Sound is a pleasant thing to study
For his master's thesis Eero-Pekka Damskägg developed a sound microscope that alters the duration of playback of sound recordings while allowing the recording to sound the same in other respects.
'With a sound microscope the frequency content of recorded sound is unchanged even if it is played back at a different speed. Sound is broken into segments of 10 - 100 milliseconds after which we examine the frequencies contained in the small pieces. We move the pieces further or closer to each other to either slow down or speed up the playback. In addition, we alter the frequency components in the pieces to make the pieces that are moved link up smoothly with each other. This technology is called a phase vocoder’, Eero-Pekka says.
Phase vocoder-based manipulation of the time scale of sound was invented already in the 1980s but it has never been perfected. Several methods produce high-quality results only with fairly small changes in playback speed. The system developed by Eero-Pekka aims at keeping the sound quality unchanged even if the playback is tens, or possibly hundreds of times slower.
'The thesis work has been a project in which I have been able to utilise both my problem-solving skills as well as matters related to signal processing that I have learned previously.'
The sound microscope was shown at the Heureka Science Centre to enthusiastic schoolchildren who were allowed to test the microscope and to guess the origins of various altered sounds.
Varied learning methods
Eero-Pekka began his studies at Aalto, studying electronics and electrical engineering at the School of Electrical Engineering. 'I became interested in acoustics in a course on signal processing that I took during my bachelor's degree studies, where our practice assignments included experiments with sound. In addition to acoustics I study Machine Learning and Data Mining as a minor. I feel that this minor has been very useful because machine learning is increasingly playing a role in audio research. A machine can be taught to recognise sounds, for instance. I have been working on a project linked with machine learning, in which a computer is taught to identify the genres of different pieces or music.'
What have you liked best about studying at Aalto?
'The courses were well organised and very interesting. They involve the use of diverse learning methods. Success in an exam is not necessarily given much weight in the grade. Instead, various exercises and group projects are used to encourage students to learn what is taught in the course', Eero-Pekka says.
Eero-Pekka has not yet thought about what he will do after he graduates.
It would be great if my work in the future were to be linked with music somehow. I could be involved in developing applications that producers and other music professionals could use in their work.'
You can study acoustics in the Master's Programme in Computer, Communication and Information Sciences - Acoustics and Audio Technology, read more ja apply now!