DatasheetQ Logo
Electronic component search and free download site. Transistors,MosFET ,Diode,Integrated circuits

STV-USB-VGA-R01 查看數據表(PDF) - STMicroelectronics

零件编号
产品描述 (功能)
生产厂家
STV-USB-VGA-R01
ST-Microelectronics
STMicroelectronics ST-Microelectronics
STV-USB-VGA-R01 Datasheet PDF : 41 Pages
1 2 3 4 5 6 7 8 9 10 Next Last
Customer Datasheet, Rev 3.1, 29 November 2000
STV0672-chipsetf-3-1.fm
2.2.6 Matrixing
This module performs a 3 x 3 matrix multiplication on the smooth RGB channels, to map pixel taking chromaticities onto nominal
display chromaticities. Depending on the characteristics, a general 3 x 3 matrix can effect the transformation; however we choose
to decompose this matrix into the form M = UT, where U is a neutral-preserving matrix, each of whose rows add to unity, and T is
a diagonal matrix of tilts, which form one component of the channel gains applied in the input processor. The combination of U
and T are coded in such a way as to require the user provision of 9 integers; three for the tilts T and six for the off-diagonal
components of U (on-diagonal coefficients are implicit in a re-ordered row computation). The strain parameter from the
housekeeper and a specially-cored version of the unsharp mask signal control the damper; a variable which desaturates the
effect of the matrix by attenuating off-diagonal components in the presence of noise (strain component, occurring in low ambient
light) or Bayer aliasing (unsharp component, occurring in image regions with high edge-content).
2.2.7 Peaking
The unsharp mask signal is added to each matrixed colour channel to compensate for edge information lost in Bayer-pattern
under-sampling and interpolation.
2.2.8 Gamma correction
Gamma correction provides a non-linear distortion of data amplitudes required for various video communication standards, as
well as cosmetic enhancement of image detail.
2.2.9 Coder
The coder module comprises a color-space convertor which takes RGB from the gamma module and produces luminance Y and
weighted colour-differences Cb & Cr. The data is now passed in YCbCr form to the Video Compression block.
2.3 Video Compression (VC) Engine
The video compression engine performs 3 main functions:
• Up scaling of input YCbCr 4:2:2 video stream from the VP (typically to scale from QVGA to CIF image formats)
• Compression and Encoding of YCbCr stream into Motion-JPEG (M-JPEG) format
• USB Bandwidth monitoring
Figure 3 below gives a block overview of the VC module. The data stream from the VP can be upto VGA size. The scaler in VC
can downsize this image. Once scaled the video stream is then converted into M-JPEG format. M-JPEG simply treats video as a
series of JPEG still images. Please note that the JPEG specific header information need not be transmitted in a MJPEG stream.
The conversion is realised via a sequential DCT (Discrete Cosine Transform) with Huffman encoding. After transfer over USB the
M-JPEG stream will be decoded in the Video-for-Windows (VfW) device driver running on the host.
The VC module is capable of compression ratios of up to 100:1 although clearly this is scene dependent. Image framerate
produced by the STV0672 chipset is fixed and furthermore the available USB bandwidth is also fixed (within the software driver).
The VC module varies the compression ratio to match the fluctuating input video data rates, that vary according to scene
dynamics, to the available USB bandwidth and required framerate.
The final stage of the VC block manages the data transfer rate from the local VC FIFO store to the USB core. STV0672 can
perform this management automatically, by employing long-term (frame-level) and short-term (block-level) compression
management. The former is achieved by varying a scalar quality-factor from frame to frame, to drive expected data rates upwards
or downwards. The latter is achieved by truncating the zig-zag sequence of AC coefficients more or less severely according to
how many preset thresholds of FIFO usage have been crossed. As FIFO usage approaches maximum, this truncation process
reduces instantaneous data rates until stability is regained, at the cost of local loss of detail in the image. The latter process is
transparent to the decoder. Statistics of threshold-crossing activity are subsequently used in the long-term quality setting
decision.
29 November 2000
7/41
Commercial in confidence

Share Link: 

datasheetq.com  [ Privacy Policy ]Request Datasheet ] [ Contact Us ]