Masanobu Miura’s research while affiliated with Kunitachi College of Music and other places

Our previous study described a procedure for calculating fluctuation strength (FS) from tremolo produced with a mandolin. However, the forms of the amplitude envelope on sounds played on the mandolin vary so little that how the forms affect the perception of fluctuation has not yet been investigated. Therefore, we present a procedure for calculating FS by considering the form of the amplitude envelope on a tremolo produced with a marimba, whose forms of the acoustic waveform on the tremolo vary due to the stiffness of the head of the mallet. We obtained acoustic features of a marimba tremolo as a set of parameters and calculated FS using the parameters. We found that the procedure observing three features concerning constant fluctuation, the deviation for time and amplitude, and deformation on the envelope can calculate FS correctly.

The motor sound on electric powered vehicle is quiet at low speeds. Thus, pedestrians have difficulty detecting the vehicles approaching them under urban noise. Although the vehicles were designed to play an alert sound to solve this problem, it has not been solved yet. Our previous studies found that characteristics of amplitude fluctuation, fluctuation frequency, non-periodic fluctuation and amplitude envelope, are effective to make them detect approaching vehicles. However, those studies were investigated under only a specific actual environment, weren't examined validity of detectability in those studies. Here, this paper investigates under another actual environment, examine the validity. Investigations were carried out by using synthesized complex sounds which were designed to have periodic and non-periodic amplitude fluctuations. Those complex sounds have characteristics of amplitude fluctuations in gasoline powered vehicle. Amplitude envelopes such as modulation wave in amplitude-modulated sound were set for deviations for time and amplitude, and amplitude-modulated complex sounds were synthesized using sine wave, sawtooth wave, and rectangle wave. Then, their effects on detectability by pedestrians were assessed in another actual environment. The results found that amplitude fluctuation enhances the ability with which people detect approaching electric powered vehicles in case of some complex sound.

The sound on electrical vehicle is quiet at low speeds, so pedestrians have difficulty recognizing electrical vehicles approaching them. Those vehicles were designed to play a warning sound to solve this problem. However, it has not been solved yet, and the introduction of warning sounds is broadly discussed on different political levels. It is necessary to design warning sound to concern detectability of approaching electrical vehicle in order to make it easier for pedestrians to recognize the vehicle. Our previous studies investigated detectability of fluctuating sound, and then found that the fluctuation frequency, non-periodic fluctuation and shape of envelope are effective to enable people to recognize approaching vehicles. However, frequency characteristic of fluctuating sound used in those studies was one pattern, namely motor sound. Here, we investigate relationship between frequency characteristic of fluctuated sound and detectability of warning sounds for electrical vehicle. Investigations were carried out by using different fluctuated sounds which were designed to have periodic and non-periodic fluctuations, and their detectability by pedestrians were assessed. The results revealed that the frequency characteristic of fluctuated sound influences the ability with which people detect approaching electrical vehicles.

Analysis of piano performance has been a great task for musicologists, in particular the “artistic deviation” on performance has been intensively discussed. Introduced here is a method to clarify what occurs on music performance by observing the eigenvector of obtained several piano playings, “Eigenperformance.” The eigenperformance is obtained for MIDI-piano playing from the tendency of onset, MIDI-velocity, duration, and instantaneous tempo curves. Eigenperformance is possible to indicate tendencies for a set of piano playing. By using eigenperformance, here investigates 1) the differences in evaluation criteria between pianists and non-pianists, and 2) the differences of deviations due to player’s skill. For 1), results of simple scale performance show that pianists evaluate short and sudden tempo changes with taking long time in ending notes along with dynamic and temporal expressions based on phrases on sheet music as good, whereas non-pianists never provide such evaluation. For 2), piano etude “For Elise” is dealt with to be investigated by four classes of proficiency. The eigenperformance revealed that the proficient performance shows same patterns of dynamics among first and second sections in sheet music, whereas the novice performance does not show any analogous on corresponding parts in sheet music.

It’s difficult to determine an chord label for acoustical signal of musical playing, in particular when playing chords with “omitting”, “inversions,” or “tension voicing” on the guitar. Additionally, “enharmonic equivalence” produces multiple possibilities. This study developed a chord estimation system that deals with an audio signal output from electric guitars considering such techniques. All of the chord types employed in this study are the sixteen patterns frequently used in guitar chord playing. Chord labels are estimated by combination of salient pitch classes (or chroma), and some of them are dealt with as “performed notes” assumed as the member of played chord. Obtained performed notes are input to the “search tree for chord labels” so as to search possible chord labels, by referring chord progression patterns included in the “chord progression database.” Sixteen chord types are triad such as major, minor, aug, dim and sus4, with four-note chords such as 6th, 7th, Maj7, aug7, 7sus4, add9, min6, min7, minMaj7, min7(b5), and dim7. This study investigated appropriate threshold and type of filter used in the judgment of performed note. Results found that chord labels are estimated as 89% of accuracy when using three types of electric guitars.

Since arrangement of sheet music for orchestral ensemble requires a certain musical knowledge and expertise, amateur orchestral musicians often have difficulties when arranging. An automatic arrangement method was proposed by employing Eigenmusic which is a set of eigenvector for many music excerpts, in order to evaluate the similarity among phrases. The difficulty in playing the arranged sheet music, however, was not evaluated; so, whether players feel difficulties on playing, or playability, was out of discussion. This study proposes a method to evaluate the degree how easy to play the phrases on sheet music. This method generates phrases for woodwind instruments using a MIDI database of backing phrases. Each musical phrases selected from the database are allocated to instrumentalists. In turn, the phrases’ playability is evaluated based on instrumentalist’s individual ratings for playing features such as cross-fingering, jumping notes, and so forth. Finally, phrases evaluated as the easiest are conjoined in each instrumentalist. We conducted an experiment to investigate the playability of generated sheet music by musician’s rating. Experimental results showed that the evaluation of playability by proposed method is effective so that proposed method generates easier phrases to play than our previous method.

The article examines a real-time emotion control system for polyphonic MIDI musical excerpts. Users are able to manipulate a pointer in the emotion plane displayed on the system using a wireless controller. The position of the pointer determines the emotion and its intensity in real time, so the performance of SMF excerpts can be continuously changed by moving the pointer in the plane. Users determine the psychological activity by indicating the position on the horizontal axis. The timbre-ID used is determined by observing the neighborhood in terms of brightness on the basis of the position of the pointer. Since this process occurs in the timbre-DB, timbre-IDs are continuously selected on the basis of brightness. This process yields the note number n and velocity value v of an SMF excerpt, and calculates S for a specified note number and velocity in accordance with the pointer position.

Psychoacoustics research has been contributed on evaluating the timbre of acoustic signals. The fluctuation strength (FS) has been respected as the important index which describes the sensory fluctuation of the acoustic signal based on its amplitude, frequency and the combination of them as well. Although FS has been calculated by tremor components on acoustic signals, here points out not only the inappropriateness of the calculation when simply based on literatures but also a method which focuses on the shape of the waveform in order to extract parameters from it, so that the FS is newly calculated by our original method with estimation results of the subjective scores of sensory fluctuation. Signals dealt with are motorbike exhaust sounds and tremolo played by marimba. The developed method is then applied to the design of an electric vehicle approaching sound for pedestrians in order to let pedestrians notice the approaching car without rising up the sensory loudness, on which designed sounds have a fluctuation with irregular pulses proven to give us the sensory fluctuation so that it is expected to be noticeable for pedestrians. This paper discusses the possibility to apply the prestigious psychoacoustic indexes to industrial and artistic sounds.