Frank Kao-ping Soong

Title: Frank Kao-ping Soong: Innovator in Audio Signal Processing

Introduction

Frank Kao-ping Soong is a distinguished inventor based in Beijing, China. He has made significant contributions to the field of audio signal processing, holding a total of 2 patents. His work focuses on enhancing the efficiency and accuracy of audio signal analysis, which has important applications in various technologies.

Latest Patents

Soong's latest patents include innovative approaches to audio signal processing. The first patent, titled "Auto segmentation based partitioning and clustering approach to robust endpointing," introduces a method for scoring possible segmentations of an audio signal based on distortions in feature vectors. This scoring system is utilized to select the most effective segmentation, which is crucial for identifying the starting and ending points of speech signals within audio data.

The second patent, "Parsimonious modeling by non-uniform kernel allocation," presents a multi-state pattern recognition model. This model is characterized by its non-uniform allocation of kernels, where different states are assigned varying numbers of kernels. The training of these kernels using relevant data allows for the formation of a robust multi-state pattern recognition model.

Career Highlights

Frank Kao-ping Soong is currently employed at Microsoft Technology Licensing, LLC, where he continues to develop innovative solutions in audio processing. His expertise and contributions have positioned him as a key figure in his field.

Collaborations

Soong has collaborated with notable colleagues, including Jian-Iai Zhou and Yu Shi, who have also contributed to advancements in audio signal processing.

Conclusion

Frank Kao-ping Soong's work exemplifies the innovative spirit in the realm of audio signal processing. His patents reflect a commitment to improving technology that impacts various industries. His contributions continue to influence the development of advanced audio processing techniques.