Two items decide the shape of our video images: the aspect ratio of the screen and the aspect ratio of the actual pixels (picture elements). Let’s tackle the oldest first – screen aspect ratio.
Screen aspect ratio has probably been kicked around since cave paintings, but it really started to matter when TV’s popularity threatened motion picture theater attendance. The 35mm film frame is essentially a 4×3 aspect ratio, much like standard definition television. That’s where early TV developers got the idea in the first place. Certainly studios had created wider-screen films before the dawn of TV, but in order to offer something different when TV first “threatened”, motion picture studios, directors and directors of photography adopted a wider aspect ratio – to make the motion picture image something grander. Since then, we have come to think of a wider screen as a more “cinematic” image.
In cinematography you generally get a wider aspect in one of three ways:
A) You may use anamorphic lenses that optically squeeze the horizontal plane of the image to fit into the 4×3 frame of the negative. This is expanded back to a normal appearance during projection by using projector lenses that reverse the effect. This method is generally used for the widest-aspect motion pictures.
B) The top and bottom of the film frame is cropped and only the center portion is used for the actual release prints. Sometimes this is cropped with an actual mask in the printing of the film and sometimes it is left up to the projection house to correctly mask the image. Cropped films are the most common ways films are produced these days.
C) Sometimes Super 35mm or Super 16mm cameras are used. Here the optical path of the camera exposes a wider image onto the negative by using part of the space on the edges of the frame, which isn’t generally exposed in standard 35mm or 16mm. The full frame is used top to bottom, plus the image goes farther to the outer edges of the negative.
When we discuss ratios of film and video, film terminology is usually applied. In film terms, standard TV uses a 1.33:1 aspect ratio (4×3). Widescreen film, exposed using method “A”, is usually 2.35:1 (when projected) – and “standard” theatrical film releases using method “B” creates a ratio of 1.85:1.
As we move into high-definition TV, a new figure has entered into our vocabulary: 16×9. When HDTV was first being developed it used a ratio of 5×3, but as it changed from analog to digital technology, the “math” worked better in a ratio of 16×9. Another reason for this ratio is that it formed a good compromise to accommodate widescreen cinematography. In “film” terms, 16×9 is expressed as 1.77:1, which is fairly close to 1.85:1, used for many “standard” theatrical films. You could fit most the image of a “standard” motion picture into the HD frame, without having to crop as much of the edges, as is currently done in NTSC and PAL broadcast.
When you work in the film world, changing between various aspect ratios is relatively easy. You can go between 2.35:1 and 1.85:1 simply by changing the camera lenses when you shoot and then changing the projection lenses and the theater curtains when the film is projected. Unfortunately, it’s not that easy in video. HDTV uses imaging that is truly 16×9, but you can also produce 16×9 projects in standard definition NTSC and PAL – a 4×3 medium.
Here’s where it gets confusing. When a standard def camera is switched into the 16×9 mode, it is doing the electronic equivalent of changing to anamorphic imaging, just like changing lenses on a film camera. The image recorded to tape is a horizontally-squeezed image that is recorded in a 4×3 aspect ratio. Any edit system and VTR can deal with standard definition 16×9 video, since the frame is really only 4×3. Your system doesn’t really “care” what it looks like. To view a 16×9 image properly, you have to use a display system (projector, monitor, etc.) that spreads the image back out to its correct appearance. For instance, display screens like widescreen 42” plasma panels will stretch these images to their proper aspect.
If you have a 16×9 master and are going to present it on displays designed for only standard 4×3 images, you will have to take an extra step and produce a “letterboxed” version for that presentation. This is produced by digitally compressing the vertical height of the image to counteract the anamorphic quality of the image. Generally 77% of the image height looks about right, which leaves you black “bars” on the top and bottom of the screen. Remember that this letterboxed version should not be considered a 16×9 tape, since it is no longer “squeezed”. If a letterboxed image is used in our plasma screen example, the screen will zoom into the image to get rid of the top and bottom black bars, thus reducing the resolution of the image.
Many productions now shoot in HDTV so that they have archival footage with some shelf value for the future, but are currently still posting in standard def. The 16×9 HDTV footage is first “downconverted” to NTSC (or PAL). In this conversion, footage can be transferred as 4×3 with cropped edges, 16×9 (4×3 anamorphic or “squeezed” image) or letterboxed. The cropped images are often problematic, because 16×9 framing may appear too tight in 4×3. If you transfer this as letterboxed, you cannot recover the full vertical resolution and must stay letterboxed. It is generally best to transfer your HDTV footage as 16×9 “squeezed” and then deal with the correct aspect and framing as part of the post-production stage.
The worst situation is if you don’t know how your final production will be viewed. For instance, shot composition and graphic design is not the same if you are producing for 4×3 as compared with a wider 16×9. If you are producing in 16×9, but the show will be shown also in 4×3, you might have to make creative compromises that you’d rather not have to make. Sometimes producers will create three versions: a 4×3 full screen, a 4×3 letterboxed and a 16×9 version. If this is the case, two sets of graphics might need to be created and 16×9 graphics must also be “squeezed” to match the video. Often “pan-and-scan” methods are used so that more-widely composed shots can be repositioned for better framing in the 4×3 full screen version of the show.
Another issue with which producers now wrestle is how to incorporate regular 4×3 footage into an otherwise 16×9 show. If the 16×9 footage is only to be used in a cropped or letterboxed 4×3 master, then fine; however, if a true 16×9 master is to be generated, then the 4×3 footage poses some challenges. Since the 16×9 footage is edited in a squeezed format, the 4×3 footage must also be horizontally squeezed to match, so that everything looks right when the master or dubs are expanded in playback. This will leave black bars on the left and right sides of the 4×3 footage. You usually cannot get away with zooming in enough to avoid this, because the amount of expansion you have to do will degrade the footage too much. Some producers use the solution of building a 16×9 graphic background for these 4×3 images. This provides a tasteful way to incorporate them into the program without calling attention to the fact that they don’t match the format of the other footage.
Generally 16×9 shows are presented in a letterbox format, so that the creative integrity is preserved, but one method employed by the BBC is the 14×9 ratio. In this approach, the show is shot in 16×9. During post, the image is letterboxed, but also “blown up” slightly, thus cropping the left and right edges a bit. This preserves most of the wider aspect, yet doesn’t leave as much of the top and bottom of the screen blank. Usually only one set of graphics need be produced.
Pixel aspect ratio
If your head isn’t swimming yet, let’s talk about pixel aspect ratio. Computer screens use square pixels. TVs and video monitors display using rectangular pixels. This difference isn’t a big deal until you go back and forth between the two.
Standard def NTSC images are 720×486 (720×480 for DV and DVD – don’t ask!), which is equal to computer images sized at 720×540. Designers and animators have to compensate when going back and forth, so that items like circular logos appear correct at one destination or the other. For instance, graphic artists working in Photoshop create images in 720×540, but then convert them to 720×486 as the last step in the process of going to video. In the other direction, editors exporting web videos in the Real, Quicktime or Windows Media formats convert their images from 720×486 (non-square pixels) to 720×540, 640×480 or 320×240 (square pixels). This results in the proper appearance when viewed on a computer display.
One of the great things about HDTV is that this whole issue is gone. HDTV uses square pixels, so that a 1920×1080 image in HDTV is the same when viewed on a computer screen. At least we’re making a little progress!
© 2003 Oliver Peters