PETER PUTMAN /
03.01.2012
Originally featured on BroadcastEngineering.com
HDMI: plug and pray
The interface's consumer drawbacks can be overcome in the professional workplace.

HDMI, ubiquitous on a wide range of televisions and other consumer electronics equipment, is becoming established in the professional video world. You'll find it on camcorders, portable VTRs, media players, professional Blu-ray players, and reference and broadcast monitors. HDMI distribution amplifiers and matrix switchers are also widely available, as are HDMI to/from format converters.

Broadcasters are accustomed to moving around high-bit-rate video, audio and metadata through traditional serial interfaces (SDI and HD-SDI), unencumbered by copy protection keys and digital display handshakes. So the emergence of HDMI as a professional display interface is a major irritation as it does not lend itself well to multipoint distribution without some technical ingenuity.

The basics

As originally conceived, HDMI would be a plug-and-play digital video and audio interface between media players and HDTVs, resembling an ad-hoc peer-to-peer network connection. In the decade since the HDMI standard was first announced, it has undergone numerous revisions and updates, with the latest being version 1.4b (October 2011).

Version 1.4 increases the maximum resolution to 4096 × 2160 pixels at 24Hz, and includes a 100Mb/s Ethernet return channel, along with support for numerous 3-D picture formats (defined in 1.4a, March 2010). HDMI also carries up to eight channels of digital audio (linear pulse code modulation, 192kHz); supports sRGB, Deep Color and xvYCC color spaces; high-bit-rate digital audio; and an audio return channel. There are five different versions of the HDMI connector, with Type A (19 pins, slide-on) being the most widely implemented.

There are two parts to an HDMI connection: The source (such as a DVD/Blu-ray player, STB, game console or computer) and the sink (TV, monitor, projector). Repeaters were also developed to retransmit the HDMI signals through additional AV equipment that could also switch between multiple HDMI inputs, such as audio receivers for home theater.

When an HDMI connection is made, the source queries the sink to find out what type of it display signal it requires. This is accomplished with extended display identification data (EDID), a set of display parameters stored in nonvolatile memory that includes pixel clock, refresh rate and timings for standard and custom resolutions.

A second handshake verifies whether copy protection is present. The HDCP layer requires an exchange of 56-bit keys between sources and sinks. Each HDCP-compliant device has a set of 40 different keys, which are exchanged when an HDMI connection is made. Compromised keys will disable the connection.

Nice try, but no cigar

This architecture works well for simple source-sink and source-repeater-sink connections, as the exchanges of EDID information and HDCP keys are straightforward. But it is not suited to professional applications that require distribution amplifiers and matrix switchers. As a result, designing a professional video display distribution system around the HDMI standard is generally inadvisable, particularly when SDI and HD-SDI interfaces can already do the job.

And yet, it is being done by numerous companies, ostensibly to support formats such as Blu-ray and HDTV set-top receivers, but also to take advantage of the high-density connector and eliminate discrete connections for video, audio and data.

For a distributed video environment in the post and broadcast world, the challenge is to connect two or more displays with different resolutions to a single HDMI source. Which EDID is to be supported, and which displays are HDCP-enabled? How are multiple sources connected to one or more displays? All of these are real problems faced by anyone designing a multipoint HDMI distribution system.

Managing EDID

The conventional approach to EDID exchanges is to insert a repeater in the matrix switch so each display communicates directly with the source. While this approach is acceptable for a peer-to-peer connection, it does not work for two or more connected displays, particularly if they have different native resolutions.

The smarter approach is to store EDID settings in memory for each connected output of the matrix switch. These settings remain active in nonvolatile memory, emulating a virtual display to ensure a media player does not go into sleep mode when it is not the selected source.

The switch then selects the highest resolution common to all connected displays. For example, a 1 × 4 matrix switch might have two 1920 × 1080p displays and two 1280 × 720p displays connected. When all four displays are active, the source is prompted to output 720p, as that is the only resolution common to all four displays.

Should both 720p displays be disconnected from the matrix, the switch automatically reflects back 1920 × 1080 as the highest common display resolution. If one or both 720p displays are reconnected, the switch then reflects back 1280 × 720 to the source. This approach ensures that every connected display will show an image, although the resolution may change from time to time to accommodate all displays.

Managing HDCP

The next step is to verify and establish HDCP connections to all sinks. Using the conventional approach, a repeater within a distribution amplifier or matrix switch would pass HDCP keys back to the source. This means the source needs to decipher a different set of keys for each connected display, and if any key is corrupted, no signal will pass to any connected display, even if the remaining displays are transmitting valid keys. (See Figure 1.)

The solution is to make the switch a sink, thereby ensuring a constant and secure connection to the source or sources. And each switch output now becomes a second source, looking for its own HDCP handshake with a connected display. (See Figure 2.)

Once the secure connection is verified and keys exchanged, video from the actual source is passed through to the display. If a connected display is noncompliant, the matrix switch or DA will not pass video to that display only; all other connected, compliant displays will continue to see video normally.

If new displays are connected to each port on the matrix or DA, the secure connection is re-established while EDID is exchanged. The copy protection is maintained at all times on all ports. Any issues with repeaters are eliminated, and multiple sources can be connected as easily as multiple displays.

Making the best of it

Like it or not, digital consumer display interfaces are here to stay. It's not easy, but the exchange of EDID and HDCP can be managed in a rational way to meet the switching and distribution requirements of a production or mastering facility, turning “plug-and-pray” into plug and play.

Peter Putman is president of ROAM Consulting LLC of Doylestown, PA.



Comments
Post New Comment
If you are already a member, or would like to receive email alerts as new comments are
made, please login or register.

Enter the code shown above:

(Note: If you cannot read the numbers in the above
image, reload the page to generate a new one.)

No Comments Found




Wednesday 9:02AM
Analysts: TV Regs 'Not as Dire as We Thought'
We feel the negatives are known and are a lot more comfortable recommending the space.


 
Featured Articles
Discover TV Technology