16:9 best achieved via camera or in post?
 

I don’t understand (believe me, I really have tried) if 16:9 is best achieved with the camera in 16:9 mode or with the camera in 4:3 mode and conversion to 16:9 in post. Most posts I have read indicate that the best result is achieved through in the latter case but I have also seen posts indicating that it doesn’t matter.

I have a Sony miniDV camera (PAL, i.e. 720*576) and a WideScreen TV-set. If the TV is set to automatic, 4:3 footage is shown full screen. Using this setting 25% of the height is lost. If I set the TV to 4:3 the whole image is visible but there are big black bars on the left and right side of the screen. Because of this I started to shoot in the cameras 16:9 mode. This is recognised by the TV as WideScreen and none of the 16:9 recorded footage is lost in viewing (I know that I have recorded 25% less than in 4:3 mode). I like this better than the 4:3 alternative since I fill the screen with 100% of the recorded image.

Now to my questions. It has been said that shooting in 16:9 with a camera that has a 4:3 chip result in loss of resolution since only 75% of the chip is used to record the footage. I understand that only 75% of the chip is used but I don’t understand why this, as such, result in loss of resolution. Isn’t the 75% of the chip that is used on the 16:9 footage the same 75% that would have been used on the very same footage should I shot in 4:3? If I connect the camera directly to my TV, the WideScreen image is the same image that I would see should I have used the 4:3 on the camera and the TV-set to automatic? I can understand (I think) that there is a loss in resolution should the footage be shown on a 4:3 set and only 75% of the chip been used but isn’t that a different thing?

PAL DV setting is 720*576 and I guess that if I record in 16:9 I will only use 720*432. If I take this to a NLE (such as Premiere) with WideScreen setting my 720*432 footage will be 720*576 and then after encoding when shown on the TV set it will show up as WideScreen. I can understand that there is a loss in resolution with using 720*432 as compared to 720*576 but isn’t this something that I have to live with as long as I don’t have a DV camera with a 16:9 chip? Doesn’t the same thing (loss of resolution) happen if I record in 4:3 and mask to 16:9 in post production?

Adam Wilt has a great article at http://www.adamwilt.com/DV-FAQ-etc.html about just this issue.
The thing is that to get the best image out of a non-16:9 CCD camera, it depends a lot on when the camera does the upscaling of the cropped image, i.e. before or after compressing the video stream.
AFAIK the XL1 does the cropping and upscaling _before_ compressing the image data, and so theoretically it would be preferable to use the in-camera 16:9 than to do the cropping and upscaling in post, as more video bandwidth is actually used in the compressed image. Think about it; scaling uncompressed (and beautiful) image data will work better than scaling 4:2:0 or even 4:1:1 DV data, as there is more information to base the interpolated image data of the upscaled 16:9 data.

However, using scaling in post will allow you to re-frame you image (though only vertically), and to many people this advantage outweighs the (barely visible) image quality advantage.

I actually just shot some 16:9 straight in the camera, and must say, I am very pleased with the result. Until now I just discarded the idea of shooting 16:9 but after this weekends experience I might do it more often (though not for larger projects).

Part of the reason why it came out so nice is because I used the ND on a medium bright day, so I could shoot at full-open iris and slow shutter speeds. This gave really nice blurry backgrounds and saturated colours...

HTH,

Kai.

I tend todo widesceen in post, primarely because I want to:

- frame the image exactly as I like it
- choose later which widescreen aspect I'm going to use
(16:9, 1.85, 2.35 etc.)
- I have options available to use the footage in different
ways (picture in picture comes to mind (without it being
widescreen)), publicity stuff etc. Heck, I might even want
to do a 4:3 output if people bang my door down for it.

I just think it gives me the best option vs. quality ratio.
But this is a personal thing ofcourse. Some people want
the 16:9 increased resolution (which without a anamorphic
attachment also means decresead resultion, talking about
confusing) and just frame more carefully.

What ever suits your need. I suggest you shoot test footage
(and actually edit it and output it to your final medium of choice)
to see what YOU like best!

ive been a user of the xl1 for about 2.5 years now and im still unsure about
the 16:9 ratio it provides,

what is the true resolution for 16:9 xl1? i use it seldom, yet when i do i squeeze the heigth to conform to 360 and leave the width at 720.
does anyone know if this is the proper way to do this? I know this is not
the official apect ratio yet it works and looks good, though i would like to know the technically correct way to display this when done.

i have shot with 16x9 anamorphic lens.

some thoughts on 16x9 anamorphic and 4x3 letterboxed.

here's is what i find. anybody viewing these true 16x9 projects are viewing a letterboxed version as only 1% of TV are 16x9. very few TV's ( 4x3) can do 16x9 .. and the ones that do are converting it into a letterbox .

i cannot see the difference on my monitor ( sony 14M2U)between the 16x9 anamorphic and the fake in camera 16x9 ... my monitor screen is just too small to see the difference ( it does have 600 line resolution) ... i do see the difference on digital projection ( 10 ft screen) PROVIDED the digital projecter can do native 16x9 !! many under 5K digital projectors do not do native 16x9 they letterbox it ( basically they project 4x3 and when projecting 16x9 they letterbox it )
when a projector has to resize 16x9 to letterbox artifacts are introduced.

which brings up LETTERBOXING ... i've noticed that when i take my 16x9 ( either fake camera or true ) and i
letterbox it so everybody can watch it on their 4x3 TV the resizing of the 16x9 ( to fit inside 4x3)
causes interlace artifacts.

from my test shooting normal 4x3 and then letterboxing it does not introduce the new interlace artifacts ( it is NOT resizing the image only putting the black at top/bottom)

true 16x9 does look better then letterboxed 4x3
when viewed on a true 16x9 TV ( 30, 40",50") but it does not stand out on 4x3 TV's letterboxed (sony 24", 27" 32") over the shot 4x3 then letterboxed clips. again IMO
the resizing of the true 16x9 to fit the 4x3 tv introduces artifacts. the more expensive TV have less artifacts.

so what is one to do if ALL viewing is going to be letterboxed?

we live in a 4x3 world. the masses are still buying 4x3 TV's
and some can handle 16x9 BUT it will be letterboxed. HD 16x9 is NOT catching on in the US. seems ABC, CBS, FOX, NBC all have different HD formats. the FCC has few rules and ALL the HD formats do meet their DTV HD standard
when you go to buy a HD TV the HD tuner is NOT included ??

if going to FILM there is NO question on shooting true 16x9 will show it's better resolution ...

DVD are 16x9 BUT 99% of viewers are watching DVD's letterboxed on a 4x3 TV NOT on true 16x9 TV's ...

just my observation on 16x9 ...

I'm not sure if I have unusual 16:9 TVs, which may explain my confusion on this.

I have two Mitsubishi 16:9 projection HDTVs. They both have THREE viewing modes:

4:3, where the TV puts a gray vertical bar on each side of the screen, and uses the central, 4:3, portion to display whatever it's given. If this is 4:3 material, then it fills that portion of the screen. If it's 16:9 material, then it occupies the full width of this central portion, with black horizontal bars above and below. Clearly, you don't watch 16:9 material in this mode.

EXPAND, where the TV stretches the image horizontally. 4:3 material is just streched in the horizontal plane, and this is how I watch most cable/satellite material. You get used to it :-). Anamorphic 16:9 material, i.e. material that looks really tall and thin in 4:3 mode, stretches out to look correct. I've found anamorphic material to be quite rare.

ZOOM, where the TV stretches the image horizontally and vertically. 4:3 material has its top and bottom cut off in this mode, but circles still look round versus elliptical. Letter box material, which is what most DVD movies provide, either fills the screen (in the case of 16:9 material), with circles looking round, or has small horizontal black bars at top and bottomm (in the case of different widescreen formats).

Is this normal for widescreen TVs? If not, it seems pretty useful. My assumption is that I can either shoot anamorphic 16:9 material on my XL1S and play it through my TV using EXPAND mode or, more usefully, shoot 4:3 and letterbox it in post, and use ZOOM mode to fill the TV screen. Although, based on some posts that I've read here, shooting 4:3 and letterboxing loses some quality versus shooting 16:9, it seems to me that shooting in 4:3 gives me more control over framing in post, and the ability to use 4:3 if I so choose. Also, letterboxing gets rid of the annoying black bar at the bottom of the picture, discussed at great length in another thread.

So, what do y'all recommend?

1. Shoot 4:3 and letterbox to 16:9
2. Shoot 16:9 and create widescreen output in Premiere, i.e. the widescreen pixel aspect ratio. My belief (which may be wrong) is that I'm still outputting 720x480, but with a different pixel aspect ratio, which my TV will show as anamorphic (tall and thin) or correct based on the TV setting. AFAIK, my TV doesn't detect pixel ratio codes embedded in the video.

Thanks! Alex.

hmm...

I know that the XL1's 16:9 pixel ratio is 640 x 360 pixels.

now when outputted to 4:3 in letterbox form, is there
a loss in resolution by the width of the picture stretching to
720? theres what seems to be a 80 pixel stretch from the original
16:9 ratio. if not then it would seem like the proper way to use the XL1's
16:9 format is to take the 80 pixel loss which in turn creates black bars on the
sides as well. this doesnt seem right? or is it?!!!!

Oh! Big surprise! I didn't realize that the XL1 (and, presumably the XL1S) output 640x360 pixels in 16:9 mode. Of course, this is DV, so, whatever the capture resolution at the CCD, the camera outputs 720x480 over Firewire. But, if it's cropping pixels at the sides, as well as the top and bottom, of the CCD array, then this is bad. It would make much more sense to capture in 4:3 mode, and then crop the top and bottom. Given a pixel aspect ratio of 0.9, my calculations indicate that you'd crop 58 pixels each off the top and bottom. This gives you a letterbox image that's 720x364, which is a ratio of 1.98. Adjusting for 0.9 pixel aspect ratio, that's 1.78, which is 16:9. Is this correct, or am I being completely stupid here?

that seems correct,

it would seem more legible to crop in post then to shoot in 16:9,

that way you done lose resolution. funny how that works.

surley the poeple at Cannon couldnt expect this ratio to be put to use
in a standard NTSC format without the loss of resolution. unless im wrong.

Is there any verification that the XL1 outputs 16:9 at only 640 accross? It seems like more people would know that since 16:9 usage has been a big question about the XL1 since its release pretty much.

F-stops and T-stops are easier to understand than all of this garbage =).

I'll just jump in with my pennies:
I shoot a show locally on the south shore here that, for many reasons is framed in a 1.85 aspect. I always shoot in 4:3 and then add letterboxing to it in post, outputting a 4:3 image. Since its for cablecast, and the majority view it on conventional tv's, we felt we would just go the safe route so that everyone could view it properly one way or another.

number, number, numbers...

When it comes to CCD matrix'es and faux 16x9 there is only one rule; test it. CCD resolutions are theory. The new Sony 1/4" camcorder has over 1 000 000 resolution...in theory. Does it produce a better image than the DSR-300 400 000 pixel CCD's? Absolutely not!

If 16x9 hold up in faux mode you should always use it if you plan to sell your program a few years down the line. DTV is a fact and it will be growing. In Europe (where I live) 16x9 television sets are big sellers. Cropping a D1 image and blowing up the middle lines to 16x9 is a bad solution. You really need those horizontal lines your codec is calculating. Many codecs don't tend well to that type of resolution scaling (especially DV). This will result in an unsharp image that will look like absolute crap on DTV in a few years.