Bob Perry

IEEE 1394 vs. DVI
A Comparative Perspective

Currently there are three digital interface standards:

© Pacific custom cable Inc.

Plugging It All Together!

We are in the midst of a gigantic transition of our country's infrastructure - the shift from analog to digital television. And as all of us are aware, digital television opens a new world of high resolution images, superb audio, and a variety of new digital services yet to come. A visit to your local professional A/V retailer reveals the vast variety of products that deliver this new technology to your living room.

Looming before the unwary consumer is a decision that will ensure the fulfillment of their digital yearnings, or total frustration. Two digital interfaces (IEEE 1394 and DVI) and one analog interface (component video) introduced to the marketplace each vie for leadership, and each have their own challenges and opportunities.

We will review each technology and its associated issues to better help you understand and make your purchase decisions

In the new world of digital television, programming is produced locally or by the networks, compressed, and transported to the local broadcaster via satellite as a compressed digital bit stream - a stream of zeros and ones representing the digital content.

It has to be compressed in order for the large signal to be transmitted

This is very much like today's Digital Broadcast Satellite (DBS) systems, where the programming is transported directly to the home as a compressed digital stream, or today's digital cable systems, or even DVD [MPEG-2], where the compressed digital bit stream is stored on a disk for playback in the home. The following diagram details the simplified process of acquiring a digital video signal and displaying it.

The steps are actually quite simple:

- As you can see from Figure 1, the digital content, which is an MPEG-2-compressed digital bit stream is either transmitted (over-the-air, on cable, or via satellite) or it is read from a storage device such as a DVD.

- The digital receiver receives the transmitted radio frequency signal, and demodulates it, resulting in a MPEG-2-compressed digital bit stream. The DVD reader mechanism reads the MPEG-2-compressed digital bit stream directly from the DVD.

- Then, the compressed bit stream can be output on an IEEE 1394 serial connection, or it can be processed further yet.

- If it is processed further, it is then decompressed (also known as decoding or expanding) into a uncompressed digital video bit stream, which can be passed over a DVI (Digital Visual Interface) monitor connection, or it can be processed further.

- If it is processed further, the uncompressed digital video bit stream can be converted to analog, in the form of RGB [Red Green Blue], or YPrPb (known as component video).

- Note that if the compressed bit stream on IEEE 1394 is connected directly to the display device, the display device must have an MPEG-2 decoder. All integrated HDTVs must have one anyway, so this is only an added cost for HDTV monitors.

Some technical details need to be understood

First, there is no qualitative difference between the signal quality of the broadcast signal and the compressed digital bit stream on the IEEE 1394 connection. They are technically identical.

There is also no qualitative difference between the DVI and analog video output-they can both be full HD quality signals without degradation. At some stage most displays need to receive an analog video signal - the difference between DVI and analog video in the graphic is the location of the digital-to-analog (D/A) conversion.

With DVI, the D/A conversion usually takes place in the display.  The analog video output shown in the graphic is the normal configuration of an external digital set-top-box that receives digital broadcasts, and outputs analog video to the display.

Of course, if we were to run the DVI and component video cables for a long distance, the analog component video cables would be more susceptible to interference and noise.

However, this tends to be a moot point, since the maximum length of a DVI cable is 3 meters.

A properly shielded component video cable should not pickup noticeable interference or noise, at even 4 to 5 meters.

Now that we have a basis for understanding the differences between the three primary connections - IEEE 1394 serial connections, DVI digital video connections, and analog video connections, we can discuss each in detail

IEEE 1394

IEEE 1394, also known as FireWire, and i.Link (Sony's name) was invented by the scientists at Apple Computer in the 1980s. They were searching for a digital connection to support networking and the transfer of large serial streams of audio and video. FireWire, the name used by Apple to describe this connection, is a marvel of technology.

IEEE 1394 is essentially an electrical specification for creating a high-speed, high-bandwidth serial network connection including product control capability. While there is a cable referred to as and IEEE 1394 cable, the specification is most accurately described as a recipe for creating this connection with a variety of unspecified physical media, including a type of wire called CAT 5 UTP (Unshielded Twisted Pairs), Plastic Optical Fiber (POF), Glass Fiber, and wireless technologies, to name a few.

Key to this recipe is the way the data is managed and transferred across the connection

First and foremost it is a serial connection. Unlike LAN technology, which creates packets of data, throws them across the network connection, and then reassembles them in the proper order, IEEE 1394 can also send the data serially, in a sequential stream of packets. Imagine a movie chopped into data blocks, then thrown across a network connection for reassembly. It's no wonder that streaming a movie from the Internet (which uses the TCP/IP protocol) is an experience to be avoided. Many have suggested that Internet video streaming should be more appropriately titled "stuttering." On the other hand, a movie sent serially arrives in the proper order and without delay.

Most devices that use IEEE 1394 fall into two categories - computing devices and peripherals, and audio/video products. Computers use IEEE 1394 by sending two types of content - TCP/IP like normal LAN technology and compressed A/V streams. (While a standard, LAN can only transfer TCP/IP packets. The IEEE 1394 is much more versatile - many data types such as TCP/IP and MPEG-2 can be accommodated.)

Audio/ video products generally just send MPEG-2 type A/V streams. Both categories of products, when sending A/V streams, can also send the control protocols that are part of the IEEE 1394 standard known as AVC - Audio Video Control. Using this control protocol, one device can control another. There is an optional higher-level software protocol that builds on AVC, which is known as HAVi (Home Audio Video interoperability).

Another unique characteristic of IEEE 1394 is that it is a self-recognizing and self-configuring connection

Simply put, when you connect two IEEE 1394 A/V products together, they send each other some data automatically, that essentially states "this is who I am, this is what I do, and this is how you control me". So each product connected together with IEEE 1394 is aware of all the other products, what they do, and how to control them. How does this work in the real world? The following shows the actual function of a IEEE 1394 product:

An HDTV equipped with IEEE 1394 displays the on-screen menu, with icons shown for products connected on the IEEE 1394 network. When a D-VCR (D-VHS) is connected to the HDTV through the IEEE 1394 cable, the icon for the D-VHS appears on the screen immediately.

No consumer configuration is necessary, no drivers load like on a PC - just connect and it works. The TV automatically recognized the D-VHS, learned its control codes, and created an on-screen icon for it. Now, using the TV remote I can just select the D-VHS icon and the menu for the device is created by the TV - using the control codes that were downloaded by the TV from the D-VHS.

And all the other devices on the network are aware of each other as well. So, for example, any TV connected to the IEEE 1394 network can see all the other products. So, you can imagine a digital television in the main living room, and one in the master bedroom, both being able to see each other, and the other products in the home.

And, each TV is able to use the D-VHS HD VCR and other devices on the IEEE 1394 network

Product designers can create their user interfaces (the on-screen control systems) for their products in any manner. Each product's control interface can be unique, but under the IEEE 1394 standard, it "hooks" to the underlying AVC and HAVi (if so-equipped) control signals.

But unlike uncompressed video interfaces such as DVI and analog video, IEEE 1394 uses a different system for creating the on-screen displays and menus. The IEEE 1394 standard provides for onscreen graphics by using a system known as Level II User Interface. Level II User Interface is specified in the European cable box standard known as MHP (Multimedia Home Platform), and in the U.S. it is specified in the CableLabs OCAP cable box standard as well.

Because the IEEE 1394 system usually carries compressed digital bit streams, these streams can not only be networked around the home, they can also be recorded using products such as the D-VHS HD VCR, hard-disk drives, and other types of recording devices.

This is a key difference between uncompressed connections such as DVI and RGB/ component video

These uncompressed connections cannot be recorded by any known consumer recording technology, and the copy protection system on DVI (known as HDCP -High Definition Copy Protection) forbids all copying - it also forbids providing a DVI input on a recording device.

The IEEE 1394 connection has DTCP (Digital Transmission Content Protection) developed by the Digital Transmission Licensing Administrator LCC (LTLA or 5C companies) copy protection, which has encoding rules. (1) These encoding rules restrict content creators from blocking the recording of content that has normally been recordable in the analog era.

DVI - Digital Visual Interface

The DVI standard was released on April 2, 1999, and was created by Intel, Silicon Image, Compaq, Fujitsu, HP, IBM, and NEC. The standard was the culmination of work done to take an underlying video signal standard, and "package" it into a standard for its interconnection and interoperability. 

The DVI interface is based on the TMDS signal format. TMDS is Transition Minimized Differential Signaling format, and is structured to allow for uncompressed digital video to be transferred from a source to a monitor, with minimal or no loss. Its primary initial application was the connection between a computer and a fixed pixel computer monitor or projector, such as an LCD monitor or projector. Because these fixed-pixel devices generally use digital signals at their basic hardware level to drive the display, by using a TMDS connection the digital display signals from the computer can be directly linked to the display. Over time, this interface has been suggested by many as an interface for consumer digital televisions as well.

The DVI standard also supports "plug-and-play" monitor recognition. The DVI connection provides for signaling to the source so that the source can adjust its signal to match the resolution and scan rate of the monitor.

Compatibility:

The connector itself has several versions - DVI-A which supports normal analog signals, DVI-D which supports the digital signals, and the combined connector, which supports both signals simultaneously. 

The DVI standard is also under evolution, as at least two separate groups are working to develop new forms of DVI. The DVI-CE standard has been under work by consumer electronics companies and other parties to create a new DVI-based interface that is more appropriate for consumer electronics applications.

And HDMI (High Definition Multimedia Interface) was recently announced by a number of companies as an effort to integrate DVI video signals, digital audio signals, and some limited control signals (such as volume up and down, channel up and down) into one connector. The HDMI companies are expected to finalize their standard in early 2003.

Regardless of the flavor of DVI, its function is essentially identical - to connect a source device to a display device, such as a computer monitor or HD monitor. (Note that HD Monitor is the specific term recognized by the Consumer Electronics Association for a display that can display HD-grade signals, but does not contain an integrated digital receiver.  This function is generally identical to the analog YPrPb and RGB signals that exist today on most HD set-top boxes and HD Monitor televisions.)

As you can see from the graphic detailing the process of receiving a digital broadcast, the compressed broadcast signal is decompressed into a digital video signal, and then output as DVI for input into the display.

© Pacific custom cable Inc.

So why change from these existing analog connections to DVI-based connections?

The primary driver is the content community. Because the analog HD signals do not contain any copy protection capability, the content community is afraid that these HD-grade signals may be stolen and transmitted across the Internet in the future This kind of distribution would play havoc on the content community's business model of releasing content into different worldwide markets at different intervals to maximize revenue.

Because DVI is uncompressed, it cannot be recorded by any known consumer recording technology. And with the application of HDCP copy protection, the connector cannot be installed on any recording device as an input.

So the content community believes that DVI is secure, and immune from any copying

Another comment cited by manufacturers regarding DVI versus IEEE 1394 is the ease of creating complex on-screen graphics and menus. Creating these menus and graphics is roughly the same process for DVI as it is for today's analog connected products - the graphics processor in the product just includes the graphics with the uncompressed program video for display.

There is no question that using IEEE 1394 requires a different approach as was noted above. Keeping in mind that the IEEE 1394 has compressed content streams, menu graphics cannot simply be overlaid on the compressed video stream. Although the approach is different, the cost and outcomes are comparable.

Analog Component Video

Analog component video has been with us for many years. Both RGB and YPrPb are types of component video - the video is broken down into components. As with so many types of signals, there are variations within the general standards.

For example, RGB (red, green, and blue) has several variations, based on how the synchronization pulses are sent. There is RGB with individual sync for each color, there is sync on green (the sync pulse is in the green video signal), and there is RGBHV, where separate horizontal and vertical sync pulses are on their own wires.

Interestingly, today the YPrPb color difference signal type is referred to as component video, and RGB is commonly thought of as not being a component video signal. This is technically incorrect, but is common practice.

Analog component video, just like DVI digital video, can provide various degrees of resolution. From yesterday's analog TV standard of 480i to far beyond HD signal resolution, it can be carried without noticeable degradation on this connection. 

Analog component video does not generally have a copy protection system in use, although there are systems for low resolution signals. Because this interface does not have copy protection for the HD "valuable" signals, the content community is loathe to support it, as previously discussed.

The analog connection between HD source components and displays is referred to as the "analog hole" by the content community, and they want to plug the hole. But today over 2.5 million consumer televisions use analog HD component video to connect to HD sources, and audio/video manufacturers do not support phasing out support for this interface over the short term, for fear of stranding early adopter consumers. 

Others, such as myself, believe that the "analog hole" was created by the content community by their refusal to support standards such as IEEE 1394 with 5C copy protection years ago, before manufacturers geared up digital television production. In the negotiations with 5C, the content community did not want to accept the encoding rules, because content providers wanted to make content that would have been recordable in the analog world un-recordable in the digital world.

This desire stemmed from the concept that by denying or limiting home recording, consumers would order pay-per view and other content repeatedly, generating multiple fees and increasing revenue

Because the consumer electronic industry could not get support for IEEE 1394 with 5C copy protection in the early years, we adopted the known and accepted interface technology of the time - component video. Of course, after component video became popular, a number of studios have now accepted IEEE 1394 and 5C.

Today, virtually every HD display has component video interfaces, in addition to IEEE 1394 and/or DVI. And the set-top boxes launched by cable so far only have component video and IEEE 1394. DBS companies have announced support for DVI output, and DISH Network has also announced support for IEEE 1394. Both DBS (Digital Broadcast Satellite) companies support component video as well.

Pulling It All Together

So in review, we have compressed interfaces, and uncompressed interfaces, each with their own individual capabilities and problems. Why is there such a debate over these interfaces? And why has this author publicly criticized DVI as being "bad for the consumer?"

In order to better understand the issues, we need to review a concept known as "Content Encoded Output Switching" or "Selectable Output Control." The concept is simple. When a movie is created in digital form, the studio can inject a code that signals the set-top box-cable or satellite, to turn off certain outputs.

So, if the IEEE 1394 interface can be shut off, the consumer cannot record, since the DVI and analog component video signals cannot be recorded as discussed earlier.

The desire for this type of output control is clearly stated in the public record, and in the specifications issued by the cable industry's laboratories - CableLabs. 

As you can see from Figure 2, disabling the IEEE 1394 interface not only disables home recording, it also eliminates the ability to route the signal around the home as part of the A/V network.

If the "good" movies and other programs cannot be networked, consumers cannot be expected to build A/V networks just to watch the "bad" stuff

This would have a dramatic chilling effect on the future of home networking in general, because consumers need simple, powerful self-configuring network technologies to have widespread home networking.

I believe this disabling of IEEE 1394 outputs is ill-conceived, and an affront to consumers' "Fair Use" home recording rights.

And disabling analog outputs would strand consumers who own HD monitors without integrated tuners, or without the ability to upgrade their HD monitor with an integrated tuner module.

It is my opinion, and the opinion of many that we should respect the investments made by consumers, particularly the early who are helping to drive the HD transition, and not allow such a scheme to exist. (2)

Many people have asked why the content community wants to stop home recording and networking

There are two basic issues that concern them - networks that would allow transfer of compressed content to the Internet, and money. The ability to transfer a compressed two-hour HD movie to the Internet does not really exist today, although there are concerns it could in the future.

This is because even with broadband, a compressed HD movie is a huge amount of data. For example, with a cable modem and the industry average upload speed of 384 kbps, a two-hour HD movie takes 129 hours to upload, which is not quite video-on-demand. And IEEE 1394 with 5C already protects against Internet retransmission. 

As for money, the content community argues it does not want to stop home recording, but just wants to control it.

As you can see from Figure 3, by eliminating the IEEE 1394 outputs from the settop box, and using DVI output, the only recording that can take place is where the bit stream is still compressed - in the set-top box, on a hard disk drive (HDD).

Both the delivery system owners (cable and satellite), as well as the content community have a stake in this, since both of them share in the revenue chain for pay-per-view and other "per-play" systems.

And as you may be aware, both cable and satellite digital set-top boxes have systems that use the cable or phone line to automatically bill you for each service or movie you watch.

So it's easy to understand how you might come home, order a movie, and record it on the HDD in the set-top box as you watch it so a family member can also watch it later. But, when you play back the HDD recording later, the system can bill you again.

Many in the cable and satellite industry argue that the ability to shut off outputs is really just a tool they will not use, but do you believe that they will run their businesses for your benefit, or theirs?

And what is the impact if cable and satellite set-top boxes simply only have DVI outputs, and no IEEE 1394 connections to shut off?

It's important to note that in my opinion, there is nothing inherently bad about DVI. It's just that DVI interfaces, combined with the ability to shut off or not include other interfaces, locks the consumer into a world that eventually replaces the "Play" button with the "Pay" button.

As you can see from Figure 4, using DVI instead of IEEE 1394 has other consequences as well. Because DVI is an uncompressed monitor connection, its signal cannot be networked.

And each source device that receives a digital stream, or plays back a digital stream from a prerecorded source requires an expensive MPEG-2 decoder.

In an IEEE 1394 environment, only one MPEG-2 decoder is needed - at the display. In 2002, an HD grade MPEG-2 decoder has a retail price impact from $700 to $1,200 depending on features.

And DVI does not allow the sharing of components throughout the home, nor the easy control of devices that is supported in IEEE 1394.

The choice of interfaces on digital television products and source components is a confusing one, and is an issue not of just technology but also business interests and long-term strategies.

These issues are also why my company, Mitsubishi, has decided not to support the D-Theater encryption feature for D-VHS HD VCRs.

The long-term strategy of D-Theater is to output the signal via DVI, and there is no commitment to continued output on IEEE 1394 [or full-resolution analog component video - Editor-In-Chief].

We simply will not expose consumers to the additional cost of the HD MPEG-2 decoder, and the risk of having outputs turned off [or "downrez" - Editor-In-Chief Gary Reber]

In the end, you the consumer get to decide the issue. You vote through your purchases and selection of products every day, and we as manufacturers always bend to your will.

So if you are considering what products to purchase, this author recommends that you research and learn the tradeoffs to be the most informed buyer possible. And supplement this research with the knowledge and help of your local specialty A/V retailer. 

Important References For More Detailed Information

DVI: www.ddwg.org
IEEE 1394: www.1394ta.org
HAVi: www.havi.org
CableLabs OCAP Standard: www.cablelabs.org
HDCP DVI Copy Protection System: www.hdcp.org
5C IEEE 1394 Copy Protection System: www.dtla.org
Digital TV Product Definitions/Classifications: www.ce.org

(1) The "5C" companies, who developed the DTCP technology, are Hitachi Ltd. Intel Corp., Matsushita Electric Industrial Co. (Panasonic), Sony Corp., and Toshiba Corp.. DTCP attaches rules to a digital signal as it is being transmitted over a high-capacity IEEE 1394 connection. Those rules can prevent a program from being recorded (termed "Copy Never") or, once it has been recorded ("Copy Once"), from being reproduced. The unhindered level of DTCP is "Copy Freely."
[Editor-In-Chief Gary Reber]

(2) As stated previously in Widescreen Review editorials and articles on the subject (see "New DVI HDMI Interface - Will Full-Resolution Analog HDTV Outputs Co-Exist" in Issue 61, Mayas a recent position statement), the magazine will be at the forefront of a petition or class-action lawsuit to prevent studio content providers from preventing consumer from displaying full-resolution analog HDTV signals and/or disabling "Fair Use" home recording. [Editor-In-Chief Gary Reber]

About The Author

Bob Perry is the Vice President of Marketing for Mitsubishi Digital Electronics America, Inc. (MDEA), and is responsible for the overall marketing function of the corporation, including product development, marketing communications, government relations, industry policy, public relations, national retailer training, and new technologies management.

Bob has also taught economics, computer science, and management at several colleges. He also spent 10 years in the U.S. Army in military intelligence, communications, and engineering. Bob holds a bachelor's degree in business administration from Columbia College and a master's degree in management from Webster University in St. Louis. 

Bob also is a director of the CEA executive board. Bob also serves as the chair of the CEA video division board, has chaired the DTV Definitions Committee (twice), as well as the DTV Transition Policy & Strategy Working Group, the CEA 1394 Interface Working Group, and DTV Picture Format Working Group.

Bob also serves as a director on the board of the Home Recording Rights Coalition, as well as a member of the steering committee.

In addition, Bob is a director of the Home Audio Video Interoperabilty (HAVi) Corporation board. 

Bob writes sporadically for a number of industry and consumer magazines on issues facing consumers and the industry, and always apologizes profusely to those who suffer through his meager writing skills. [I happen to think Bob is a terrific communicator and writer. [Editor-In-Chief Gary Reber]

© 1384 Trade Association

www.pseudology.org