-12, gain_y=-10)

[smartel]

i added the line ConvertToRGB(interlaced=true)
i use TE4XP

looks more like this :

Tweak(sat=1.5)
crop( 64, 0, 0, 0)
ColorYUV(off_y=-12, gain_y=-10)
ConvertToRGB(interlaced=true)

i don't know if it really helps ..

[smartel]

i'll redo my tests all over , hope i can show some results soon

[jagabo]

I usually adjust levels first because some other filters will clamp levels. Tweak() for example, unless you specify coring=false.

[smartel]

every new terms that is added to this thread i have to search for what it is , when and how to use it ...

what is coring=false used for ?

i'm french speaking , english is my 2nd language
and avisynth is a whole different language altogether

and i know that like audio , video workflow (filters, modification) has to be in a specific order , unless you want a special fx

and i want to use this in conjonction with the convolution3d dll
did some tests looked very good

[jagabo]

YUV video is supposed to have Y (luma, the grayscale image) values from 16 to 235. Coring, or clamping, will eliminate values outside that range. All values less than 16 become 16, and all values higher than 235 become 235. Your source video has luma peaks between 236 and 255. By using Tweak() first you lost the details between 236 and 255.

The situation is similar for the U and V (color) channels -- except their valid range is 16 to 240. I don't think your video has values outside that range. It might if you increase the saturation too much.

[smartel]

what about the ConvertToRGB32(matrix="PC.601") command
for serving to tmpge4xp , is it the right one ?

[jagabo]

what about the ConvertToRGB32(matrix="PC.601") command
for serving to tmpge4xp , is it the right one ?

I don't use TMPGEnc4, but with the older TMPGEnc Plus you would use ConvertToRGB32(matrix="rec601"). Actually, that is just the default in TMPGEnc Plus. You can specifiy PC.601 levels if necessary. The option is called "Output YUV data as Basic YCbCr not CCIR601".

[smartel]

i'll keep on testing
i can't find that option

[smartel]

here is another one just a test a little more croped and i use another encoder (tmpgenc plus)

script used

ColorMatrix(hints=true, threads=0)
#deinterlace
AssumeTFF()
ConvertToYUY2()
crop( 74, 0, 0, 0)
SeparateFields()
odd=SelectOdd.Convolution3D (0, 6, 10, 6, 8, 2.8, 0)
evn=SelectEven.Convolution3D (0, 6, 10, 6, 8, 2.8, 0)
Interleave(evn,odd)
Weave()
LanczosResize(640,480) # Lanczos (Sharp)
#denoise
Tweak(sat=1.1)
ConvertToRGB32(interlaced=true)

i don't know if i'm doing better
i should crop a little less the details are not there
still got some more testing to do

[jagabo]

You should use ConvertToYUY2(interlaced=true).

Why are you cropping 74 pixels off the left edge? And resizing to 640x480? And resizing again to 720x480 in TMPG? You should keep the original 720x480 width (no crop, no resize) and set the source aspect ratio in TMPG.

I think your using too much noise reduction.

[smartel]

i wanted to center and zoom in a little

i should not resize since the crop already make it 640X480 (crop (64, 0, 0, 0)) you're right my error

resizing in tmpg crops top and bottom making the zoom effect

maybe i should use less convolution3d ?

and maybe crop mod16 in avisynth and resizing in avisynth or tmpg

i got to know the better way (for resizing) and use the tool aspect maybe

it's trial and error most of the time ` i don't really know . trying to learn

just discovered fitcd maybe i could use that for the script

[smartel]

i'll use a very basic script for the color conversion in avisynth and do the rest in tmpge
otherwise i'm lost

[jagabo]

It's usually not a good idea to resize interlaced video. It will develop artifacts on sharp horizontal edges. And the way you have done it is messing up the aspect ratio. If you want to resize you can do it while the fields are separated.

ColorMatrix(hints=true, threads=0)
#deinterlace
AssumeTFF()
ConvertToYUY2(interlaced=true)
crop(64, 20, 0, -20)
SeparateFields()
odd=SelectOdd.Convolution3D (0, 6, 10, 6, 8, 2.8, 0).BilinearResize(720,240)
evn=SelectEven.Convolution3D (0, 6, 10, 6, 8, 2.8, 0).BilinearResize(720,240)
Interleave(evn,odd)
Weave()
#denoise
Tweak(sat=1.1)
ConvertToRGB32(interlaced=true)

After opening that in TMPGEnc, set the source aspect ratio to 4:3 NTSC.

[smartel]

the resulting file is this

and i would do any color ajustments within tmpge
i'm really more comfortable when i see beforehand the tweaking

hope that my tests can help others too

this sample is just a quickie that i did

[smartel]

is BilinearResize my best option !?,

i read that lanczos (3 or 4) may introduce ringing but sharper

what about spline 16 or 36 ??

maybe bilinearresize is good because it's line based !?

[jagabo]

Bilinear is the least sharp of the resizing filters and reduces noise slightly. Yes, Lanczos Spline16, and Spline36 are increasingly sharper and increasingly likely to create oversharpening artifacts and increase noise.

Your last sample looks like it was stretched vertically but not horizontally. Again, distorting the aspect ratio.