AVS編解碼

AVS編解碼,北京奧維視訊科技有限責任公司(Beijing AVSolutiontech Technology Co., Ltd)是國內為數不多的幾家自主研發數字音視頻壓縮算法的高新技術企業之一。公司的核心成員均具有碩士學歷,是一支富有激情和創新精神的年輕團隊,我們將憑藉敏銳的市場洞察力,本著對客戶高度負責、對科學認真求實的態度,致力將“奧維視訊”打造成行業中具備技術領先優勢的知名品牌。

基本介紹

  • 中文名:AVS編解碼
  • 外文名:Beijing AVSolutiontech Technology Co., Ltd
  • 產品型號:   AVS-P2 Decoder on DM64x+
  • 產品類別 :   算法
AVS編解碼AVS編解碼
奧維視訊通過對行業的深入認知、辨析,選擇了美國德州儀器公司(Texas Instruments,簡稱TI)作為自己的技術合作夥伴。奧維視訊的標誌性產品——商用AVS音視頻編解碼算法,便是基於TI公司的DaVinci系列處理器開發的,該算法完全符合我國自主智慧財產權的第二代信源編碼標準——《中華人民共和國國家標準 信息技術 先進音視頻編碼》。
基於TI的DaVinci系列處理器,奧維視訊還可以為客戶提供各種數字音視頻產品的解決方案,以滿足樓宇監控、可視對講、智慧型交通、視頻電話、DVR(數字視頻錄像機)和視頻伺服器等不同領域的多媒體通訊產品的技術要求,並能夠迅速的為客戶搭建功能樣機平台。根據產品的不同技術特點,奧維視訊可以為客戶定製具有高附加值的一對一解決方案,避免產品同質化現象,提高產品生命周期和競爭力,並且可以為需求客戶提供各種套用模組的OEM、ODM服務。
奧維視訊已成功推出了DaVinci系列處理器的評估系統、多種配套板卡和套用軟體。針對廣大用戶的需求,還特別在每個季度舉辦專業的DaVinci產品技術開發培訓班和研討會。
AVS-P2解碼器
產品名稱   AVS-P2解碼器
產品型號   AVS-P2 Decoder on DM64x+
產品類別   算法
公開報價  

產品概述   The AVS standard is the series of 'information technology advanced Audio and Video Coding' standard abbreviation. The AVS standard includes four main technical standards: system, video, audio and digital copyright management.
The fully Intellectual Property of second generation Source Coding-Decoding Standard – AVS is owned by China.
The AVS source coding technology mainly focuses on mass video and audio (original data and sources) compression technology, also known as Digital Video and audio coding technology. Obviously, as the premise of digital information transmission, storage and broadcast it is also the general foundation character standard of the digital video& audio industry.
On 30th April, 2005, AVS standard video part officially passed the public show and became the national standard.

特性   eXpressDSPTM Algorithm Interface Standard (XDAIS) compliant
eXpressDSP Digital Media (XDM) interface compliant
Validated on the DM6446 EVM
AVS-P2 JiZhun Profile up to level 4.0 supported
Progressive, interlaced and Picture Adaptive Frame Filed (PAFF) picture type decoding supported
Multiple slices and multiple reference frames supported
2D-VLC decoding supported
Intra-prediction and inter-prediction modes supported
Up to 4 MV per MB supported
Frame size being non-multiples of 16 through frame cropping supported
Frame width of the range 16 to 1920 pixels supported
Byte-stream syntax for the input bit stream supported
Outputs are available in YUV 420 planar and 422 interleaved little endian formats
In-loop filter supported
All Intra macro block mode supported
All P and B macro block mode supported
Long term reference frames supported
Reference picture list reordering supported

功能概要   This section describes performance of the AVS-P2 jizhun profile decoder.
Table 1. Configuration Table
CONFIGURATION
ID
Level 1.0 Jizhun Profile
AVS_DEC_001
Level 2.0 Jizhun Profile
AVS_DEC_002
Level 3.0 Jizhun Profile
AVS_DEC_003
Table 2. Cycles Information – Profiled on DM644x EVM With Code Generation Tools Version 6.0.15
CONFIGURATION ID
PERFORMANCE STATISTICS(1) (MEGA CYCLES PER SECOND)(2)
TEST DESCRIPTION
AVERAGE
AVS_DEC_001
Foreman_1Mbps_CIF_ GOP=IBBP…BBP_352x288@30fps
4MV, QPI, LPF On, MB-Level RC, 300Frames
54
AVS_DEC_002
Football_2Mbps_D1_GOP=IPPP…PPP_704x576@25fps
4MV, QPI, LPF On, MB-Level RC, 250Frames
240
AVS_DEC_003
Transformer_5Mbps_D1_GOP=IBBP…BBP_720x480@30fps
4MV, QPI, LPF On, MB-Level RC, 2500Frames
280
(1) There could be a variation of 1-2% in the values
(2) Measured with program memory, stack, I/O buffers in external memory, 32K-bytes L1P Cache, 64K-bytes L1D Data memory, 16K-bytes L1D Cache, 64K-bytes L2 Cache, DDR speed at 162 MHz
Table 3. Memory Statistics - Generated With Code Generation Tools Version 6.0.15
CONFIGURATION ID
MEMORY STATISTICS(1)
TOTAL
PROGRAM MEMORY
DATA MEMORY
INTERNAL(2)
EXTERNAL(3)
STACK(4)
AVS_DEC_003
125
64
2660
1
2850
(1) All memory requirements are expressed in kilobytes (1K-byte = 1024 bytes) and there could be a variation of approximately 1-2% in values
(2) Internal data memory is used for system heap
(3) External memory is calculated under 720x576 with 422Interleave format
(4) Maxim stack size is 669 bytes under AVS-P2 conference test
Table 4. DMA Resource Statistics
CONFIGURATION ID
DMA RESOURCE
QDMA
MAX TRANSFER
MAX TRANSFER SIZE
MAX PARAM SET
AVS_DEC_003
0
1
829440
1
1
1
32224
1
2
8
338
8
3
2
882
2
Table 5. VICP Memory Resource Statistics
CONFIGURATION ID
MEMORY STATISTICS(1)
seq DATA
SEQ PROGRAM
IMX CMD
IMAGE BUF
IMX WORKBUF
AVS_DEC_003
0
0
0
0
0
(1) The AVS-P2 algorithm does not need any VICP resource
NOTES:
(1) L1P : 32K-bytes program cache (32 bytes cache line width, direct mapped cache)
(2) L1D : 64K-bytes data memory and 16K-bytes data cache (64 bytes cache line width, 2-way set associative cache)
(3) L2 : 64K-bytes cache (128 bytes cache line width, 4-way set associative cache)
AVS-P2編碼器
產品名稱   AVS-P2編碼器
產品型號   AVS-P2 Encoder on DM64x+
產品類別   算法
公開報價  

產品圖片  

產品概述   The AVS standard is the series of 'information technology advanced Audio and Video Coding' standard abbreviation. The AVS standard includes four main technical standards: system, video, audio and digital copyright management.
The fully Intellectual Property of second generation Source Coding-Decoding Standard – AVS is owned by China.
The AVS source coding technology mainly focuses on mass video and audio (original data and sources) compression technology, also known as digital video and audio coding technology. Obviously, as the premise of digital information transmission, storage and broadcast it is also the general foundation character standard of the digital video& audio industry.
On 30th April, 2005, AVS standard video part officially passed the public show and became the national standard.

特性
eXpressDSPTM Algorithm Interface Standard (XDAIS) compliant;
eXpressDSP Digital Media (XDM) interface compliant;
Validated on the DM6446 EVM;
AVS-P2 JiZhun Profile up to level 4.0 supported;
Quarter-PEL interpolation for motion estimation supported;
In-loop filtering which can be switched off for whole picture and slice boundaries supported;
User controllable multiple slices per picture supported;
User controllable quantization paramter range supported;
unrestricted motion vector search which allows motion vectors to be outside the frame boundary supported;
Frame size being non-multiples of 16 through frame cropping supported;
Arbitrary resolutions up to PAL D1 (720x576), including standard image sizes such as SQCIF, QCIF, CIF, QVGA, and VGA supported;
User configurable group of pictures (GOP) length supported;
YUV422 interleaved and YUV420 planar color sub-sampling formats supported;
Controls the balance between encoder speed and quality by using the user definable motion estimation settings and RDO (Rate-Distortion-Optimize) complexity option;
All Intra modes in I and P slices supported;
Run time changes in bit-rate, frame rate, and resolution supported;
Byte stream format supported.

功能概要  
Performance Summary
This section describes performance of the AVS-P2 jizhun profile encoder.
Table 1. Configuration Table
CONFIGURATION
ID
Level 1.0 Jizhun Profile
AVS_ENC_001
Level 2.0 Jizhun Profile
AVS_ENC_002
Table 2. Cycles Information – Profiled on DM644x EVM With Code Generation Tools Version 6.0.15
CONFIGURATION ID
PERFORMANCE STATISTICS(1) (MEGA CYCLES PER SECOND)(2)
TEST DESCRIPTION
AVERAGE
AVS_ENC_001
Foreman_512kbps_CIF_ GOP=IPPP…PPP_352x288@30fps
1MV, QPI, LPF On, Picture-Level RC, 300Frames
170
AVS_ENC_002
Substation_4Mbps_D1_GOP=IPPP…PPP_704x576@25fps
1MV, QPI, LPF On, Picture -Level RC, 800Frames
480
Transformer_4Mbps_D1_GOP=IPPP…PPP_720x480@30fps
1MV, QPI, LPF On, Picture -Level RC, 200Frames
580
(1) There could be a variation of 1-2% in the values
(2) Measured with program memory, stack, I/O buffers in external memory, 32K-bytes L1P Cache, 64K-bytes L1D Data memory, 16K-bytes L1D Cache, 64K-bytes L2 Cache, DDR speed at 162 MHz
Table 3. Memory Statistics - Generated With Code Generation Tools Version 6.0.15
CONFIGURATION ID
MEMORY STATISTICS(1)
TOTAL
PROGRAM MEMORY
DATA MEMORY
INTERNAL(2)
EXTERNAL(3)
STACK(4)
AVS_ENC_002
65
64
1860
1
2850
(1) All memory requirements are expressed in kilobytes (1K-byte = 1024 bytes) and there could be a variation of approximately 1-2% in values
(2) Internal data memory is used for system heap
(3) External memory is calculated under 720x576 with 422Interleave format
(4) Maxim stack size is 669 bytes under AVS-P2 conference test
Table 4. DMA Resource Statistics
CONFIGURATION ID
DMA RESOURCE
QDMA
MAX TRANSFER
MAX TRANSFER SIZE
MAX PARAM SET
AVS_ENC_002
0
3
512
3
1
3
2304
3
2
3
13680
3
3
3
2304
1
Table 5. VICP Memory Resource Statistics
CONFIGURATION ID
MEMORY STATISTICS(1)(2)
SEQ DATA
SEQ PROGRAM
IMX CMD
IMAGE BUF
IMX WORKBUF
AVS_DEC_003
1
2.5
1
8
4
(1) The AVS-P2 algorithm does not need any VICP resource
(2) The AVS-P2 algorithm has a VICP option which can be turned on when used on DM644x SoC
Notes
(1) L1P : 32K-bytes program cache (32 bytes cache line width, direct mapped cache)
(2) L1D : 64K-bytes data memory and 16K-bytes data cache (64 bytes cache line width, 2-way set associative cache)
(3) L2 : 64K-bytes cache (128 bytes cache line width, 4-way set associative cache)
AVS算法評估平台
產品名稱   AVS算法評估平台
產品型號   CodecEVP
產品類別   算法
公開報價   ¥9,800.00

產品圖片

AVS編解碼
AVS編解碼AVS編解碼

產品優勢   奧維視訊的CODECEVP開發套件包含了集成數個成熟穩定算法的SDK軟體包,節省了客戶開發、集成、測試算法的時間,也使客戶方便地開發出基於Linux系統的各種應用程式,全面滿足不同市場的需求,並適時推出極具競爭力的的特色產品。

產品概述   The AVST_CODECEVP module (160mm x 108mm x 29mm) can be directly used for mass products and it also enables developers to start immediate evaluation of DaVinciTM processors and begin building digital video applications such as surveillance digital video servers / recorders and many other Stadard-Definition video applications.
The AVST_CODECEVP allows developers to write production-ready application code for the ARM and provides almost fullest codec combo which access the 64x+ DSP core and Video / Imaging Co-Processor core using DaVinci APIs to begin immediate application development for the TMS320DM6446 digital media processor.

硬體資源 1、處理器:TMS320DM6446 DaVinciTM processor
2、DDR2:256MB
3、Nor flash:16MB
4、複合視頻input/output
5、S-Video input/output
6、立體聲音頻input/output
7、兩個RS232接口(雙列直插)
8、USB2.0 接口
9、10/100Mbps Ethernet 接口
10、+5V DC 單排直插電源
11、JTAG 接口

軟體資源   CODECEVP SDK 包含:
1、AVS-P2 JZP / MPEG-4 SP / H.264 BP Video Encoder
2、AVS-P2 JZP / MPEG-2 MP / MPEG-4 SP / H.264 BP / Video Decoder
3、AAC / MP3 Audio Decoder
4、G.711 / G.729AB Voice Codec
5、JPEG Image Codec
6、Demo version of Montavista Linux Pro 4.0
7、MontaVista Pro Demonstration Tools
8、Drivers for UART, I2C, SPI, EDMA, USB Host/gadget, Video Ports, etc.
9、OSS Audio (ASP), GPIO, PWM, WDTIM
10、U-Boot Loader
11、IR remote
12、DM6446 Getting Started Guide and Hardware Technical Reference
13、DM6446 DVEVM schematics in PDF format

熱門詞條

聯絡我們