Ampex Knows Video

Ampex has been at the forefront of video recording technology since 1956, when we invented the video tape recorder. We’ve won awards and accolades (including an Oscar and 12 Emmys) for our video processing and recording capabilities, and today offer a comprehensive range of products for video acquisition, storage and manipulation based on our miniR™ and AMux product families.

Definition, Rate, Color Depth, Compression and Transports

There is a vast array of acronyms, standards, and technologies connected with contemporary digital video systems. Broadly, though, these can be categorized as controlling one of three aspects of the video: the characteristics of the image being captured, the mechanism by which the captured images are communicated, and how the data is transformed for storage or onward delivery. Table 1 shows some combinations of technologies and related keywords supported by the miniR recorder using standard AMux modules.

Multiple Channels

The design of the miniR recorder from Ampex allows for the recording of as many channels as the internal bandwidth and physical constraints permit. The recording format used is IRIG 106 Chapter 10, supporting close time correlation between multiple channels as well as options for including additional (non-video) data, such as location and position information from avionics sources.

Standard Definition

Standard Definition (SD) video is typically defined by the NTSC and PAL standards used for analog broadcast television, and the digital standards that have by and large succeeded them for TV broadcasting. It is by far the most widespread type of video in use today, and is in general what is intended by the word “video” without any other qualifier. The key image characteristics are a 4x3 aspect ratio, 525 or 625 interlaced vertical lines (of which about 10% are not visible), and is interlaced at 60 or 50 fields per second (i.e. 30 or 25 frames per second, with each frame being transferred as two fields, each with half the number of lines). There are several different physical interfaces, the most common of which are known as “Composite” and “S-Video,” the latter offering better quality by using two distinct signals, one for color information and the other for the luminance (black/white) signal.

AM-122 Dual SD Video Acquisition Module

The workhorse of miniR video configurations, the AM-122 handles two channels of Standard Definition video, through either Composite or S-Video connections. In addition to the video signals, the AM-122 has a total of two audio inputs, and an audio “event marker” tone generator. Each signal, audio or video, is “looped through” to permit daisy chaining or monitoring of the recorded signal. The module provides both MPEG-2 and MPEG-4 Part 2 encoding, the latter offering slightly better compression performance at the expense of being a far less common approach, and thus potentially less compatible. The user can configure the data rate to between 1.5Mb/s and 15Mb/s, and select from 0 to 18 “P-Frames” after each “I-Frame” (“P-Frames” are the “deltas” based on the preceding “I-Frame.”) The output from the module is an MPEG-2 “Main Profile @ Main Level” (MP@ML) Transport Stream, packetized according to IRIG 106 Chapter 10 as a Video Data Packet. The AM-122 supersedes the single channel AM-111.

AM-132 Dual SD Video Acquisition Module

The AM-132 video acquisition module adds the option to select a Video Time Insertion (VTI) overlay on the acquired signal, so that a time code can be seen on the recorder video. In all other respects it is identical to the AM-122 module.

AM-11x4 Quad SD Video Acquisition Module

This module is a bit of a hybrid: it takes four Standard Definition signals and “mattes” them into a single High Definition frame, which it then provides to an attached Compression Module (see later under “High Definition”). As the “matting” happens at the point the video is acquired, a single stream that contains the four time-synchronized input channels is created, which can generally be easier to manage than four independent channels. As with the AM-122 and AM-132, each Standard Definition input channel is selectable between Composite and S-Video, and the AM-11x4 also provides two audio input channels. The audio signals, but not those of the video channels, offer a buffered “looped through” connection within the module, which “recreates” the audio signal. Video Time Insertion is also selectively available on any or all of the four channels.

Enhanced Definition

Unlike the term “Standard Definition”, which is understood to mean signals covered by a specific set of standards (e.g. RS-170A), “Enhanced Definition” is a rather nebulous term. Typically, the enhancements are in the scan type and frame rate, so instead of 60 interlaced fields per the NTSC standard (i.e. 30 frames per second), you might have a 60 progressive frames with same Standard Definition frame size. Another characteristic of Enhanced Definition is that it is frequently used in applications where the raw signal is required, without compression. Because these formats are typically used in industrial imaging (“machine vision”) and rarely, if ever, used in broadcast applications, no dominant transport interface has emerged, with IEEE1394, Ethernet and Camera Link being the three most prevalent options.

AM-261 IEEE1394 D/CAM Acquisition Module

The IEEE1394 (Firewire) Interface is defined by a family of standards which include physical and logical specifications. One of the popular choices for Enhanced Definition devices is to use the IEEE1394 Digital Camera Specification, which defines a mechanism for transferring uncompressed image data without audio. The AM-261 module is designed to interact with a 1394 D/CAM source (i.e. camera) and capture the video stream, and also provides a “pass through” connection to allow the source image data to be “daisy chained” to another device.

High Definition

High Definition (HD) video is generally defined as having significantly more than 625 vertical lines of resolution. While it is typically associated with broadcast television, HD video also includes the display formats used by computer displays, which often use 5:4 and 4:3 aspect ratios compared to the 16:9 of HDTV.

Due to the significantly larger amounts of data – there can be up to six or more times the image data in an HD frame compared to an SD frame – data compression techniques are a critical aspect of high definition video. While the ubiquitous MPEG-2 algorithms can, and sometime are, used for HD compression, two alternative approaches, MPEG-4 Part 10 (also known as AVC and H.264) and Motion JPEG-2000 (MJPEG2K) can offer significant benefits.

Ampex’s HD video solutions are designed around a two module set: one module provides the interface and such features as Video Time Insertion, while the other handles the compression and packetization task. This allows a “mix and match” approach, since the two compression alternatives are optimized for different tasks. Ampex high-definition video modules incorporate a connector that may be located on either side of the circuit board, or on both sides. This connector provides the physical interface between the input module and the compression module(s). Model numbers for these modules use a letter designator to indicate the connector location: “a” indicates a connector on the top (component) surface, “b” on the bottom surface, and “c” for both connectors. For an input module to mate with a compression module, they must have compatible connector locations, so e.g. an AM-12b1 will mate with an AM-170a, and an AM-13c1 would mate with two modules, an AM-160a and an AM-160b. In all other regards, the modules are identical, so we use the letter “x” (as in AM-170x) to indicate the presence of these possible variations.

AM-12x1 SMPTE-292M HD-SDI Acquisition Module

HD-SDI, as defined by the SMPTE-292M standard, is the preferred interface for professional broadcast digital TV, and has become the preferred connection technology for digital signals, being used by both the L-3 Wescam “True HD” series and the FLIR Systems, Inc. SAFIRE HD series of imaging turrets, amongst others. The AM-12x1 receives a single 1.5Gb/s HD-SDI signal, which can carry video in several frame rates and sizes. The two most common choices are a 1280x720 progressively scanned image at 60 frames per second (720p60, defined by SMPTE-296M) and a 1920x1080 progressively scanned image at 30 frames per second (1080p30, defined by SMPTE-274M). Together with the video data, metadata in a format defined by SMPTE 336M (“KLV”) is correctly interpreted by the AM-12x1, in accordance with the recommendations of the Motion Imagery Standards Board (MISB), part of the US National Geospatial-Intelligence Agency. In addition to the video signal, the AM-12x1 can acquire two channels of audio, for which it provides a buffered “loop through” connection. As with the AM-132 and AM-11x4, the AM-12x1 provides configurable Video Time Insertion into the image frame.

AM-13x1 RGB/DVI Video Acquisition Module

The AM-13x1 provides two complementary types of video input: from an analog source, and from a digital source. The former is suitable for both “VGA” (actually, up to “SXGA”) video output sources as from a computer graphics card, as well as HD analog video such as RS-343 or “Component Video”; STANAG 3350 defines a HD signal based on RS- 343. The digital input supports the DVI signal, which means it will also support the video (but not audio) capabilities provided by HDMI. As with the AM-12x1, support is provided for two audio channels and Video Time Insertion. The base image format is up to 1280x1024, progressively scanned, at 30 frames per second. However, by using two AM- 160x Motion JPEG2000 compression modules, the AM-13c1 module can support 1280x1024 at 60 frames per second by splitting the compression task into two. (N.B: this approach cannot work with the AM-170x AVC module, because that depends on inter-frame compression while Motion JPEG2000 is limited to intra-frame techniques).

AM-160x Motion JPEG2000 Video Encoder Module

Motion JPEG2000 excels at compressing video streams where each frame may contain a small transient signal, such as a blip on a radar. Because each frame is compressed without regard to those that precede or follow it, every frame has the same quality as the next, and the video stream may be cut or frozen at literally any point – which is why Motion JPEG2000 is the preferred compression technique of Digital Cinema. The AM-160x typically delivers approximately a 20:1 compression ratio. As noted previously, two AM-160x modules may be used with the same input in order to deliver twice the frame rate.

AM-170x MPEG4 AVC Video Encoder Module

MPEG4 Part 10 Advanced Video Coding (AVC), also known as H.264, is a highly sophisticated motion compensation-based encoding standard which provides extremely high levels of compression while preserving good visual fidelity in the recorded signal. This compression technique is recommended by the Motion Imaging Standards Board (MISB) for video recording over older techniques such as MPEG2.

Other Technologies

Ethernet has become a powerful technology in video systems. In addition to its ability to capture video traffic placed on a network by an “IP Camera” or similar acquisition device, it is often more convenient to use Ethernet as a playback transmission technology, since once the video had been digitized and probably compressed, there is often no benefit to recreating the signal that was originally recorded. Further, digitized video can easily be played on a general-purpose computer rather than specialized video monitors, etc. The same connection can be used to control the recorder, too, via the built-in Ethernet port on the miniR system’s base module. In addition to playback, the design of the miniR recorder allows one or more compressed video channels to be monitored over a network link while being recorded. And because it is a network, the video can be as readily transported through a tactical data link as it is displayed at an on-board operator’s console. The Ethernet connection can also be used to implement a NADSI connection in accordance with STANAG 4575; a dedicated CPU translates the miniR recorder’s Ethernet interface to the Fibre Channel connection required by NADSI.

AM-801 Gigabit Ethernet Module and mR-ES2 Gigabit Ethernet Switch

The AM-801 Ethernet module supports a single 1000Base-T connection, which can be used to capture image data at up to 640Mb/s, or support replay and monitor applications. Note that the same module is used for all capabilities, with the desired functionality being selected via a setting accessed through the miniR recorder’s built-in web server. The mR-ES2 Ethernet switch provides convenient gigabit network switch with jumbo frame functionality within the rugged miniR recorder package.

mRMM and mRMM-1394 Memory Module with Firewire

Ampex offers two ranges of removable memory module to work with the miniR 700. The first provides excellent storage density, currently providing up to 448GB in a very compact form factor. The second provides half the maximum capacity, but adds a Firewire (IEEE 1394b) download port using a standard connector, facilitating access to the recorded video without a dedicated download station or cradle.