Douglas Crockford

Blog

Books

Videos

2019 Appearances

JavaScript

JSLint

JSON

Github

How JavaScript Works

About

DisplaySearch HDTV 2005

I attended DisplaySearch's annual HDTV Conference at the Beverly Hilton Hotel on August 23-24. There were 288 attendees. The picture of the industry presented this year was more stable than last year's conference. The industry has settled on common vocabulary and patterns. It is understood that there are seven display technologies in the market, and the relative advantages of each is understood.

This conference does not cover front projection. There was no discussion about new technologies, such as SED, OLED, and laser.

The Transition

Most consumers are still unaware that we are in the process of completely replacing the television system. This can be seen in their buying patterns. Only 25% of last year's TV sets were digital. Of those, only 85% were HD capable, and only 70% were wide screen.

Prices are continuing to drop, at a rate that is too fast for the manufacturers, and too slow for consumers. HDTV is not creating much demand. Mostly we are seeing the pattern of people moving up to a larger screen as part of their normal replacement cycle. This is particularly clear in markets such as Europe. People are buying HDTV sets even though there is virtually no HDTV programming available on that continent.

1080p

The next big thing is sets with 1080p60 support. This will be a key feature in the next round of large screens in all technologies. It is recognized that there is no programming available in 1080p, and isn't likely for some time. A 1080p set is thought to be better at displaying 1080i material, although that depends on the quality of your deinterlacer. 1080p also justifies larger displays.

Vertical lines screen height/viewing distance recommended screen size at 7 feet in inches
480i 7.1 20
720p/1080i 4.6 40
1080p 3.1 55

The Sony PS3 will have 1080p output.

Mark Cuban of HDNet once again called on the CE industry to force the Content Industry to deliver higher quality programming. He wants 1080p content with less compression damage. All of the delivery systems are overcompressing, which leads to an uncorrectable loss of image quality. The bigger the display, the more objectionable the damage becomes: Big screens look worse, not better. He is looking at distributing video material on terabyte hard drives. He recommended AVS Forum as a way of getting in touch with media geeks.

Many of the networks are now producing and archiving their shows in 1080p, anticipating that that will give them more distribution choices in the future.

Broadcasting

It is doubtful that broadcast can ever do 1080p because ATSC only supports MPEG2 and cannot be upgraded. This is another reason to be worried about the survival of broadcasting. Broadcasters have been overcompressing in order to do multicasting (transmitting multiple programs in a single channel). Broadcasters have been insisting that cable systems carry their extra channels. The CE industry should go to the FCC to force the broadcasters to stop multicasting so that all of their bandwidth can be applied to HDTV.

Rear Projection

Rear Projection has the advantage that cost increases slowly as screen size increases (unlike the panel technologies), so it is finding a niche at the large screen end of the market. The number one market for large displays is NA (North America). Most of the ROW (rest of the world) thinks that 30 inches is large, so RP does not sell well outside of NA. There might be a market for very large displays in China and Korea.

The most popular of the RP technologies is RP-CRT because of its low cost, but it is declining rapidly and will probably be obsolete in a few years. The RP-MD technologies are growing, but not quickly enough to match the decrease in RP-CRT, so RP as a group has been declining.

I think this is because retailers do not like RP. They prefer to sell flat panels. The flat displays have better field of view, so they tend to look better in the store. They take up less store space than the bulkier RP sets. Consumers are excited by the idea of hanging a TV on the wall, although most flat sets will not be wall mounted.

There is also a problem in the way retailers display sets. They favor ultrabright displays because they catch the eye in the store. However, such displays wear out faster, have less color fidelity, and can even be painful to watch. The industry needs to be much better at self-enforcing standards of image quality.

RP will be facing more competition from the flat panels. A 100" LCD panel and 102" PDP panel are announced by Samsung. Large panels will put more price pressure on RP.

Companies are working on thinner cabinets, alternative light sources (like LED and laser), and removable shields to protect the screen.

Dell

Dell gave a very good presentation on direct marketing of TVs. They provide free delivery and set up, and free replacement if there is a problem, and free return if the customer is not satisfied. Their website is clear and authoritative. Good online material can significantly reduce the return rate, especially when compared to retail. Their prices are good. In the industry, 88% of sets are sold at retail, but there could be growth in the direct channel.

Dell has 125+ kiosks around the country so that people can preview the sets. It is estimated that a consumer will look at a set 5 times before buying.

Misc

The battle between BluRay and HD-DVD continues. It is not clear now that either format is good enough to support 1080p without overcompression damage.

EchoStar has bought the programming half of Voom, after having bought its satellite. It will add 10 channels of HDTV this year, and another 11 channels next year.

Channel changing times on MPEG2 are quite long. With H.264, they get even worse. Channel surfing, which used to be one of the most popular ways to enjoy television, becomes virtually impossible.

There is PSIP metadata that can be used to help scalers do a better job. Broadcasters are not including it. They say it is because the sets do not recognize it. The set makers say it is because the broadcasters do not provide it. They point fingers at each others and consumers get bad pictures.

CBS is putting the Broadcast Flag in their programs, even though consumer equipment is not required to recognize it.

Consumers do not know what they bought. For example, they don't reliably know if they bought LCD or PDP.

Consumers want simplicity. Industry does not provide it.

Consumers think that an HDTV set will make everything HD. No one has told them that HDTV is a completely new and separate system.

Networks misidentify programs as HDTV.

EICTA has made an HD Ready logo program for Europe. I think it is a good idea, although they permit 15x9.

Many brands are experimenting with extra features, but I think a simple monitor is still the best choice for consumers.

Consumers think these are the most important criteria:

  1. Picture Quality
  2. Ease of Use
  3. Trusted Brand
  4. Price

But they do not know how to judge picture quality, and ease of use cannot be determined in the store. This leads to a high return rate.

Syntax gave a good presentation on bad interfaces. They showed an OSD that their own engineers proposed and showed why it was very bad. This is an extremely common problem. Most engineers are incapable of understanding the customer experience.

HP showed a couple of design features:

JPEG2000 is a very interesting compression format. It can cover the range from lossless to very lossy. It is more error resistant than MPEG, and is easier to compress with. Its noise characteristics are much nicer. It is royalty free.

There were presentations by two image processing technology vendors: Gennum and Silicon Optix. Good image processing is really important.

HDMI 1.2 was just released. It allows for RGB as well as YCbCr. It also provides support for low voltage sources. They are working on a 1.3 specification. I don't understand how it is reasonable for a connector standard to be going through revisions like this.

There was almost no mention of IEEE 1394. It seems that FireWire has lost and will not be an important interconnect technology.