Sony Corporation Global Headquarters

Point to Point Video Interfaces

Returning to the system's level discussion of infotainment system communications, it is also entirely reasonable to consider the option that all video transmissions between a source and a display could take place over a point-to-point interface. This is what is happening in practice today, and so it is also worthwhile reviewing these interfaces.

Point to point video interface standards today fall into two broad categories: analog and digital. Two analog video interfaces commonly employed in automotive OEM and aftermarket products include: composite video and S-Video. Composite video makes use of a single 75 ohm coaxial cable and connector, and consists of the voltage sum of the black-and-white information (Y) and the color information (C). S-video makes use of a multi-wire cable, dividing the video information into two separate (75 ohm coax or twisted pair cables) signals: one for luminance (Y) and one for chrominance (C). Each signal is then sent shielded, enclosed in a 4-pin Mini-DIN.

One issue associated with analog interfaces is the difficulty encountered when trying to overlay graphics such as touch screen control or other status information. To achieve visually acceptable results, the video mixing source must be able to synchronize the digital video graphics with the analog video signal and to modulate the graphics color components onto a chroma subcarrier (in the case of composite video) whose phase and frequency track those of the video signal. Time synchronization errors may distort the graphics overlay and/or video image, and subcarrier phase errors will alter the two signals colors. Typical solutions to this issue require digitizing the incoming analog video signal using a pixel rate sampling clock that has been locked to the incoming horizontal sync pulses.

Another quality issue associated with composite video is cross-color artifacts. These are residual elements of the chrominance and luminance information that remain during the process of separating the video into R, G, & B for display. These artifacts become noticeable with higher quality displays, and so for this reason composite video is viewed more as a legacy interface format, with S-Video preferred to composite video. The highest performance analog video interface, which sees use in home electronics and older computer displays, is Component (or RGB) video. However, due to the large number of shielded cables required (3 to 5 depending on implementation), component video is not acceptable for automotive applications due to size and weight restrictions.

Mitigation of some of the mixing and color artifact issues is possible if a digital video interface is selected. Digital point to point video interfaces include: HDMI/DVI, LVDS, and GVIF.

HDMI/DVI

HDMI (High-Definition Multimedia Interface) was the first industry-supported, uncompressed, all-digital audio/video interface. It is targeted at both high-definition video and multi-channel, digital audio for consumer Audio Visual entertainment equipment (HDTV, Amplifiers, etc). HDMI is backward compatible with the DVI interface, but without the more advanced upgrades and no audio. More than 400 companies have become adopters, and more than 60 million devices featuring HDMI are expected to ship in 2006. The HDMI 1.0 specification was released in December 2002 and supported data rates up to approximately 5Gbps and a 24-bit color depth. The HDMI 1.3 recently released specification boosts the data rate to 10.2Gbps and the supported color depth to 48 bits. HDMI is usually paired with HDCP content protection.

LVDS SERDES

LVDS is an open electrical standard developed in 1995, based on technology which give it many attractive qualities. LVDS typically uses a current mode driver terminated by a 100-ohm resistor resulting in 350mV swings across the receiver input. LVDS supports data rates up to 2.5Gbps when using four differential shielded twisted pairs of conductors, one being assigned to carry the clock. Available IC's supporting the LVDS interface and targeting the automotive market currently do not support content protection. The signal standard of LVDS is composed of 4 channels of video data transmission lines (R, G, B, and display control signal) and 1 channel of clock transmission line, totaling 5 channels of data transmission lines transmitted by the host system. LVDS is based on the physical layer definition of the IEEE1596.3 standard and ANSI/TIA/EIA-644. The LVDS transmission lines have a capability of 7 x fckbps as a transmitter serializes the video data (8 bits each), horizontal and vertical clock signals, data enable signal, and control signal totaling 28 signals into 4 channels.

GVIF

GVIF is a base-band 1-bit serial digital video interface typically using a single shielded twisted differential pair of conductors. GVIF supports color depths up to 24 bits and a maximum data rate of 1.95Gbps, enough to support XGA screen resolutions. For SXGA resolutions and higher, two pairs of conductors are required. GVIF exhibits excellent EMI performance due to the small differential signal (400mVp-p) with good DC balancing. GVIF transceivers include high speed clock generators, clock and data recover circuits, data coding logic, synchronizing controller, and an automatic cable equalizer. The most recently released GVIF transmitter/receiver pair (CXM4017R/4018R) also include HDCP as a built-in function. As previously mentioned, almost all current infotainment systems, including Rear Seat Entertainment (RSE), make use of a dedicated link for video interfaces. Most of these interfaces are analog due to cost considerations. For the system designer, this means that for these cases the video interface is uni-directional. However, with the newer digital video interfaces, the link is often bidirectional. Bidirectional communication is mandatory for HDCP.