From NTSC DV 720 x 480 to 1280 x 720 for web. Start to end so far . . .

[Petermgr]

Hi all,

I have researched converting my Sony Handycam videos to a format I will be posting to my Vimeo account. After a LOT of reading posts and testing, I think I've accomplished what I want. Before I continue with my other videos, I want opinions regarding what I perhaps should tweak, should do or not do, or maybe rethink my approach.

My Goal: To take all my old Hi8 and Digital8 videos from my Sony Handycam and post them in a 16:9 format to my Vimeo account, while retaining the integrity and dimension ratio of my original SD videos. I want to display these 4:3 ratio videos within a 16:9 frame for layout purposes on my website.

Process:

1. Capture all my videos using firewire and Scenalyzer as avi files.
2. Import the avi files into Vegas Pro 15 with the project properties set to match the source the file. The source file being 720 x 480, interlaced (Lower field first), Pixel ratio 0.9091, Frame Rate 29.970.
3. Export as avi using Frameserver
4. Run the following script through MeGUI

SetMTMode (5, 10)
AviSource("C:\1-Vegas tools\Video\fs.avi")
ConvertToYV12(interlaced=true, matrix="PC.709")
AssumeBFF
QTGMC(Preset="slow", EZDenoise=1.0)
Distributor()
AddBorders(120,0,120,0)
Spline64Resize(1280,720)

I used the AddBorders to adjust the ratio of my avi video to better match a 1.33 ratio to reflect the actual 4:3 of the original video. In using this method, I'm hoping I'm not degrading the quality of the final video, although it seems to look fine.

Here is a side by side of my Vegas Pro preview, and the rendered video:


[Attachment 44664 - Click to enlarge]

And what it looks like in the 16:9 (1280 x 720) format on my Vimeo page:


[Attachment 44665 - Click to enlarge]

I welcome your feedback!

Thanks,

Peter

[jagabo]

Instead of Spline36Resize() try using nnedi3_rpow2(2, cshift="Spline36Resize", fwidth=1280, fheight=720). That will give less aliasing artifacts.

I'm only seeing small thumbnail images. Why not full size?

And actually, DV uses rec.601 specs so the 4:3 image is in a 704x480 portion of the frame. That is, the 720x480 frame is slightly wider than 16:9 (by a factor of 720/704). So you should upscale to 982x720 then add borders to make a full 1280 pixel wide frame.

[poisondeathray]

Vegas' preview looks wider . The island looks "fatter" and taller in the output pic compared to vegas. Normally you would resize to 960x720 (which is 4:3 exactly), then add borders 160 to each of left, right

[Petermgr]

Instead of Spline36Resize() try using nnedi3_rpow2(2, cshift="Spline36Resize", fwidth=1280, fheight=720). That will give less aliasing artifacts.

And I'm only seeing small thumbnail images. Why not full size?

Thanks jagabo. I will give that a try!

[Petermgr]

Vegas' preview looks wider . The island looks "fatter" and taller in the output pic compared to vegas. Normally you would resize to 960x720 (which is 4:3 exactly), then add borders 160 to each of left, right

Thanks poisondethray. For resizing, do your recommend I do that in Vegas before exporting through Frameserver? Or resize to 960x720 via MeGui, then add the 160 borders, and then resize again to 1280x720?

[poisondeathray]

I would resize, then add borders. Once you add borders the dimensions will be 1280x720 (you don't have to resize again)

But see jagabo's edit - he's correct in that DV uses ITU aspect ratio (it's not 4:3, it's actually slightly wider) . So resize to 982 x 720 , then add borders. It will have to be uneven because of the chroma subsampling 150 and 148 or vice-versa (because 1280 - 982 = 298 , you need to add 298 worth of borders. 298/2 =149, but it has to be evenly divisible by 2, because of the chroma subsampling 4:2:0 . )

[Petermgr]

. . . I'm only seeing small thumbnail images. Why not full size? . . . And actually, DV uses rec.601 specs so the 4:3 image is in a 704x480 portion of the frame. That is, the 720x480 frame is slightly wider than 16:9 (by a factor of 720/704). So you should upscale to 982x720 then add borders to make a full 1280 pixel wide frame.

I just uploaded small screen shots to show the relative size of my video within the 1280x720. So the image links, when selected show a slightly larger image. Re, the 982x720, thanks jagabo!

[Petermgr]

I would resize, then add borders. Once you add borders the dimensions will be 1280x720 (you don't have to resize again)

But see jagabo's edit - he's correct in that DV uses ITU aspect ratio (it's not 4:3, it's actually slightly wider) . So resize to 982 x 720 , then add borders. It will have to be uneven because of the chroma subsampling 150 and 148 or vice-versa (because 1280 - 982 = 298 , you need to add 298 worth of borders. 298/2 =149, but it has to be evenly divisible by 2, because of the chroma subsampling 4:2:0 . )

Thanks poisondeathray. I'll give it a try!

[jagabo]

You might consider softening the boundary between the video and the added black border. This reduces DCT artifacts at the edges of the picture. I made a SoftBorders() function that does this. It has the same calling convention as AddBorders():

###################################################
#
# Add borders with a soft edge.  Like AddBorders().
#
###################################################

function SoftBorders(clip source, int left, int top, int right, int bottom, int "border_color")
{
    border_color  = default(border_color, color_black)

    mask = BlankClip(source, color=color_white)
    mask = left!=0 ? Crop(mask, 2, 0, -0, -0).AddBorders(2, 0, 0, 0) : mask
    mask = top!=0 ? Crop(mask, 0, 2, -0, -0).AddBorders(0, 2, 0, 0) : mask
    mask = right!=0 ? Crop(mask, 0, 0, -2, -0).AddBorders(0, 0, 2, 0) : mask
    mask = bottom!=0 ? Crop(mask, 0, 0, -0, -2).AddBorders(0, 0, 0, 2) : mask
    mask = AddBorders(mask, left, top, right, bottom)
    mask = mask.Levels(16, 1, 235, 0, 255, coring=false)
    mask = mask.Blur(1.0).Blur(1.0)

    black=BlankClip(mask, color=border_color)
    AddBorders(source, left, top, right, bottom)
    Overlay(black, last, 0, 0, mask)
}

[Petermgr]

You might consider softening the boundary between the video and the added black border. This reduces DCT artifacts at the edges of the picture. I made a SoftBorders() function that does this. It has the same calling convention as AddBorders():
}[/CODE]

Thanks for this jagabo. I have copied your function to my .avs file. However, I am unclear as to where I put the information regarding the border specifications, etc.

[jagabo]

You can put the function anywhere. Replace your call to AddBorders(148,0,150,0) with SoftBorders(148,0,150,0).

[Petermgr]

You can put the function anywhere. Replace your call to AddBorders(148,0,150,0) with SoftBorders(148,0,150,0).

Great. Looks good, thanks!