Figure 1. Image displayed in its correct aspect ratio on a 16:9 screen with a center cut for 4:3 viewing. That is the same anamorphic image displayed on a 4:3 screen.
Today, we are experiencing a rapidly evolving television viewing experience. The analog-to-digital evolution from SD to HD system adoption is occurring. It is believed that 100 to 120 stations per year add — and ultimately transition their operations to — digital production and distribution methodologies. We are in the early stages of the changeover cycle. Broadcast entities tend to go in 10-year cycles, and this transition is proving to be no different. Two years down, eight to go.
SD and HD adoption
We are in the amazing midst of high-quality, lower-cost HD video appearing in the marketplace at both the consumer and professional broadcast levels. There is already a significant use of HD video for broadcast documentary work. And the availability and economic hurdles of shooting some form of HD are no longer barriers to adoption.
Similarly, as more stations offer HD broadcasts, we will naturally see the creation of HD-originated programming. Whether or not we experience the immediate cutoff of SD broadcasts, today's television program makers will constantly be reminded of the complexities of feeding both SD and HD audiences.
Anyone who has watched a classic epic, such as “Lawrence of Arabia,” shot in an amazing aspect ratio of 2.20:1 on a 4:3 television screen can attest that the experience is less than desirable. A number of objectionable artifacts are present:
Conversations take place between people while you see only one person.
People are much taller and skinnier.
Titles and credits disappear off-screen.
The wide vistas of such a classic are terribly compromised.
Figure 1 on page 88 is an example of the obvious effects of improperly displaying the original aspect ratio of an image. For the viewer, the solutions are to either view the letterboxed material in its original aspect ratio on the 4:3 screen or to enjoy it with considerably less letterboxing on a 16:9 screen.
Challenge: Today's programming
Some of the pertinent questions to consider when one is faced with original productions for today's television marketplace include:
What is the application? News, documentary or film?
What resolution is required?
What format should be used?
What frame rate should be used?
What aspect ratio should be used?
How will I derive my HD and SD masters?
How will graphics, titles and credits be created?
How will program interchange with rest of world be accomplished?
These questions demand television content producers closely examine the distribution format necessary for both a domestic and worldwide audience. When one takes into account the number of distribution formats that are necessary and that may have different aspect ratios, frame rates, resolution requirements and language requirements, it is not unusual to find that upwards of a dozen actual distribution masters become necessary.
Current acquisition and post methodologies for news programming do not demand a change in frame rate. It is expected that content will continue to be acquired at 29.97fps for news applications. Certainly, there will be a migration from 4:3 to 16:9 aspect ratio support.
Creative challenges and issues
The producer who is faced with creating a program for 16:9, 14:9 (in the United Kingdom) and 4:3 aspect ratio consumption must consider what origination format and aspect ratio to use. If the program is shot in a 16:9 aspect ratio, how will the 4:3 master be derived? What happens to editorial timing? If we see a character enter the frame in the 16:9 version, it could be several moments (seconds) later that we see the character in the 4:3 version. What aesthetic issues suddenly become important considerations?
The woman cut out of the right image in Figure 1 on page 88 is an example of what can occur when the original aspect ratio of the acquired image does not take into account the display format. The solution for the producer is to shoot for the 16:9 aspect ratio but protect the important aspects of the image for the 4:3 aspect ratio. In so doing, the producer can then derive a 4:3 center cut from the 16:9 image for 4:3 SD viewing. While pan and scan versions of feature films exist, they are decreasing in frequency, and it is extremely rare to find HD-originated television programming with a pan and scan version for obvious economic reasons.
Graphics, titles and credits
There are similar aspect ratio and placement considerations for the creation of graphic elements and titles that are often overlooked. If a production has originated in 16:9 and is deriving a 4:3 center cut, the creation of all graphic elements should ideally be accomplished in two separate stages.
Figure 2. The figure at right is an obvious distortion of shape and elements that occurs when a graphic originally created for 4:3 is stretched to fit a 16:9 display.
Originating titles and graphics in 16:9 and then resizing to 4:3 (or vice versa) has been the method budget- or time-challenged producers have relied on. (See Figure 2) However, it is much more desirable to generate two sets of original graphics elements for both 16:9 and 4:3 transmissions.
The four dimensions of a format
The ultimate goal for producing content for television is to master once and derive all required formats from a single master. There are four dimensions for us to examine:
Whenever possible, acquire content in the highest possible resolution matched to the resolution that will be required for distribution. If the ultimate goal is an uncompressed 1920×1080 HD master, all attempts should be made to acquire at that resolution.
- Aspect ratio
For simultaneous television delivery of content in both 16:9 and 4:3 aspect ratios, it is desirable to acquire in 16:9 and protect for 4:3 transmission via a center cut.
- Frame rate
Which frame rate should we acquire at? 24fps, 25fps or 29.97fps? For purposes of program interchange, at the non-news level, the most logical choice for program mastering is 24fps. It is interesting to note that while news production will continue to maintain its 29.97fps heritage, this may undergo re-examination. As videotape acquisition for news decreases, and as the broadcast infrastructure migrates to a data-centric one, will 24fps adoption be considered?
- Scan method
Interlaced or progressive? A progressive scan method allows for the superior derivation of interlaced deliverables. The converse, deriving a progressive master from an interlaced master, requires conversion methods that may yield objectionable spatial and temporal artifacts.
Through the use of a 16:9 HD 24-frame progressive master, one can then derive all NTSC, PAL and 4:3 deliverables. Graphics should be originally composed for both 16:9 and 4:3 aspect ratios. Through the use of template-driven proxies, even on-the-fly substitution of aspect ratio correct graphics can be accomplished. A template-driven proxy allows a content distributor to insert the appropriate graphics (either the 4:3 set or the 16:9 set) as the program is being aired. This is accomplished through the use of triggers within the template that refer to a lookup table of graphics.
The changing delivery and viewing experience
As we create content for the HD and SD television marketplace, we will certainly face issues as we encounter the changing distribution and viewing experience. When we begin to view content delivered to us over IPTV and when we view content on portable devices such as telephones and digital signage venues, a new host of challenges becomes apparent. What we call the television marketplace today will change rapidly in concept and definition in the near tomorrow.
Tom Ohanian is senior director for Autodesk Consulting Media and Entertainment.
The preferred format for HD and SD television delivery
In order to meet the needs of today's and tomorrow's demanding market for television program delivery, it is desirable to:
Acquire in the highest possible resolution.
Acquire in a 16:9 aspect ratio.
Acquire at 24fps for TV and convert to 29.97fps. News remains at 29.97fps for acquisition.
Acquire in progressive scan.