How to downgrade NTSC 1080p to SD 480p widescreen?

[mlong68]

Hi,

I need help in trying to create a AviSynth script that will allow me to down grade a NTSC 1080p video to SD 480p widescreen.

My source file is from a Canon 5d Mark II, 1080p video, and I want to down grade to 480p for certain customers. I also shot the video with a wide lens.

The following is my source file from the Canon 5D Mark II

General
Complete name : K:\Dvd Projects\Dorothy_Class_Reunion\Friday\MVI_0390.MOV
Format : MPEG-4
File size : 3.36 GiB
Overall bit rate : 47919470

Video
Count : 185
Count of stream of this kind : 1
Kind of stream : Video
Kind of stream : Video
Stream identifier : 0
ID : 1
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format/Url : http://developers.videolan.org/x264.html
InternetMediaType : video/H264
Codec ID : avc1
Codec ID/Info : Advanced Video Coding
Codec ID/Url : http://www.apple.com/quicktime/download/standalone.html
Codec : AVC
Codec/Family : AVC
Codec/Info : Advanced Video Codec
Codec/Url : http://developers.videolan.org/x264.html
Codec/CC : avc1
Codec profile : Baseline@L5.0
Codec settings : 1 Ref Frames
Codec settings, CABAC : No
Codec_Settings_RefFrames : 1
Duration : 602602
Duration : 00:10:02.602
Bit rate mode : VBR
Bit rate mode : Variable
Bit rate : 46380433
Bit rate : 46.4 Mbps
Width : 1920
Width : 1 920 pixels
Height : 1080
Height : 1 080 pixels
Original height : 1088
Original height : 1 088 pixels
Pixel aspect ratio : 1.000
Display aspect ratio : 1.778
Display aspect ratio : 16:9
Original display aspect ratio : 1.765
Original display aspect ratio : 16:9
Rotation : 0.000
Frame rate mode : CFR
Frame rate mode : Constant
Frame rate : 29.970
Frame rate : 29.970 fps
Frame count : 18060
Resolution : 8
Resolution : 8 bits
Colorimetry : 4:2:0
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8
Bit depth : 8 bits
Scan type : Progressive
Scan type : Progressive
Interlacement : PPF
Interlacement : Progressive
Bits/(Pixel*Frame) : 0.746
Stream size : 3493621216
Stream size : 3.25 GiB (97%)

[Blue_MiSfit]

Truth be told, this belongs in the AviSynth usage subforum, but that's okay. I'll give you some suggestions.

1) When you say 480p - what do you mean exactly? Do you need to make a DVD? I'm guessing since you put this in the MPEG-4 AVC subforum you want to encode H.264 - possibly for web streaming?

2) The actual downconversion part is quite easy. I'd start by remuxing your MOV to an MKV file, using mkvmergegui. Once you have the MKV files, simply use an AviSynth script like this:


dss2("foo.mkv")

bicubicresize(720,480)


Now you have a 480p30, 16x9 anamorphic video, which is pretty standard stuff. If you want a simple square-pixel stream for web distribution, switch 720,480 for 854,480 or 720,404, whichever you prefer. You can swap out bicubic for something more exotic like spline36, but I usually use bicubic for speed reasons.

You can use eac3to on the mkv to encode an audio stream (AC3 for DVD, AAC for basically anything else).

More suggestions beyond this will require more input on your behalf! :) Tell us what you're targeting. You can't go wrong with x264 for H.264 video, or HCenc for MPEG-2 video.

Derek

[TheRyuu]

...but I usually use bicubic for speed reasons...

You should be able to hit at least 100fps with Spline36 although I do agree with you that Bicubic will be faster in some situations will be faster although I prefer something a tad more sharper to the standard bicubic settings:

BicubicResize(704,400,0.1,0.61) #or even 0 for b although that might halo a tad more
#obviously adjust your resolution to wherever you need it

Gets pretty close to Spline in terms of sharpness although sometimes the slight blurriness/anti-aliasing it provides can be beneficial (only when downsizing). I actually sometimes use bicubic for that reason sometimes even if I don't need the speed.

[mlong68]

1) When you say 480p - what do you mean exactly? Do you need to make a DVD?
2) The actual downconversion part is quite easy. I'd start by remuxing your MOV to an MKV file, using mkvmergegui.
Derek

I'm trying to create a new workflow to down grade HD 1080p video to SD to put on regular DVD. I need to make a SD DVD for customers who attended a family reunion I video taped a while back. Once my video is downgraded I wanted to import it into Sony Vegas 10, and add lower thirds to the video. I've had issues with Sony Vegas for example, it doesn't downgrade well, so this is why I wanted to use Avisynth to maybe resolve some of the issues I'm experiencing with Vegas.

In order to use Avisynth, do I have to convert all files to a MKV? Not sure if Sony Vegas can handle that file format.

I guess this would be my workflow.

1. Use Avisynth script to convert to 480p as a MKV file. (Not sure if Sony Vegas can handle .mkv files) Not sure if it will keep my colorspace the same?
2. Import into Sony Vegas 10, add lower thirds to video
3. Not sure what my Pixel ratio should be set to in Vegas: 1.2121 (NTSC DV Widescreen) or 1.0000 (Square) or 0.9091 (NTSC DV) for 480p DVD?
4. Set compression to Lagarith lossless codec using YUY2 colorspace. (should I use another codec like panasonic?)
5. Render from Vegas as .AVI file
6. Place .AVI file into Cinema Craft Basic
7. Set to progressive source
8. Set Aspect ratio to 16:9
9. Set Luminance to (16 to 235) or (0-255)? Not sure which one I should check?


I'd appreciate any help you can give me to improve/solve my workflow for the best quality I can get.

Thanks!

Matt

[Blue_MiSfit]

Regarding your proposed workflow

1) I'd suggest encoding your avisynth script into a HuffYUV / Lagarith coded AVI. Do it in YUY2. I'd suggest adding "ConvertToYUY2" before the resize statement in your script
2) Sounds good
3) I think the NTSC DV Widescreen setting will get you close enough.
4) Yes, Lagarith or HuffYUV again
5) Yes
6) You're going to use CCE, eh? Okay - but I'd at least take a look at what HC can do for you.
7) Yes, but keep in mind you're encoding 480p30 to DVD. I think you can encode at 480p30 and add the pulldown flags to 480i60.
8) Yes
9) 16-235 most likely, but I'd double check both what you feed into Vegas, and what comes out of Vegas, using Avisynth's Histogram() function.

Derek

[Audionut]

AFAIK canon cameras do TV levels (16-235).

Honestly, your doing this the long way. What problems are you having with Vegas downgrading?
What exactly do you mean by, downgrading?

Vegas can import Canon MOV files and output MPEG2 ready for dvd authoring.

Your wanting to add all these extra steps that imo are unnecessary.

CCE and HC are better MPEG2 encoders, but with the low motion footage you are dealing with, it won't matter.

[mlong68]

Honestly, your doing this the long way. What problems are you having with Vegas downgrading?

When I say downgrade, I mean 1080p video to 480p video. The quality looks like 100p or worst when using vegas, and the image on the TV looks very narrow. When I resized to 480p using VirtualDub, The video looks great. My 1080p footage is over is over 1 hour long, I can't place this many video clips into Vegas at one time, my machine doesn't have the resources. I also noticed that vegas was changing my colorspace.

Matt

[mlong68]

dss2("foo.mkv")


Where do I get the function dss2?

[J_Darnley]

Where do I get the function dss2?
Haali Media Splitter, avss.dll

[mp3dom]

When I say downgrade, I mean 1080p video to 480p video. The quality looks like 100p or worst when using vegas, and the image on the TV looks very narrow.

Since you're working with true progressive source, you need to have both project and clips setup as "Not interlaced" (check your settings). This will avoid the use of deinterlacer for scaling.
You can work with square PAR or DV PAR, it depends by your footage. If you only add effects or trim clips there's no real difference as long as both project and clips match the same aspect ratio. If you have a square graphics (PSD/TGA) and you want to keep the right aspect ratio in the clip, you need to have the graphics in square pixel and clips/project in DV PAR. This will adjust the graphics aspect ratio. It's more or less the same as the pixel aspect ratio correction that you can find in Photoshop.

[setarip_old]

@mlong68

Likely the simplest way to downconvert your .MOV to a compliant standard DVD "package" (.IFOs, .BUPs, .VOBs) is to use "multiAVCHD"...

[Blue_MiSfit]

@mlong68:

You're still not clear about what you intend to do with this 480p30 video. You can scale it any way you please in AviSynth, but what happens next is highly dependent on your ultimate goal.

Help us help you :)

Derek

[mlong68]

@mlong68:

You're still not clear about what you intend to do with this 480p30 video. You can scale it any way you please in AviSynth, but what happens next is highly dependent on your ultimate goal.

Help us help you :)

Derek

I'll create the DVD using DVD-lab PRO2

[henryho_hk]

1920x1080 = 704x480 NTSC 16:9 .... not 720x480

704x480 is a valid DVD resolution

[Blue_MiSfit]

I don't consider a 2.2% AR difference to be visible, but in this case, why not? :)

[mp3dom]

There's probably a Colormatrix setting missing that you should use in the downconversion process prior to import into Vegas.

[henryho_hk]

704 vs 720 .... not only better AR.... we can also save 2.2% bitrate for complex scenes

[Sharc]

704 x 480 => PAR = 5760:4739 = ITU compliant
720 x 480 => PAR = 32:27 = Generic

Which one to choose depends on how your player will stretch the anamorphic picture. (A standalone DVD player is likely to follow the ITU standard).