AES Tokyo Convention 2009
Poster Session P2

P2 — Measurement and Signal Processing

Friday, July 24, 12:00 — 16:00 (Core Time for Odd Numbers: 14:00-15:00, Core Time for Even Numbers: 15:00-16:00)
Chair: Yoshinobu Kajikawa (Kansai University)

P2 - 1   Vocoder-based morphing tool demonstrations for flexible voice manipulations

Hideki Kawahara (Wakayama University), Masanori Morise (Ritsumeikan University), Toru Takahashi (Kyoto University), Hideki Banno (Meijo University), Ryuichi Nisimura (Wakayama University), Toshio Irino (Wakayama University)
A flexible framework for voice manipulations based on a high-quality vocoder, TANDEM-STRAIGHT and temporally variable multi-aspect morphing was introduced. This framework was made accessible by introducing graphical user interfaces with supporting tools. The tools provide means to modify existing speech materials interactively and intuitively by modifying parameters directly or interpolating between examples on arbitrarily designed morphing trajectories. Demonstrations of flexible voice manipulations suitable for game applications also will be introduced.

P2 - 2   Methods for Sound Improve Articulation and Discrimination Performances

Yasuo Sano
Improve the articulation (clarity) and discrimination performances of audio equipment or audio signal easily by simple circuit constitution to provide a high definition audio signal and audio equipment inexpensively. The way of this solution, that it only using series connected two kinds harmonics generator with asymmetrical difference operation without harmful distortion. Small level harmonics have same rising time to original input signals what hearing white noise or impulse on aural system where asymmetrical difference operation will reduce these components with phasing harmonics generation by Haas effect same time together. Furthermore, these components are reducing harmful distortion components by slow slope filtering with summing to input signals. Consequently, the sounds articulation and discrimination performance of audio signals or audio equipments are improved.

P2 - 3   A new upmixing algorithm based on frequency domain independent component analysis

Sungyoung Kim (Yamaha Corporation) and Makoto Yamada (Yamaha Corporation)
n this paper, we propose a new noble upmixing algorithm that manipulates front and rear components of the five-channel application from the conventional stereo (two-channel) contents. The proposed method is based on a statistical analysis called Frequency Domain Independent Component Analysis (FDICA). As its name implies, FDICA extracts independent (not necessarily orthogonal) components from a stereo signal, which differs from conventional methods such as Principal Component Analysis (PCA). We created an extended surround imagery by placing a relative independent source to rear channels. The subsequent subjective evaluation showed that the proposed method was preferred to some conventional processors due to the similarity in spatial attributes of the its original multichannel mix.

P2 - 4   Evaluation of Synchronized Significant Multi-bits Acoustic Steganography Method

Xuping Huang(The Graduate University for Advanced Studies(SOKENDAI)), Isao Echizen(SOKENDAI/National Institute of Informatics), Yoshihiko Abe(Iwate Prefectural University)
Recently, steganography using multiple media content has been proposed for enhancing information security along with encryption. Research areas range from hidden capacity enlargement, to robustness enhancement of stego data towards attacks and so on. In this paper, model and algorithm of real-time steganography scheme are proposed. This method is implemented to embed secret acoustic data stream which is recorded as synchronously as it is embedded into another acoustic cover data stream. In addition, embedding positions among [1st, 8th] bit in cover data with sampling size of 16-bit can be arbitrarily specified. Experimental results demonstrate that the secret bit stream can be interspersed into significant bit locations in cover without drawing suspicion even though some certain performance degradation is caused.

P2 - 5   Sound Field Evaluation by using Auditory Filter:Application of Dynamic Compressive Gammachirp Filter

Yuki Matsumoto (Graduate School of Design, Kyushu University), Masahiro Suzuki(Graduate School of Design, Kyushu University), Akira Omoto(Graduate School of Design, Kyushu University/ONFUTURE Ltb.)
We attempt to apply various auditory models to evaluation of room acoustics. In this report, we examine the decay curve which is calculated from the impulse response filtered by using dynamic compressive gammachirp filter. Reverberation time which is observed from the obtained decay curve could provide early decay time which is observed by traditional method and have been said to be correlate closely with subjective reverberation time. We apply dynamic compressive gammachirp filter to the evaluation of the impulse responses which are measured at many points in the hall. The results show that the proposed method might evaluate sound field taking into consideration the auditory property of human.

P2 - 6   Analysis of head-related transfer functions based on spatio-temporal frequency characteristics

Yasuko Morimoto (Graduate School of Information Science, Nagoya University), Takanori Nishino (EcoTopia Science Institute, Nagoya University), Kazuya Takeda (Graduate School of Information Science, Nagoya University),
This paper describes a new method for representing a head-related transfer function (HRTF), which is an acoustic transfer function between a sound source and the ear canal entrance. An HRTF is defined as a function on time and space. The spatio-temporal frequency characteristics can be visualized and analyzed by showing the spectrum computed by two-dimensional Fourier transform on time and space. In our experiments, we investigated the basic property of the spatio-temporal frequency characteristic and the difference between all data for the HRTFs obtained by numerical analysis and actual measurements. The reverberation and HRTF individuality were also examined. From the results, the characteristics were mostly concentrated in a specific band frequency, and there were also differences among databases.

P2 - 7   Loudspeaker evaluation with directional cross-correlation function

Kazuhiko Kawahara(Faculty of Design, Kyushu University), Akihiro Sonogi(Graduate School of Design, Kyushu University)
Quantitative evaluation measure was proposed for di®used property of distributed mode loudspeaker(DML). Full width at half maximum angle was introduced to evaluate cross-correlation function, Gontcharov had proposed Graphically. We showed DML radiate spatially less coherently than conventional loudspeaker does. We could separate DMLs from other pistonic radiators with proposed measure.

P2 - 8   A Proposed Method of Characterizing Audio Distortion Induced by Power Supply Ripple in Audio Amplifier

Yang Boon Quek (Texas Instruments Inc)
Digital input Class-D amplifiers will be the predominant amplifier technology enabling consumer audio systems in the future. The traditional Power Supply Rejection Ratio (PSRR) measurement method cancels supply ripple in Bridge-Tied-Load (BTL) amplifiers, thus is unable to measure audio distortion induced. The proposed method employs innovative measurement of both Intermodulation Distortion (IMD) and Total Harmonic Distortion Plus Noise (THD+N) to more accurately represent audio quality. A new term known as Power Supply Ripple Distortion Factor (PSRDF) is introduced as a figure of merit for audio quality. Examples of how the proposed method effectively characterizes different levels of distortion induced by power supply ripple in closed-loop and open-loop BTL Class-D amplifiers are also presented. The proposed method is also applicable in characterizing all audio power amplifiers.

All Sessions
Last modified: Mon Jun 22 20:47:56 2009