Grass Valley Imagestore 750 v.3.0.1 - Manual (Page 26)

10
Introduction
Architectural Summary
The Imagestore 750 has 8 video inputs, HD or SD. Each SDI video input feeds the internal video
buses (program, preview, monitor, and clean-feed) and one parallel audio-processing block. The
Imagestore 750 has line FIFOs for all 8 SDI inputs for timing adjustment. Embedded audio can
be de-embedded from any 4 of the 8 inputs, and then be manipulated as required by the
Imagestore 750’s audio engine.
The program video bus (PGM) has an A/B mixer, four keying layers and an optional dual-window
2D DVE. The A/B mixer allows the PGM output to be transitioned between two SDI video
streams. Each of the keying layers may be fed graphics content from its own dedicated store or
any of the three available external fill and key pairs, and then keyed over the background video.
The stores can be loaded with a still image, animation, Easytext template, analog/digital clock,
or Emergency Alert System (EAS) crawl. All available media files are stored in the media library
(compact flash or hard disk, depending upon the option purchased) and then loaded into the
internal stores on demand. The dual-window 2D DVE is used to create picture-in-picture
squeeze effects which can be dynamically positioned between different keying layers on the
program video bus.
The preview video bus (PVW) has four keying layers and an optional dual-window 2D DVE. Each
keying layer shares graphics content with the corresponding program keying layer, but can be
controlled independently in order to preview graphics before they are viewed on-air. Similarly
the dual-window 2D DVE can be used to preview DVE moves before they are played on-air.
The clean-feed (CLN) and monitor (MON) video outputs allow a number of different internal
points within the video system to be sampled and viewed for partial branding or monitoring
purposes. For example, the CLN output can take its output from DSK3 of the program bus to
show a partially branded version of the program output, without a station logo.
A fail-safe mechanical video relay bypass connects the A video source to the program output
directly and connects the C video input to the preview output (PVW) directly when a power
supply problem occurs. The bypass mechanism is shown in Figure 2-1.
The Imagestore 750’s audio engine accepts audio inputs from four embedded SDI feeds, 16 AES
pairs, and from the multi-channel Easyplay2 audio clip player, which gives simultaneous playout
of up to four audio clips sharing a total of 16 channels. The audio engine provides 16-channel A/
B mixing and multiple voice-overs. During voice-overs, background audio is automatically
ducked. Silence detection automates control of multiple independent voice-overs fed from the
same Easyplay clip. (Silence detection is typically used in a multi-language broadcast
environment.)
Optional audio modules can be added that provide integrated support for Dolby encoding and
decoding, and stereo-to-5.1 up-mixing. The audio engine complements Dolby and up-mixing
with advanced metadata processing. Audio description (AD) allows a background commentary
track to be mixed with the background audio for people who are visually impaired.
One of up to 16 possible tap points from the audio engine can then be dynamically assigned to
each of the available unit outputs, to the four embedded outputs (PGM, PVW, CLN, MON), to the
external AES outputs, and to the level metering.
 The audio engine is configured using the ‘Audio Graph’ page of the Imagestore 750
Configurator.
An Ethernet interface provides rapid transfer of images and animations using Media Conversion
Software (MCS) and other third-party media management applications. This interface also
allows the transfer of images between a PC and the Imagestore 750 using Miranda’s Vertigo
Xplorer.