Resolution Question

I'm working with several SD, 4:3 ratio clips, plus some stills. I've been asked to create a 16:9 ratio, 1920x1080 movie from these. First, when I'm dropping the SD into the sequence, the ratio changes, & looks like I'm cropping some of the image. I know I've created 16:9 sequences with pillaring, but can't remember how.

I've never changed the pixel ratio. For SD video, I use the Sequence preset DV NTSC 48 kHz, & the same for Capture.

I haven't done any editing for about 3 months; I probably need to take a refresher course at the Apple store. tia.

david

Tell whoever asked you to do this that you can't do it. Or get ready for some seriously poor video. There are hardware and software tools that can uprez SD to HD, but I don't think anything, except maybe a Teranex box, can take 4:3 and blow it up to widescreen HD.

All the best,

Tom

Have a look at this to see the size of the upres you're trying to achieve.




Going from 4:3 SD NTSC to 1920 x 1080 means blowing it up around five times bigger than the original. Not pretty.

What's DALSA?

[en.wikipedia.org]

All the best,

Tom

That chart's a bit disingenuous. Regardless of the number of photosites on the sensor, the Genesis only outputs 1920x1080 frames.

Not that it matters, really.

Yeah I just nabbed it to show the NTSC SD : 1080 difference. It was probably originally from an ad for the Genesis or RED.

Jeff, if the Genesis has loads of photosites but only outputs 1920x1080, why does it have more photosites? Are they used up in color space?

And if that is so, does additional color space add to perceived resolution?

Harry The Idiot.

Genesis uses an RGB stripe pattern of rectangular photosites - and uses a pair vertically - so (1920x3) x (1080x2), which on the F35 (same sensor as Genesis) leads to funky rainbow moire horizontally because although the RGB are not co-sited they're treated as if they are, and vertically the binning and low levels of optical low pass filtering lead to visible luma moire / aliasing.

The simple answer is of course to always measure the resolution of a complete camera system. Although there is some correlation between sensor photosites and measured resolution, different systems of achieving the final image lead to vastly different results.

You can start with few photosites and make the best of them, or start with many and massacre them.

Graeme

Crikey!

Well, that settles it then, Graeme.

Silly of me not to spot that.

Best

Harry.

(PS. hey guys, is it me, or is this incredibly complicated and full of - like - logarithms and stuff?)

Only thing to do with resolution is actually measure it.