Digital signal

Digital signal amplitude value refers to is discrete, amplitude says is limited in a finite numerical within. Binary code is a digital signal. The influence of binary code by noise is small, easy to have digital circuit handling, so get a wide range of applications.

characteristics

Strong anti-jamming capability, without noise accumulation

In simulation of communication, in order to improve signal-to-noise ratio, need in signal transmission process promptly to the transmission signal attenuation amplification, signal during transmission inevitably superimposed on the noise was also at the same time amplification. Along with the increase of the transmission distance, noise accumulated more, to the transmission quality seriously deteriorating.

For digital communication, as a result of digital signal amplitude for a finite discrete values (usually takes two amplitude), although also during transmission by the noise, but when SNR deterioration to certain degree, namely, the proper distance using judgment regeneration methods, regeneration into no noise and the original sender as digital signal, so can realize long distance of high quality transmission.

Facilitate encryption processing

Information transmission, the security and privacy of more and more important, digital communication encryption processing than analog communications easier to voice signal as an example, through digital tansformation signals usable simple digital logic operation for encryption and decryption processing.

Facilitate the storage, processing and exchange

Digital communication signal type and the computer USES signal consistent, are binary code, thus facilitating and computer networking, also facilitate computer auto-calculating on digital signal for storage, processing and exchange, can make the network management and maintenance realize automatic and intelligent.

Equipment facilitate integration, miniature

Digital communication using tdma, do not need bulky filter. Equipment for most of the circuit is digital circuit, usable massive and vlsi implementations, so small volume, low power consumption.

Facilitate constitute integrated digital network and integrated services digital network

Adopting digital transmission mode can be programmed through digital switching equipment for digital transmission and exchange, in order to realize the comprehensive. In addition, telephone service and all kinds of words business can achieve digitization, constitute integrated services digital network.

Occupy channel bandwidth wider

All the way to 4kHz band simulating telephone bandwidth, all the way the digital telephone accounted for about 64kHz, this is an analog communications are still have the vitality of the major reasons. As wideband channel (cable, digital microwave) using (a pair of cable can be opened thousands of road phone) and digital signal processing technology development (can be a digital telephone digital rate by 64kb/s compressed to 32kb/s even lower digital rate), digital telephone bandwidth problem has already is not main problem.

Introduced above knowable, digital communication technology has many advantages, so various countriesall positively development of digital communication. In recent years, China's digital communication is rapidly developing, is heading for a fast pace, intelligent, broadband is changed and integration in that direction.

produce

Analog signals

Signal waveform simulation with information changes, its characteristic is analog signal amplitude continuous (continuous meaning is in a certain scope inside can take infinite multiple number). Analog signals, its signal waveform in time is continuous, so it is a continuous signal. Analog signals according to certain time interval T sampling after the waveform sampling signal, because in time is discrete, but this signal amplitude is still continuously, so is still analog signals. Telephone, fax, television signals are analog signals.

Signal sampling time discrete, but radii of discrete values. Common sampling signal is periodic rectangular pulse and cycle impulse pulse sampling. Analog signals in the whole time axis are defined, in "no convergence value" regional significance are radii value is zero. And discrete time signal only in discrete time only on have a defined and other places no definitions, and rotate the value is zero is different concepts, these two kinds of signal in time axis looks similar, although in different types of system based on two essential difference between signal. Intuitive said, discrete time signal of the horizontal axis can think already do not represent time.

Digital signal

Its characteristic is a digital signal amplitude is limited in a finite numerical within, it is not a continuous but discrete. Binary code, each code yuan only takes two amplitude (0, A) : 4 into yards, each code element take four (3, 1, - 1, one of - 3). The amplitude is discrete signal called digital signals.

Signal digital process

Signal digital requires three steps: sampling, quantitative and coding. Sampling refers to use every certain time signal samples values sequence to replace in time in the continuous signals, namely in time admiral analog signal discretization. With a finite amplitude quantification is approximate continuous variation of the original value of the analog signal amplitude values, continuous amplitude into a limited number of a certain intervals of discrete values. Coding is according to certain rules, the quantified values in binary numerals, then converts binary or more values of digital signal flow. The resulting digital signal by cable, microwave, trunk, satellite channel etc digital circuit transmission. In the receiver has to do with the analog signal digital process instead, repass rearmounted filtering and return to the original analog signals. The above digital process called pulse code modulation.

sampling

The speech signal is analog signal, it not only in amplitude values is continuous, and that in time also continuously. To make the speech signal digital and realize the tdma, must first in time for voice signal discretization of, a process called sampling. So-called sampling is every certain time interval T, extraction of a voice signal instantaneous magnitude value (sampling value), sample after a series of gains from discrete sampling in time value is called sample data sequences, as shown in figure 2-4 shows. Sampling of the sample value after in time sequence is discrete, can undertake tdma, will also be each sample value after quantization, coding transform into a binary digital signals. Theory and practice proves that as long as sampling pulse interval T acuities 1 / (2fm) (or f p 2fm) (FM is the highest frequency voice signal), then a sampling of the sample data sequences after not distortion in reductive into the original voice signal.

For example, all phone signals for 300 ~ band 3400Hz FM = 3400Hz, the sampling frequency fs p 2 x 3400 = 6800Hz. If press 6800Hz sampling frequency to 300 ~ 3400Hz telephone signal sampling, the sampling data sequences of sample after not distortion in reductive into the original voice signal, the speech signal sampling frequency usually take 8000Hz/s. For PAL TV signal. Video bandwidth for 6MHz, according to CCIR601 Suggestions, brightness signal sampling frequency for 13.5 MHz, chroma signal for 6.75 MHz.

quantitative

Sampling the analog signal into a time discrete pulse signal, but the pulse amplitude is still simulation, they must conduct discretization of, can eventually with digital prond. This will be for amplitude on the processing LingQu shekels, the process is called quantification. Quantitative there are two ways, which is shown in figure 2-5. Figure 2-5 (a), shown quantitatively, only integer shebna does not enter the, namely 0 ~ 1 volt between all the input voltage output 0 volts, 1 ~ 2 volts all input voltage output between 1 volts, etc. Use this quantitatively, input voltage is always more than the output voltage, thus creating quantizing error always positive, maximum quantizing error equals two adjacent quantization interval Δ level. Graph (b), shown quantitatively in integer tome give up a into, namely 0 ~.5 volt between the input voltage are output 0 volts, 0.5 ~ 1? 5 volt between the output voltage of all output 1 volts, etc. Use this quantitatively quantizing error has a negative, quantizing error is the absolute value of the maximum Δ / 2. Therefore, use have give up a into law quantitatively, small errors.

The signal can be thought of as quantitative output signal and quantizing error sum, so only quantification output signal to replace the original signal will have distortion. Generally speaking, can put the quantizing error range of probability distribution as in Δ / 2 ~ + - Δ / 2 between evenly distributed. Can proof, quantification distortion power? , namely with minimal quantization interval is proportional to the square. Minimum quantization interval is smaller, distortion is smaller. Minimum quantization interval is smaller, used to represent certain scope analog signal needed for the quantitative series, so the processing and transmission is more complicated. So, quantification both want to reduce quantization series, and to make quantitative distortion doesn't show. Usually use a binary number to indicate whether a certain quantitative series, after transmission at the receiving end again according to the binary number to restore the original signal amplitude. So-called quantification bits means distinguish all quantitative level required several binary number. For example, there were eight quantitative level, then available three binary number will distinguish, because, says eight quantitative level quantification for three bits quantification. Eight bits quantitative means were a quantitative level quantification.

Quantizing error and noise is have substaintial distinction. Because any time of quantizing error can be from the input signal and noise and find out, there is no such relationship between signal. Can be proven, quantizing error is higher-order nonlinear distortion of the product. But quantification distortion in the performance of similar to signal noise, also have a wide spectrum, it is also called quantization noise and signal-to-noise than measure.

The above-mentioned using uniform spacing quantitative level quantitative method as the uniform quantitative or linear quantification, the quantitative way to create large signal when SNR is willing, but the small signal when SNR insufficient faults. If making small signal quantitative level between width is small, and the big signal quantitative level between width some bigger, can make small signal and the large signal when the signal-to-noise ratio to converge. This kind of non-uniform quantization levels arrangements called non-uniform quantitative or nonlinear quantification. Digital TV signal is used mostly non-uniform quantitatively, this is because analog video signal should pass calibration, and calibration is similar to nonlinear quantitative characteristic, can reduce the influence of small signal error.

For audio signal of non-uniform quantification is also adopts compressed and expansion method, namely after the sender to the input signal compression processing again even quantification, again at the receiving end dilate in processing.

Current international generally used easily implemented A law 13 line pressure expansion characteristics and muon law 15 multiline pressure expanding properties. China's regulation adopts A law 13 line pressure expanding properties.

Adopt 13 line pressure expansion characteristics after small signal when quantification SNR improve quantity can amount to 24dB, and this is by sacrifice large signal quantification signal-to-noise ratio (loss 12dB) request.

coding

Sampling, quantified signal is not digital signals, need to convert it to digital coding pulse, the process called encoding. The simplest coding method is binary code. More specifically, even with a n bits to represent binary codes of the sample value has been quantified, each binary digits corresponding to a quantitative values, and then give them arrangement, gets by binary pulses of digital information flow, the whole process (see figure 2-7. Encoding process in the receiver, can according to information received composed of original sample value, repass low-pass filter restore the original signal. In this way the pulse sequence consisting of frequency equals sampling frequency and quantitative bits of integrated, called transferred digital signal digital rate. Obviously, the higher the sampling rate, quantification bits, the greater the digital rate is higher, the needs of transmission bandwidth more wide

In addition to the natural binaries, and other forms of binary code, such as gray code and folding binaries etc, table 2-1 shows the three kinds of binary code. These three yards has its advantages and disadvantages: (1) natural binary code and binary number one-to-one, simple, it is right weight code, each one has a definite size, from the top bits in order to position, can directly size comparison and arithmetic. Natural binary code can be directly by several/mode converter converted into analog signals, but in some cases, for example from decimal 3 converts binary code-and 4 when each all want to change, make the digital circuit generates a lot of peak current pulse. (2) gray code is not this one defect, it in adjacent level switching between, only a born change, gray code is not right weight code, each code without fixed size, not directly compare the size and arithmetic operations, also cannot directly converted into analog signals, should pass a code transformation, become natural binary code. (3) fold binaries down center level fluctuation symmetrical, suitable for said plus or minus symmetrical dual polarity signal. It's the top bits used to distinguish signal amplitude of positive and negative. Folding yards of anti miscode ability.

Table 2-1 various binaries quantization interval

The quantization interval natural binaries gray yards folding binary code

0000000011

1001001010

2010011001

3011010000

4100110100

5101111101

6110101110

7111100111

In communication theory, coding into the source coding and channel coding two kinds big. So-called source coding "shall refer to a source of extra information removed, forming a suitable for transmission signal. In order to restrain the interference of the signal channel noise, often need to signal again coding, encoded in the receiver hard to jamming the mistake form, this is called channel coding. In order to deal with interference, must spend more time, transfer of unnecessary repeat signals, thus occupied more frequency band, this is communication theory of a basic principle.