1、Introduction to Digital Signal Processing Signal and digital signal processing as a branch of information processing, has infiltrated into the scientific research, technological development, industrial production, national defense and national economy in various fields has achieved fruitful results.
2、 Of the signal in time domain and transform domain analysis of the characteristics, processing, enable us to the nature of the signal features and a more clear awareness and understanding of the signals need to get our form of information of use to improve the extent, then in a broader and higher le
3、vel the access to information. DSP chip, also known as digital signal processors, is particularly suitable for digital signal processing microprocessor. The emergence and development of DSP chips, digital signal processing technology for the improvement of many of the new system, the new algorithm c
4、ame into being, the application field is expanding. Currently, DSP chips have been widely used in communications, automation, aerospace, military, medical and other fields. Digital signal processing is that the signal to digital form and deal with the theory and technology. Digital signal processing
5、 and analog signal processing is a subset of signal processing. The purpose of digital signal processing is a continuous real-world analog signals to measure or filter. Therefore needed before carrying out digital signal processing the signal from the analog domain into the digital domain, this is u
6、sually achieved through the ADC. The digital signal processing often have to transform the output to the analog domain, which is achieved through the DAC. The core digital signal processing algorithm is the discrete Fourier transform (DFT), is the DFT to the signal in the digital domain and frequenc
7、y domain are realized discrete, general-purpose computer which can handle discrete signal. Er Shi digital signal processing theory to practical from the fast Fourier transform (FFT), FFT has significantly reduced the DFT computation to make real-time digital signal processing as possible, greatly pr
8、omote the development of the discipline.The advantages of digital signal processing system as follows: 1. Good flexibility: when the processing method and parameters change, the processing system simply by changing the software design to meet the appropriate changes. 2. Precision: Signal processing
9、systems can be A / D transformation of the median, the processor word length and the appropriate algorithm to meet the accuracy requirements. 3. Reliability: processing system affected by environmental temperature, humidity, noise and electromagnetic interference caused by the less affected. 4. Can
10、be a large-scale integration: With the technological development of semiconductor integrated circuits, digital circuits can be used for very high integration, small size, low power consumption, and good product consistency.1.The development of DSPThe late 70s early 80s, AMIs S2811 chip, Intels 2902
11、chip, the birth marks the beginning of DSP chips. With the rapid development of semiconductor integrated circuits, high-speed real-time digital signal processing and digital signal processing applications demands the continuous extension of the field, in the 80 years since the beginning of ten years
12、, DSP made epoch-making development. View from the operation speed, MAC (multiply and accumulate) time of 400 ns from 80 down to 40 ns below, data-processing capacity several times. MIPS (million instructions per second) from the early 80s 5MIPS to the current 40 MIPS or more. Key components within
13、the DSP chip multiplier accounts from the early 80s about 40% of mold Area dropped to less than 5%, on-chip RAM for more than an order of magnitude increase. View from the manufacturing process, the early 80s with 4m of NMOS technology and is now using sub-micron CMOS technology, DSP chips, pin numb
14、ers from the early 80s up to 64 to more than 200 now, so the increase in the number of chip pins increase the flexibility of application, so that all external memory expansion and more convenient communication between processors. And earlier than DSP chips, DSP chips now have two floating point and
15、fixed-point data format, floating-point DSP chips for floating-point operations, so computing precision greatly enhanced. DSP chip cost, size, voltage, weight and power consumption earlier DSP chip with a large degree of decline. In the DSP development system, software and hardware development tools
16、 continue to improve. At present some of the chip with the corresponding integrated development environment, which supports breakpoint settings and program memory, data memory and DMA access and the program running and tracking a single department, and can use high-level language programming, some m
17、anufacturers and some software development business application software for the DSP to prepare a common subroutine library and various algorithms and various interface program, which makes application development easier, shorten development time, thus increasing the efficiency ofGeneral characteris
18、tics of DSP chips(1) In a cycle to complete a multiplication and an accumulation.(2)Harvard architecture, program and data space separated, you can also access instructions and data.(3) Chip with fast RAM, typically by an independent data bus simultaneously in two visits.(4) A low cost or no cost re
19、cycling and jump hardware support.(5) Fast interrupt handling, and hardware I / O support.(6) With single-cycle operation in a number of hardware address generator.(7) Can perform multiple operations in parallel.(8) To support pipeline operations, fetch, decode, and execution operations may overlap.
20、2.DSP Applications in Communication System DSP technology has been widely used in communications. Mainly in the following areas. (1)Software Defined Radio Software radio technology and computer technology is constantly touch merged into the 3rd generation mobile communication systems provide a good
21、user interface. DSP hardware technology and its algorithm is the key to realization of software radio. Software radio system flexibility, openness and compatibility features of the signal processor is mainly centered through a common hardware platform and software to achieve. It is mainly to complet
22、e the internal data processing station, modem and codec and so on. As the internal data flow and a large radio filtering, frequency and other processing operations more often, to high-speed, real-time, parallel digital signal processor module or ASIC to meet the demand. To complete such an arduous t
23、ask, to increase the processing speed required hardware, chip capacity expansion and requested algorithm for optimization and improvement of the processor. The two aspects of the continuous improvement requirements will be the development of digital signal processing technology unremitting power. Th
24、e only way to achieve high-speed internal software radio operation and multiple functions of the flexible switching and control. Software implementations are generally two types of devices that use DSP to implement and field-programmable gate array (FPGA) to implement. (2)Speech Coding The purpose o
25、f voice data compression is that in the transmission rate as low as possible gain high-quality audio performance, that is, that narrow bandwidth of voice signals Ke Yi in the channel transmission, the voice of quality and Xiajiangdebu more Huo dropped as much as possible . Speech coding system for t
26、he early use of the waveform coding method, also called waveform coding. Essentially follows the Nyquist sampling theorem, strong adaptability, good quality synthetic speech, but the high rate coding, coding efficiency is very low. The parameter coding is different from the efficient encoding of wav
27、eform coding method, which is based on the mechanism in speech production, mainly on the extraction characteristics of speech signal encoding parameters, you can achieve very low encoding rate. But only to the effect of synthesized speech, voice quality than waveform coding. Over the past decade, sp
28、eech coding technology has made breakthrough progress. ITU, one after another through a series of low bit rate speech coding standard telephone band. By the parameter coding and waveform coding the hybrid coding method that is of a synthetic coding, can get better sound quality at the same time redu
29、ce the coding rate, the most representative is the linear predictive coding (LPG) and Code Excited Linear Prediction Coding (CELP). This encoding can be 4-16kbit / s of the low coding rate on high-quality reconstructed speech, but the complexity of the algorithm, the computing speed of the processor
30、 demanding. On speech processing, the higher the compression ratio, the more complex encoding algorithm, real-time compression can not use the logic circuit, it will not use the bulky, slow, high cost of Microcomputer. The DSP is a suitable choice, the web conferencing, voice communications, surveil
31、lance systems are important areas of components. The use of DSP voice compression algorithm not only provides a broad application prospects, and the system design simple, reliability is also greatly improved. (3)GPS System GPS was developed by the United States to receive navigation satellite signal
32、s based on non-autonomous navigation system. Is widely used in various military and economic fields. With GPS technology in various fields to promote and popularize the application of the receivers small size, intelligence and algorithms to meet user needs studies is necessary. There are two global
33、positioning system components: the satellite constellation, ground control / monitoring network and user receivers. In GPS applications, GPS receivers often require reprocessing the data collected, or use GPS receivers to provide some information for the development of an industry. DSP small size, h
34、igh speed, low power, high reliability characteristics. Suited to the complexity of real-time GPS signal processing. OEM version of its composition with the GPS information systems, not only satisfy the real-time GPS signal processing and high complexity, and because of the powerful DSP processing c
35、apability, the system can be further extensions. Typical Application: (1) General signal processing: convolution, correlation, FFT, Hilbert transform, adaptive filtering, spectrum analysis, waveform generation and so on. (2) Communications: High-speed modulator / demodulator, encoder / decoder, adap
36、tive equalizer, simulation, honeycomb network of mobile phones, echo / noise cancellation, fax, telephone conference, spread spectrum communications, data encryption and compression . (3) voice signal processing: speech recognition, speech synthesis, text changed voice, speech coding vector. (4) gra
37、phics image signal processing: two, three-dimensional graphics transformation and processing, robot vision, digital map, image enhancement and recognition, image compression and transmission, animation, desktop publishing systems. (5) control: robot control, engine control, automatic driving, voice
38、and so on. (6) Instruments: function occurs, data acquisition, aerospace wind tunnel testing. (7) Consumer Electronics: Digital TV, digital voice synthesis, toys and games, digital answering machine. To DSP chip as the core structure of the digital signal processing system, data acquisition, transmi
39、ssion, storage and high-speed real-time processing as one can fully realize the advantages of digital signal processing system, can satisfy the precision equipment in the field of manned space flight , reliability, channel bandwidth, power consumption, voltage and weight requirements. Currently, DSP
40、 chips are the high-performance, high integration and low cost of direction, the various kinds of common and dedicated DSP chips has introduced new, applied technology and development tools continues to improve. Such as real-time digital signal processing applications - especially in the field of ma
41、nned space applications in a wider space. We have reason to believe, DSP chips will further the development and application of signal processing on the manned space far-reaching impact.数字信号处理器简介 数字信号处理作为信号和信息处理的一个分支学科，已渗透到科学研究、技术开发、工业生产、国防和国民经济的各个领域，取得了丰硕的成果。对信号在时域及变换域的特性进行分析、处理，能使我们对信号的特性和本质有更清楚的认识
42、和理解，得到我们需要的信号形式，提高信息的利用程度，进而在更广和更深层次上获取信息。DSP芯片，也称数字信号处理器，是一种特别适合进行数字信号处理运算的微处理器。DSP芯片的出现和发展，促进数字信号处理技术的提高，许多新系统、新算法应运而生，其应用领域不断拓展。目前，DSP芯片已广泛应用于通信、自动控制、航天航空、军事、医疗等领域。 数字信号处理是将信号以数字方式表示并处理的理论和技术。数字信号处理与模拟信号处理是信号处理的子集。数字信号处理的目的是对真实世界的连续模拟信号进行测量或滤波。因此在进行数字信号处理之前需要将信号从模拟域转换到数字域，这通常通过模数转换器实现。而数字信号处理的输出经
43、常也要变换到模拟域，这是通过数模转换器实现的。数字信号处理的核心算法是离散傅立叶变换(DFT)，是DFT使信号在数字域和频域都实现了离散化，从而可以用通用计算机处理离散信号。而使数字信号处理从理论走向实用的是快速傅立叶变换(FFT)，FFT的出现大大减少了DFT的运算量，使实时的数字信号处理成为可能、极大促进了该学科的发展。 数字信号处理系统的优越性表现为：1.灵活性好：当处理方法和参数发生变化时，处理系统只需通过改变软件设计以适应相应的变化。2.精度高：信号处理系统可以通过A/D变换的位数、处理器的字长和适当的算法满足精度要求。3.可靠性好：处理系统受环境温度、湿度，噪声及电磁场的干扰所造成
44、的影响较小。4.可大规模集成：随着半导体集成电路技术的发展，数字电路的集成度可以作得很高，具有体积小、功耗小、产品一致性好等优点。 1、DSP的发展 70年代末80年代初，AMI公司的S2811芯片，Intel公司的2902芯片的诞生标志着DSP芯片的开端。随着半导体集成电路的飞速发展，高速实时数字信号处理技术的要求和数字信号处理应用领域的不断延伸，在80年代初至今的十几年中，DSP芯片取得了划时代的发展。从运算速度看，MAC（乘法并累加）时间已从80年代的400 ns降低到40 ns以下，数据处理能力提高了几十倍。MIPS（每秒执行百万条指令）从80年代初的5MIPS增加到现在的40 MIP
45、S以上。DSP芯片内部关键部件乘法器从80年代初的占模片区的40%左右下降到小于5%，片内RAM增加了一个数量级以上。从制造工艺看，80年代初采用4m的NMOS工艺而现在则采用亚微米CMOS工艺，DSP芯片的引脚数目从80年代初最多64个增加到现在的200个以上，引脚数量的增多使得芯片应用的灵活性增加，使外部存储器的扩展和各个处理器间的通信更为方便。和早期的DSP芯片相比，现在的DSP芯片有浮点和定点两种数据格式，浮点DSP芯片能进行浮点运算，使运算精度极大提高。DSP芯片的成本、体积、工作电压、重量和功耗较早期的DSP芯片有了很大程度的下降。在DSP开发系统方面，软件和硬件开发工具不断完善。
46、目前某些芯片具有相应的集成开发环境，它支持断点的设置和程序存储器、数据存储器和DMA的访问及程序的单部运行和跟踪等，并可以采用高级语言编程，有些厂家和一些软件开发商为DSP应用软件的开发准备了通用的函数库及各种算法子程序和各种接口程序，这使得应用软件开发更为方便，开发时间大大缩短，因而提高了产品开发的效率。 通用DSP芯片的特点 (1)在一个周期内可完成一次乘法和一次累加。 (2)采用哈佛结构，程序和数据空间分开，可以同时访问指令和数据。 (3) 片内有快速RAM，通常可以通过独立的数据总线在两块中同时访问。 (4) 具有低开销或无开销循环及跳转硬件支持。 (5) 快速中断处理和硬件I/O支持
47、。 (6) 具有在单周期内操作的多个硬件地址产生器。 (7) 可以并行执行多个操作。 (8)支持流水线操作，取指、译码和执行等操作可以重叠进行。2、DSP在通信系统中的应用 DSP 技术已广泛应用于通信领域。主要集中在以下几个方面。 （1）软件无线电 软件无线电技术与计算机技术正在不断触合为第3 代移动通信系统提供了良好的用户界面。DSP 的硬件技术及其算法是实现软件无线电的关键所在。软件无线电系统的灵活性、开放性和兼容性等特点主要是通过以信号处理器为中心的通用硬件平台及软件来实现的。它主要完成电台内部数据处理、调制解调和编码解码等工作。由于电台内部数据流且大进行滤波、变频等处理运算次数多,必
48、须采用高速、实时、并行的数字信号处理器模块或专用集成电路才能达到要求。要完成这么艰巨的任务,必须要求硬件处理速度不断增加,芯片容量扩大,同时要求算法进行针对处理器的优化和改进。这2 个方面要求的不断提高将是数字信号处理技术发展的不懈动力。只有这样,才能实现电台内部软件的高速运行和多种功能的灵活切换和控制。软件的实现方式一般有两种, 即采用DSP 器件来实现和现场可编程门阵列(FPGA)来实现。 (2)语音压缩编码 语音数据压缩的目的是能在尽可能低的传输速率上获得高质量的语音效果,即希望语音信号可以在带宽较窄的信道中传输,而语音的质且下降得不多或尽可能不下降。语音编码系统早期使用的是波形编码方法
49、,也叫波形编码。其本质上遵循奈奎斯特采样定理,适应能力较强,合成语音质量较好,但是编码速率高,编码效率极低。而参数编码是不同于波形编码的高效编码方式,它是从语音产生的机理出发,主要是对提取的语音信号特征参数进行编码,可以达到极低的编码速率。但是只能达到合成语音的效果,语音质量不如波形编码。近十年来,语音编码技术取得了突破性的进展。ITU 等陆续通过了一系列低码率的电话频带语音编码标准。由参数编码和波形编码结合的混合编码方式即分析一合成编码,可在获得较好音质的同时有效降低编码率,其中最具有代表性的是线性预测编码(LPG)和码激励线性预测编码(CELP)。这种编码方式能在4-16kbit/s的中低编码速率上得到高质量的重建语音,但算法复杂,对处理器的运算速度要求很高。对语音处理来说,压缩倍率越高,编码算法也越复杂,实时压缩就不可能用逻辑电路实现,也不会用体积大、速度慢、成本高的微机实现
陕公网安备 61072602000132号 违法和不良信息举报：0916-4228922