A View from Dresden onto the History of Speech Communication
Part 8: Collectibles from the Electronic Era of Speech Technology
We have shown in this series about the HAPS that there are numerous exhibits from the era, where research tools were restricted to mechanic constructions. The situation changed when electronic devices arrived at the laboratories of the phoneticians and engineers. Nevertheless, many devices of that time are interesting for us, because they had to work without computer control, which frequently required very special solutions.
A typical example is spectral analysis, which is nowadays performed in real time with virtually low effort. Without computer control, a very sophisticated hardware was required to produce spectrograms by blackening special prepared paper, which was fitted to a revolving drum (Figure 1). The U.S. company “Kay electric” produced these devices as “Sonagraphs” since the 1950s. They were improved step by step following the availability of more advanced components. The HAPS includes different sonagraphs from the early “Model R” over several decades, which vividly demonstrate their development.
Turning to an example of early electronic speech synthesis now, we have to stress the fact that the idea of the so-called concatenative synthesis, which received practical importance in the 1990s, is surprisingly not new. It emerged with the invention of magnetic storage of audio signals. Single sounds which were naturally spoken could be stored and re-ordered into a new sequence. The synthesizer “Lora” from the former FTZ Darmstadt (now a unit of the German Telekom) is an early example. It consisted of a stapled series of storage elements like that in Figure 2. The different elements were equipped with pieces of magnetic tape storing the particular sounds. All elements were arranged in parallel, and the selection of the proper element was controlled in a complicated way using a camshaft. The main problem, however, was the production of naturally sounding sound transitions. It is reported that the transitions were implemented using the Schwa as intermediate sound.
Despite of the existence of this early example of speech synthesis in time domain, parametric approaches have been predominant for several decades. In Part 7 of this series, we presented two early demonstrators from that era. Formant synthesis was the most important synthesis method for about 40 years. In this rather long period, the available active components changed from electron valves over discrete semiconductors to integrated circuits with growing complexity. This development influenced, in turn, the evolution of the computer from huge machines in separate centres to process computers in the laboratories, where they could be applied to control the synthesizers immediately. Further miniaturization enabled introducing speech synthesis as an embedded technology.
The HAPS is able to demonstrate this dependency of the speech synthesizers from the electronic components as well as the available processors by means of different synthesis devices, which were developed under the guidance of Walter Tscheschner. Figure 3 is summarizing the different prototypes, which can be traced as exhibits of the collection and/or as technical documents and audio files.
Today, most of the products in speech synthesis and recognition are pure software-based solutions. But research is going on, and we look forward to new ideas, which will leave their traces in collections like the HAPS for future generations.
Photographs Copyright TU Dresden / HAPS