DTV Aspect Ratio Signaling

One result of DTV is that we now frequently find ourselves dealing with multiple aspect ratios. A multiplicity of picture aspect ratios has always existed in film and cinema, but decades of television engineers were relatively well insulated from this reality. This is no longer the case.

In the United States, television had one aspect ratio – 4:3 – and the FCC regulations on vertical and horizontal blanking made it mandatory that the screen be filled with picture. When the wide-angle theatrical formats were shown on television, they were always panned and scanned so that the TV screen was filled. When the blanking regulations were removed in the 1980’s, it became legal to display pictures on television that did not fill the entire screen, thus making it possible to broadcast letterboxed images.

Although wide-screen theatrical material has for some time been routinely transmitted in letterboxed format on European television, this practice has taken some time to become established in the United States. It is still resisted by many U.S. viewers and broadcasters.

DUAL-ING ASPECT RATIOS

Thanks to DTV, we now have two native video aspect ratios. In addition to the traditional 4:3, 16:9 has become a fact of life. It is the native aspect ratio of all HDTV formats and it became established as a standard-definition aspect ratio as well, particularly in Europe. This, like many other things, complicates DTV.

Right off the bat, DTV displays – both HD and SD – come in both aspect ratios, even though 16:9 is the native HDTV aspect ratio worldwide.

Further, set-top boxes that are intended to display HD and SD DTV images on NTSC TV sets must necessarily display on a 4:3 screen. When a 16:9 DTV picture is displayed on a 4:3 display, three possibilities present themselves.

If the picture fills the width of the screen, there will be black areas above and below it.

If it fills the height of the screen, portions of the right and left edges of the image will be cropped off.

The third option is to distort (squeeze) the image so that it fits the screen.

The first two of these options result in the images being smaller than optimum, and the third produces "Goyaesque" people. Similarly, when a 4:3 image is displayed on a 16:9 screen – unless the image is cropped vertically or stretched horizontally – black areas will exist on the left and right sides of the screen. A term that has been frequently used to describe this is "pillarbox".

One of the display problems that HDTV broadcasting has spawned occurs when 4:3 material is up-converted to HD. We know that this is done frequently, as a substantial portion of the broadcast day of many DTV stations consists of an up-conversion of the NTSC programming.

And up-converted commercials and other non-program material are often integrated into HDTV programs. This typically produces a pillarboxed image in a 16:9 field that, in turn, is letterboxed onto a 4:3 display. The result: black bars all around the image with the viewers seeing a "postage-stamp"-sized picture afloat in a sea of black.

RATIO-NALIZATION

It would be advantageous to have a mechanism by which to signal a DTV receiver about aspect ratio information. For example, in cases where 4:3 material is up-converted to HD – as described above – if the receiver knew that this material was really 4:3, it could display it as 4:3 so as to fill the screen rather than make a postage stamp out of it.

There are, in fact, some standards in existence that do define ways to transmit aspect ratio signaling data to the receiver. There is a method described in EIA 608B whereby the aspect ratio data is carried in the extended data services packets of Line 21, but there has been no implementation of this capability on either the transmitting or the receiving end, and no reason to believe that there ever will be, as the aspect ratio problem is not a big one in NTSC.

In the DTV world, the DVB standards that are used in Europe and other places in the world have a standardized data structure called "active format description," or AFD. The AFD data is carried in the user data portion of the MPEG-2 video elementary stream.

AFD provides for a 4-bit field describing the active aspect ratio format. The formats supported by AFD include some that we would find useful in the United States.

The formats that we would find most beneficial include:

• Active format is the same as the coded frame. No letterboxing or pillarboxing is used.

• 4:3 (centered). 4:3 video is pillarboxed inside a 16:9 frame.

• 16:9 (centered). 16:9 video is letterboxed inside a 4:3 frame.

• 16:9 with protected 4:3 center: This could be used to tell the receiver how to extract and display a 4:3 image from a 16:9 transmission.

Other formats include:

• 14:9 (centered). 14:9 video pillarboxed inside a 16:9 frame or letterboxed inside a 4:3 frame.

• >16:9 box (centered).

• 16:9 box and 14:9 box at the top of the screen (all vertical black area is below the active image).

• 4:3 with protected 14:9 center.

• 16:9 with protected 14:9 center.

Most of the above "other" formats are relevant to formats and practices that are not found in the United States. An exception is the "greater than 16:9 box".

Wide-screen cinematic formats such as Cinemascope are sometimes transmitted letterboxed in a 16:9 frame in the United States, but it is difficult to imagine what a receiver with either a 4:3 or a 16:9 display would do with such an image other than treating it as a 16:9 image.

The 14:9 aspect ratio, a compromise that produces "equal pain" when boxed in 4:3 or 16:9, is a development of the BBC that has not found favor here. The practice of pushing letterboxed pictures to the top of the screen is not done in the United States either.

The incorporation of the DVB active format description into the ATSC standards is currently under study. If it is standardized and implemented in DTV transmissions and receivers, it would facilitate the optimal use of the display screen for any transmitted aspect ratio.

Randy Hoffner