Audio Research Team conducts research in many areas related to audio, speech, and music signals. The research is mainly application-oriented, but includes also basic research not targeting specific applications. The research is funded by different sources, including EU,
Academy of Finland,
Tekes, and projects funded by individual companies.
More detailed description of the research areas can be found in their separated pages:
Digital Filtering Group
Group Leader: Professor Tapio Saramäki
The Digital Filtering Group focuses on developing efficient algorithms and structures for design and implementation of digital filters and filterbanks. The application areas of filters and filterbanks, considered by the group, include, among others, multimedia, communications, analog-to-digital and digital-to-analog converters based on sigma delta modulators, audio and speech, instrumentation, image processing, biomedical engineering, antenna arrays, switched capacity circuits, etc.
Current research of the group can roughly be divided into the following 5 topics:
- Recursive digital filters
- Linear-phase FIR digital filters
- Sampling-rate conversion systems
- Filterbanks
- Optimization tools for digital filter and filterbank design.
Each of these topics is discussed in more detail in the following sections.
Recursive digital filters
Following topics have been investigated:
- Low-sensitivity and low-noise recursive filter structures – some of which are directly synthesized in the z-plane – and optimization tools for their design.
- Systematic approaches for designing multiplication-free recursive digital filters, which are implemented either as a parallel connection of two all-pass filters (lattice wave digital filters) – for both the conventional case and for case where phase linearity is of great importance – or as a cascade of these filters.
- Deriving structurally symmetric ladder filter structures, which include as special cases properly modified lossless discrete integrator and wave digital ladder filter structures.
- Deriving Remez-type algorithms for the design of arbitrary-magnitude recursive filters with the same or different denominator and numerator orders and for the design of various kinds of recursive filters composed of all-pass sections.
Linear-phase FIR digital filters
Following topics have been investigated:
- Deriving computationally-efficient linear-phase FIR filter structures (including multiplication-free structures), which exploit a relatively strong correlation between the neighboring impulse-response samples.
- Developing efficient algorithms for the optimization of FIR filters synthesized using multistage frequency-response masking approaches.
- Deriving approaches for the synthesis of computationally-efficient FIR and all-pass filter based IIR filters with an adjustable fractional delay.
- Deriving a Matlab routine strictly obeying the basic principle of the Remez multiple exchange algorithm, which Parks-McClellan algorithm does not follow.
Sampling-rate conversion systems
Following topics have been investigated:
- Developing structures (and optimization tools for their design) for efficiently performing decimation and interpolation by an integer factor.
- Implementation of polynomial-based interpolation filters – for efficiently performing the sampling-rate conversion by an arbitrary factor.
- Deriving polyphase structures that exploit coefficient symmetry of linear-phase FIR filters used in rational sampling-rate converters.
Filterbanks
The group is considering the design and implementation of various types of filterbanks and their use in several areas of signal processing – the emphasis being in communication and audio processing applications. Current emphasize is laid on following types of filterbanks:
- Two-channel FIR and IIR filterbanks.
- Critically sampled modulated filterbanks (e.g. cosine-modulated, GDFT).
- Oversampled filterbanks.
- Nonuniform filterbanks.
Optimization tools for digital filter and filterbank design.
In most cases, a good optimization algorithm (tool) is required when designing a filter, approximating a desired frequency characteristic, minimizing number of coefficients in a structure, etc. Therefore, our group actively seeks new optimization techniques that can be used for optimization of various aspects of filters and filterbanks.
DSP Application Examples
Acoustic Acquisition & Presentation
Signal Coding & Compression
Machine Synthesis of Signals
Cellular Phone
Discrete Multitone Transmission
High Power RF Amplifier
Digital Camera
Digital Sound Synthesis
Digital Television
Software Radio
No comments:
Post a Comment