Why not capture VHS in 640 x 480?
 

Hey everyone,

My friends are and I are starting to archive our old VHS/S-VHS tapes with the awesome set up I bought from Lord Smurf. We have been testing different settings in virtualdub and doing a lot of research on this site. When capturing 720x480 the aspect ratio seemed a little off, so we started capturing in 640x480. That looks pretty good but now (after reading a couple of threads on aspect ratios) I'm curious if I am losing any video information by not capturing in 720x480? Is 640x480 an actual 4:3 ratio? Or are we just perceiving it to be close enough?

Thanks!

Somebody who is more knowledgeable than me on the standards can chime in but I believe the standard set for digitized analog tape material was 720x480 ,and other factors like YUV color type and Rec.601 colorspace, okay it's more like 704x480 and 720x480 is meant for anamorphic widescreen.

With 720x480 you can more easily crop and resize to a resolution that fits your desire aspect ratio, so 640x480 for 4:3 and 854x480 for 16:9

720x480 by itself is 3:2, so that's why raw footage looks weird, the aspect ratio has not been corrected so everything looks too stretched.

720x480 is also the standard resolution for DV and DVDs.

Basically it's the standard digital resolution for SD material from my understanding.

Thanks. I’m still trying to wrap my head around the colorspace stuff and how it relates to codecs and translating analog to digital. I shot on DV a lot back in the ‘10s, and all of that was 720X480.

My current archiving quest is the analog VHS tapes I did in the ‘90s. So if I need to crop/resize that analog footage to 640x480 anyway, is there a disadvantage to capturing that way? Like, am I losing some chroma or luma video data by not capturing in 720?

You don't need to resize to 640x480 to get 4:3 aspect ratio. You just need to use a file format with the 4:3 aspect ratio flag, and software which abides by that.

Unfortunately some modern video editors have forgotten how to deal with non-square-pixel video and/or interlaced video, or make you jump through hoops to manually correct it. (For example, OpenShot considers de-interlacing to be an "effect" which must be manually applied to each clip in the timeline, and you have to manually correct the aspect ratio as well.)

If you must have non-interlaced, square-pixel output, first de-interlace (preferably using something like QTGMC), and then upscale to 960x720, 1440x1080, or 1920x1440.

The rec.601 standard doesn't govern just color, but resolution and other parameters. When you set your capture software to capture @ 640 you're basically telling it to resize from 720 to 640 on the fly resulting in a slight loss of details, Capture cards capture at 720 samples per scan line, Any change from that is a loss.

Thanks. This answered my question:cool::D