You are on page 1of 20

AN INTRODUCTION TO

Da Vinci
Supervised by:

Technology

Dr. D.D.Doye

Presented by: 1/11/2010

SYED MUFFASSIR M.S.ALI


SGGS I & ET, Nanded SGGS I & ET,Nanded,INDIA 1

DA VINCI and TI are the trademarks of Texas Instruments Inc, USA.

Contents
Introduction Block diagram of DaVinci Processors Important Features Innovation Made Easy Implementation examples Impact References
1/11/2010 SGGS I & ET,Nanded,INDIA 2

What is DaVinci Technology?


Texas Instruments DaVinci Technology is a new software technology that has combined software, hardware, and development tools to form a digital signal processing (DSP) system that enhances digital multimedia.

1/11/2010

SGGS I & ET,Nanded,INDIA

Why was it developed?


To eliminate CODEC errors between different processors. To bring digital video to the component level, enabling developers to introduce digital video to an application without requiring developers to program a single line of DSP code.
1/11/2010 SGGS I & ET,Nanded,INDIA 4

Why not FPGAs or ASICs?


FPGAs require more Power and are not compact. ASICs fails to achieve best quality under wide range of operating conditions and limits their ability to adapt to multiple applications. Development time required and time is more.
1/11/2010 SGGS I & ET,Nanded,INDIA 5

system-onProcessors: Digital video system-on-chips


TMS320DM6443 Video decode: Video Accelerator and Networking for display TMS320DM6446/7 Video encode/decode: Video Accelerator and Networking for (HD) capture and display
DM6443

DM6446

Tools:DaVinci Hardware and Software Development

DVEVM

Digital Video Evaluation Module

DVSDK
Digital Video Software Development Kit

1/11/2010

SGGS I & ET,Nanded,INDIA

TMS320DM644x Processors
DM6443 /6
ARM Subsystem
Video-Imaging Coprocessor (VICP)

Video Processing Subsystem Front End


Resizer Histogram/3A Preview Video Processing Subsystem CCD Controller Video Interface

DSP Subsystem

Video-Optimized TMS320C64x+ DSP @ 810MHz


H.264 MP@L3, 30fps SD Decoding VC1/WMV9 Full D1 SD Decoding MPEG-2 MP@ML SD Decoding MPEG-4 ASP Full D1 SD Decoding H.264 BP D1 Encoding Simultaneous H.264 BP CIF Coding

ARM926EJ-S 405 MHz CPU

C64x+TM DSP 810 MHz Core

Back End
On-Screen Display (OSD) Video Enc (VENC) 10b DAC 10b DAC 10b DAC 10b DAC

Switched Central Resource (SCR)

Dedicated video processing subsystem


Back end - Integrated OSD, four video DACs, 24-bit digital RGB output Front end Resizer, Image processing engine, 16-bit digital input
2006 10KU Price: TMS320DM6443 $29.95 TMS320DM6446 $34.95

Peripherals
EDMA 3.0 USB 2.0 PHY

Connectivity VLYNQ EMAC With MDIO

System GeneralPurpose Timer Watchdog Timer

Serial Interfaces Audio Serial Port I2 C SPI UART UART UART DDR2 Controller (16b/32b)

Program/Data Storage ATA/ Async EMIF/ Compact NAND/ Flash SmartMedia MMC/ SD

= DM6446 additions = New for DM644x


7

1/11/2010

SGGS I & ET,Nanded,INDIA

Important Features of DaVinci Processors.


ARM Core + DSP Core.. Advanced Very-Long-InstructionWord (VLIW) DSP Core. Image Processing Pipeline. Eight Functional Units : 6 ALUs(32/40 Bit),2 Multipliers of 16X16 bit Multiplies per clock cycle. 64 32-Bit general Purpose Registers. Video Processing Subsystem viz VPFE &VPBE. Speed of DSP:6480 MIPS. Memory

ARM: 16K I-Cache, 8K D-Cache, 16KB RAM, 8K ROM DSP: 32K L1 P-Cache, 32K L1 D-Cache, 128K L2 Cache, 64K Boot ROM
1/11/2010 SGGS I & ET,Nanded,INDIA 8

Video Processing Subsystem.


EndVideo Processing Front End1. CCD Controller(CCDC), a Preview Engine (Previewer), Histogram Module, 2. Auto Exposure/White Balance/Focus Module (H3A), and 3. Resizer. EndVideo Processing Back End1. On-Screen Display Engine (OSD) 2. Video Encoder (VENC).
1/11/2010 SGGS I & ET,Nanded,INDIA 9

How it becomes easy to Innovate?


Using Hardware and Software Toos like DVEVM and DVSDK . 1. DVEVM has both hardware and software. 2. DVSDK is designed to tune complex Da Vinci based digital video systems quickly & efficiently. Open Source and Linux based. Fast Development with DaVinci On Screen Display.
1/11/2010 SGGS I & ET,Nanded,INDIA 10

How it becomes easy to Innovate?


Software Tools 1. Software development tool Code Composer Studio (CCS) 2. Two modes, Simulator and Emulator. 3. Collection of optimized image/video processing functions

1/11/2010

SGGS I & ET,Nanded,INDIA

11

Exploiting SoC (DaVinci ) for Reduced BOM.


Example: -PLUTO 3.0 PLUTO Single Channel, Micro Networked Video Security Processing Platform
Courtesy: YMagic Inc, USA.

1/11/2010

SGGS I & ET,Nanded,INDIA

12

Value: DaVinci Processors


with Reduced BOM.
Example:---PLUTO 3.0 Single Channel, Micro Video Processing Platform
Ethernet 10/100MBit Ethernet PHY CMOS Camera I/F Video Encoder (A/D)

Digital ITU656 Media Processor

Video Decoder (D/A)

Analog Video Out

Analog Video In

GLUE Logic (FPGA IDE controller)

GLUE Logic (CPLD for I/F) USB Controller

GLUE Logic (FPGA for MMC/SD Controller)

GLUE Logic (CPLD for GPIO addition) Peripheral Hardware Interfaces

Hard Disk

MMC/SD Socket

Courtesy: YMagic Inc,USA


1/11/2010

YMagic is the trademark of YMagic Inc,USA.


SGGS I & ET,Nanded,INDIA 13

Value: DaVinci Processors


with Reduced BOM.
Example:--PLUTO 3.0 New Generation Single Channel, Micro Networked Video Security Processing Platform

Ethernet 10/100MBit Ethernet PHY CMOS Camera I/F Video Encoder (A/D)

ITU656 Media Processor

Digital DM6446

Video Encoder (D/A)

Analog Video Out

Analog Video In

GLUE Logic (FPGA IDE controller)

GLUE Logic (CPLD for I/F) USB Controller

GLUE Logic (FPGA for MMC/SD Controller)

GLUE Logic (CPLD for GPIO addition) Peripheral Hardware Interfaces

Hard Disk

MMC/SD Socket

Courtesy: YMagic Inc,USA


1/11/2010

YMagic is the trademark of YMagic Inc,USA.


SGGS I & ET,Nanded,INDIA 14

Exploiting Computational Speed of DaVinci


Experimental Results and Analysis from Implementation of Contourlet Transform for Image Fusion using DaVinci Platform.
By: Y.Wei,Y.Zhu,Feng.Z,,Y.Shi,Mo and et al.

1/11/2010

SGGS I & ET,Nanded,INDIA

15

Contourlet Xform for Image Fusion


It is a multi-scale and multidirectional,two dimensional transform. It is a combination of Laplacian pyramid and directional filter bank. Image Fusion:Image C = Inv C (C(Image A) +C(Image B))

1/11/2010

SGGS I & ET,Nanded,INDIA

16

Experimental Results and Analysis:


Program of contourlet-based image fusion consumes 8 seconds only on DaVinci platform instead of 67 seconds on ARM core. Loop of Convolution operation 1000 times requires 46 sec while it takes 1698 sec on ARM.

1/11/2010

SGGS I & ET,Nanded,INDIA

17

Impact!!
Decrease time-to-market turnover of new products. Enable OEM product differentiation with flexible, complete solution. Lower system cost Fast application development.

1/11/2010

SGGS I & ET,Nanded,INDIA

18

References:
www.focus.ti.com/docs/prod/folders/print/tms320dm6446.html/ www.thedavincieffect.com/ Y.Wei,Zhu & et al. Implementing Contourlet Transform on Heterogenous Platform.IEEE Eigth International Conference on Embedded Computing. Deepu Talla and et al.Using DaVinci Technology for DV Devices. IEEE Computer Society Micro, vol. 24, no. 2, Oct/Nov. 2009, pp. 53-61. IEEE Review I October 2005 www.ieee.org/review TIs Da Vinci targets Video Market.Page No:9 X.Kong,H.Lin,J.Lin,S.Chen & et al. Implemetation of G.729 Codec on DaVinci Technology,IEEE 2008 International Conference on Multimedia and IT. http://www.ti.com/lit/pdf/sprue26 Digital Signal Processors by B.Venkataramani, M.Bhasker. Published byTata McGraw Hill Ltd,New Delhi.

1/11/2010

SGGS I & ET,Nanded,INDIA

19

THANKS!!
1/11/2010 SGGS I & ET,Nanded,INDIA 20

You might also like