TMC2081KBC Просмотр технического описания (PDF) - Fairchild Semiconductor

Номер в каталоге

Компоненты Описание

производитель

TMC2081KBC

Digital Video Mixer

Fairchild Semiconductor 

TMC2081

PRODUCT SPECIFICATION

Table 8. Alpha Channel Gains

a value (hex) 8-bit Gain

000

0/256

001

1/256

..

07F

127/256

080

128/256

..

0FE

254/256

0FF

256/256

100

256/256

1XX

256/256

9-bit Gain

0/256

1/256

127/256

128/256

254/256

255/256

256/256

256/256

Fill Color Registers

Three registers, 03, 04, and 05, store a solid fill color, F.

Either GBR values or YCBCR values may be stored but the

format must match the data format of the A- and B-channels

at the input to the crosspoint switch.

Fill color registers are accessed through the D7-0 micropro-

cessor port. Fill color may be used as an alternative video

source for fades.

a-Mixer

There are three sources of data for the mixer: A-channel pix-

els, B-channel pixels, and the stored fill color, F. One pair of

inputs, either AB, BF or FA are selected by the Crosspoint

Switch to be passed to the V1 and V2 inputs of the a-Mixer.

Prior to mixing, V1 and V2 data formats must be matched

(see Table 1).

Within the a-Mixer are three dual input 9-bit mixers which

mix each of the component channels of V1 and V2. By vary-

ing the value on the a-channel from 000h to 100h, the Mixer

performs a 256-step transition from one digital video source

to the other.

Six dissolve transitions are supported: A-to-B, A-to-F,

B-to-A, B-to-F, F-to-A, and F-to-B. Type of dissolve is

selected by directing the A-, B-, or F pixels to the V1 or V2

mixer input via the ABF Crosspoint Switch. This is done

either by internal Control Registers via the microprocessor

port or directly through the SMX2-0 inputs. SMX2-0 input

pins are enabled via SMX Control Register bits. When

enabled, SMX2-0 directly control the ABF Crosspoint Multi-

plexer on a pixel-by-pixel basis, for externally derived wipe

patterns.

Rate of dissolve is controlled directly through the a-channel.

Transfer function of the mixer is:

M = ( a ) V1 + ( 1-a ) V2

where V1 and V2 are two of the three inputs A, B or F

selected by the crosspoint switch.

For an A-to-B dissolve transition, as the value of the eight

LSBs of the a-channel change from 00h to FFh, (or 000h to

100h in the 9-bit mode), an increasing level of A-channel

contribution and a decreasing level of B-channel contribu-

tion becomes evident at the output, M.

Bit a8 of the a-channel can correct for the 255/256 gain

factor in the A-channel that occurs when the 8-bit a value is

FFh. When a8 =1, bits a7-0 are ignored, A-channel gain is set

to 256/256 and B-channel gain is set to 0/256.

Modified transfer functions may be selected for background/

foreground and drop-shadow effects by programming con-

trol register bits, MIXTFN.

A Foreground Key may be created such that:

M = ( a ) V1

A Background Key may be created such that:

M = ( a ) V1 + V2

By using foreground and background mixers in series, drop

shadow effects can be implemented.

a may change at pixel rates up to 40 Mpps on a pixel-by-

pixel basis, allowing smooth transitions from one video

source to another. Transition time interval may vary from

many frames to only a few or a single pixel depending upon

the a-channel data rate.

a8 may be used like a key input. Either unity gain V1 or

( 1-a )V2 may be selected. A- and B-channel pixels may be

mixed by switching a8 on a pixel-by-pixel basis. Pipeline

latencies of the a-, A- and B-channels are matched.

Passing of Non-Pixel Data

In the PASSON mode, the TMC2081 is transparent to data

accepted during the PASSEN = LOW period (see Figure 10

and Figure 11). Either PDA or PDB data may be selected to

pass on reference signals containing time codes, subcarrier

phase and frequency data from upstream video processors.

Digital Outputs

Data at the M23-0 output port, may be selected from either

the mixer or, for digital preview, the A or B crosspoint switch

inputs.

The 444-to-422 formatter may be bypassed for 24-bit output.

To convert 24-bit YCBCR data to the 16-bit YCBCR422 for-

mat, the formatter needs to be enabled.

Except for color index, all data formats shown in Figure 1

are available:

• YCBCR444

• YCBCR422

• GBR

6

Share Link: