Floor vibration serviceability is a primary design consideration for steel framed floors. Designers in North America typically use the AISC Design Guide 11 methods to check this limit state, but its methods are difficult to apply to atypical floor framing. Finite element analysis is a logical choice for predicting vibration response to walking, but simplified designer-friendly procedures are not available. Three relatively simple, experimentally verified methods of predicting low frequency floor vibration due to walking are presented in this dissertation. The methods are based on finite element analysis of the floor system, are applicable to a wide range of situations, and are intended to be no more complicated than is justified by the current ability to predict modal properties. The first method is to predict the acceleration response using response history analysis with individual footstep forces as the loading function. The second method also uses response history analysis to predict the acceleration waveform, but with a Fourier series representation of the load. The third method is a simplified frequency domain method in which the predicted accelerance frequency response function is used to predict the steady-state response to walking which is reduced to account for incomplete resonant build-up. A two year experimental program including three laboratory specimens, a four bay full-scale mockup, and two steel-framed building floors, was completed at Virginia Tech. These floor systems represent a wide cross-section of the steel framed floor systems used in North America. Modal tests were performed using an electrodynamic shaker and experimental modal analysis techniques were used to estimate the modal properties: natural frequencies, mode shapes, and damping ratios. Responses to walking excitation were measured several times in each tested bay for individuals walking at subharmonics of natural frequencies. During each test, the walker crossed the middle of the bay using a metronome to help maintain the intended cadence. The test with maximum response represents the maximum peak acceleration that can be reasonably expected to occur due to a single walker. The proposed methods were used, with measured damping ratios and walker weights, to predict the modal properties and responses to walking for comparison with measured values. The methods were found to be reasonably accurate, contain significant data dispersion, and be on the conservative side. The results of these comparisons were used to develop design recommendations, including reduction factors to account for the conservatism. The design methods were used to predict the modal properties and responses to walking in a "blind" manner using only information that would be available to a designer. Comparisons of measurements and predictions were used to determine the accuracy of the proposed prediction methods, which were found to be sufficiently accurate for design usage. / Ph. D.

The subject of this thesis is perceptual sensitivity with respect to subtle frequency-based and temporal manipulations in speech, music and mixed stimuli. We hypothesize that an individual's sensitivity to variation in all three types of stimuli should be similar (i.e. a correlation should exist), seeing that findings in evolutionary biology, neurosciences, psy- chology and experimental phonetics are pointing towards a relatively strong link between the mechanisms of perception in speech and music. Our listening experiment revealed mostly intermediate correlations; additionally, we argue that by employing syntactically less complicated stimuli, which would target specifically fundamental sensitivity without requiring a complex syntactic analysis in parallel, even more robust correlations could be obtained. While the influence of prior formal linguistic education on performance in the test was negligible, the influence of musical experience was considerable, which lends further support to the idea of simplifying especially the music stimuli in future research. Key words: music, speech, perception, sensitivity, correlation

Este trabajo propone una serie de métodos eficientes para convertir una señal de audio musical polifónica (WAV, MP3) en una partitura (MIDI).

Improved cochlear implant (CI) designs and surgical techniques have allowed CI patients to retain acoustic hearing in the implanted ear post-operatively. These EAS (electric-acoustic stimulation) CI users listen with a combination of acoustic and electric hearing in the same ear. While electric hearing alone improves speech recognition in quiet, preserved acoustic hearing allows EAS CI users to outperform traditional CI users in speech recognition in noise and demonstrate “speech masking release,” an improvement in speech recognition in temporally fluctuating noise relative to steady noise. Masking release is arguably an ecologically valid metric, as listeners often attend to target speech embedded in fluctuating competing speech. Improved speech recognition outcomes have been attributed to the spectral and temporal resolution provided by acoustic hearing. However, the relationship between spectral and temporal resolution and outcomes in EAS CI users is not clear. This study evaluated speech masking release, spectral ripple density discrimination thresholds, and fundamental frequency difference limens (f0DLs) in EAS CI users. Both the ripple and f0DL tasks are thought to measure underlying spectral resolution and temporal fine structure. EAS CI subjects underwent testing in three listening modes: acoustic-only, electric-only, and acoustic+electric. Comparisons across listening modes allowed the benefit provided by acoustic hearing to be quantified. It was hypothesized that speech masking release, spectral ripple density discrimination thresholds, and f0DLs would be poorest with electric-only hearing and would improve in the acoustic-only and acoustic+electric listening modes. This would reflect the benefit of preserved acoustic hearing. It was also hypothesized that speech masking release would correlate with spectral ripple density discrimination thresholds and f0DLs, reflecting the roles of spectral and temporal fine structure cues. Lastly, it was hypothesized that EAS CI users with more residual hearing (lower audiometric thresholds) would perform better on all three tasks. Speech masking release was evaluated using a 12-alternative-forced-choice (AFC) spondee recognition in noise task. The noise was a two-talker and a ten-talker babble presented at -5 dB SNR, and masking release was quantified as the difference in spondee recognition in two-talker babble relative to ten-talker babble. Spectral ripple density discrimination thresholds were assessed in a 3-AFC task using a broadband stimulus that contained spectral peaks and valleys logarithmically spaced on the frequency axis. The spacing between spectral peaks (ripple density) was varied to determine the threshold at which listeners could no longer resolve the individual spectral peaks. F0DLs were assessed via a 3-AFC task using a broadband harmonic complex with a baseline f0 = 110 Hz. The f0 of the test intervals was varied to determine the smallest change in f0 that the listener could detect. Results showed that performance in all three measures was poorest when EAS CI users were tested using electric-hearing only, with significant improvements when tested in the acoustic-only and acoustic+electric listening modes. F0DLs, but not spectral ripple density discrimination thresholds or audiometric thresholds, significantly correlated with speech masking release. Speech masking release also significantly correlated with open-set AzBio sentence recognition in noise scores obtained from clinical records. Results indicated that preservation of residual acoustic hearing allows for speech masking release, likely due to access to temporal fine structure cues provided by residual hearing. The significant correlation between speech masking release and sentence recognition in noise indicates that the ability to extract target speech embedded in temporally fluctuating competing speech is important for speech recognition in noise. Funded by National Institutes of Health/National Institutes on Deafness and Other Communication Disorders (NIH/NIDCD) P50 DC000242, American Speech-Language-Hearing Foundation Student Research Grant, and American Academy of Audiology Student Investigator Research Grant.

Bei der Überführung eines wissenschaftlichen Laborsystems zur automatischen Spracherkennung in eine reale Anwendung ergeben sich verschiedene praktische Problemstellungen, von denen eine der Verlust an Erkennungsleistung durch umgebende akustische Störungen ist. Im Gegensatz zu additiven Störungen wie Lüfterrauschen o. ä. hat die Wissenschaft bislang die Störung des Raumhalls bei der Spracherkennung nahezu ignoriert. Dabei besitzen, wie in der vorliegenden Dissertation deutlich gezeigt wird, bereits geringfügig hallende Räume einen stark störenden Einfluss auf die Leistungsfähigkeit von Spracherkennern. Mit dem Ziel, die Erkennungsleistung wieder in einen praktisch benutzbaren Bereich zu bringen, nimmt sich die Arbeit dieser Problemstellung an und schlägt Lösungen vor. Der Hintergrund der wissenschaftlichen Aktivitäten ist die Erstellung von funktionsfähigen Sprachbenutzerinterfaces für Gerätesteuerungen im Wohn- und Büroumfeld, wie z.~B. bei der Hausautomation. Aus diesem Grund werden praktische Randbedingungen wie die Restriktionen von embedded Computerplattformen in die Lösungsfindung einbezogen. Die Argumentation beginnt bei der Beschreibung der raumakustischen Umgebung und der Ausbreitung von Schallfeldern in Räumen. Es wird theoretisch gezeigt, dass die Störung eines Sprachsignals durch Hall von zwei Parametern abhängig ist: der Sprecher-Mikrofon-Distanz (SMD) und der Nachhallzeit T60. Um die Abhängigkeit der Erkennungsleistung vom Grad der Hallstörung zu ermitteln, wird eine Anzahl von Erkennungsexperimenten durchgeführt, die den Einfluss von T60 und SMD nachweisen. Weitere Experimente zeigen, dass die Spracherkennung kaum durch hochfrequente Hallanteile beeinträchtigt wird, wohl aber durch tieffrequente. In einer Literaturrecherche wird ein Überblick über den Stand der Technik zu Maßnahmen gegeben, die den störenden Einfluss des Halls unterdrücken bzw. kompensieren können. Jedoch wird auch gezeigt, dass, obwohl bei einigen Maßnahmen von Verbesserungen berichtet wird, keiner der gefundenen Ansätze den o. a. praktischen Einsatzbedingungen genügt. In dieser Arbeit wird die Methode Harmonicity-based Feature Analysis (HFA) vorgeschlagen. Sie basiert auf drei Ideen, die aus den Betrachtungen der vorangehenden Kapitel abgeleitet werden. Experimentelle Ergebnisse weisen die Verbesserung der Erkennungsleistung in halligen Umgebungen nach. Es werden sogar praktisch relevante Erkennungsraten erzielt, wenn die Methode mit verhalltem Training kombiniert wird. Die HFA wird gegen Ansätze aus der Literatur evaluiert, die ebenfalls praktischen Implementierungskriterien genügen. Auch Kombinationen der HFA und einigen dieser Ansätze werden getestet. Im letzten Kapitel werden die beiden Basistechnologien Stimm\-haft-Stimmlos-Entscheidung und Grundfrequenzdetektion umfangreich unter Hallbedingungen getestet, da sie Voraussetzung für die Funktionsfähigkeit der HFA sind. Als Ergebnis wird dargestellt, dass derzeit für beide Technologien kein Verfahren existiert, das unter Hallbedingungen robust arbeitet. Es kann allerdings gezeigt werden, dass die HFA trotz der Unsicherheiten der Verfahren arbeitet und signifikante Steigerungen der Erkennungsleistung erreicht. / Automatic speech recognition (ASR) systems used in real-world indoor scenarios suffer from performance degradation if noise and reverberation conditions differ from the training conditions of the recognizer. This thesis deals with the problem of room reverberation as a cause of distortion in ASR systems. The background of this research is the design of practical command and control applications, such as a voice controlled light switch in rooms or similar applications. Therefore, the design aims to incorporate several restricting working conditions for the recognizer and still achieve a high level of robustness. One of those design restrictions is the minimisation of computational complexity to allow the practical implementation on an embedded processor. One chapter comprehensively describes the room acoustic environment, including the behavior of the sound field in rooms. It addresses the speaker room microphone (SRM) system which is expressed in the time domain as the room impulse response (RIR). The convolution of the RIR with the clean speech signal yields the reverberant signal at the microphone. A thorough analysis proposes that the degree of the distortion caused by reverberation is dependent on two parameters, the reverberation time T60 and the speaker-to-microphone distance (SMD). To evaluate the dependency of the recognition rate on the degree of distortion, a number of experiments has been successfully conducted, confirming the above mentioned dependency of the two parameters, T60 and SMD. Further experiments have shown that ASR is barely affected by high-frequency reverberation, whereas low frequency reverberation has a detrimental effect on the recognition rate. A literature survey concludes that, although several approaches exist which claim significant improvements, none of them fulfils the above mentioned practical implementation criteria. Within this thesis, a new approach entitled 'harmonicity-based feature analysis' (HFA) is proposed. It is based on three ideas that are derived in former chapters. Experimental results prove that HFA is able to enhance the recognition rate in reverberant environments. Even practical applicable results are achieved when HFA is combined with reverberant training. The method is further evaluated against three other approaches from the literature. Also combinations of methods are tested. In a last chapter the two base technologies fundamental frequency (F0) estimation and voiced unvoiced decision (VUD) are evaluated in reverberant environments, since they are necessary to run HFA. This evaluation aims to find one optimal method for each of these technologies. The results show that all F0 estimation methods and also the VUD methods have a strong decreasing performance in reverberant environments. Nevertheless it is shown that HFA is able to deal with uncertainties of these base technologies as such that the recognition performance still improves.

Syftet med föreliggande studie var att undersöka sambandet mellan subglottalt tryck och ljudtrycksnivå, och undersöka inverkan av kön och tonhöjd. Röststyrkan är starkt beroende av subglottalt tryck och sambandet har analyserats i ett flertal studier, som alla visar på ett linjärt förhållande mellan logaritmen av subglottala trycket och ljudtrycksnivån (SPL). Schutte (1980) analyserade sambandet för 21 kvinnor och  24 män som producerade ett stort antal mätdata med olika SPL och tonhöjd. Tryck mättes med ballong i esofagus. Tanaka & Gould (1982) analyserade 4 kvinnor och 6 män på tre ljudnivåer på bekväm tonhöjd; svag, normal och stark ljudstyrka. Subglottalt tryck uppmättes med pletysmograf, med försökspersonen sittandes i lufttät box. Pressad fonation karaktäriseras av ett högt subglottalt tryck och en förhållandevis låg SPL, så sambandet mellan tryck och SPL torde påverkas av glottal adduktion och troligen också av tonhöjd. Därför borde normativa data från röstfriska personer vara av intresse. I denna studie producerade 16 kvinnor och 15 män sekvenser av stavelsen [pæ] på fyra tonhöjder, jämnt fördelade över en oktav. Regressionsanalys användes för att approximera sambandet mellan SPL och logaritmen av subglottalt tryck samt för att beräkna genomsnittlig SPL-ökning för dubblat subglottalt tryck och beräknad SPL vid 10 cm H2O. Resultatet visar som väntat ett mycket starkt samband mellan subglottalt tryck och ljudtrycksnivån, med en korrelationskoefficient på 0.835 respektive 0.826 för kvinnor respektive män. En fördubbling av subglottalt tryck gav en genomsnittlig SPL-ökning av 11,5 dB (SD 3.8) för kvinnor och 10,0 dB (SD 2,7) för män. Skillnaden mellan kvinnor och män var här signifikant, vilket ger stöd för att använda separativa normativa värden för kvinnor och män. Genomsnittlig beräknad nivå vid 10 cm H2O var 83,6 dB (SD 3,9) för kvinnor och 82,2 dB (SD 4,6) för män beräknat till 15 cm mikrofonavstånd. Sambandet mellan subglottalt tryck och SPL skiljde sig en aning beroende på tonhöjd, men skillnaden var ej signifikant. Trots de relativt höga standardavvikelserna tyder resultatet på att det vore värt att studera i vad mån avvikelser från de potentiellt normativa värdena skulle kunna vara tecken på någon slags fonatorisk dysfunktion. / The purpose of this study was to examine the relationship between subglottal pressure and sound pressure level (SPL), and to study the importance of gender and fundamental frequency in this relationship. Vocal loudness is strongly dependent on subglottal pressure. The relation between them has been analyzed in several investigations, all showing a linear relationship between the SPL and the log of the pressure. For example, Schutte (1980) analyzed the relation in 21 female and 24 male subjects who produced a great number of samples at different degrees of vocal loudness and at the subjects’ preferred pitch. Pressure was measured by means of an esophageal balloon. Tanaka and Gould (1982) analyzed 10 subjects each producing vowels at three loudness levels at comfortable pitch. Pressure data were obtained from a plethysmograph, with the subject sitting in an airtight box. Pressed phonation is characterized by a high subglottal pressure producing a comparatively low SPL, so hence the pressure – SPL relationship would be affected by glottal adduction and possibly also by F0. Therefore normative data from healthy voices should be of interest. In the present study 16 female and 15 male normal voices were asked to produce diminuendo and crescendo sequences of the syllable [pæ] at four pitches, equidistantly spaced within an octave. Trendlines were used to approximate the relation between SPL and the log of subglottal pressure. The resulting regression equations were used to calculate the average SPL increase for doubling of pressure and the SPL produced by a pressure of 10 cm H2O. The results showed an average correlation coefficient of 0.835 and 0.826 for female and male subjects. A doubled pressure produced an SPL increase of 11.5 dB (SD 3.8) and 10.0 dB (SD 2.7) for the female and the male voices. The difference between female and male voices was significant, which supports use of separate normative values for female and male voices. On average, a subglottal pressure of 10 cm H2O produced an SPL @ 0.15 m of 83.6 dB (SD 3.9) and 82.2 dB (SD 4.6) for the female and the male voices. The relationship between subglottal pressure and SPL depended somewhat on fundamental frequency, but the difference was not significant. In spite of the relatively high standard errors the results indicate that it would be worth to study in what extent differences from the potentially normative values of this study may be a sign of some sort of phonatory dysfunction.

This diploma thesis with estimation of pitch period of the human voice. The paper listed some of the methods for Estimation of pitch period and method for preprocessing and final processing of the signal after application of functions to determine the frequency of the pitch period in graphical user interface.

Diploma’s thesis is focused on determining person’s height from spoken utterance. First part of the work evaluates present situation and refers to the published studies. Knowledge gained in these studies was used in this thesis. Study with the best results according to estimated height of the speakers was chosen. The experiment realized in the chosen study was performed in this work. The system for the estimation of the height of the speakers based on the speech signal was created. This system was successfully tested by using several acoustic features on spoken utterances from TIMIT database.

No description available.

The presented research examined three areas: best practices in high quality dynamic testing of in-situ floor systems, extensive dynamic testing of three bare (non-fit out) in-situ multi-bay steel composite floors to estimate their dynamic parameters/response and to identify trends in dynamic behavior, and development of a set of fundamental finite element (FE) modeling techniques to adequately represent the dynamic response of steel composite floors for the purpose of evaluating vibration serviceability. The measurement, analysis, and computation of a floor's accelerance frequency response function (FRF) is the core premise linking all areas of the presented research. The burst chirp signal using an electrodynamic shaker is recommended as the most accurate and consistent source of excitation for acquiring high quality measurements suitable for use in parameter estimation, operating deflection shape animation, and calibration/validation of FE models. A reduced mid-bay testing scheme is recommended as a time-saving alternative to modal testing over a full coverage area, provided the only desired estimated parameters are frequencies, damping, and mid-bay acceleration response. Accelerance FRFs were measured with an electrodynamic shaker located within 23 unique bays on the three tested floors. Dominant frequencies ranged from 4.85 Hz to 9 Hz and measured estimates of damping varied considerably, ranging from 0.44% to 2.4% of critical (0.5%-1.15% was typical). Testing showed several mode shapes were localized to just a few bays and not all modes were adequately excited by forcing at a single location. The quality of the estimated mode shapes was significantly improved using multi-reference modal testing. FE models for the tested floors were developed based on high quality measured data and were shown to provide adequate representations of measured floor behavior. Fundamental techniques are presented for modeling mass, stiffness, boundary conditions, and performing dynamic analysis. A method of evaluating vibration serviceability was proposed using the FE model's computed accelerance FRF for comparison with a design accelerance curve that represents an acceleration response threshold in the frequency domain. An example design accelerance curve is presented based on current serviceability guidelines for acceleration tolerance and effective harmonic forces due to human activities such as walking. / Ph. D.