Jason Lukasiak

Learning objects and learning designs: an integrated system for reusable, adaptive and shareable learning content

This paper proposes a system, the Smart Learning Design Framework, designed to support the development of pedagogically sound learning material within an integrated, platform-independent data structure. The system supports sharing, reuse and adaptation of learning material via a metadata-driven philosophy that enables the technicalities of the system to be imperceptible to the author and consumer. The system proposes the use of pedagogically focused metadata to support and guide the author and to adapt and deliver the content to the targeted consumer. A prototype of the proposed system, which provides proof of concept for the novel processes involved, has been developed. The paper describes the Smart Learning Design Framework and places it within the context of alternative learning object models and frameworks to highlight similarities, differences and advantages of the proposed system.

An analysis of the limitations of blind signal separation application with speech

Blind Signal Separation (BSS) techniques are commonly employed in the separation of speech signals, using Independent Component Analysis (ICA) as the criterion for separation. This paper investigates the viability of employing ICA for real-time speech separation (where short frame sizes are the norm). The relationship between the statistics of speech and the assumption of statistical independence (at the core of ICA) is examined over a range of frame sizes. The investigation confirms that statistical independence is not a valid assumption for speech when divided into the short frames appropriate to real-time separation. This is primarily due to the quasi-stationary nature of speech over the temporal short term. We conclude that employing ICA for real-time speech separation will always result in limited performance due to a fundamental failure to meet the strict assumptions of ICA.

Blind speech separation using a joint model of speech production

We propose a new blind signal separation (BSS) technique, developed specifically for speech, that exploits a priori knowledge of speech production mechanisms. In our approach, the autoregressive (AR) structure and fundamental frequency (F0) production mechanisms of speech are jointly modeled. We compare the separation performance of our joint AR-F0 algorithm to existing BSS algorithms that model either speechs AR structure or F0 individually. Experimental results indicate that the joint algorithm demonstrates superior separation performance to both the individual AR algorithm (up to 77% improvement) and F0 (up to 50% improvement) algorithms. This suggests that speech separation performance is improved by employing a BSS model with a more realistic description of the speech production process.

Extending waveform interpolation to wideband speech coding

This paper investigates the extension of waveform interpolation (WI) to wideband speech coding. Included is an analysis of the evolutionary behaviour of wideband speech and the consequences for WI. We highlight problems associated with direct application of the classical WI algorithm applied to wideband speech.

Linear prediction incorporating simultaneous masking

Whilst linear prediction is the cornerstone of most modern speech coders, few of these coders incorporate the perceptual characteristics of hearing into the calculation of the linear predictor coefficients (LPCs). This paper proposes a method of incorporating simultaneous masking into the calculation of the LPCs. This modification requires only a modest increase in computational complexity and results in the linear predictor removing more perceptually important information from the input speech signal. This results in a filter that better models the formants of the input speech spectrum. The net effect is that an improvement in quality is achieved for a given bit rate or alternately a bit rate reduction can be achieved while maintaining perceived quality. These results have been confirmed through subjective listening tests.

Two Channel, Block Adaptive Audio Separation Using the Cross Correlation of Time Frequency Information

TIFCORR is a Blind Signal Separation technique that is well suited to separating audio signals, requiring each signal to be sparse in only a local time-frequency region of their representation [1]. TIFCORR can suffer from inconsistencies in mixing system estimation, thus we present a modified algorithm incorporating k-means clustering [2] to improve estimation robustness. To improve the data efficiency ofTIFCORR, we also include an adaptive weighting function for mixing column estimates. These modifications transform our algorithm into a block adaptive algorithm with the ability to track time-varying mixtures.

Exploiting simultaneously masked linear prediction in a WI speech coder

This paper uses a method of incorporating simultaneous masking into the calculation of a linear predictive filter (SMLPC) as the front end to a 2 kbps waveform interpolation (WI) speech coder. A modification to the masking threshold calculation used in SMLPC is proposed. This modification improves the performance of SMLPC in noise like sections by placing greater emphasis on strongly voiced speech. MOS test results reveal that the modified SMLPC improved the perceptual quality of the WI coder. The improvement is significant for female speakers whilst the quality for male speech is virtually unchanged. This result conflicts with previous results reported for SMLPC where only male speech was improved. The change is attributed to the modification of the masking threshold and confirms that adapting the masking threshold according to the pitch of the speech will allow SMLPC to remove more perceptually important information from all input speech than standard LPC.

Performance of MPEG-7 low level audio descriptors with compressed data

This paper presents a detailed analysis of lossy compression effects on a set of the MPEG-7 low-level audio descriptors. The analysis results show that lossy compression has a detrimental effect on the integrity of practical search and retrieval schemes that utilize the low level audio descriptors. Methods are then proposed to reduce the detrimental effects of compression in searching schemes. These proposed methods include multi-frame searching and machine learning derived prediction. The proposed mechanisms greatly reduce the effect of compression on the set of MPEG-7 descriptors; however, future scope is identified to develop new audio descriptors that account for compression effects in their structure.

Low bit rate wideband WI speech coding

This paper investigates waveform interpolation (WI) applied low bit rate wideband speech coding. An analysis of the evolutionary behaviour of wideband characteristic waveforms (CWs) shows that direct application of the classical WI algorithm may not be appropriate for wideband speech. We propose a modification whereby CW quantisation is performed using classical WI decomposition for the low frequency region and noise modelling for the high frequency region. Wideband WI coders incorporating this modification and operating at 4 kbps and 6 kbps are described. Subjective testing of these coders shows that WI is a promising approach to low bit rate wideband speech compression.

A Sequential Approach to Sparse Component Analysis

A sequential approach to sparse component analysis (SeqTIF) is proposed in this paper. Although SeqTIF employs the estimation process of the simultaneous TIFROM algorithm, a source cancellation and deflation technique are also incorporated to sequentially estimate speech signals in the mixture. Results indicate that SeqTIFs separation performance shows a clear improvement upon the simultaneous TIFROM approach, due to the less restrictive assumptions it places upon the signals in the mixture. In particular, the analysis indicates SeqTIFs data efficiency is high, enabling the sequential approach to track a time-varying mixture with much greater accuracy than the simultaneous algorithm. Furthermore, SeqTIF is a more flexible approach, free from the constraints that a simultaneous approach places upon the mixing system