Wolfgang von Kempelen's Speaking Machine (1769)

From MSc Voice Technology
Jump to navigation Jump to search

Introduction[edit | edit source]

In 1769, Hungarian inventor Wolfgang von Kempelen began developing a mechanical device capable of mimicking human speech, known as the "Speaking Machine"[1]. In 1791, Kempelen shared his insights and the inner workings of his "Speaking Machine" in a book titled "Mechanismus der Menschlichen Sprache" ("Mechanism of Human Speech and Language")[2]. His Speaking Machine consisted of a bellows to simulate the lungs, a reed to act as vocal folds, and a leather tube representing the vocal tract (including the mouth, tongue, throat, and other structures involved in speech production). By manipulating the shape of the tube, a variety of vowel sounds could be produced. Consonants were generated through the use of keys that controlled openings analogous to the lips, tongue, and palate.[3]


The Speaking Machine represented the first complete physical modeling of the human vocal system capable of synthesizing speech sounds.[4] Wolfgang von Kempelen’s pioneering work played a foundational role in the early exploration of artificial speech production. Beyond its significance in understanding vocal mechanisms, the Speaking Machine serves as an exemplar of human ingenuity in addressing complex challenges, even with the technological limitations of its time.[2] Sir Charles Wheatstone resurrected the work of Wolfgang von Kempelen, creating an improved replica of his Speaking Machine in 1837, inspiring Alexander Graham Bell to build his own speaking machine, which led to his invention of the telephone in 1876.

Historical Context[edit | edit source]

In the 18th century, there was a flourishing of mechanical engineering and automation, marking the era of automata, which also included advancements in speech automation technology. [5]Around 1780, Wolfgang von Kempelen produced the first ever fully functional mechanical speech synthesizer, known as the speaking machine. During this time, he became interested in the issue of speech among the deaf-mute community. More fundamentally, the prevailing importance and significance of speech motivated his work, reflecting the pervasive focus on this aspect of human communication during that period.[6]


In the century preceding von Kempelen, linguists, physicists, psychologists, and speech teachers engaged in extensive speculation and exploration of the mechanics of speech.[6]Meanwhile, the knowledge of sound, acoustics, and phonetics was relatively limited, so rather than basing his speaking machine on acoustic theories, Kempelen opted for an engineering approach of analysis-by-synthesis or trial and error. He aimed to achieve an audible result through a simple mechanism closely resembling our articulatory apparatus (the various parts of the human speech mechanism, including the lips, tongue, vocal cords, and so on), yet playable like a musical instrument.[7]


However, the sounds produced by von Kempelen's synthesizer failed to meet the expectations of the 20th-century human ear, given the advancement in technology leading to the accustomed perfection or near-perfection in the reproduction of the human voice. Nevertheless, von Kempelen's contemporaries, unaccustomed to such technological advancements, were satisfied with the speaking machine, which adequately met their needs and expectations.[8]This era marked the initial explorations in speech technology, laying the groundwork for the subsequent development of sound synthesis and speech technologies.

Key Innovations[edit | edit source]

The historic invention of Kempelen's Speaking Machine represents an early attempt and starting point for speech synthesis technology. Although this technology is considered outdated by today's standards and cannot directly influence the modern field of speech synthesis, it holds significant historical importance within the context of the 18th century. It marks the early exploration of speech synthesis, providing a foundation for the field's development.

1. Mechanical simulation:[edit | edit source]

Kempelen's Speaking Machine garnered significant attention in its time because it was a mechanical device capable of mimicking human speech, which was considered a notable technological achievement in the 18th century. Although the technical principles of this machine significantly differ from modern speech synthesis technology, it served as an inspiration for subsequent research in speech synthesis and marked the early exploration of mechanical simulation of speech.[9]

2.Artificial Vocal Organ Simulation:[edit | edit source]

Kempelen's Speaking Machine included the simulation of human vocal cords, tongue, lips, and other sound-producing organs with mechanical components. While Kempelen was not the pioneer in this field, his Speaking Machine incorporated a more complex mechanical structure, which allowed for more precise simulations. Additionally, Kempelen's machine had the capability to control sound through mechanical devices, rather than merely passively imitating sounds. This was a significant technological achievement given the technological standards of the time.[10]

3.Multilingual support:[edit | edit source]

Multilingual support in Kempelen's Speaking Machine was a significant innovation with important historical implications. The machine possessed the capability to simulate different languages and dialects, which meant it could produce various sounds and pronunciations, not limited to a specific language. This was considered a technological achievement in the 18th century as it aimed to replicate the production of multiple speech sounds, thereby opening possibilities for research into different language communities and dialects.[11]

Impact[edit | edit source]

Von Kempelen's travail was not for naught and his impact far from nil.[4] Kempelen's speaking machine together with the book "Mechanismus der menschlichen Sprache" ("Mechanism of Human Speech and Language") he published in 1791 in which he described his observations on human speech production and his experiments with his speaking machine had a profound and multifaceted impact on the development of new ideas in linguistic research and advancements in audio engineering.

1. Contributions to linguistic research:[edit | edit source]

While before Von Kempelen the larynx was considered central to speech production, his simple and successful demonstration drew the attention of 19th-century scientists to the vocal tract, the cavity between the glottis and the lips, as the main site of acoustic articulation.[4] In his book, he classifies the vowels according to her width of the lip channel giving a ranking of A >E>I>0>U and the width of the so called tongue channel that can be interpreted as horizontal tongue position. Although Kempelen isn't very explicit here, the observation clearly resembles the perceptual analysis of the second formant in whispered vowels described a century before by Reyher (1679) and the vowel tunes of von Helmholtz (1862). [12]This innovative vowel classification (classifying vowels based on characteristics like the width of the lip channel and tongue channel) represented a novel approach to the study of speech sound physical properties.

2. Development of Speech Synthesis and Audio Engineering:[edit | edit source]

Development of Speech Synthesis: Kempelen's research findings represented a crucial step in the development of speech synthesis. His insights into the resonating properties of the vocal tract and the impact of tube width on vowel production provided the foundational knowledge that subsequent researchers needed to simulate human speech mechanically.

Influence on Later Innovators: Later innovators built upon Kempelen's ideas, gained insights into resonating properties and developed theories of multiple resonance (the amplification or reinforcement of specific sounds produced by the vocal tract). Christian Kratzenstein, who made notable contributions to the study of speech and acoustics in 1779 and made early attemps at mechanical speech synthesis, built upon the groundwork laid by Kempelen. In 1830, it was Willis, starting from the ideas of Kratzenstein and von Kempelen, who first gained resonable insight in the resonationg propgerties of neutral tubes that would be able to give the illusion of different vowels. And in 1838 Charles Wheatstone constructed his famous version of von Kempelen's speaking machine. It was a bit more complicated and was capable to produce vowels and most of the consonant sounds. Some sound combinations and even full words were also possible to produce.[13]Inspirations in research and experiments with mechanical and semi-electrical analogs of vocal system in speech synthesis technologies were directly influenced by Kempelen's early work, making it a key milestone in the history of speech technology.

Kempelen's Speaking Machine represented an early mechanical approach to imitating speech sounds, and a few decades later the Voder from the Bell Telephone Laboratory was a pioneering electronic device that introduced new possibilities in generating speech-like sounds. Kempelen's Speaking Machine and the Bell Telephone Laboratory's Voder were foundational in the history of speech synthesis, paving the way for more advanced and realistic speech synthesis technologies that followed in the decades and centuries thereafter.

Future Research[edit | edit source]

Although Kempelen's speaking machine did not achieve fully realistic artificial speech synthesis, it provided significant inspiration for subsequent developments in speech synthesis technology. Over the years, academia has produced numerous research achievements, yet there remain several areas warranting further exploration:

1. Studies on the Reproduction and Invention Process of the Machine:[edit | edit source]

While Kempelen's work "Mechanism of Human Speech" appears to provide a fairly detailed description of the appearance, illustrations, and the quality of synthesized speech of his final version speaking machine, there are numerous ambiguities and inaccuracies in these descriptions, casting doubt on their ability to faithfully replicate the machine's original form.

For instance, Kempelen's use of the Latin alphabet sometimes results in unclear associations between letters and phonetic qualities. For example, the distinction between whether the letter "O" represents a semi-open or semi-closed rounded back vowel remains unclear. Additionally, the reason behind the absence of front vowels between "E" and "A" is ambiguous. Kempelen's use of "CH" for palatal, velar, and possibly uvular voiceless fricatives ("higher" and "lower Ch") further obscures the specific "CH" phonetic qualities that the speaking machine can synthesize. Moreover, Kempelen's claim that opening only one "nostril" of the machine can synthesize an /n/ phoneme, although intriguing, lacks conclusive evidence.

And for the evolution of Kempelen's invention process. The question of why Kempelen modified the machine's construction after publicly demonstrating a seemingly convincing version, and whether these modifications aimed to simplify the machine to align with his work "Mechanism of Human Speech," remains unanswered.

Besides that, Kempelen introduced a "membranous pipe model" in his work "Mechanism of Human Speech," suggesting that it could simulate human vocal cords. However, he chose a distinctly unnatural model for the machine's glottis. This decision presents a puzzling question for future scholars to delve into.[14]

2. Optimizing and Improving Speech Machine Technology:[edit | edit source]

Kempelen demonstrated great attention to detail when it came to addressing secondary issues, but he approached the core issue of throat dimensions and related apertures in a relatively casual and superficial manner. Future researchers may choose to conduct experiments in this particular area.

Furthermore, an intriguing avenue of research involves exploring the use of double reeds as a replacement for the vocal cords, as Kempelen attempted in his initial prototype. These double reeds are closer to simulating human vocal cords, and researchers can investigate the feasibility of this approach and whether it significantly alters the machine's "tonal quality." Additionally, studying the impact of different reed materials on the machine's speech sounds presents an engaging research area.

Lastly, researchers can draw inspiration from Kempelen's work to delve further into the relationship between human voice production and organ construction. In Kempelen's research, he proposed the use of the organ register "vox humana" as the foundation for a speech synthesizer. This is why he solely employed reeds to mimic vocal cords, akin to the process of crafting an organ. This register, in conjunction with the so-called "tremulant," can generate sounds resembling the vibrato in human singing.[15]

These issues present intriguing avenues for future research, offering the potential for a deeper understanding and development of Kempelen's speaking machine.

References[edit | edit source]

  1. Tarnóczy, T. H. “The Speaking Machine of Wolfgang von Kempelen.” The Journal of the Acoustical Society of America 21, no. 4_Supplement (July 1, 1949): 461–461. https://doi.org/10.1121/1.1917078.
  2. 2.0 2.1 Trouvain, Jürgen, and Fabian Brackhane. "Wolfgang von Kempelen’s speaking machine as an instrument for demonstration and research." In Proceedings of the 17th International. Congress of Phonetic Sciences. 17-21 August 2011, Hong Kong, pp. 164-167. 2011.
  3. Brackhane, Fabian, Jürgen Trouvain, and Richard W Sproat. “KEMPELEN Mechanismus der menschlichen Sprache,” n.d.
  4. 4.0 4.1 4.2 Schroeder, Manfred R. “A Brief History of Synthetic Speech.” Speech Communication 13, no. 1–2 (October 1993): 231–37. https://doi.org/10.1016/0167-6393(93)90074-U.
  5. Trouvain J, Brackhane F. The relevance of today Wolfgang von Kempelen's speaking machine[J]. Phonus. Research Reports of the Institute of Phonetics at the University of the Saarland, 2011, 16: 149-166.
  6. 6.0 6.1 Dudley H, Tarnoczy T H. The speaking machine of Wolfgang von Kempelen[J]. The Journal of the Acoustical Society of America, 1950, 22(2): 151-166.
  7. Pompino-Marschall B. Von Kempelen et al.: remarks on the history of articulatory-acoustic modelling[J]. ZAS Papers in Linguistics, 2005, 40: 145-159.
  8. van den Broecke M P R. Wolfgang von Kempelen's speaking machine as a performer[J]. Sound Structures. Studies for Antonie Cohen. Dordrecht: Foris, 1983: 9-19.
  9. Grassegger, Hans. “Von Kempelen and the Physiology of Speech Production,” n.d.[1]
  10. Trouvain, Jürgen, and Fabian Brackhane. “THE RELEVANCE TODAY OF WOLFGANG VON KEMPELEN’S ‘SPEAKING MACHINE,’” n.d. [2]
  11. Liénard, Jean-Sylvain. “FROM SPEAKING MACHINES TO SPEECH SYNTHESIS,” n.d.[3]
  12. Pompino-Marschall, Bernd. “Von Kempelen’s Contribution to the Theory of Acoustic Articulation,” n.d.
  13. Lemmetty, Sami. “Review of Speech Synthesis Technology,” n.d.
  14. Brackhane, Fabian. “DIE SPRECHMASCHINE WOLFGANG VON KEMPELENS – VON DEN ORIGINALEN BIS ZU DEN NACHBAUTEN,” n.d.
  15. Trouvain, Jürgen, and Fabian Brackhane. “THE RELEVANCE TODAY OF WOLFGANG VON KEMPELEN’S ‘SPEAKING MACHINE,’” n.d.

LLM Review[edit | edit source]

We use the following prompts to ask different GPTs to give us feedback for the Wikipedia page:

Prompt 1 on ChatGPT3.5:[edit | edit source]

Suppose you are a professor in a master's program in speech technology. Now your students have edited a Wikipedia interface about the Wolfgang von Kempelen's Speaking Machine. This assignment mainly tests the students' understanding of the history of speech Synthesis technology and their conceptual understanding, ability, information retrieval ability, logical thinking ability and resource integration ability, Please give feedback on this assignment from the following aspects:

1. Evaluate this assignment in terms of page structure, clarity of concept explanation, depth and breadth of content involved, language fluency, etc.

2. Correct errors in conceptual terminology, historical facts, format, etc.

3. Provide further opinions on the improvement of the assignment, such as whether the structure needs to be adjusted, whether the content needs to be supplemented, whether the expression needs to be more refined, etc.


Here is the feedback:

  1. Page Structure and Clarity of Concept Explanation: The assignment starts with a well-structured introduction that provides a clear overview of Wolfgang von Kempelen's Speaking Machine and its historical significance. The content is logically organized, with headings like "Historical Context," "Key Innovations," "Impact," and "Future Research," making it easy to follow. The language used is clear and concise, which enhances the overall readability. The use of headings and bullet points helps break down complex information into digestible sections.
  2. Depth and Breadth of Content: The assignment covers a wide range of topics related to Wolfgang von Kempelen's Speaking Machine, including its historical context, key innovations, and impact on linguistics and speech technology. It delves into details about how the machine worked and its significance in its historical context. The inclusion of historical context and comparisons to modern speech synthesis technology provides a comprehensive view of the subject matter.
  3. Language Fluency: Overall, the language used in the assignment is fluent and well-written. It effectively conveys complex concepts in a clear and concise manner. However, there are a few minor grammatical and punctuation issues that can be addressed to enhance fluency further.
  4. Correct Errors in Terminology and Facts: The assignment provides a generally accurate account of Wolfgang von Kempelen's Speaking Machine and its historical context. Still, there are some minor issues with terminology and historical accuracy that should be corrected:
    1. In the introduction, "Sir Charles Wheatstone improved Wolfgang von Kempelen's Speaking Machine in 1837, inspiring Alexander Graham Bell to build his own speaking machine" might be misleading. Charles Wheatstone did work on speech synthesis devices, but it's not entirely accurate to say he "improved" Kempelen's machine, and there is no direct link between his work and Alexander Graham Bell's invention of the telephone.
    2. The section on Kempelen's classification of vowels mentions a ranking (A > E > I > O > U) that may require clarification. It's not entirely clear how this relates to modern phonetics or what it represents in terms of vowel sounds.
  5. Suggestions for Improvement:
    1. Correct the errors mentioned above to ensure the assignment's accuracy.
    2. Consider providing more information about Charles Wheatstone's work in speech synthesis and its role in the development of the field.
    3. The assignment mentions that Kempelen's machine could simulate different languages and dialects. It would be helpful to provide more details or examples of how this multilingual support worked in practice.
    4. The assignment could benefit from citations and references to specific sources, especially when making claims about historical facts and technological details. This would enhance the credibility of the information presented.
    5. While the "Future Research" section provides interesting areas for further study, it could be more concise and focused. The assignment might benefit from summarizing the potential research areas more succinctly.

Overall, this assignment provides a comprehensive and informative overview of Wolfgang von Kempelen's Speaking Machine. With some minor corrections and enhancements, it can serve as an excellent educational resource for students in a master's program in speech technology.

Prompt 2 on Claude.ai:[edit | edit source]

As a group of master's students, we're collaborating on a project to create a wiki page about Wolfgang von Kempelen's Speaking Machine (1769) in the field of speech synthesis. Wikipedia pages focused on specific topics, especially those related to science, should be easily comprehensible to the general public and not laden with technical jargon. Please help to identify any terms or phrases that may require further explanation to ensure that our content is accessible to a broad readership.


Here is the feedback:

Your introduction and explanation of Wolfgang von Kempelen's "Speaking Machine" and its historical context are well-structured and provide a clear understanding of the topic. However, you've included some terminology and concepts that may need further explanation for general readers:

  1. Resonance: When discussing Kempelen's insights into the resonating properties of the vocal tract, you might want to briefly explain what resonance is. In this context, resonance refers to the amplification or reinforcement of specific sounds produced by the vocal tract.
  2. Vocal Tract: Define the vocal tract as the anatomical part of the human body responsible for producing speech sounds. Mention that it includes the mouth, tongue, throat, and other structures involved in speech production.
  3. Vowel Classification: Explain that Kempelen classified vowels based on characteristics like the width of the lip channel and tongue channel. Provide a simple explanation of how vowels are produced and distinguished from each other.
  4. Articulatory Apparatus: Mention that this refers to the various parts of the human speech mechanism, including the lips, tongue, vocal cords, and so on.
  5. Mechanical Simulation: When discussing mechanical simulation of speech, briefly explain the concept of simulating human actions or functions using mechanical devices.
  6. Speech Synthesis: Define speech synthesis as the process of generating artificial speech using machines or software. Mention that modern speech synthesis involves computers and advanced algorithms.
  7. Resonance and Multiple Resonance: Provide a concise explanation of the concept of resonance and its importance in speech production. Mention how multiple resonance can affect the quality of speech.

Incorporating these explanations will ensure that readers who are not familiar with the terminology can still grasp the significance of Kempelen's work and its impact on speech synthesis technology.

How We Utilized the Feedbacks:[edit | edit source]

We have adopted most of GPT's suggestions and made adjustments to our wiki page accordingly. However, we have also found that GPT's feedback may not be entirely accurate (e.g. some part of the first advice in the 'Correct Errors in Terminology and Facts' section). We believe that critically using large language models instead of accepting suggestions verbatim is an important skill in the age of artificial intelligence.

Group Members[edit | edit source]

Yinqiu Wang (Historical Context)

Yanhua Liao(Key Innovation)

Ziyun Zhang (Future Research)

Cantao Su (Introduction)

Chenyu Li (Impact)