Søren Bech

On the relative importance of spatial and timbral fidelities in judgments of degraded multichannel audio quality

Mean opinion score ratings of reproduced sound quality typically pool all contributing perceptual factors into a single rating of basic audio quality. In order to improve understanding of the trade-offs between selected sound quality degradations that might arise in systems for the delivery of high quality multichannel audio, it was necessary to evaluate the influence of timbral and spatial fidelity changes on basic audio quality grades. The relationship between listener ratings of degraded multichannel audio quality on one timbral and two spatial fidelity scales was exploited to predict basic audio quality ratings of the same material using a regression model. It was found that timbral fidelity ratings dominated but that spatial fidelity predicted a substantial proportion of the basic audio quality.

Spatial aspects of reproduced sound in small rooms

This paper reports on the influence of individual reflections on the auditory spatial aspects of reproduced sound. The sound field produced by a single loudspeaker positioned in a normal listening room has been simulated using an electroacoustical synthesis of the direct sound, 17 individual reflections and the reverberant field. The threshold of detection was measured using the method of adjustment for five reflections using three subjects for noise and speech. The thresholds have been measured for two simulated situations (1) a loudspeaker with a frequency independent directivity characteristics and frequency independent absorption coefficients of the room surfaces and (2) a loudspeaker with directivity similar to a standard two-way system and absorption coefficients according to measurements of real materials. The results have shown that subjects can reliably distinguish between timbre and spatial aspect of the sound field, that the spectral energy above 2 kHz of the individual reflection determines the importance of the reflection for the spatial aspects, and that only the first order floor reflection will contribute to the spatial aspects.

Timbral aspects of reproduced sound in small rooms. I

This paper reports some of the influences of individual reflections on the timbre of reproduced sound. A single loudspeaker with frequency-independent directivity characteristics, positioned in a listening room of normal size with frequency-independent absorption coefficients of the room surfaces, has been simulated using an electroacoustic setup. The model included the direct sound, 17 individual reflections, and the reverberant field. The threshold of detection and just-noticeable differences for an increase in level were measured for individual reflections using eight subjects for noise and speech. The results have shown that the first-order floor and ceiling reflections are likely to individually contribute to the timbre of reproduced speech. For a noise signal, additional reflections from the left sidewall will contribute individually. The level of the reverberant field has been found to have an effect on the contribution of the individual reflections. An increase in the level of individual reflections are most likely to be audible for the first-order floor and ceiling reflections, and certain reflections from the sidewalls.

Relationships between experienced listener ratings of multichannel audio quality and naïve listener preferences.

The preferences of a large number of naïve listeners were elicited in response to a selection of multichannel audio items that had been degraded in quality by using band-limiting and down-mixing algorithms. Relationships were sought between these preference ratings and the quality judgements of experienced listeners in an attempt to determine whether one could be predicted from the other. Results suggest that a simple regression model can be used to do this with adequate results, but that a better prediction can be successfully based on experienced listener ratings of timbral and spatial fidelity. There is a difference between naïve and experienced listeners in the weightings of the fidelities and their relationship to overall quality.

Audio coding in wireless acoustic sensor networks

In this paper, we consider the problem of source coding for a wireless acoustic sensor network where each node in the network makes its own noisy measurement of the sound field, and communicates with other nodes in the network by sending and receiving encoded versions of the measurements. To make use of the correlation between the sources available at the nodes, we consider the possibility of combining the measurement and the received messages into one single message at each node instead of forwarding the received messages and separate encoding of the measurement. Moreover, to exploit the correlation between the messages received by a node and the node?s measurement of the source, we propose to use the measurement as side information and thereby form a distributed source coding (DSC) problem. Assuming that the sources are Gaussian, we then derive the rate-distortion function (RDF) for the resulting remote DSC problem under covariance matrix distortion constraints. We further show that for this problem, the Gaussian source is the worst to code. Thus, the Gaussian RDF provides an upper bound to other sources such as audio signals. We then turn our attention to audio signals. We consider an acoustical model based on the room impulse response (RIR) and provide simulation results for the rate-distortion performance in a practical setup where a set of microphones record the sound in a standard listening room. Since our reconstruction scheme and distortion measure are defined over the direct sound source, coding and dereverberation are performed in a joint manner. HighlightsWe treat the problem of source coding for wireless acoustic sensor networks.We consider vector sources to make use of the time correlation in the audio sequences.We use the measurements at the receiving nodes as side information using distributed source coding.We derive local rate-distortion functions to be used for rate allocation for an optimal sum-rate.Our encoding/decoding process is joint with dereverberation.

Image quality assessment by expert and non-expert viewers

The difference between expert and non-expert viewers in assessing image quality is evaluated in two experiments. The assessment performance in terms of discrimination ability and reproducibility is measured for both groups. The results of these experiments suggest that both groups of viewers exhibit the same assessment behavior when judging the level of a given image quality attribute, such as e.g. sharpness. When judging overall image quality, however, expert viewers seem to weight various attributes differently as compared to non-expert viewers.

Sharpness, sharpness related attributes and their physical correlates

In this paper we present a metric that predicts differences in perceptual sharpness from physical parameters extracted from CCD camera recordings of sine-grating and step-profile test images. Sharpness differences were induced by using five different high-end television sets to display natural color images as well as by using blurred versions of these images. The metric is a two-step model that uses the perceptual strengths of three sharpness-related attributes on an intermediate level between the physical parameters and perceptual sharpness. The data obtained from category scaling experiments indicate that perceptual strength of sharpness is a linear combination of the perceptual strengths of the attributes detail rendering and contour rendering. It was found that differences in detail rendering correlate with differences in modulation depth of the sine gratings and that differences in contour rendering can be explained from the slope and overshoot parameters of the step-profiles.

Distributed Remote Vector Gaussian Source Coding for Wireless Acoustic Sensor Networks

In this paper, we consider the problem of remote vector Gaussian source coding for a wireless acoustic sensor network. Each node receives messages from multiple nodes in the network and decodes these messages using its own measurement of the sound field as side information. The nodes measurement and the estimates of the source resulting from decoding the received messages are then jointly encoded and transmitted to a neighbouring node in the network. We show that for this distributed source coding scenario, one can encode a so-called conditional sufficient statistic of the sources instead of jointly encoding multiple sources. We focus on the case where node measurements are in form of noisy linearly mixed combinations of the sources and the acoustic channel mixing matrices are invertible. For this problem, we derive the rate-distortion function for vector Gaussian sources and under covariance distortion constraints.

Modeling the Subjective Quality of Highly Contrasted Videos Displayed on LCD With Local Backlight Dimming

Local backlight dimming is a technology aiming at both saving energy and improving visual quality on television sets. As the rendition of the image is specified locally, the numerical signal corresponding to the displayed image needs to be computed through a model of the display. This simulated signal can then be used as input to objective quality metrics. The focus of this paper is on determining which characteristics of locally backlit displays influence quality assessment. A subjective experiment assessing the quality of highly contrasted videos displayed with various local backlight-dimming algorithms is set up. Subjective results are then compared with both objective measures and objective quality metrics using different display models. The first analysis indicates that the most significant objective features are temporal variations, power consumption (probably representing leakage), and a contrast measure. The second analysis shows that modeling of leakage is necessary for objective quality assessment of sequences displayed with local backlight dimming.

Distributed Remote Vector Gaussian Source Coding with Covariance Distortion Constraints

In this paper, we consider a distributed remote source coding problem, where a sequence of observations of source vectors is available at the encoder. The problem is to specify the optimal rate for encoding the observations subject to a covariance matrix distortion constraint and in the presence of side information at the decoder. For this problem, we derive lower and upper bounds on the rate-distortion function (RDF) for the Gaussian case, which in general do not coincide. We then provide some cases, where the RDF can be derived exactly. We also show that previous results on specific instances of this problem can be generalized using our results. We finally show that if the distortion measure is the mean squared error, or if it is replaced by a certain mutual information constraint, the optimal rate can be derived from our main result.