Widescreen?

[Matthew Overstreet]

All right, this is a stupid question to those in the know, but I pride myself in stupid questions.

The XL2 is a 16:9 native camera. I've heard that its image sensors are somewhere in the range of 900somethingx 480. Correct? I've also heard that regardless of the resolution of the CCD sensors, it'll always be recorded to 720x480, because that is the DV standard. Correct, again?

So, what I'm wondering is . . . if a fullscreen image is recorded to 720x480 and its fullscreen--how do you get widescreen without matting it, and thus losing resolution? Does the XL2 record an anamorphic image that you have to unstretch in post?

The reason I'm asking is . . . I shot a film (not with the XL2) with an HD110 in 720p. I can't handle the JVC codec right now, so I had the DP downconvert the footage to SD--I'm assuming it'll be 24p 16:9. So, I'm just wondering how I'll be receiving this standard def fil that will hopefully be 16:9.

Thanks!

[Mathieu Ghekiere]

I *know* the XL2 records to oversized 4/3 CCD's to take an 16/9 area out of them.

What I don't know for sure, but I *think* it indeed records it anamorphic and you have to unstretch it, I think you have to put 16/9 in your project settings in your NLE.
But I could be wrong!

[Kyle Ringin]

In a word it's PAR (pixel aspect ratio). Both 16:9 and 4:3 DV use the same resolution.

Neither DV 4:3 or 16:9 are square pixel formats (like 720p). The camera sets a flag in the DV stream to say whether it is 16:9 or 4:3, the playback device then knows which PAR to use to display the video (or you have to tell it if it gets it wrong).

If the camera was set to record in 16:9 the footage will be fine - you just have to make sure you are using a 16:9 project.

Regarding your 720p footage - the DP should definately render it to 16:9, if he knows you want to edit it (that's probably obvious), you'll probably get a DV file (avi or qt) with a 16:9 Aspect Ratio - exactly what you want.

Kyle.

[Matthew Overstreet]

Thanks, I think I understand the subject a bit more now. I tried creating a preset sequence in FCP and used 23.98 as the frame rate, and I chose a standard DV ratio (3:2) that is 720x480. That should work, right? I also went into an advanced menu option and changed it from interlaced to progressive. Am I doing thing correctly?

[Justine Haupt]

Originally Posted by Matthew Overstreet

So, what I'm wondering is . . . if a fullscreen image is recorded to 720x480 and its fullscreen--how do you get widescreen without matting it, and thus losing resolution? Does the XL2 record an anamorphic image that you have to unstretch in post?

I'm not sure the part about how the sensor handles 16:9 vs 4:3 has been specifically answered, so I'll go ahead anyway.

If you look at the diagrams Canon has of the XL2 sensors on their website it becomes clear how the mechanics and optics of a higher res 16:9 image than 4:3 are achieved, in the camera: The aspect ratio of the physical sensor is 4:3. The 16:9 image is recorded from a cropped area of the 4:3 sensor. The 4:3 image, though, is actually recorded from a cropped area of the 16:9 part of the sensor. So in fact, when recording in 4:3 the sensor is recording a cropped area of the larger 4:3 sensor. It's a rectangle inside a rectange, with wasted pixels.

So really, Canon just sacrificed the 4:3 resolution to be able to say 16:9 has more pixels. The method is debateable, but it does take care of field of view issues with a given focal lenth for 4:3 vs. 16:9.

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

[Jarrod Whaley]

Originally Posted by Justin Haupt

So really, Canon just sacrificed the 4:3 resolution to be able to say 16:9 has more pixels. The method is debateable, but it does take care of field of view issues with a given focal lenth for 4:3 vs. 16:9.

We've been round and round on this debate several hundred times in this forum. There's no "sacrifice" in 4:3 resolution when you consider the fact that the usable 4:3 area of the chips still has more pixels packed into it than any other mini-DV camera ever built.

Think of it this way: if Canon had somehow been able to find actual 16:9-shaped chips of the exact same size, no one would be complaining about "wasted pixels." So what's the difference between the two methods at the end of the day? Nil. Every 16:9 shape of any size will fit into a larger 4:3 shape, and vice-versa. On into infinity. :D So why not let us use the extra real estate, some ask? Maybe because widescreen should be wider than 4:3? :)

I really think Canon should have kept this stuff a secret (like they have done with their 24f technology) just to silence all the complaints. In fact, I'm guessing that the lessons learned from this "wasted pixels" debate are exactly the reason why they've kept the way that 24f works a secret.

[Ash Greyson]

Nothing is wasted, those are non-firing pixels. The 960X720 sensor has 960X480 active and 960X576 for PAL. The 4:3 image is cropped from the 16:9 portion just like every other 16:9 native camera on earth.




ash =o)

[Richard Hunter]

Originally Posted by Jarrod Whaley

the usable 4:3 area of the chips still has more pixels packed into it than any other mini-DV camera ever built.

Hi Jarrod. I agree with the rest of your post, but I thought this part should say 16:9 rather than 4:3 as many DV cameras have 720x480 pixel 4:3 sensors (but not many have 960x480 pixel 16:9 sensors).

Richard

[Justine Haupt]

Well, I understand that perfectly about the portion of the chips that are non-firing.. I'm just saying it would have been possible to create a higher res 4:3 mode, would it not?

As an XL2 owner my self, its native 16:9 is one of the things I love about it, but I still have to admit that it would have been possible to make higher res 4:3.

[Chris Hurd]

Originally Posted by Justin Haupt

I'm just saying it would have been possible to create a higher res 4:3 mode, would it not?

It would have been possible, but not preferable. You could be looking at potential moire patterns and other undesirable image artifacts that might occur during the re-sizing process that scales it back to 720x480 where it belongs. Also consider the marketing angle: the way it is now, it's native 16:9 and native 4:3. A higher res 4:3 mode would have the 16:9 camp crying "foul" and wondering where is their high res 16:9 mode... a scandal of staggering proportions, with claims of conspiracy and a mad scrambling for an immediate class-action lawsuit. The best solution is the way it's actually done. It is a full-resolution, native 4:3 image at 720x480. And it is a full-resolution, native 16:9 image at 960x480. Jarrod Whaley has spelled it out perfectly in his post above. Hope this helps,