I noticed that Canon removed the 16:9 guidelines on the XL2. I thought this was kind of a cool feature on the XL1s. At the very least they could have replaced them with 4:3 guidelines when shooting in 16:9 mode. I mean this is all software driven.

Alot of indie filmmakers like to frame for both when shooting.

Truthfully though I was hoping they would put in guidelines for other aspect ratios as well. Like 1.85:1 and 2.35:1 but I guess it's such a small thing to gripe about.

Here's a list of other XL1 / XL1S features (http://www.dvinfo.net/canonxl2/articles/article02.php) not retained in the XL2. Plus, some accessory compatibility info.

Thanks Chris. That's where I got the info from in the first place. Great job on all of that. :)

That's indeed a bit strange that the guides are missing. I always
liked to crop 4:3 to 16:9 and be able to recompose the shots.
There isn't a memory card slot for overlays either (like there is
on the GL2). Perhaps it is something we could add through the
SDK, although I doubt that.

But if you look at Chris's excellent page on the CCD block, I think it would make very little sense to shoot 4:3 and crop to 16:9 on this camera. You would be using much less of the CCD area, more like a 1/4" CCD.

Do the math for a 16:9 crop: 345,600 pixels in 4:3 mode less 25% that you crop off equals 259,200 (720x360 pixels). Why would you do this instead of using 460,800 pixels (960x480)? They clearly optimized this design for 16:9, bravo Canon! Seems like the XL-1, DVX-100a or PD-170 will make more sense if your main need is 4:3 or if you like to shoot cropped 16:9...

Oh I totally agree with that, Boyd. It's more like I was wondering
why they removed it. Probably to encourage people to actually
shoot in 16:9, probably. Which is basically what this camera will
be best at.

<<<-- Originally posted by Boyd Ostroff : But if you look at Chris's excellent page on the CCD block, I think it would make very little sense to shoot 4:3 and crop to 16:9 on this camera. You would be using much less of the CCD area, more like a 1/4" CCD.
-->>>

That's why they should do it in reverse on the XL2 use 4:3 guidelines when shooting in 16:9 mode as I said. That way while you're shooting 16:9 you can also frame for 4:3.

Would it be too radical to letterbox the 16:9? ;-) Yeah, I know that some people still need real 4:3. But cropping 16:9 to 4:3 is still throwing away information, since your 16:9 as recorded to tape will be anamorphically squeezed down to 720 pixels wide. Similar to shooting cropped 16:9, you would be throwing away 25% of the horizontal resolution.

The way the CCDs are described, there would be no difference (in terms of pixel count) on the XL2 between shooting 4:3 or shooting 16:9 and cropping it down to 4:3. IF the MiniDV format can store the full 16:9 resolution (I'm not sure it can store that high), then theoretically, you should get exactly the same results. To me, that would suggest that the best thing for a filmmaker to do would be to shoot 16:9 all the time and crop that down to get his 4:3 shot. That would give you the ultimate flexibility if you decide you wanted something ever so slightly off from what you originally intended (slight left/right misframing).

As for losing resolution, the 720x480 resolution of the "cropped" 4:3 mode is exactly the resolution supported by MPEG-2 DVD (NTSC, anyway) encoding, and the resolution defined as "480p" in the digital world. So if you're shooting for a 4:3 target, this is probably all you'll ever need. That's why Canon can claim that it's 16:9 native but doesn't lose any resolution for 4:3, either. In fact, the 16:9 is probably going to have to be downsampled for any final TV broadcast or DVD/Video master.

I give many kudos to Canon for this one.

Ok, just looked it up. MiniDV only supports 720x480 anyway, so if you're going to finalize something in 4:3 you'd want to shoot that way, NOT in 16:9 mode, to get the best resolution out of the XL2. But that means that the XL2, even though it's cropping its CCDs, is still spitting out full MiniDV resolution, and that the 16:9 mode is actually going to have to be downsampled from the CCDs.

Oversampling is still good - you'll theoretically get a better picture out of that than you would by just sampling exactly what you're recording - but you should still get a pretty darn good picture in their 4:3 mode - theoretically, a much better one than the XL1s.

Anybody got any frame grabs yet? I'm dying to see them.

ALL footage on DV and DVD is stored at 720x480. So the XL2's
16:9 960x480 is sampled back to 720x480 (as Russell indicates
as well).

You would be far better of to just shoot in 4:3 with the XL2 in
this case. You will get a better resolution (than shooting 16:9
and then downconverting to 1.0 pixel aspect ratio) but you will
also get a longer focal range. That might be an "issue" ofcourse.

If you look into Chris's "Exploring the XL2 Menu System" article and under "Display Setup" there is a switch for TV screen. Could that be for 4:3 guidelines or something else?

Forgive my lack of technical terminology, but wouldn't it make more sense to shoot 16:9 and "transform" it down to a letterboxed 4:3? Much like if you were to, in Adobe Photoshop, make 2 images; One being 960x480 and the other being 720x480. The first being the imported 16:9 footage, then in post shrinking the width with locked ratio down to 720, keeping it centered hence letterboxing 4:3?

Is what I just described what was previously called "downsampling"?

Holly: I've just explained in my post above how you would
actually loose resolution this way. Think about it! You have a
wider image that gets squashed into a square image. Then
you need to unsquash this square image back to a wide image
to then chop off the edges! This will lower and soften your
4:3 image!

Peter: I'm pretty sure that function controls either how much
information you seen on the monitor out (most probable) or
how much information you see on the viewfinder. As Chris
clearly stated in the XL2 Watchdog, the 16:9 guidelins do not
exist on the XL2.

Rob: I was thinking in terms of... squashing the image down and in.

Like so:
Captured image:
http://www.kmdstudios.com/holly/169capture.jpg

Then, to output 16:9 to a 4:3 screen... shrinking it down into a 4:3 ratio so that the result is like this:
http://www.kmdstudios.com/holly/169in43.jpg

I do not get how once it is squeezed into a 4:3 composition that it would then have to be unsqueezed this way?

Rob, I think Holly is just talking about a simple 16:9 letterbox: "shrinking the width with locked ratio down to 720, keeping it centered hence letterboxing 4:3" . Or at least, that's what you would get if you lock the aspect ratio and set the width at 720, with a black bar above and below. And it's why I asked at earlier: "Would it be too radical to letterbox the 16:9?"

Yeah, you could call that downsampling, but it's generally just called letterboxing. If you edit in FCP all you need to do is create a 4:3 sequence and drop your 16:9 clip into it. But some applications (like the 6 o'clock news) will probably want fullscreen 4:3 and not letterboxed 16:9...

I'm lost. I've been doing this XL2 thing for 13 hours straight now.
I hope Boyd understands what is going on <g> Letterboxing is
no problem indeed!

Yeah, if you just wanted to letterbox it you'd be better off shooting in 16:9 mode and then shrinking it down. You'd have a higher resolution master to work with before putting out your finished product.

But you'd be better off keeping it just the way you filmed it (16:9) and distributing it as an animorphic DVD - assuming DVD is your target medium.

By cropping to 4:3, I (and I assume Rob, also) was referring to actually trimming the edges back off to get a fullscreen picture. If you're going to do that from an XL2 picture, you'd be better off shooting it 4:3 to begin with.

That's what I meant indeed Russell.

Russell, Boyd, Rob: Thank you. The animorphic DVD would yes be the best option for that medium. For regular VHS the method I stated above seems most usefull, and for web 4:3 applications as well. If blowing up to film, then it's fine the way it was. So, for all applications still it seems the best way to shoot would be 16:9.

If you're shooting 16:9 for 4:3 all you'd really have to do is put the 16:9 clip into a 4:3 composition and then stretch it to it's normal 16:9 width.. putting the extra image information outside of the viewable 4:3 composition area. Makes the most sense to me because the actual hieght is the same for both ratios... correct?

Much like in Vegas with the pan and crop keyframing features. I've found those features usefull in the ability to keyframe the footage's vertical position.. to help stabilize and compose the images within an overlaying letterbox. Just in case it wasn't shot within the guides correctly in the first place.

So in this instance, you'd have left and right adjustability.

This of course, might be too much of a painfull process for some people.

Hey Rob... GET SOME SLEEP! You guys have done a fantastic job here and have really put DVinfo on the bleeding edge this time. Thanks for all your efforts!

i wonder if this means we can get REALLY REALLY crazy and go for 2.35 =).

Boyd: and that's what I did, 5 hours later. A problem at work
reared its head. Oh well. Thanks!!

Well, I think that the 16:9 mode should offer guidelines for 1.85, for theatrical blowup. Even though it's only a slight difference, one is still chopping off a bit of information for that eventuality.

Charles: that might be possible with the SDK (who knows yet).
When I return from my vacation I will see if I can take a look at
this. I know you would probably want an outline anyway
(dotted?) so it would be interesting to see if we could add such
a feature. I'll keep you posted on this.