XL-1H 1080x1440 - How is this 16:9?
 

Just looking at specs of XL-1H (and others including XDCamHD) and they're all 1080x1440 rather than 1080x1920. The first one is 4x3 not 16:9, so did some investigating and found that they're rectangular pixels, so in effects strecthed in the horizontal to make it 16x9. Surely this affects the detail level, just like strectching a 4x3 image to 16x9 in post.

Explained at http://www.dvinfo.net/canonxlh1/articles/article06.php second paragraph.

The CCD pixels are not "stretched," but are rectangular in the first place. In other words, there's no stretching going on, they were made that way to begin with. Instead of building a wall with square bricks, you're building it with rectangular bricks. Any "loss of detail" is more than made up for by Pixel Shift (the green CCD is offset horizontally by one-half pixel to provide more sampling points per photosite). And this is exactly how the HDCAM format works as well.

Thanks Chris,
Obviously though you'll get less detail from a 1440 rectangular pixels than from 1920 square ones. Wondering now whether my Digibeta has square or rectangular pixels?

From an academic perspective, you're entirely correct, Steve. But if you look at how well a 1440x1080 16x9 image holds up even when projected on a large screen, I don't think there's much to be worried about in practice. The optical resolution is not quite up there with modern 35 mm film, but it's not too far off either. In my experience, optical film resolution (@ISO100) tops out around 2700 dpi - beyond that, you're mostly sampling grain when scanning to digital. Since the camera aperture in 35mm anamorphic film cameras is 22.0mm (0.866") wide, this optical resolution would transfer into something like 2340 sampling points in the horizontal axis. Still a difference, but not too far off...

FWIW,

Ron