Analyzing the nature of all sounds, from human speech to Noh chanting

Pioneer of sound research, challenging research that others do not do

I am involved in research related to sound and language media.

The starting point of my research is the study of natural language processing (words used by humans to communicate with each other). I have been deeply involved in the development of speech recognition technology and spoken dialogue systems since I was a student, working on research that uses computers to analyze the language of words spoken by humans in conversation.

Developing cutting-edge technology is not without its difficulties. At that time, research could only be conducted in limited environments, such as working with supercomputers. Even if we came up with an effective system, we could not conduct verification experiments without a place to store audio data for analysis and a computer capable of high-speed processing. Furthermore, it was considered very difficult to conduct research on music with large amounts of data, even if one wanted to do so.

Over time, as processing speeds have increased, the amount of data that can be handled has grown, and network environments have become faster, speech recognition research has made great strides toward practical application. Speech data is stored on servers, and analysis data is instantly exchanged over the network. I am amazed at the advances in this environment, which would have been unthinkable in my student days.

As for my approach to research, I feel it is meaningful to create a new framework in an undeveloped field.

Research in fields that few people do is a continuous process of trial and error, but I am able to challenge myself with free ideas. I got involved in voice recognition technology because it was a research field that few people were involved in at the time.

Nowadays, my research has expanded to include not only speech but also all kinds of sounds, including music. At our laboratory, we are engaged in unique research projects such as "voice conversion for animated voice actors" and "automatic conversion system for orchestral scores for small ensembles," both of which were initiated by our students.

Taking advantage of Hosei's diverse research fields to analyze chants

In recent years, with the cooperation of the Nogami Memorial Hosei University Institute of Nohgaku (hereafter, Nohgaku Institute), an affiliated facility of Hosei University, we have been conducting research on the digital analysis of Noh chanting.

In order to analyze sound, the target sound source is recorded and then captured into a computer to be converted into data. Although the existence of the sound is not visible, the sound data can be visualized in graphs such as sound waveforms, and can be used for various types of analysis, such as extracting a portion of the sound.

Noh chanting has a unique style that cannot be conveyed simply by transcribing the melody in musical notation. Moreover, even within the same piece of music, different schools of Noh have different nuances such as vibrato, emphasis, and inflection that make the voice tremble, making the music difficult to understand for those who are not familiar with it.

Therefore, we are attempting to visually clarify what kind of expression is added to the melody of a chant by estimating the melody part of the chant. I hope that this approach will lead to an objective understanding of the traditional performing arts and provide a foundation for further research on Nohgaku.

Hosei University has 15 undergraduate and graduate schools as well as numerous research facilities. However, I have always felt that it is a shame that there are so few opportunities to interact across them in research and learning. Nowadays, information technology (IT) is a technology that everyone uses, and I feel that more meaningful research and study can be conducted if we cooperate with each other.

I believe that the connection we have established with the Nohgaku Research Institute this time in the form of asking for cooperation in research is the first step toward building a horizontal relationship. I would like to use this opportunity to expand our ties in many directions.

I want to foster engineers who will lead the future of Japan.

There are many companies where knowledge of information science can be utilized. In particular, we have a close relationship with electronics manufacturers, and students who have studied technology at university have played an active role in the development of the industry by going out into the world and playing an active role.

In recent years, however, electrical manufacturers have been struggling to secure human resources. I feel a sense of crisis that if this situation continues, there will not be enough engineers to lead society, and I feel that it is an urgent need to foster high-quality engineers.

I sometimes feel frustrated by the reserved attitude of the students, but I believe that they have the potential to achieve their goals if they focus on them and do their utmost. I would like to push them so that they will be able to play an active role in society.

• He has created scores for irregular bands including songs, keyboards, saxophones, violins, and basses, using an automatic arrangement program created by his students. The band performed at a Christmas party held in the Faculty of Computer and Information Sciences.