Digital Music Processing
This group researches technologies for analysing and processing music, either vocal or instrumental, to achieve a range of musical effects. It uses a range of Digital Signal Processing (DSP) techniques on musical applications, and works closely with the Digital Image and Video Processing research group, which employs similar algorithms and techniques in its own work.
Research staff include: Dr Zlatko Baracskai, Yonghao Wang, Izzy Ali-MacLachlan, Ryan Stables, Dominic Ward, Alan Dolhasz and Sean Enderby.
Current projects
Adaptive digital audio effects: a variety of signal processing techniques are being implemented to create high-end novel audio effects for use in production and performance. SuperCollider programming language is used and the resulting prototype effects are Audio Unit format. The novelty in our effects library is that each of the parameters/sliders can be instantly linked to an LFO or to a feature extraction based signal. A demonstration will be held at the ICMC2011. To download and read about our free and open-source devices please follow the ADMOD.
Music processing library: A range of MIDI and audio signal processing techniques are developed for the Max for Live environment. Tools for algorithmic composition, cross-adaptive strategies and education are just some of the chapters in this versatile collection. To download and read about our free and open-source devices please follow the M4LIB link.
Singing voice processing: Perceptually defined processing of a singing voice is being developed based on the STFFT orthogonal transform and its derivatives. Assessment of different analysis/resynthesis schemes will be followed by research into the qualitative perception of the singing voice and the multi-dimensional mapping to the digital audio processing parameters.
Real-time psychoacoustic models: in collaboration with Midas/Klark we are developing psychoacoustic models for real-time operation. These will form a part of novel audio display techniques also to include a variety of perceptual cross-correlation visualisations to aid the sound engineer in mixing audio.
Intelligent music mixing systems: have been developed by researchers to perform fundamental mixing procedures automatically, providing engineer support. This research uses models of the peripheral hearing system to investigate the relationship between the mixing decision and psychoacoustic phenomena such as partial masking.
Analysis and processing of traditional Irish flute music: focusing on different regional styles of playing musical variation and ornamentation techniques are sought to be detected and produced to aid the understanding and the learning process of authentic folk music styles.
Low Latency Audio Converter: We measure the delay between input and output of audio band ADCs and DACs. We identify the major cause of the delay within Sigma Delta Modulator (SDM) based converters and we designed a new architecture of SDM based ADC/DAC to minimise the input/output delay.
Embedded audio processing: We designed an embedded High Definition audio controller using FPGA and DSP chip to evaluate multi-channel, low latency, high resolution audio processing in portable platforms as well as providing research platform for multichannel sample based audio processing.
Musical control signal processing: Connected to the joint EPSRC bid with BIAD and the Birmingham Conservatoire we are developing a flexible control signal processing framework and widely applicable processing modules. This will allow for development of various gesture following and interface analysis implementations within a user focused framework Integra Live.
Recent projects
LICKBACK - Voice to singing conversion: we have developed an engine to create customised greetings by the conversion of names entered as text to a sung name that is seamlessly inserted into gaps within an existing greeting track (in collaboration with Lickworx Ltd via a Shorter KTP scheme).
Latency measurement and analysis: in digital audio conversion, codecs and operating systems has been the research focus of Yonghao Wang. The findings are relevant to the future research in real-time digital audio processing applications in embedded, desktop and mobile systems.
Digital subharmonic generation: Novel signal processing techniques have been invented to aid the creation of sub-bass audio. Following up on the published paper on subharmonic distortion in the Proceedings of the 128th Audio Engineering Society Convention further developments are planned in collaboration with Michael Woodford of the C4DM at the Queen Mary’s University in London.
Recent publications
Trutzschler J., Baracskai Z., “A Graphical User Interface for Supercollider Audio Units”, International Computer Music Conference, Huddersfield, August 2011.
Baracskai Z., “New Trends in Algorithmic Composition”, 7th International Symposium of Musicology in Sarajevo, October 2010.
Wang Y., Stables R., “Latency Measurement for Desktop OSes and
Adaptive Audio effects”, Audio Engineering Society (AES) 128th Convention in London, May 2010
Baracskai Z., “Algorithms for Digital Subharmonic Distortion”, Audio Engineering Society (AES) 128th Convention in London, May 2010
Baracskai Z., “The Max for Live platform”, Digital Music Research Network, Queen Mary London, December 2009
Baracskai Z., “Instrument Extensions”, Symposium of Performing Technologies, Sonic Arts Research Centre Belfast, May 2009.
Poster presentation proposal accepted for Audio Engineering Society (AES) 128th Convention in London, May 2010