A brief history of cold digital sound emitters or why we still use analog speakers

8 min

Over the course of the century, dynamic loudspeakers have been the most popular type of sound emitters. Traditional analog speakers are used everywhere. They remain the last analog device in the sound reproduction path familiar to the modern person. But if archaeologists of some civilization of the distant future had discovered analog dynamic loudspeakers through, they would probably have wondered why their ancestors needed such illogical heating devices. The speaker turns most of the energy into heat and this is not its only problem.

At the same time, digital emitters of various types have been produced in limited quantities for a long time. The latter are little known to a wide range of consumers, roads and are used relatively rarely. Further, a brief history of digital sound emitters, the devices in which they were used and are used, as well as considerations about their prospects.

Prerequisites for the emergence

Since the mid-1920s, undivided dominance in electroacoustics has remained with the electrodynamic loudspeaker, in its various variations. Neither electrostats, which at first burned coolly, disrupting the sessions of the first sound films in the 30s, and then became simply fabulously expensive, could not displace it. No ionophones that are not capable of adequate bass reproduction. Neither piezoelectric emitters that could not withstand competition due to the small frequency range.

burnt subwoofer speaker coil

At the same time, the dynamics can hardly be called a technically perfect solution. So, for tweeters, the coil temperature of 100 degrees Celsius is not a limit, the efficiency for this reason rarely exceeds 1%, and the coil temperature of the bass drivers of dynamic drivers can easily exceed 150 and even 200 degrees when working at rated power. Distortions, both frequency and non-linear, as a rule, leave much to be desired and require correction or technologies that will significantly reduce them. A similar story occurs with a transition characteristic, which in expensive solutions makes it constantly chase a large frequency range, which ideally should go much further than the spectrum heard by the human ear.

But, despite all the shortcomings of the speaker, it was he who became the most demanded in the aggregate of advantages. At the same time, tireless researchers did not stop looking for something more productive, energy efficient, and also more manageable. Engineers began to look for a way to convert a digital signal into sound directly, without using a DAC.

Bell Labs acoustic experiments in the 1920s

Theoretically, digital speakers were described by Bell Labs in the 1920s. Their principle was simple enough. The least significant bit controls the speaker, in which the value “1” drives it with a maximum amplitude, the value “0” completely stops the signal. Further, the least significant bit doubled the initial radiation area, the next bit doubled its area, etc., in accordance with the number of bits. In the 1920s there was no urgent need for this type of conversion of digital signals into sound, and theoretically, the works laid the foundation for many years.

Bell Speaker

In earlier versions, the radiation area of ​​the next bit was located concentrically around the segment of the previous bit, but this rule is not mandatory. Theory was first put into practice in 1980. The developer also became the company Bell Lab. It was a disk-shaped electrode on which a thin film membrane was fixed. The electrode was divided into isolated segments, with the area ratio described above, by the number of discharges 4.3, 2.1.0. The segments were excited by a rectangular digital signal in accordance with its value.

For telephone communications fidelity of reproduction was enough, but this radiator was unsuitable for reproduction of music. The fact is that for speakers to obtain sufficient volume, the area of ​​the corresponding emitter was unacceptably large. Also a problem was the conversion distortion, which in classic DACs can be eliminated using filters. But in digital emitters, their use is impossible, since the conversion takes place directly and they are the final link in the reproduction.

Japanese experiments

The next step in the development of digital sound emitters was the creation of electret and piezoelectric digital speakers by SONY. The principle of operation was not very different from that used at Bell Lab, but the design was different. The electrodes of such emitters were concentric sections with equal area. Sections were connected in groups, the number of groups depended on the bit depth of the emitter.

A fundamentally different method of dividing the sections of a digital loudspeaker was proposed by engineers at Matsushita Electric Corporation (today Panasonic Corporation). In the patents, and today owned by the company, it is proposed to combine the segments that emit sound in groups, in accordance with the weight coefficient of the discharge.

None of the developments described in the section has been developed in view of production costs, high distortions, low manufacturability, and other specific problems of newborn technology.

Digital speakers

Attempts to create an electrodynamic digital emitter began almost immediately after the appearance of piezo and electret speakers of this type. The problems of the latter consisted in a narrow frequency range and a kind of frequency response, which did not allow them to be used effectively anywhere, except for communication devices for voice and HF sections of speakers.

Philips patent drawing

Philips and Sony began experiments to create a digital speaker back in 1982. The principle was that the number of coils in the emitter increases, the number of sections in this case corresponds to the capacity. The result was a Philips patent No. 4612420 (https://patents.justia.com/patent/4612420), shortly before that, No. 58-31699 was registered in Japan, demonstrating a similar design of a digital speaker.
We can assume that a digital speaker with a multi-link coil was one of the most long-lived versions of a digital emitter. The last mention of a similar development dates back to the 2000s when a similar principle was applied by B&W, the flagship Audiophile development.

University piezo emitters

In addition to corporations creating electronics, the topic of digital emitter was actively developed at universities. A group of scientists from Shinzu University in Nagano in the 1990s focused on creating piezoelectric digital speakers. They got their first result in 1993, and by 1999 they showed a radiator designed for a 16-bit signal with a sampling frequency of 48 kHz.
We can say that this development was the first digital emitter, the characteristics of which were sufficient for limited multimedia use. The characteristics of the device were as follows:

  • Frequency range: 40-10000 Hz;
  • Frequency response uneven within 4dB.
  • THD 3.5% at 50 Hz and 0.1% at 10000 Hz
  • Sensitivity 84 dB

Quantization noise and other artifacts of this type of digital-to-analog conversion associated with low bit depth in such emitters were strong enough to speak of any high fidelity of reproduction. It was obvious that loudspeakers of this type can be used in multimedia devices only to a limited extent, mainly for communication and sound alerts, but not for high-quality music playback.

Brighton lattice or Helsinki algorithm

The notorious British scientists have applied a fundamentally new principle. A group of researchers from Brighton University, with the financial support of B&W, developed an AS in which they did not try to shove a digital emitter into one housing, but presented it as a distributed array of many separate dynamic emitters, which were grouped in accordance with the discharge of the signal. Thus, two directions were opened for the development of digital speakers. The first is an increase in the quantization bit depth, which made it possible to reduce noise, and the second is a signal correction to compensate for distortions of dynamic (or other) emitters.

The creation of a new type of digital emitter aroused keen interest in the academic community. As a result, the Finnish company Audio Signal Processing Espoo and the University of Helsinki have created an algorithm that optimizes the operation of the Brighton sectional grill. The algorithm made it possible to equalize the phase and amplitude in the entire spectrum of reproducible frequencies. The algorithm also appeared in the year 2000.

The Digital Sound Projector

The developments described above were used by 1..limited to create The Digital Sound Projector, a device that was introduced in 2002. We can say that this is the first full-fledged product in the history of electroacoustics that uses a digital emitter to play music with high fidelity.

The creation of The Digital Sound Projector was attended by manufacturers of microprocessors ARM Ltd, an interdisciplinary research company Cambridge Display Technology, a chip manufacturer Analog Devices. Later, the small-scale release of the product was continued by Pioneer.

The device used 256 small emitters, each of which reproduced a single pulse. Like the pixels on the monitor, the system added up the big picture of many signals. The processor, in accordance with the Finnish algorithm, controlled the playback parameters and carried out noise elimination and distortion compensation. The compensation process took into account both decoding artifacts and the interference of waves from various emitters.

One of the significant achievements was the efficiency, which reached 10%, which significantly exceeded the value of the classic analog speakers. The principle of distributed controlled digital radiation has also significantly reduced harmonic and intermodulation distortion. Perhaps the most significant and obvious drawback of the system was its complexity, low manufacturability, and, consequently, high cost. At the beginning of the 2000s, the world was not ready to accept something so complex and, obviously, it was not ready to accept so far. Tangible problems in the form of complexity and cost did not make the technology of mass gratings and buried it in the cemetery of ideas that did not shoot.

Modern stage of development

Despite the obvious difficulties, the technology of digital radiation has received an unexpected development. So in 2015, it was announced the creation of a MEMS emitter, which is based on the complementary metal-oxide-semiconductor (CMOS) structure. We are used to MEMS microphones and MEMS accelerometers, the turn of loudspeakers has come.

Audio Pixels announced the creation of MEMS emitters, which said it was close to creating digital emitters that can outperform analog speakers. Limiters are the small amplitude, as well as the limitation of the low-frequency range that most innovators encounter in the field of sound emitters.

Another example of the use of digital emitters is the Audio-Technica ATH-DSR9BT headphones, which are devoid of the usual DAC and are equipped with Pure Digital Drive digital speakers. The manufacturer does not disclose the essence of the technology in detail, however, judging by the available information, this is the reincarnation of a digital speaker with many coils, however, unlike Philips radiators of the mid-80s, Pure Digital Drive operates with a multi-bit signal.

How I solved the problems of ultrasonic radiation, quantization noise, as well as the correction of distortions introduced by the mechanical parts of the device, I do not know. But judging by the fact that the device is positioned as the company’s flagship wireless, there is a chance that the solution is effective. It is also known that the speaker was created in partnership with Trigence Semiconductor.

Warm analog near future

I’ll try to play wanga with my grandmother and summarize all of the above. The hope of digital radiation is MEMS, but it has formidable physical limitations, which means that it will limit their use to a predominantly wearable form factor. Another problem is the speed of development of MEMS technologies that make plans, as joking among developers, in the “dog years”, i.e. where other industries conditionally need a year, MEMS will take seven years.

Another issue is cost. And until manufacturability grows, the cost does not decrease, and it does not grow rapidly due to the already mentioned MEMS development speed. The simplicity and reticence of the production of speakers was so fond of manufacturers that in order to change them for something, very good arguments are needed, and increasing the efficiency is clearly not one of them. Therefore, supporters of techno-archaic and other analog audiophile “steampunk” can not worry. Tube amplifiers, of course, will not return after the resurrected vinyl, but warm and even hot (in the literal sense) true-analog speakers will live another ten or another years. Unfortunately, the fate of digital loudspeakers today is still relatively expensive rare experimental products and scientific research.


Leave a Reply