Aims

The aims of the course are to:

• To gain understanding of the main processes in an image compression system, and the typical trade-offs in designing such a system:
• An input filtering (or transformation) process, which compacts most of the energy of the image data into a relatively small number of filter output samples;
• A quantisation process, which represents these samples to some desired accuracy;
• ##A lossless entropy coding process, which codes the quantised samples into the minimum number of bits that still allows the samples to be recovered to their quantised accuracy in the decompressor.

Content

This project introduces you to some of the essential design tradeoffs which must be made during the design of image data compression systems. The main purpose of such systems is to compress as far as possible the size of data file required to store an image (typically a real-world scene) while still preserving the quality of the decompressed image at an acceptable level.

The project covers techniques which to some extent reflect the compression inherent in the JPEG, JPEG2000 and JPEG-XR standards. JPEG (Joint Photographic Experts Group) is the image compression standard from 1992 still commonly used today. JPEG2000 and JPEG-XR are more modern versions which are gradually becoming more widespread. The images above are examples of the same data compressed to the same size but using three different schemes.

FORMAT

Students will work in pairs. Each student will write interim reports by the end of weeks 1 and 2 and a final report by the end of week 4.

ACTIVITIES

The project introduces you to each of these processes in turn and allows you to make a number of inter-related design decisions. New concepts are introduced as the project progresses, rather than by trying to introduce too much theoretical material at the beginning.

At the end of the project all groups will use their final design solutions to compress a small set of images to given file sizes, and the quality of the reconstucted images will be assessed both subjectively and objectively in a competition (complete with a prize!) to select the best design.

Aims

The aims of the course are to:

• To introduce non-parametric signal analysis and processing metthods in the transform domain
• To apply non-paremetric processing methods to real audio datasets
• To study and apply parametric methods to signals, with an emphasis on probabilistic techniques (likelihood and Bayesian) and incuding model choice
• To carry out an extended application in Matlab in audio processing for noise reduction, packet concealment, declipping or interpolation.

Content

This project introduces signal modelling/processing techniques and applies them to audio and musical signals. First non-parametric methods based on transforms such as the Discrete Fourier Transform (DFT) and its fast variant the FFT are studied and experiments are carried out with windowing, frequency resolution etc. These methods are then applied in an audio noise reduction setting, processing sound signals using overlap-add analysis/processing/synthesis to perform noise reduction. Then parametric models are introduced, with estimation using least squares., maximum likelihood and Bayesian techniques. The autoregressive (AR) model is used as an example here, and model choice is studied within likelihood and Bayesian probabilistic settings. Once again, the techniques are applied in audio signals, using models to perform packet loss concealment and interpolation of missing data in audio, as well as constrained interpolation for clipped and/or heavily quantised signals. Students will have the opportunity to incude their own ideas into the applied schemes and there will be a competition for the best noise reduction performance from test audio datasets, evaluated using mean-sqaured error, perceptual criteria and (informal) listening tests.

FORMAT

Students will work in pairs.

ACTIVITIES

Week 1:Non-parametric analysis processing in the transform domain – DFT, FFT, windows, window length, overlap-add processing and application to denoising of audio signals.

Week 2:Parametric modelling of signals using maximum likelihood and Bayesian techniques, including Bayesian model choice, autoregressive models and sinusoidal models.

Weeks 3 & 4:Extended application in audio to denoising, packet loss concealment de-clipping or interpolation. Investigation of several of the week 1 and 2 techniques in real signal application. Competition for best noise reduction performance (Matlab code must be original and run-able by the markers!).

Aims

The aims of the course are to:

• To introduce non-parametric signal analysis and processing metthods in the transform domain
• To apply non-paremetric processing methods to real audio datasets
• To study and apply parametric methods to signals, with an emphasis on probabilistic techniques (likelihood and Bayesian) and incuding model choice
• To carry out an extended application in Matlab in audio processing for noise reduction, packet concealment, declipping or interpolation.

Content

This project introduces signal modelling/processing techniques and applies them to audio and musical signals. First non-parametric methods based on transforms such as the Discrete Fourier Transform (DFT) and its fast variant the FFT are studied and experiments are carried out with windowing, frequency resolution etc. These methods are then applied in an audio noise reduction setting, processing sound signals using overlap-add analysis/processing/synthesis to perform noise reduction. Then parametric models are introduced, with estimation using least squares., maximum likelihood and Bayesian techniques. The autoregressive (AR) model is used as an example here, and model choice is studied within likelihood and Bayesian probabilistic settings. Once again, the techniques are applied in audio signals, using models to perform packet loss concealment and interpolation of missing data in audio, as well as constrained interpolation for clipped and/or heavily quantised signals. Students will have the opportunity to incude their own ideas into the applied schemes and there will be a competition for the best noise reduction performance from test audio datasets, evaluated using mean-sqaured error, perceptual criteria and (informal) listening tests.

FORMAT

Students will work in pairs.

ACTIVITIES

Week 1:Non-parametric analysis processing in the transform domain – DFT, FFT, windows, window length, overlap-add processing and application to denoising of audio signals.

Week 2:Parametric modelling of signals using maximum likelihood and Bayesian techniques, including Bayesian model choice, autoregressive models and sinusoidal models.

Weeks 3 & 4:Extended application in audio to denoising, packet loss concealment de-clipping or interpolation. Investigation of several of the week 1 and 2 techniques in real signal application. Competition for best noise reduction performance (Matlab code must be original and run-able by the markers!).

Aims

The aims of the course are to:

• To introduce non-parametric signal analysis and processing metthods in the transform domain
• To apply non-paremetric processing methods to real audio datasets
• To study and apply parametric methods to signals, with an emphasis on probabilistic techniques (likelihood and Bayesian) and incuding model choice
• To carry out an extended application in Matlab in audio processing for noise reduction, packet concealment, declipping or interpolation.

Content

This project introduces signal modelling/processing techniques and applies them to audio and musical signals. First non-parametric methods based on transforms such as the Discrete Fourier Transform (DFT) and its fast variant the FFT are studied and experiments are carried out with windowing, frequency resolution etc. These methods are then applied in an audio noise reduction setting, processing sound signals using overlap-add analysis/processing/synthesis to perform noise reduction. Then parametric models are introduced, with estimation using least squares., maximum likelihood and Bayesian techniques. The autoregressive (AR) model is used as an example here, and model choice is studied within likelihood and Bayesian probabilistic settings. Once again, the techniques are applied in audio signals, using models to perform packet loss concealment and interpolation of missing data in audio, as well as constrained interpolation for clipped and/or heavily quantised signals. Students will have the opportunity to incude their own ideas into the applied schemes and there will be a competition for the best noise reduction performance from test audio datasets, evaluated using mean-sqaured error, perceptual criteria and (informal) listening tests.

FORMAT

Students will work in pairs.

ACTIVITIES

Week 1:Non-parametric analysis processing in the transform domain – DFT, FFT, windows, window length, overlap-add processing and application to denoising of audio signals.

Week 2:Parametric modelling of signals using maximum likelihood and Bayesian techniques, including Bayesian model choice, autoregressive models and sinusoidal models.

Weeks 3 & 4:Extended application in audio to denoising, packet loss concealment de-clipping or interpolation. Investigation of several of the week 1 and 2 techniques in real signal application. Competition for best noise reduction performance (Matlab code must be original and run-able by the markers!).

Aims

The aims of the course are to:

• To introduce non-parametric signal analysis and processing metthods in the transform domain
• To apply non-paremetric processing methods to real audio datasets
• To study and apply parametric methods to signals, with an emphasis on probabilistic techniques (likelihood and Bayesian) and incuding model choice
• To carry out an extended application in Matlab in audio processing for noise reduction, packet concealment, declipping or interpolation.

Content

This project introduces signal modelling/processing techniques and applies them to audio and musical signals. First non-parametric methods based on transforms such as the Discrete Fourier Transform (DFT) and its fast variant the FFT are studied and experiments are carried out with windowing, frequency resolution etc. These methods are then applied in an audio noise reduction setting, processing sound signals using overlap-add analysis/processing/synthesis to perform noise reduction. Then parametric models are introduced, with estimation using least squares., maximum likelihood and Bayesian techniques. The autoregressive (AR) model is used as an example here, and model choice is studied within likelihood and Bayesian probabilistic settings. Once again, the techniques are applied in audio signals, using models to perform packet loss concealment and interpolation of missing data in audio, as well as constrained interpolation for clipped and/or heavily quantised signals. Students will have the opportunity to incude their own ideas into the applied schemes and there will be a competition for the best noise reduction performance from test audio datasets, evaluated using mean-sqaured error, perceptual criteria and (informal) listening tests.

FORMAT

Students will work in pairs but they will submit individual, independent reports and write their own code.

ACTIVITIES

Week 1:Non-parametric analysis processing in the transform domain – DFT, FFT, windows, window length, overlap-add processing and application to denoising of audio signals.

Week 2:Parametric modelling of signals using maximum likelihood and Bayesian techniques, including Bayesian model choice, autoregressive models and sinusoidal models.

Weeks 3 & 4:Extended application in audio to denoising, packet loss concealment de-clipping or interpolation. Investigation of several of the week 1 and 2 techniques in real signal application. Competition for best noise reduction performance (Matlab code must be original and run-able by the markers!).