Passing Voice Digitally

Pablo Oropin
BS in Computer Engineering BS in Computer Science Harding University poropin@harding.edu

ABSTRACT
The purpose of this paper is to describe how voice is passed digitally across a given network: telephone or internet.

result is an electrical representation proportional to the frequency and amplitude of the sound that caused the diaphragm to vibrate.

Categories and Subject Descriptors

E.4 [Coding and Information Theory]: Formal model of communication.

General Terms
Standardization, and Theory.

Keywords
Amplitude, Frequency, A/D Converter, Sampling, Nyquist Theorem, Noise, Bandwidth, T1 line, Bit

Figure 1. Representation of a diaphragm

2.1 Characteristics of Analog Signals That Transmit Voice

An analog signal that transmits voice has two constantly changing characteristics: the amplitude and the frequency of the signal. The first one refers to the strength of the electrical signal which varies with the loudness of the voice. Meanwhile, the frequency changes with the pitch or tone of the voice. Those two variables will be limited to humans ability to generate voice in a given range which is approximately 300 to 3300Hz in the frequency domain and up to 70dB in amplitude domain.

1. INTRODUCTION
Because passing voice by using analog signals was a problem for the network suppliers and the telcos alike, both of them turned to passing voice digitally. Digitizing voice requires the integration of many devices that go from a simple analog transmitter to a more complex analog-to-digital converter. Even though passing voice digitally requires more equipment across a given network, many advantages come from doing that such as eliminating noise when people talk, and being capable to repeat a voice signal across long distances without any problems. Thus, it is necessary to know how voice is digitized by understanding how a technique called sampling is works and how an analog-to-digital (A/D) converter employs sampling to represent voice in a digital fashion.

3. ANALOG-TO DIGITAL CONVERSION

In order to convert voice into a digital format, it has to be converted into an analog format first as explained above. Once the analog representation of the voice is obtained, the next step is to convert it to digital by using a device called an analog-to-digital (A/D) converter which employs a technique called sampling. Sampling refers to the process of measuring the value of an analog signal at regular rates called samples and assign a digital value to that sample (See Figure 2). If the samples of an analog signal that transmits voice are taken frequently enough (using an A/D converter) and those samples are played back genuinely, the ear will not be able to notice any difference from the original sound.

2. ANALOG DEVICES FOR TRANSMITTING VOICE

Before fully jumping into how an A/D converter is used to pass voice digitally, it is important to understand how voice is converted into an electrical signal by using an analog device. One of the most common analog devices to accomplish this task is a transmitter which has a diaphragm (see Figure 1) composed of loosely packed carbon particles. The sound pressure causes the diaphragm to move back and forth which at the same time causes the carbon particles to move back and forth as well which produces an electrical resistance to the charge. What is interesting about this is that such electrical resistance happens at the same amplitude and frequency as the sound pressure that caused the diaphragm to move back and forth does. In other words, the final
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Please communicate with my supervis Jon White professor at Harding University. A small fee may be incurred due to the nature. Pablo Oropin, Passing Voice Digitally Harding University

3.1 Sampling
The bandwidth of the voice signal that needs to be passed digitally is 3000Hz (3300Hz-300Hz) which is the range of frequencies that humans are able to speak. However, to establish a good sample rate to measure the analog voice signal at a frequently enough rate, the Nyquist theorem has to be use to determine such rate. This theorem states that the minimum sampling rate must be at least twice the highest frequency. Therefore, the sampling rate for an analog voice signal has to be 6600 (2*3300Hz, the highest frequency) Hz or samples / second. As a matter of fact, the official sampling rate is 8000 samples /second which is what a T1 line uses and it is a little bit higher to

what humans need in order to convert their voice into a digital format. This sampling rate is defined to address the higher range of frequencies that may occur in a conversation when people say S and F sound; furthermore, it was standardized by ANSI (American National Standard Institute). Each sample taken every 1/8000th of a second measures the amplitude level of the analog voice signal at a given time and it contains 8 bits of information where each of those bits represents a 0 or a 1. 8 bits can represent a range of values from 0 to 255(See figure 2). Thus, a total number of 256 levels are possible, which is enough to sample an analog voice signal and play it back faithfully enough that the ear will not be able to notice any difference.

employed to transform the digital signal into an audible form that the human ear can recognize. In fact, the digital format of the human voice is converted back to the analog one (See Figure 3). Figure 3. Voice is converted from analog to digital and back to analog again

4. REMARKS
To represent voice in a digital fashion still requires converting it into an analog electrical signal. Thus, the human voice is passed digitally by sampling the analog representation of the sounds at a given rate using an A/D converter. Nonetheless, all of this is done because digital technology makes easier the manipulation and combination of signals for carriers giving them more flexibility to move digital signals from source to destination (From human to human).

5. ACKNOWLEDGMENTS
Figure 2. Sampling 8000 times/second and assigning a value from 0 to 255. Obviously, the more samples are taken per second the higher the quality of the sound is when it is played back, but the problem is that the ear is not sensitive enough to notice those changes. That is the reason why voice is digitized using only 8000 samples per second. I am really thankful for the Telecommunication class taught by Dr. White, it helped me to understand better the concept of passing voice digitally

6. REFERENCES
[1] Arcomano, R. 2002. VoIP Howto. DOI= http://tldp.org/HOWTO/VoIP-HOWTO.html#toc4 . [2] Bates, Gregory, D. 2001. Voice & Data Communications Handbook. The McGraw-Hill, California, 53-90. [3] Sheldon, T. 2001. ADC (Analog to Digital Conversion). DOI=http://www.linktionary.com/a/adc.html. [4] Torres, G. 2007. How a Analog-to-Digital Converter(ADC) works. DOI=http://www.hardwaresecrets.com/article/317. [5] Voice Over IP. Wikipedia 2007. DOI= http://en.wikipedia.org/wiki/Voice_over_IP

3.2 Digital-to-Analog
Every bit obtained from sampling done by the A/D converter is in the form of a square wave as opposed to the analog voice signal which is a continuously sinusoidal signal. The square wave signal from the bits travels in a pair of wires across the network where other devices such as repeaters are found to strength the digital signal. Ultimately, the digital samples of the voice are played back at the destination point where an analog-to-digital converter is