Analog and Digital Signaling
The human voice generates sound waves; a telephone converts the sound waves into analog signals. However, analog transmission is not particularly efficient. Analog signals must be amplified when they become weak from transmission loss as they travel. However, amplification of analog signals also amplifies noise.
The PSTN is a collection of interconnected voice-oriented public telephone networks, both commercial and government-owned. The PSTN today consists almost entirely of digital technology, except for the final link from the central (local) telephone office to the user. To obtain clear voice connections, the PSTN switches convert analog speech to a digital format and send it over the digital network. At the other end of the connection, the digital signal is converted back to analog and to the normal sound waves that the ear picks up. Digital signals are more immune to noise, and the digital network does not induce any additional noise when amplifying signals.
Signals in digital networks are transmitted over great distances and are coded, regenerated, and decoded without degradation of quality. Repeaters amplify the signal, restore it to its original condition, and send this clean signal to the next network destination.
The Analog-to-Digital Process
Pulse code modulation (PCM) is the process of digitizing analog voice signals. Several steps are involved in converting an analog signal into PCM digital format, as shown in Figure 8-1 and described here:
■ Filtering: Filters out the signal's nonspeech frequency components. Most of the energy of spoken language ranges from approximately 300 hertz (Hz) to 3400 Hz; this is the 3100 Hz bandwidth, or range, for standard speech. Analog waveforms are put through a voice frequency filter to filter out anything greater than 4000 Hz.
■ Sampling: Samples the filtered input signal at a constant frequency, using a process called pulse amplitude modulation (PAM). This step uses the original analog signal to modulate the amplitude of a pulse train that has a constant amplitude and frequency. The filtered analog signal is sampled at twice the highest frequency of the analog input signal (4000 Hz); therefore, the signal is sampled 8000 times per second, or every 125 microseconds (Mu-sec).
Sampling Frequency
Analog speech is filtered at 4000 Hz before being sampled. The Nyquist theorem states that a signal should be sampled at a rate at least two times the input frequency to obtain a quality representation of the signal. Therefore, the input analog signal is sampled at a rate of 8000 times per second.
■ Digitizing: Digitizes the samples in preparation for transmission over a telephony network; this is the PCM process. PCM takes the PAM process one step further by encoding each analog sample using binary code words. An analog-to-digital converter is required on the source side, and a digital-to-analog converter is required on the destination side.
Figure 8-1 Analog-to-Digital Conversion Process
PCM—Pulse Code Modulation
Figure 8-1 Analog-to-Digital Conversion Process
"0010111010101
Filtered Analog Audio Signal
Sampling Stage
Filtered Analog Audio Signal
Sampling Stage
"0010111010101
Digitizing Stage
The digitizing process is further divided into the following steps:
■ Quantization and coding: A process that converts each analog sample value into a discrete value to which a unique digital code word can be assigned. As the input signal sample enters the quantization phase, it is assigned to a quantization interval. All quantization intervals are equally spaced throughout the dynamic range of the input analog signal. Each quantization interval is assigned a discrete binary code word value. The standard word size used is 8 bits, enabling 256 possible quantization intervals.
KEY POINT
Because the input analog signal is sampled 8000 times per second and each sample is assigned an 8-bit-long code word, the maximum transmission bit rate for telephony systems using PCM is 8000 samples per second * 8 bits per sample, which results in 64,000 bits per second, or 64 kilobits per second (kbps).
■ Companding: The process of first compressing an analog signal at the source and then expanding (decompressing) this signal back to its original size when it reaches its destination. (Combining the terms compressing and expanding creates the term companding.) During the companding process, input analog signal samples are compressed into logarithmic segments, and each segment is quantified and coded using uniform quantization. The compression process is logarithmic, meaning that the compression increases as the sample signals increase. In other words, larger sample signals are compressed more than smaller sample signals, thereby causing the quantization noise to increase as the sample signal increases. This results in a more accurate value for smaller-amplitude signals and a uniform signal-to-noise ratio across the input range.
Two basic variations of logarithmic companding are commonly used: The a-law companding standard is used in Europe, and Mu-law is used in North America and Japan. The methods are similar—they both use logarithmic compression to achieve linear approximations in 8-bit words—but they are not compatible.
A-law and Mu-law Companding
Following are the similarities between a-law and Mu-law companding:
■ Both are linear approximations of a logarithmic input/output relationship.
■ Both are implemented using 8-bit code words (256 levels, one for each quantization interval), resulting in a bit rate of 64 kbps.
■ Both break a dynamic range into 16 segments: eight positive and eight negative segments. Each segment is twice the length of the preceding one and uses uniform quantization within each segment.
■ Both use a similar approach to coding the 8-bit word. The first bit (the most significant bit) identifies polarity; bits 2, 3, and 4 identify the segment, and the final 4 bits quantize the segment.
The differences between a-law and Mu-law include the following:
■ Different linear approximations lead to different lengths and slopes.
■ The numerical assignment of the bit positions in the 8-bit code word to segments and quantization levels within segments is different.
■ A-law provides a greater dynamic range than Mu-law.
■ Mu-law provides better signal/distortion performance for low-level signals than a-law.
■ A-law requires 13 bits for a uniform PCM equivalent. Mu-law requires 14 bits for a uniform PCM equivalent.
■ An international connection should use a-law; Mu-law-to-a-law conversion is the responsibility of the Mu-law country.
This information was adapted from Cisco's Waveform Coding Techniques document, available at
http://www.cisco.com/warp/public/788/signalling/waveform_coding.html#subrstsix.
Time-Division Multiplexing in PSTN
Time-division multiplexing (TDM) is used in networks that are commonly deployed by telephone companies, including the PSTN. As illustrated in Figure 8-2, TDM is a digital transmission technique for simultaneously carrying multiple signals over a single trunk line by interleaving octets from each signal into different time slots.
Figure 8-2 Circuit-Switched Networks Use Time-Division Multiplexing Conversations
Silence Silence Silence
Silence Silence Silence
|
Voice |
Voice |
Voice |
|
|
Silence |
Silence |
Silence |
|
|
Voice |
Voice |
Voice |
|
|
Silence |
Silence |
Silence |
|
Multiplexing Silence
Multiplexing Silence
Silence
Voice
|
Voice |
Voice |
Silence |
The PSTN allocates a dedicated 64-kbps digital channel for each call. Although TDM cannot allocate bandwidth on demand as packet switching can, TDM's fixed-bandwidth allocation ensures that a channel is never blocked because of competition for bandwidth resources on another channel, and that performance does not degrade because of network congestion.
With time slot allocation, the number of simultaneous calls cannot exceed the number of TDM slots in the trunk. One call always allocates one TDM slot, regardless of whether silence or speech is transmitted. Time slot allocation ensures that connections always have access to a trunk, thereby resulting in low delay. However, because of the allocation method, the overall trunk utilization, also known as trunk efficiency, becomes relatively low.
The low trunk efficiency of circuit-switched networks is a major driver for the migration to unified packet-switched networks in which bandwidth is consumed only when there is traffic.
Continue reading here: PBXs and the PSTN
Was this article helpful?