Internship recruitment:
We are looking for internship students in the 4th year of undergraduate or graduate studies from local or international countries. The duration is from 3 months to 6 months.
The project works for the internship students are listed below:
Using the zero-aperture microphone array, the project aims to develop 3D audio techniques including 3D spatial audio acquisition and playback, 3D sound source localization, and sound source recovery. The interested applications include teleconference, virtual reality, multimedia, social robotics, military surveillance etc. Specifically, we are looking for internship students for
High-quality microphone preamp on PCB: We are targeting to design high-quality microphone preamp on PCB for clean signal acquisition. The microphone that we are using is the standard hearing-aid microphone. We form a microphone array using four of the microphones. The microphone array will be served as the input component for sound acquisition. For all the following processing and development, the signal quality of the microphone is one of the key factors. You are required to design the microphone preamp for a four-channel microphone array. A good amplitude gain and high signal-to-noise-ratio (SNR) is expected.
Sound data acquisition through USB interface: The USB connection has been widely used in the Computer-related interfaces. It is simple and very low cost. We are planning to use the USB connection for sound data acquisition from external microphones with a PC. The student is expected to develop programme for sound data acquisition using Microsoft Visual Studio 2008 (VC++). The programme should be able to support multiple microphone channels. The students may reference other online available APIs such as Microsoft Kinect for Xbox 360.
Real-time implementation of 3D audio demo system for acquisition and playback: A testing system for the 3D audio acquisition and playback through headphone has been implemented in Matlab. The system deployed a standard HRTF (head-related-transfer function) database measured from a model of human head, which were mapped to the responses applied on the output of microphone array. The map was done by a least-squares minimization method. The IA/IO student is required to familiarize with the offline testing system for 3D audio acquisition and playback and implement it in C/C++ code for a real-time system. A framework of the system in C/C++ has been implemented using the MS Visual Studio 2008 and will be provided to student.
Onset detection for sound source localization in reverberant environment: A sound source localization system is used to identify origins of sound sources using microphone array signal processing. It has wide applications in robotics, surveillance systems. In many cases, the localization system has to be deployed in very reverberant environment. Due to multipath of sound propagation, performance of the sound source localization system can become unreliable since the reflected paths influence the used direction estimation methods. To get rid of the reverberation effects, it might consider using the signal portion for the estimation that contains only direct path. As we know, the direct path of sound source often arrives at the microphone first. After a very short duration, the reflected paths then arrive at the microphone. If the signal portion of the direct path can be detected (which is usually called onset detection), the sound source localization based on this portion will be much more reliable. The IA/IO student is required to study methods for the onset detection and implement them in Matlab for performance evaluation.
For any enquiry or application, please email shengkui.zhao@adsc.com.sg.