Native aspect ratio on XL2

[Guest Zulkifli Yusof]

Hi all,

 

what's the native aspect ratio of the XL2 in 16:9 mode? Is it 1.66 or 1.78?

[David Mullen ASC]

Divide 16 by 9 and you get 1.77777777......

 

However, note that the pixel ratio stays the same for 4x3 and 16x9 in standard def video, which is confusing. For example, in NTSC, it's 720 x 480 pixels for either a 16x9 or 4x3 picture, and if you divide 720 by 480, you get 1.5, not 1.33 (4x3) and certainly not 1.78 (16x9.) The simple explanation is that the pixels are not square in either format.

 

The XL2 uses an oversized 4x3 sensor from which to extract 4x3 or 16x9 pictures using the same number of sensor sites total for either, more or less, so there is no loss in switching to 16x9 mode.

[Josh Bass]

Doesn't it start out higher res in 16:9?

 

It has a 960x720 chip, and uses the 720x480 part for 4:3, and the 960x480 part for 16:9, which gets compressed (or whatever) to 720x480.

[David Mullen ASC]

Doesn't it start out higher res in 16:9?

 

It has a 960x720 chip, and uses the 720x480 part for 4:3, and the 960x480 part for 16:9, which gets compressed (or whatever) to 720x480.

 

Don't confuse recorded pixels with number of sensor sites (and neither is really a meaurement of resolution.)

 

I thought the 4x3 and 16x9 image areas on the XL2 sensors would form a cross-shape if overlaid (like I believe the Sony PD10 did), but maybe I'm wrong and both share the same number of vertical sensor sites and the 16x9 images uses more of the width of the sensor -- if so, I don't see the point of having a 4x3 920 x 720 "pixel" sensor if you never use more than the 16x9 area inside of it for either 16x9 or 4x3, unless this was just a cheap way of taking a 4x3 sensor from another camera and making use of it.

 

If you're right, it makes more sense to call the XL2 a "16x9 native" camera since it extracts 4x3 from 16x9.

[Jim Keller]

Don't confuse recorded pixels with number of sensor sites (and neither is really a meaurement of resolution.)

 

And, at the end of the day, the quality of your lens has far more to do with the quality of your finished product than whether you shot it 4:3 or 16:9.

[Josh Bass]

I refer you to this diagram:

 

 

from this article:

 

http://www.dvinfo.net/canonxl2/articles/article06.php

 

It describes the CCD block layout of the XL2.

 

 

To me, it looks noticably sharper in 16:9 than 4:3. I did a short film recently where I had to do a lot of cropping/blowing up to keep unwanted elements out of the frame, and the 16:9 holds up much better to blowup than the 4:3. Maybe that's just 'cause I'm looking at a letterboxed image on my monitor instead of one that fills the whole screen (when in 16:9).

Edited June 13, 2007 by Josh Bass

[David Mullen ASC]

OK, I believe you though it seems a waste of a 4x3 sensor... Rather than really make a good 16x9 sensor, they basically made a worse way of creating 4x3 so that the 16x9 image would look better. It makes sense if you have a 16x9 sensor to do it this way, but if you have a 4x3 sensor, then what's the point of not using a quarter of it for anything?

[Josh Bass]

You got me. It makes pretty pictures (I've seen 'em projected on a larger screen, and it holds up nicely, even in 4:3), so that's all that matters to daddy.

[Guest Zulkifli Yusof]

Thanks guys for the replies, especially to Josh for bringing up the diagram.

 

I was grabbing still frames off my NLE timeline and I was wondering how to unsqueeze the footage; but after seeing the diagram, I know now.