Martin Spiertz

A probability-based combination method for unsupervised clustering with application to blind source separation

Unsupervised clustering algorithms can be combined to improve the robustness and the quality of the results, e.g. in blind source separation. Before combining the results of these clustering methods the corresponding clusters have to be aligned, but usually it is not known which clusters of the employed methods correspond to each other. In this paper, we present a method to avoid this correspondence problem using probability theory. We also present an application of our method in blind source separation. Our approach is better expandable than other state-of-the-art separation algorithms while leading to slightly better results.

Inversion of short-time fourier transform magnitude spectrograms with adaptive window lengths

In this paper, we extend the Real-Time Iterative Spectrogram Inversion method (RTISI) for generating a time-domain audio signal from a magnitude spectrogram such that it can handle changing spectrogram window lengths. For each desired window length, we use a separate buffer structure and synchronize the buffers each time the window length changes. This way, the proposed method helps to improve the time/frequency-resolution trade-off for algorithms that operate on magnitude-only spectra.

Transient detection with absolute discrete group delay

This paper presents a new transient detection algorithm which uses the average absolute discrete group delay as a measure for the transient characteristic of sound. It shows that a bell-shaped window function performs a high-pass effect to the spectral coefficients, leading to a concentration of group delay values in the πarea for steady-state signals. This concentration is violated if a transient occurs. From this phenomenon, we derive a new transient detection method, improve it by a maximum order-filter, and show that it works well on percussive and tonal-percussive sounds.

Iterative monaural audio source separation for subspace grouping

Monaural blind audio source separation usually separates a mixture into more signals than active sources. Therefore, a clustering of the separated signals is needed to reconstruct the sources. We propose a new iterative clustering and show that this approach outperforms classical clustering approaches which use features of the separated signals for clustering. The iterative clustering starts with the separation into two source estimates. Based on this, at each iteration the squared error between the source estimates of the former iteration and a linear superposition of the separated signals of the current iteration is minimized. The corresponding linear superposition generates new source estimates. The algorithm is evaluated on a large test set regarding melodies of different instruments, singing, and speech from the EBU.

H.264/AVC Compatible Scalable Multiple Description Video Coding with RD Optimization

We present a multiple description video coding scheme based on H.264/AVC. Each description is independently decodable by a standard compliant decoder, and mutually refining decoding of more than one description is possible with moderate pre-processing on the bit stream level. The scheme allows for fine-granular control of redundancy. Scalability is added by enabling the tools defined in the scalable extension of H.264/AVC. Then, the redundancy can be dynamically adjusted according to the network conditions. Furthermore, based on a formulation for the error propagation in the proposed system, we demonstrate that the redundancy can be generated in a rate-distortion optimized way

Central Decoding for Multiple Description Codes based on Domain Partitioning

Multiple Description Codes (MDC) can be used to trade redundancy against packet loss resistance for transmitting data over lossy diversity networks. In this work we focus on MD transform coding based on domain partitioning. Compared to Vaishampayan’s quantizer based MDC, domain based MD coding is a simple approach for generating different descriptions, by using different quantizers for each description. Commonly, only the highest rate quantizer is used for reconstruction. In this paper we investigate the benefit of using the lower rate quantizers to enhance the reconstruction quality at decoder side. The comparison is done on artificial source data and on image data.