Best Way to Mask XL2 Viewfinder for "Fake" 16:9!

[Greg Boston]

I just went through this learning experience having shot my first true 16x9 24p project.

I just recently started using the Mac and have the production suite. In FCP, you need to set your capture pref to NTSC DV with the ALL IMPORTANT anamorphic box CHECKED! Then, make sure your sequence preset is set the same way as your capture preset. You will be working in a nice 16x9 screen. Save the edited sequence to self contained QT movie with the same settings above. Since I shot in regular 3:2 24p, I left the time base at 29.97.

Bring the movie into DVDSP as an asset. Right click(or command-click) on that movie in the asset pane and be sure to check the anamorphic flag in the encoder preference. This will make sure that DVDSP knows that you want the anamorphic flag set in the mpeg2 render. For your menus, you will have to set the checkbox in the inspector pane for 16x9.

The only thing that has to be set for a pixel size other than 720x480 is graphics from other software such as livetype or motion. For these, use a setting of 854x480. Then burn your finished project to DVD.

Now, as stated above, the DVD player SHOULD have a set up menu where you tell it which type of tv you have. When it plays the DVD, you will see a full screen image if you own a 16x9 or a letterboxed 16x9 on a 4:3 screen if that's what you have.

It worked for me! If anything, Ram should be asking how to set up 4:3 framing guides on his 16:9 VF. I do wish Canon would have included those. Otherwise, you can always sit though your footage in the edit doing a pan and scan which you can also tell your project to use in DVDSP. My DVD player also offers the pan and scan option if I set it for a 4:3 display output.

The main thing I wanted to stress after 'doing it wrong' myself is that you need to keep those video files at 720x480 all the way through. If you start trying to change the pixel dimensions to fit a widescreen format, you will wind up with a severe quality hit and time hit cause it trys to reformat all your footage when all you have to do is make sure to check the anamorphic button whereever you find it from capture through finished DVD.

I was very happy with the footage as displayed on a 65" WS RP HDTV.

Have fun...

-gb-

[Pete Bauer]

David,

Your reminder to import a project rather than export/import uncompressed files is excellent. And of course in PPro you have unlimited nesting of sequences within a project. I just verified that it works fine to import a 16:9 PRPROJ (and all its associated media files) into a 4:3 project. No problem...I don't know why you're not able to do so, but I'm guessing some buried setting...

I'm embarrassed that I didn't think of that to begin with! I'm sure I must have posted that before I finished my coffee this morning!
-----
Richard,

Yes, we both contributed to a painfully long discussion on picket fences. ;-)

http://www.dvinfo.net/conf/showthrea...ght=resolution

I understand your point about the in-camera downsampling, but then what's the net effect of that along with horizontal pixel shift and 12-bit processing prior to DV compression? In truth, none of us know just how the electronic guts are engineered. The bottom line is the stellar image that's firewired out of the camera, so about the only way to really know for sure would be a carefully done resolution comparison of the 720 widescreen pixels to the 720 narrow screen pixels (...or was that picket fences?!!!!)

If any discernable difference is found, I'm sure it would be slight anyway. Shooting 16:9 and letterboxing is easier in terms of framing shots, better for minimizing quality loss from rendering various effects, and for later re-purposing of edited 16:9 video. Pending any surprising outcomes in a 16:9 vs 4:3 shoot-out or other reasons to shoot 4:3 that none of us has yet considered, I still personally would shoot in 16:9. But then, it'll all work out fine if Ram decides to go with black tape and straight 4:3 all the way, too.

[Richard Hunter]

<<<-- Originally posted by Pete Bauer :

If any discernable difference is found, I'm sure it would be slight anyway. Shooting 16:9 and letterboxing is easier in terms of framing shots, better for minimizing quality loss from rendering various effects, and for later re-purposing of edited 16:9 video. Pending any surprising outcomes in a 16:9 vs 4:3 shoot-out or other reasons to shoot 4:3 that none of us has yet considered, I still personally would shoot in 16:9. But then, it'll all work out fine if Ram decides to go with black tape and straight 4:3 all the way, too. -->>>

Hi Pete. First, let me say that I agree with everybody who says the XL2 images are great, and that I'm not in any way trying to put down what this camera can achieve. If someone prefers the workflow of shooting 16:9 and then resizing it for 4:3 output I am sure the results will be OK (in fact I am very sure, because I do this myself).

The point I am trying to make is that I believe the 4:3 mode on the XL2 produces the best image quality, from a technical viewpoint. Even though 16:9 can produce great results, if the end result is a 4:3 video (with or without letterbox bars) then 16:9 can never be as good as shooting 4:3 to begin with. Talking about shooting 16:9 and then using uncompressed for intermediate steps in order to reduce the resampling degradation doesn't make much sense to me when you can easily avoid the resampling simply by shooting 4:3. The generation loss of a good DV codec (like Vegas or Canopus) is very low.

Regarding Rams' other point about rendering, adding black bars to a 4:3 project should give faster renders than resizing a 16:9 video, but I suppose it depends on the NLE. I will give it a try in Vegas and see what the difference is.

Richard

[David Lach]

David,

Your reminder to import a project rather than export/import uncompressed files is excellent. And of course in PPro you have unlimited nesting of sequences within a project. I just verified that it works fine to import a 16:9 PRPROJ (and all its associated media files) into a 4:3 project. No problem...I don't know why you're not able to do so, but I'm guessing some buried setting...

Are you saying you're able to import a nested sequence coming from a 16:9 project and have it properly letterboxed in a 4:3 project? Because I sure can't. I can import a 16:9 video file and when the "Scale clips to project dimensions when adding to sequence" option is checked, it will be automatically letterboxed.

But if I drag the nested sequence into the timeline instead of a video file, it does not letterbox it, it just crops the sides and there is no way I'm aware of to get the full 16:9 image back in the 4:3 monitor window.

[Ed Pierre]

<<<-- Originally posted by Steev Dinkins : Not to enter into platform battle, but Final Cut Pro can easily nest sequences (drag and drop one edited sequence into another). After Effects is able to do this easily as well. It's hard to believe Premiere doesn't support this.

By nesting a 16:9 anamorphic sequence inside a 4:3 sequence, FCP automatically does the right thing, and doesn't require exporting first. Just make sure to purge renders if you're not working uncompressed, so it's not double compressing DV25 format.

Steev is it possible to shoot 16:9 using the xl2 with final cut pro and output as 4:3 without the letterbox. I am shooting a project that I am still undecided as if I should shoot 16:9 or 4:3 maybe you can help me decide.
It's intented for tv, but I want the film look and I think by shooting 16:9 it will more like film.


Also, if quality and hard drive space are both a concern, you can export an uncompressed-processed sequence straight out to Mpeg-2, without exporting a file first. This way you can apply color smoothing to rid the blocky artifacts in DV25 and get a better quality Mpeg-2 encode.

Again, I would definitely not shoot 4:3 and crop that to widescreen in post, when you have an XL2 with it's gorgeous 16:9 imaging path (with proper framing in your viewfinder - IMPORTANT), from which you can scale to a 4:3 image in post if need be.

If you were really paranoid about this scaling process, do the scaling on an uncompressed export of the edited piece in After Effects with it's superior scaling. -->>>

[Steev Dinkins]

<<<-- Originally posted by Ed Pierre :Steev is it possible to shoot 16:9 using the xl2 with final cut pro and output as 4:3 without the letterbox. I am shooting a project that I am still undecided as if I should shoot 16:9 or 4:3 maybe you can help me decide. It's intented for tv, but I want the film look and I think by shooting 16:9 it will more like film. -->>>

Well, sure, you could shoot 16:9, then crop off the left and right sides to make it 4:3, but I wouldnt' do that. If you are wanting the final output to fill a 4:3 screen, shoot 4:3 in the camera.

The decision to shoot 4:3 vs 16:9 is a pretty lengthy subject. I'm shooting my creative projects 16:9 now since it just looks nicer to me, even though it ends up letterboxed on most TVs. I think 4:3 would be a smart decision if your target audience is 4:3 broadcast and it's a commercial endeavor. I don't have anything airing on broadcast, so I pretty much go with whatever I want. Also, I think 16:9 looks nicer on the web too. One more argument would be that the XL2 16:9 footage would uprez to HD a lot better, once that becomes mainstream. I am archiving work in uncompressed format for that purpose.

[Ram Purad]

Wow... I have created a monster! Good to see quite a few ppl are participating in this. But now with opinions coming from all directions, I'm confused! I want the "DV God" come and settle things!

To shoot in 16:9 or to shoot in 4:3 and letterbox it... that's the question.... (output being 4:3).

My intial idea was to just shoot in 4:3 but to frame it to 16:9 and then add letterbox in post. I was gonna buy some electric tape from local hardware store. Looks like I have to hold that until this comes to an end. Another reason I like doing this so that I can put my logo on lower right corner of the video.... I like part of the logo to overlay on letterbox area. That sort of give a stylish look. But if I take shooting in 16:9 route, I guess I can't do that... the logo has to be in the video within the safe area.

Main thing for me is minimizing rendering time as much as possible. If I can take 16:9 route and still don't have to do much rendering then I don't mind going this way.

I have 3 4:3 weddings filling up my hard drive.... will finish them soon. After that I'm gonna clear up my hard drive and gonna do some test shoots myslef. Let u guys know with the result.

[Steev Dinkins]

Tis what discussion threads are all about - hear different views, and make up your own mind.

About putting logo in corner, you could still do that with the nested 16:9 inside 4:3.

For me, I am happy to not have to tape up my viewfinder with electrical tape and other crap anymore.

[Pete Bauer]

I finally got around to doing a resolution comparison and posted some frame grabs:

http://www.geosynchrony.com/scratchpad.htm

The short version is that unless I screwed something up, the XL2 bumps right into the theoretical horizontal resolution limit in 16:9.

Looks like it is pretty much a tie for getting letterboxed 4:3 ... shooting 16:9 and importing into a 4:3 project looks pretty much the same as shooting 4:3 and masking (for my fairly crude test anyway).

Cheers,

[Richard Hunter]

Hi Pete. It's good to see someone actually carrying out tests for this. Thanks for posting the results.

I have a question - why are the heights of the charts different for the 16:9 and 4:3 frame grabs? I would have expected that the chart size and geometry would remain fixed and only the amount of space at the sides would change. I checked on my XL2, and the visible height does not change when I switch between 16:9 and 4:3 modes.

My wife is quite tolerant too, but anyway I have a separate room for editing that she doesn't come into very much. So if I could get a soft copy of the resolution charts I'd be happy to repeat your tests using a PAL cam. Are these available somewhere for download?


Richard

[Pete Bauer]

I probably should have explained more about the images, but it was a beautiful Saturday outside as I sat obsessively trying to get this done -- as quickly as possible!

I did the frame grabs as *.bmp exports directly off the timeline in PPro 1.5.1, so each file was square pixel 720x480. That gives an image aspect ratio of 1.5 regardless of the pixel aspect ratio that was being used on the timeline. The 16:9 gets squished from 1.2 PAR to 1.0, and the 4:3 gets stretched from 0.9 PAR to 1.0.

To see it as in the timeline, the 16:9 picture needs to either be oversampled and saved at about 854x480 square pixels, or viewed with an application like Photoshop CS that can display the image in non-square (1.2 PAR) pixels. Similarly, the 4:3 pictures would appear in correct image aspect ratio if resampled to 640x480, or viewed at 720x480, 0.9 PAR. But, of course, the effective NUMBER of pixels in the images won't change, just how much they are stretched or squeezed.

The rez charts are here (the one I used is near the bottom of the page):
http://www.bealecorner.com/trv900/respat/

If you're willing to go to the bother, I'd welcome a repeat of the experiment. Even though the vertical resolution will be different in PAL, the horizontal direction is still 720px. I'm not at all an expert in these things, so getting a result indicating the camera is putting out right at, or even just a nit better than the theoretical limit of 360 line pairs across leaves me wondering a bit. Is it some electronic wizardry or despite doublechecking my math, did I not do this quite right?

In any case, this may be all but moot before long. As fine a camera as the XL2 may be now, HiDef is nipping at its heels!

[Richard Hunter]

Hi Pete. Thanks for the chart files. I'll give it a go soon (before my XL2 is obsolete :) ).

Richard