Keiichi Zempo

Direction of Arrival Estimation Based on Delayed-Sum Method in Reverberation Environment

Microphone arrays have been used for estimating the direction of arrival (DOA). Owing to intercorrelation among direct and reflected sounds, there are difficulties in estimating DOA using microphone array in highly reverberant environments. The purpose of this research is to estimate the correct DOA in such environments using a small number of microphones with the aid of signal processing. The proposed signal processing uses the relationship that direct signals always arrive earlier than reflected signals. By comparing the delayed-sum signals corresponding to candidate direct signals, the DOA of direct sounds could be determined correctly. We have confirmed the usefulness of the proposed method by conducting several experiments. This proposed method is different from conventional methods because it has high tolerance to the effect of reverberation; it not only enables estimation of DOA but can also be applied to the measurement of the delay of time of flight in reverberant environments.

Localization of Acoustic Reflective Boundary Using a Pair of Microphones and an Arbitrary Sound Source

Compassing the spatial shape is sometimes performed to develop a surround speaker system or enhance a signal. Although the spatial shape is reconstituted from the boundary position, conventional methods offer arrays consisting of a large number of elements (microphones and loudspeakers). In this research, we developed a method of localizing the acoustic reflective boundary based on the direction-of-arrival (DOA) of direct and reflected sounds using a pair of microphones. This method offers the DOA values of direct and reflected arbitrary sound sources. Although the estimation involves some estimation errors caused by approximations in the calculation, the correction function was obtained from several conditions numerically. Results indicate that it is possible to localize the reflective boundary almost accurately (the average error ratio from the true value is below 10%) through the presented method.

Auditory lateralization steering independent of loudspeaker position using sound from ultrasound

There are a lot of impaired vision people and they are planning physical access by information of positional relationships of auditory cues or by oneself. In this paper, we propose auditory lateralization guide independent of loudspeaker position by using super-directional speaker system for pinpoint presentation and suppressing spread sounds because the individual who do not offer the support would feel it as noise. Through the experiments, it was confirmed that we could steer the auditory lateralization that the supported individual would percept, by using the acoustic signal which was convoluted with the impulse response recorded on dummy head corresponds to the direction which want to present.

Analysis of sound propagation in human head for bone-conduction headphones using finite element method

In this paper, we evaluated the characteristics of the bone-conducted sound propagation in the human head through computer simulation using a reduced head model with finite element method. The model was made of three spheres, which represent the skin, skull and brain, and the brain was represented by the water for simplicity. The step response of the head was investigated to show that the sound propagation would be generated in the form of longitudinal and bending wave through the head. From the simulation results, the sound speed of the longitudinal and bending waves were approximately 1500 m/s and 260 m/s, and the latter was compatible with the previously measured sound speed of bone-conduction.

Measurement of the differential transfer function between bone-conduction and air-conduction for sound localization

In this paper, a measurement of the difference between air-conduction and bone-conduction transfer functions was performed. The transfer function was measured by canceling air-conduction and bone-conduction sound each other. The transfer function was measured for 30 kinds of frequencies ranging from 100 Hz to 15 kHz for 6 subjects. From the results, the frequency characteristics of amplitude difference measured using all subjects have the common dip at approximately 3 kHz and peaks at 3 to 7.5 kHz in this case, and the cepstrum insists that the BC sound arrives to the inner ear 0.2 ms prior to the AC sound. The frequency characteristics of phase differences have the linear part, and it would indicate that AC and BC sound would have different propagation time. With these differential transfer functions, the spatial audio can be realized by compensating BC sound as if it is AC sound.

Phonoscape: Auralization of Photographs using Stereophonic Auditory Icons

In this paper, we developed an auditory display method which improves the comprehension of photograph to apply the support system for person with visual impairment. The auralization method is constructed by object recognition, auditory iconization and stereophonic techniques. Through the experiments, the enhancement of intelligibility and discriminability was confirmed compared to the image-to-speech reading machine method.

The Alteration of Gustatory Sense by Virtual Chromatic Transition of Food Items

When we visually recognize food items, their color influences the impression of their taste, because the gustatory sense is affected by other senses such as smell and eyesight. In this study, in order to control the gustatory sense of human beings, we examined the influence on gustatory sense by giving a virtual chromatic transition of food through Head-Mounted Display. The experiment was conducted under six conditions that changed three factors: “before or after eating”, “viewing through HMD or naked eyes”, “with virtual chromatic transition or without”. As a result, we confirmed that the difference in the amount of visual information influenced the impression of food items before eating. It is expected to realize more applicative control of the gustatory sense by analyzing this difference.

Speech balloon system using single-channel microphone array on see-through head-mounted display

In this paper, we developed a portable caption support system consisting of a see-through head-mounted display (HMD), a microphone array, and a speech recognition system to visualize the conversation information for person with hearing impairment. Through the system, the speech signal was emphasized by signal processing for single input channel, and the speech balloon was generated through HMD.

Estimation of contact force and amount of hair between skin and bone-conducted sound transducer using electrical impedance

Noninvasive bone-conducted hearing aids require the consistent fitting of transducers for consistent hearing. In this paper, a method of estimating the contact force and amount of hair between a persons skin and a bone-conducted sound transducer using electrical impedance is proposed. Experiments are conducted with a human surface model consisting of hair, skin, and bone. The estimator is implemented with a three-layered neural network. Ten measurements for 70 conditions are conducted by changing the contact force from 0 to 5 N and the amount of hair from 0 to 169.6 mm3. With the trained estimator, it is possible to estimate contact force and the intermediate material thickness with mean errors of 0.025 N and 0.424 mm3. This result supports the feasibility of the proposed method and contributes to the reproducible placement of the bone-conducted sound transducer.

Effect of parameters of phase-modulated M-sequence signal on direction-of-arrival and localization error

In this paper, signals for acoustic beacons with M sequence are evaluated with various parameters for self-localization of mobile robot. The effect of (i) baud rate, (ii) carrier frequency and (iii) order of M sequence on DOA estimation error and localization error are evaluated by computer simulation. As a result, localization is possible with same class or more precision compared to the result from sound beacon with three chirp signals within 5 kHz-18 kHz by choosing the appropriate parameters, which are (i)baud rate is more than around 8 kHz and (ii)carrier frequency is 10 kHz. Particularly, (i) baud rate and (ii) carrier frequency have influence on the localization error. However, the localization error do not change into even a further baud rate in this condition if baud rate is higher than carrier frequency. In this simulation, no influence is confirmed from (iii)order of M sequence. The result indicates that M sequence can be used as signal of acoustic beacons instead of chirp, which occupies frequency band.