Magazine Archive

Home -> Magazines -> Issues -> Articles in this issue -> View

Hi-Fi

Something Old? Something New?


Sony have taken out a patent on a speaker system. Nothing new in that you may think but there you'd be wrong! This speaker is the first of a new generation of 'digital' speakers.

To understand how it works though, we have to understand how audio signals can be digitally encoded.

Sound consists of a pressure wave that travels in air. If a suitable transducer, a microphone, picks these signals up, its output will consist of an electrical voltage whose amplitude will rise and fall in sympathy with the pressure wave (the sound) that impinges on it. If this signal is examined on an oscilloscope, a complex ever-changing voltage level will be seen. Conventional audio equipment operates with amplified levels of this original voltage.

There is, however, another way in which this signal can be encoded. Before describing this in detail it will be as well to define just what we are trying to encode. Sound levels in a live concert can vary by a ratio of 10,000:1 in amplitude. This is known as the signal's dynamic range.

The frequency range that must be covered extends from 20Hz to 20kHz, a thousand to one ratio. Digital recording works by sampling the waveform at a very high frequency and encoding its instantaneous amplitude.

Now computers cannot count in tens like us, they only know two numbers, 0 and 1. 1 represents a high voltage level and 0 an absence of voltage.

At first sight this seems extremely limiting but in fact a computer can count quite easily with just these two levels. In our conventional number system (decimal) large numbers are represented by thousands, hundreds, tens and units, i.e. powers of ten. The computer equivalent uses powers of two (binary). For example, consider the computer number 1001100. Note that only 0 and 1 appear. The key to translating this back to a decimal number is shown below:

26 25 24 23 22 21 20
64 32 16 8 4 2 1
1 0 0 1 1 0 0

Note that the number is read from left to right and contains 1-64, no 32's or 16's, one 8, one 4, no twos and no 1's. By adding up what we have we find that the number is 64+8+4=76. Notice that instead of tens, hundreds, etc. (i.e. powers of ten) each column advances by a power of two.

Our music signal has a dynamic range of 10,000:1 and to encode any of the possible levels between these limits we must be able to count up to 214, 16,384. Now each of our columns contains a bit of information which is either 0 or 1, so we can say that each ot our samples requires 16,384 bits of information. Having settled this, all that has to be established is the sampling rate.

Luckily this can be readily determined because of the work of a gentleman named Nyquist who showed that the maximum frequency that can be completely recovered when modulated on a carrier is exactly half that of the carrier. So to encode our highest frequency audio signal we need a carrier of 40kHz minimum.

Our digitally encoded signal then consists of a 14-bit binary code that changes 40,000 times per second. The most successful commercial digital recording equipment, the Soundstream, uses a 16-bit code and a sampling rate of 50kHz.

Once in digital form, the program has to be translated back to analogue in order for it to be heard, which raises the natural question 'Why bother in the first place?' The answer is simple. For a start, once the signal has been recorded in digital form, it won't deteriorate in any way. No compression is required since the full dynamic range can be accommodated; and last, but by no means least, the S/N ratio is greatly improved.

Unfortunately the problem of converting the digital code back to analogue form remains. At present this is done with D/A converter chips and the resulting signal is then recorded on a conventional analogue disc. All this brings us back to where we started with Sony's new speaker.

The innovative idea behind this system is to dispense with moving coil speakers and conventional amplifiers and replace them by compressed air drivers. Basically a tank of compressed air is fed via electromechanical valves to 'Horns' which vent into the listening room. There are 14 of these horns and the control for the air flow valves is obtained directly from the incoming digital 'word'.

When a logical one is applied to the valve it opens for the duration of the pulse allowing air to enter the horn. If things are arranged so that the air pressure pulse produced at the horn connected to the most significant digit is 214 times stronger than that at the least significant digit, then the combined pressure valve produced by the horns will be an exact replica of the input before it was digitally converted. In other words, the whole system operates as a powerful digital to analogue converter without the need for a separate amplifier.

A fascinating idea and one which we will all no doubt hear more of in years to come. It is not altogether new though.

At the turn of the century, a mechanical amplifier using compressed air was invented, which worked on similar principles. The 'drive unit' consisted of a slotted, fixed plate with a comb which was free to slide across it. Compressed air was blown through the plate and the comb was mechanically linked to the record stylus. As the stylus moved the comb across the fixed plate, the air pressure varied in sympathy with the groove modulations producing sound. The device was coupled to an acoustic horn to provide amplification.

Commercial versions of this equipment were on sale in 1906 and for reasons that are somewhat obscure, it was named the 'Autexa-phone'.

A commentator of the time reported that the sound could be heard for two or three miles in calm weather which must have meant a hefty SPL was being generated.

The analogy to an electronic amplifier can be more readily appreciated if one remembers just how these work. A conventional amplifying device modulates an external power source, a DC supply. This modulation is controlled by the input signal. Thus an amplifier functions as a copying device. In the air driven speaker, the pressurised air is analogous to the DC supply. The input signal is the mechanical movement of the comb and the output is modulated air pressure changes (sound).

Hi-Fi, Lo-Cost



There are several ways one can acquire a decent hi-fi system. The most obvious (and expensive) is to trail around hi-fi stores and, after listening to various combinations, part with your well earned cash. Another way, is to assemble the projects featured in this and other magazines. A third, very much underestimated method, is to assemble a stereo system from ready made modules, many of which are excellent in value and performance.

One such module I have recently come into contact with is an interesting example of this approach. Bi-Pak have established an excellent reputation over the years for the reliability and performance of their amplifier modules so it was with great interest that I received the new S453 FM tuner module.

Bi-Pak FM main board and tuning panel.


Like it's predecessor, the S450, this module consists of two PCBs, one of which holds the guts of the tuner, and the other the four 100k multiturn presets which control the varicap tuning.

Also in common with the S450, tuning is accomplished by means of four pushbuttons. Although this means that only four FM stations can be selected, this isn't a major drawback. For those who wish to experiment it is not too difficult a problem to add further stations by using a separate switch bank.

One of the great things about this tuner is its small size, the large PCB is only 125 x 80mm. The tuning PCB is 45 x 80mm. Adding the unit to an existing installation is simple since the unit requires a supply voltage between 18-25V and has a current consumption of 45mA. In many cases a suitable supply rail will already be available and the tuner can then be run via a simple network consisting of a dropper resistor and decoupling capacitor.

So much for the mechanics. Now, how does it sound? To find out I built the module together with a simple PSU into a Verocase in which it was a fairly tight fit. After the usual, debugging exercise, in my case the elimination of an earth loop, I was most impressed with the performance. While it is not the most sensitive tuner that I have ever tested, adequate stereo reception of all the major stations with good signal to noise ratio was obtained. Stereo separation was extremely good (Bi-Pak quote 30dB). What is more important though is a tuner's ability to produce a good stereo sound stage with the subtle depth of effects intact. This the tuner can do. Technically the unit employs a discrete front end with an IC limit/demodulator stage. Stereo decoding is carried out by the ubiquitous PLL IC. Stereo operation incidentally can be bypassed by fitting an external switch.

In conclusion, a fine product which can be recommended for those who wish to obtain a good tuner and are prepared to do some mechanical work themselves. The price is £19.53 exclusive of VAT and should be available during August from Bi-Pak Limited.



Previous Article in this issue

Organ Talk

Next article in this issue

Advanced Music Synthesis


Electronics & Music Maker - Copyright: Music Maker Publications (UK), Future Publishing.

 

Electronics & Music Maker - Jul 1981

Feature by Jeff Macaulay

Previous article in this issue:

> Organ Talk

Next article in this issue:

> Advanced Music Synthesis


Help Support The Things You Love

mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.

If you value this resource, you can support this project - it really helps!

Please Contribute to mu:zines by supplying magazines, scanning or donating funds. Thanks!

We currently are running with a balance of £100+, with total outgoings so far of £1,036.00. More details...
muzines_logo_02

Small Print

Terms of usePrivacy