In this thesis two versions of a single filter frequency masking narrow-band high-speed recursive digital filter structure, proposed in [1], have been implemented and evaluated considering the maximal clock frequency, the maximal sample frequency and the power consumption. The structures were compared to a conventional filter structure, that was also implemented. The aim was to see if the proposed structure had some benefits when implemented and synthesized, not only in theory. For the synthesis standard cells from AMS csx 0.35 mm CMOS technology were used.

A common problem of hearing aids is whistling caused by feedback from the loudspeaker back to the microphone. A method of reducing the negative effects, caused by the feedback, is called feedback cancellation. A variant of feedback cancellation uses a probe signal, which is applied to the speaker of the hearing aid and is used to continuously estimate the feedback. Oticon A/S has suggested a master's thesis with the purpose of designing and evaluating an algorithm generating a probe signal for feedback cancellation systems. The challenge was to find an inaudible probe signal with as much energy as possible. Two approaches have been investigated for generating a probe signal. In the first approach the psychoacoustic principle of masking was used to estimate how much noise that could be added to a signal without being heard. Psychoacoustic models, including masking, are used in MPEG (Moving Pictures Expert Group) audio coding and one of these models has been examined in the thesis. In the second approach a standard LPC (Linear Prediction Coding) algorithm was used. In both the MPEG and the LPC approach, warped signal processing has been utilized improving the methods. A listening test was performed, evaluating the methods generating the probe signal. The purpose of the test was to determine whether the noise, generated using the MPEG and LPC approach, was inaudible. A hearing aid system with feedback cancellation, using the probe signal, was also simulated. The listening test showed that the noise (probe signal) had to be lowered, much more than expected, to be inaudible. As a consequence, shown in the simulations, the feedback cancellation system, using the probe signal, had trouble identifying the feedback of the hearing aid.

Syftet med studien var att undersöka hur hög precision äldre personer har när de ålderskattar maskerade och omaskerade ansikten. 21 kvinnor och 19 män, totalt 40 deltagare med genomsnittlig ålder på 57,7 år fick skatta åldern på 30 maskerade och 30 omaskerade ansiktsbilder. Samma stimuluspersoner användes med och utan mask. Stimulusbilderna delades upp i grupperna yngre (18-32 år) samt äldre (39-72 år). Resultatet visar att omaskerade ansikten skattas med högre precision än maskerade ansikten och att kvinnor hade en högre precision än män. Inga signifikanta skillnader fanns i deltagarnas förmåga att skatta yngre eller äldre ansikten. Tre interaktionseffekter hittades. (1) Mellan ålder och kön, (2) mellan ålder och maskering och (3) mellan mask, ålder och kön. De systematiska avvikelserna visar att de yngre och omaskerade stimuluspersonernas ålder överskattades samt att de äldre personernas ålder underskattades. Resultatet diskuteras i relation till åldersskattnig och självförankringseffekt. / The purpose of this study was to examine how precise older people can estimate the age of masked and unmasked faces. 21 women and 19 men, 40 participants in total, were shown pictures of 30 masked and 30 unmasked faces, and asked to estimate their age. The same stimulus persons were both masked and unmasked. The pictures were divided into age groups, younger (18-32) and older (39-72). The results showed that unmasked faces were better estimated than masked faces, and that women were more precise than men. There were no significant difference between participants ability to estimate the age of young and old faces. Three interaction effects was found. (1) Between age and sex, (2) between age and mask and (3) between mask, age and sex. The systematic divergences show that the age of the younger and unmasked stimulus persons were overestimated, and that the age of the older persons were underestimated. The results are discussed in relation to age estimation and self-anchoring effects.

Denna experimentella studie undersökte om maskerande brusljud på ovidkommande tal påverkar inlärning av visuell information. Experimentet genomfördes i ett laboratorium med 32 försöksdeltagare. Visuella ord presenterades för försöksdeltagarna samtidigt som auditivt tal från samma semantiska kategori, med eller utan maskerande brusljud, presenterades. De visuella orden skulle återges i valfri ordning. Resultatet av studien visar att ett maskerande brusljud på ovidkommande tal har positiv effekt på inlärningsförmågan. Detta visades av att försökspersonerna mindes fler visuellt presenterade ord samt att de lyckades ignorera det ovidkommande talet bättre då talet maskerades av ett brusljud jämfört med om talet inte maskerades av ett brusljud. Resultaten av studien kan med fördel tillämpas i öppna kontorslandskap. Detta då medarbetare i öppna kontorslandskap ofta utför kognitivt krävande uppgifter i en bullrig miljö innehållande bland annat bakgrundstal. Ett maskerande brusljud kan minska störningen av ovidkommande kontorsljud och ovidkommande tal och på så sätt positivt påverka arbetsprestationen. / This study examined if a masking white noise on irrelevant speech affects the encoding of visual information. An experiment was carried out in a laboratory with 32 participants. The participants were presented to a series of written words and were prompted to recall these words in any order. While the participants studied the written words, irrelevant speech from the same semantic category was presented with or without a masking noise. The participants were told to ignore the irrelevant speech. The results of this study showed that the number of intrusions from the irrelevant speech decreases and the number of recalled written words increases when the irrelevant speech is masked by a white noise compared to irrelevant speech without a masking noise. The findings of this study could be applied in the acoustic design of open-plan offices where cognitive tasks, such as reading comprehension and proofreading, are performed in a noisy environment. A white noise can reduce the intelligibility of office noise and irrelevant speech, which have positive effect on work performance.

The human auditory perception system is much more noise-robust than any state-of the art automatic speech recognition (ASR) system. It is expected that the noise-robustness of speech feature can be improved by employing the human auditory based feature extraction procedure. In this thesis, we investigate modifying the commonly-used feature extraction process for automatic speech recognition systems. A novel frequency masking curve, which is based on modeling the basilar membrane as a cascade system of damped simple harmonic oscillators, is used to replace the critical-band masking curve to compute the masking threshold. We mathematically analyze the coupled motion of the oscillator system (basilar membrane) when they are driven by short-time stationary (speech) signals. Based on the analysis, we derive the relation between the amplitudes of neighboring oscillators, and accordingly insert a masking module in the front-end signal processing stage to modify the speech spectrum. We evaluate the proposed method on the Aurora 2.0 noisy-digit speech database. When combined with the commonly-used cepstral mean subtraction post-processing, the proposed auditory front-end module achieves a significant improvement. The method of correlational masking effect curve combine with CMS can achieves relative improvements of 25.9% over the baseline respectively. After applying the methods iteratively, the relative improvement improves from 25.9% to 30.3%.

none

Response times have been used to explore cognitive and perceptual processes since 1850 (Donders, 1868). The technique has primarily been applied to humans, birds, and terrestrial mammals. Results from two studies are presented here that examine response times in bottlenose dolphins (Tursiops truncatus) and white whales (Delphinapterus leucas). One study concerned response times to stimuli well above the threshold of perceptibility of a stimulus, and the other concerned response times to stimuli near threshold. Two white whales (Delphinapterus leucas) and five Atlantic bottlenose dolphins (Tursiops truncatus) were presented stimuli well above threshold. The stimuli varied in type (tone versus pulse), amplitude, duration, and frequency. The average response time for bottlenose dolphins was 231.9 ms. The average response time for white whales was 584.1 ms. There was considerable variation between subjects within a species, but the difference between species was also found to be significant. In general, response times decreased with increasing stimulus amplitude. The effect of duration and frequency on response time was unclear. Two white whales (Delphinapterus leucas) and four Atlantic bottlenose dolphins (Tursiops truncatus) were given audiometric tests to determine masked hearing thresholds in open waters of San Diego Bay (Ridgway et al., 1997). Animals were tested at six frequencies over a range from 400 Hz to 30 kHz using pure tones. Hearing thresholds varied from 87.5 dB to 125.5 dB depending on the frequency, masking noise intensity and individual animal. At threshold, median response time across frequencies within each animal varied by about 150 ms. The two white whales responded significantly slower (∼670 msec, p<0.0001) than the four dolphins (∼410 msec). As in terrestrial animals, reaction time became shorter as stimulus amplitude increased (Wells, 1913; Stebbins, 1966). Across the two studies, the dolphins as a group were faster in the abovethreshold study than in the nearthreshold study. White whales had longer response times than bottlenose dolphins in both studies. Analysis of response time with an allometric relation based on weight shows that the difference in weight can explain a significant part of the difference in response time.

The human auditory perception system is much more noise-robust than any state-of theart automatic speech recognition (ASR) system. It is expected that the noise-robustness of speech feature vectors may be improved by employing more human auditory functions in the feature extraction procedure. Forward masking is a phenomenon of human auditory perception, that a weaker sound is masked by the preceding stronger masker. In this work, two human auditory mechanisms, synaptic adaptation and temporal integration are implemented by filter functions and incorporated to model forward masking into MFCC feature extraction. A filter optimization algorithm is proposed to optimize the filter parameters. The performance of the proposed method is evaluated on Aurora 3 corpus, and the procedure of training/testing follows the standard setting provided by the Aurora 3 task. The synaptic adaptation filter achieves relative improvements of 16.6% over the baseline. The temporal integration and modified temporal integration filter achieve relative improvements of 21.6% and 22.5% respectively. The combination of synaptic adaptation with each of temporal integration filters results in further improvements of 26.3% and 25.5%. Applying the filter optimization improves the synaptic adaptation filter and two temporal integration filters, results in the 18.4%, 25.2%, 22.6% improvements respectively. The performance of the combined-filters models are also improved, the relative improvement are 26.9% and 26.3%.

<p>In this thesis two versions of a single filter frequency masking narrow-band high-speed recursive digital filter structure, proposed in [1], have been implemented and evaluated considering the maximal clock frequency, the maximal sample frequency and the power consumption. The structures were compared to a conventional filter structure, that was also implemented. The aim was to see if the proposed structure had some benefits when implemented and synthesized, not only in theory. For the synthesis standard cells from AMS csx 0.35 mm CMOS technology were used.</p>

Auditory evoked potentials (AEPs) have become commonly used to measure hearing thresholds in fish. However, it is uncertain how well AEP thresholds match behavioral hearing thresholds and what effect variability in electrode placement and tank composition has on AEPs. In the first experiment, the effect of testing tank composition and electrode placement on AEPs was determined by recording AEPs in the same individual fish in a steel and PVC cylindrical testing tank, and simultaneously recording AEPs from four locations and two different depths on each of 12 goldfish, Carassius auratus. Results from these studies show that tank composition has an effect AEP strength and hearing thresholds, with steel producing lower thresholds for all frequencies. Electrode placement and depth showed no significant effect on hearing thresholds.In the second experiment, the hearing sensitivity of 12 goldfish was measured using both classical conditioning and AEPs in the same setup. For behavioral conditioning, the fish were trained to reduce their respiration rate in response to a 5s sound paired with a brief shock. Once the behavioral audiogram was completed, the AEP measurements were made without moving the fish. The same sound stimuli were presented and the resultant evoked potentials were recorded for 1,000-6,000 averages. AEP input-output functions were then compared to the behavioral audiogram to compare techniques for estimating behavioral thresholds from AEP data. Results show a large range in variability between behavioral and evoked potential thresholds between fish, with the linear regression evoked potential analysis method producing closer thresholds to behavioral methods. In the third study, the effects of masking were examined on the behavioral and evoked potential audiograms. Behavioral thresholds were first determined with a constant masking noise for two frequencies, followed by threshold measurements with no masking noise.