Multimedia
Blogs
Forum
Affordable HD
Whitepapers
Newsletters
Events
Advertisers
DCP Directory
MillimeterIntegrating DV Source Footage into HDV Projects, Part 1
Dec 10, 2007 12:00 PM, By Jan Ozer
Figure 3: HDV is top field first, DV bottom field first.
Click here for a larger image
Interlaced and Progressive
Most DV footage was shot in interlaced mode, with a sprinkling of progressive, while HDV is split more evenly between the two. If you’re converting from interlaced DV to interlaced HDV, obviously, you don’t want to de-interlace. However, you should know that with DV video, the bottom field is first, while with HDV, the top field is first. Most of the output presets that you’ll work with know this, but if you’re creating an Apple ProRes 422 intermediate file, as discussed below, you may need to enter this manually (Figure 3). If field order gets hosed, you’ll see obvious interlacing artifacts, frames playing out of order, jerkiness, or all of the above.
On the other hand, when converting from interlaced DV to progressive HD, deinterlacing is a huge component of the quality equation, which could and perhaps should change your workflow, particularly for Final Cut Pro editors. That’s because Final Cut Pro deploys different deinterlacing algorithms than Compressor, which uses the glacial but ultra-high quality Optical Flow algorithm from Shake. If converting from interlaced source to progressive, you may produce noticeably better quality by scaling and particularly deinterlacing in Compressor. This means creating a separate sequence for your SD footage, rendering it out through Compressor, then re-importing the file back into Final Cut Pro. On the other hand, I found scaling quality relatively close between Final Cut Pro and Compressor, so if you don’t need to de-interlace, scaling in Final Cut Pro should work just fine.
Figure 4. Your creative options when converting 4:3 footage to 16:9 (letterboxing, enlarging, and stretching).
Click here for a larger image
Scale/Fit/ Distort
If you start with widescreen DV, it should fit perfectly into HDV once scaled to about 225 percent. On the other hand, with 4:3 DV, you’re trying to fit a 4:3 peg into a 16:9 hole. You have three options, as shown in Figure 4.
To the extreme left is letterboxing, which displays the entire 4:3 frame in the 16:9 window, with black letterboxes on each side. In the middle, I’ve enlarged the image to fit the frame, which crops off pixels from the top and bottom, but doesn’t distort the video. On the right, I’ve stretched the image to fit the frame, which noticeably distorts the video.
Obviously, content will dictate this decision. If you need to present all the information in the frame, and can’t distort the video, letterboxing may be your only alternative. I use the middle alternative when it doesn’t cut off information that I need to retain in the frame, although it zooms the video an additional 30 percent or so, which can contribute to graininess. I’ve never stretched the video, but I could see that as a viable alternative when enlarging isn’t available.
