Best method for importing & upscaling clips & footage

[rhkamra]

Hi Everyone,

I am in a situation where I have a project which is primarily in HD, but I need to upscale some lower resolution footage/clips from various sources into the sequence. The footage I'm importing includes: captured media from VHS, h264 & flv files from web sourced archive footage, and SD DVD clips.

I understand that I should have my sequence at the highest resolution (the HD setting of the primary camera in this case), but have a few questions:

1) Should I convert the h264 mov and FLV flash files to prores 422 via compressor before importing them? Is there a better way? Should I be using compressor to only change the codec, or should I also use it to deal with the upscaling?

2) When upscaling to fill the frame (example: remove pillarboxing and fill my 16:9 project with a 4:3 image), is the best way to (in the canvas) grab the corners of the footage, and stretch the video to the size I want?

3) For the DVD clips I am capturing with them MPEG streamclip, which saves of h246 by default, then going to move that into prores with compressor, and import into final cut--is there a better way?

4) Are there any free plugins which improve upscaling?

Any help greatly appreciated, thanks!

 

[CaptainChunk]

1. If you have FCP 7, I would use the ProRes 422 (LT) codec. 422 and 422 HQ are WAY overkill for video that's low bitrate to begin with. If you're on FCP 6, use ProRes 422 (not HQ).

2. Cropping in the canvas like you described would be the easiest way to see how the lower-res video will look upscaled in the 16:9 frame, because you can do this on the fly without exporting anything.

3. You can actually export directly to ProRes right in MPEG Streamclip. As long as the codecs are installed, they should show up on the list. I'd avoid encoding as H.264 and then back to ProRes in Compressor. Not only is that a waste of time, but the quality would probably suffer even more.

4. There's really no way to improve upscaling, outside of using specialized hardware scalers that use algorithms to intelligently add additional information to "missing" pixels. Any way you look at it, video coming from an NTSC source (like a DVD) is going to have to be stretched and distorted by FCP to look correct in a square-pixel 16:9 frame.