Speech and Music, Perception
This work presents preliminary results of a study examining the spherical directivity of the Greek singing voice. Unlike previous works focusing mainly on the horizontal plane, this study reports results on three elevations (+30o, 0o, and -30o). Six professional singers in Greek Classical and Byzantine music were recorded signing in a sound-treated space using a 28-microphone array mounted on a semi-spherical thin-shell structure. The collected dataset consists of short song excerpts and vowel sounds at different pitches. Directivity results across all elevation angles are reported based on overall and per third-octave band RMS levels.
This work investigates the possibilities of using serious games and gamification technologies as means of entertainment and informal education. The focus is on Speech Emotion Recognition Systems (SER), which can be a useful tool both in the industry and in the academia. Most commonly, SER systems rely on machine learning algorithms that require huge amounts of ground-truth data to be trained and achieve a decent performance. In this direction, the proposed game “Raise your Voice” anticipates to provide a game-based approach for collecting of emotionally loaded audio recordings, based on a five-class classification schema (anger, disgust, fear, happiness and sadness). The outmost target of the game is to attract a wider audience and collect data fast by providing also an entertaining experience.
Vision is one of the human senses that plays a dominant role in maintaining one's mental and not only health. It is very important to recognize objects in the space and navigate avoiding dangers and possible injuries. In in this paper we developed an audio application for navigation and orientation in enclosed spaces, helping mainly people with vision impairments but (possibly) also people with other disabilities and diseases related to age or even for the general public. The application is connected with an indoor localization system and with the help of the mobile sensors it can determine the position and orientation of the person in the space. Through voice messages, using a variety of technologies such as databases (Database - DB), the Internet of Things (IoT), Bluetooth Low Energy (BLE) technologies, but also artificial intelligence algorithms (Artificial Intelligence - AI), the application returns the corresponding audio message that corresponds to our position and orientation.
In this study, a differentiable implementation of an auditory model is presented, that can be parameterized to account for hearing loss based on the pure-tone audiogram. This model is employed as an objective function for training a deep learing-based speech enhancement system. The proposed criterion outperforms existing objectives in terms of objective intelligibility and quality metrics, when evaluated for moderate and severe hearing loss. Additionally, significant adaptation of the enhanced speech dynamics to the hearing loss characteristics is observed. Finally, experimental results indicate the potential of deep learning methods end-to-end personalized speech processing schemes for hearing aids.
Η ασθένεια του Parkinson (PD) είναι μια εκφυλιστική ασθένεια του κεντρικού νευρικού συστήματος που επηρεάζει κυρίως το κινητικό σύστημα . Επιπλέον προκαλεί εξασθένηση της φωνής στους ασθενείς σε ποσοστό άνω του 90%, κυρίως σε άτομα με ηλικία άνω των 50 ετών. Συγκεκριμένα η ομιλία των ατόμων που πάσχουν από τη συγκεκριμένη ασθένεια χαρακτηρίζεται από μειωμένη ένταση, βραχνάδα, ατελή άρθρωση, μειωμένο άγχος, σύντομες βιασύνες στον λόγο, δυσαρμονία και δυσκολία στην αναπνοή . Αρκετές έρευνες έχουν πραγματοποιηθεί με σκοπό την αναγνώριση της νόσου μέσω της φωνής, χρησιμοποιώντας είτε τεχνικές μηχανικής μάθησης είτε νευρωνικά δίκτυα. Η εργασία αναδεικνύει την προσπάθεια να διαγνωστεί η νόσος σε πρώιμο στάδιο καθώς επίσης και τη χρήση των παραμέτρων που χρησιμοποιούνται για να τροφοδοτήσουν τις τεχνικές μηχανικής μάθησης. Στα αποτελέσματα θα παρουσιαστούν τα ποσοστά ακρίβειας εκτίμησης των τεχνικών μηχανικής μάθησης που έχουν χρησιμοποιηθεί για την αναγνώριση ομιλίας ατόμων με Parkinson. Ο αλγόριθμος χρησιμοποιεί δυο τεχνικές για την αναγνώριση των συμπτωμάτων της ασθένειας, τον μετασχηματισμό κυματιδίων(wavelet transform) σε συνδυασμό με την ανάλυση πολλαπλών επιπέδων(Multiresolution Analysis or MRA) και διάφορα χαρακτηριστικά της φωνής που θα αναλυθούν παρακάτω. Σκοπός αυτής της έρευνας είναι η πιθανή έγκαιρη διάγνωση της ασθένειας, έτσι ώστε να ξεκινήσει η αντιμετώπιση της.
In this study, the validity of a Machine Learning multiclass classification process is examined, as to classify a speaker in a cognitive decline stage, aiming to develop a simple screening test. The target classes comprise Cognitively Healthy controls, Subjective Cognitive Decline and Early & Late Mild Cognitive Impairment. Speech data was collected from structured interviews on 84 people, split in stages of increasing required levels of cognitive difficulty. Audio features were extracted based on Silence, Prosody and Zero-Crossings, as well as on the feature vectors’ differences between stages, and were evaluated with the Random Forest, Extra-Trees and Support Vector Machines classifiers. The best classification was achieved using models trained with stage differences features (on SVM), resulting in a mean accuracy of 80.99±3.29%.
The present work examines the relationship between sound timbre and the visual representation of the concepts that characterize it. Based on the LTM standard (Luminance - Texture - Mass) and the work compares the current experimental findings with results of previous experiments in which the listener responded using verbal scales and descriptions. Through an online platform created for this purpose, 24 sounds from 61 participants were examined, who did not report any particular hearing problems. During the experimental process, participants had to match the sound stimulus they heard with the given images. Our results showed that the correlation between visual and verbal responses was quite large, mainly for the qualities of luminance-brightness and texture. This relationship is confirmed by previous experiments concerning the same qualities, but the response of the listeners was done in a different way.
In this work, we propose an Automated Audio Detection (AAC) model, which produces captions in Greek language. Initially we introduce a new AAC dataset, based on the Clotho dataset. We also present the extraction of acoustic and semantic information from sound files and captions respectively. The architecture is based on deep learning techniques, more specifically, a sequence to sequence architecture with Gated Recurrent Units as Encoders and Decoders.
Sound Editing and Reproduction
During the development of an educational platform on fire fighting in a CAVE (Cave Automatic Virtual Environment) , special multichannel audio reproduction techniques were used. The virtual sound sources are dynamic in time, and they create the soundscape that is reproduced by the sound system implemented in the CAVE. At the current study it is described the sound system, the techniques used for creation of the soundscape (VBAP and HOA) , and finally the results of the perceptual quality tests of the sound system reproduction accuracy are presented.
Extrusion is a reliable industrial process for the manufacture of laminates of any cross-sectional shape and with a practical length of several tens of metres. Laminatess can now be made from any malleable metal, including aluminium. It is a simple, highly reliable industrial technique based on hydraulic ram extrusion. A successful industrial practice is also based on the elimination of any unplanned maintenance and rapid repair of faults, so their detection is a task of the utmost importance. The Achilles heel of hydraulic rams can be considered the mechanical aging of its seal. Early detection of impending total gasket failure is critical to the efficiency of industrial extrusion. One of the most widespread and proven in practice techniques for damage detection and localization is Acoustic Emission, with hundreds of applications including damage detection in structures. This project aims to use the Acoustic Emissions on a piston extrusion as generated and recorded during a typical extrusion, for the purpose of detecting mechanical wear of the piston seal, as a measure of early and accurate damage prediction.
This study presents the design and development of an automated speech audiometry application in the Greek language. The testing procedure employs algorithmically synthesized nonsense monosyllabic speech stimuli that are presented via commercial earphones or hearing aids. Ten participants repeatedly conducted speech recognition tests in quiet and in noise using the proposed application. Test-retest reliability results and stimulus discernibility analysis indicate the limitations and potential extensions of employing synthetic speech stimuli and mobile devices for conducting automated speech audiometry tests in Greek.
This work focuses on the large-scale analysis and evaluation of headphone responses. For the purposes of this work, data was collected consisting of measured headphone frequency responses. This analysis involves extracting characteristic statistical parameters to record systematic behaviors for all types of headphones. The work also extends to the evaluation of responses with reference to "targets" curves, interpreting the error that is formed between them. The main purpose of this work is the mass analysis, evaluation, and comparison of headphones, studying their behavior in the spectrum of audio frequencies. This is achieved by comparing the responses via the calculation of spectral parameters and comparing their curves in the frequency spectrum, through an application developed for this purpose.
This work constitutes a review of the field of the reproduction of complex continuous acoustic signals via laser optoacoustic transduction achieved by plasma generation in the atmospheric air or on solid targets by strong short or ultrashort optical pulses. A prototype system for reproducing complex continuous audio signals from laser pulse trains is presented as well as a computational tool for describing the acoustic response of the system, developed by the University of Patras (UP) in collaboration with the Karlsruhe Institute of Technology (KIT). Moreover, experimental results along with two phenomenological models are presented for the description of the acoustic emission of laser plasma sources. The models were developed by the collaboration of the UP with the Hellenic Mediterranean University (HMU). Finally, discussion is carried out about the possibilities of the proposed new technology and its possible scientific and technological areas of application.
Singing in a foreign language can pose an important challenge to a singer. Vowel sounds that do not exist in one’s native tongue are usually the most frequent “trap” for somebody to sound foreign, while the erroneous production of such phonemes might lead to crucial intelligibility issues. In this preliminary study, it is examined to which degree the substantial knowledge of the foreign language assists with the more authentic depiction of the foreign - to the Greek language - vowel sounds in singing. To this end, native Greek-speaking classical singers provided audio excerpts from the standard French-language repertoire of their voice type. Formant analysis revealed in which special cases lie the differences. As expected, the general conclusion is that the substantial knowledge of a foreign language assists in better pronunciation delivery of the foreign-language sung text. The study results provide additional information in regards to the necessity for a classical singer to study foreign languages, while imply the imperativeness of introduction of “lyric diction” courses in the Greek university or conservatory programs of study.
This paper introduces a Convolutional Neural Network (CNN) machine-learning approach for robust detection of the direction of arrival of any single source via binaurally recorded signals inside reverberant rooms. The proposed method relies on compacting the auditory model binaural parameters of interaural time and interaural level differences, into short one-dimensional vectors allowing the length reverberant effects to be incorporated into the feature vectors. The method has been tested in a very large database of rooms with varying acoustic characteristics and achieved robust localization with different audio source signals. The results indicate a high degree of generalization of the proposed method and a promising, flexible machine learning framework.
Acoustics in Educational and Cultural Activities
Augmented Reality Audio is a field of research and development with continuous development in recent years. However, a major problem encountered is the static way of mixing the virtual with the real acoustic environment. The problem lies in the constant change of the real acoustic environment which creates the need either for the listener to redefine the level, or for a dynamic change of the virtual environment to adapt to the real one. To address such a problem, a prototype dynamically changing mixing model was implemented which aims to acoustically adapt the virtual environment to the real one by using rules based on the extraction of loudness features according to a prototype model for calculating binaural loudness in real-time mode. The positive results from the use of the proposed model against the legacy mixing model in an augmented reality audio game in a laboratory environment, confirm the improvement of the listener's auditory perception.
The haunted house story arc is prominent in horror stories, from the folklore and literature to cinema. A film’s sound design in general is critical in creating the desired mood and is thus used as a tool to highlight the haunted house in the films that take place in that. In this paper, some examples of haunted house films and their sound design are presented and analyzed according to their similarities and differences. There is also mention on reports from the folklore around the world regarding haunted houses and their sounds.
The proposed announcement refers to the diachronic development (from 1950 to 2018) and the spatial distribution of summer cinemas in Thessaloniki. It is part of a wider university research, where the social, economic and urban planning data that shaped the "cinematic landscape" of the city are recorded.
The announcement was based on the processing of archival material and interviews. Τhe location, distribution and density of the cinemas in the wider area of Thessaloniki along with the technological developments contributed to the rise and fall of cinemas, from the first screening to the spread and eventually the shrinking of this outdoor entertainment.
This paper presents a prototype software application, which is designed to automatically find audio recordings that match a given image. The application facilitates the ordinary workflow of sound design. In this direction, the work includes an initial analysis of the users’ requirements, along with the design of a prototype application, which extracts captions from images, and performs natural language processing to maximise the match between images and sound recommendations. The application aspires to provide a user-friendly graphical interface, maximising the user experience. The current version processes images only and features a small audio library, but later versions will support video and data exchange with large online audio repositories.
The present study examines the possibility of integrating voice technologies into traditional journalistic processes, specifically the use of voice dialog assistants (voicebots) both for the processes of sharing news and for gathering information within the journalistic profession. The creation of a proposed modular model gives the possibility to adapt the implementation depending on the usage scenario and the needs of the journalist. Technologies such as topic modeling and sentiment analysis can be leveraged to support existing journalistic tasks, while combining the advantages of automation with those of a fully auditory interface.
A coupled-mode model will be presented for the prediction of low-frequency shipping noise in marine areas with complex bathymetry. For the operational application of the model, the local environments are categorized according to water depth, and for each discrete water depth the vertical eigenvalue problem is solved, and the modes and coupling matrices corresponding to subsequent depths are calculated and stored. Then, depending on the location of each noise source (ship) and the propagation environment, the appropriate modes and coupling matrices are retrieved and combined, and the acoustic field is synthesized. The preliminary solution of the vertical problem, the archiving and selective retrieval of the relevant results allows for rapid calculation of the total acoustic field, which is particularly important for the coverage of large sea areas and large numbers of ships. This work was supported by the program NSRF 2014-2020 (OPS 5010880 and OPS 5045792).
A pilot acoustic deep-water observatory for detection and 3-dimensional localization of sperm whales in real time was designed and developed the framework of the SAvEWhales project. The sperm whales are detected and localized from the pulsed sounds (clicks) that they produce while diving. The observatory was deployed and successfully tested for a total of 6 months in the summer of 2020 and 2021 off SW Crete. A detailed description of the detection and localization system as well as results and comparisons with independent observations were recently published and can be found in the open-access journal Frontiers in Marine Science, DOI: 10.3389/fmars.2022.873888. The SAvEWhales project was funded by Ocean Care (CH).
The paper presents a scheme for the recognition of the lower order normal modes of acoustic signals recorded in the marine environment. The identification is based on the spectrogram of the signal. A suitable Convolutional Neural Network (CNN) is developed and used. The adjustment of the parameters of the neural network will be performed with synthetic spectrograms which are produced using a standard code to calculate the acoustic field at sea from a known source based on the theory of normal modes in selected environments. A simple parameterization of the dispersion curves is adopted to characterize the corresponding normal modes. The main goal of the identification scheme is to determine the values of the parameters that describe the small-order dispersion curves. These parameters constitute the outputs of the neural network (labels). The fitting of the parameters for the input signals (label generation) is achieved through a training process using a standard least-squares method and using the theoretical arrival times of the normal modes. In the paper we present the first results for a simple geometry of the marine environment with controlled uncertainties of the environmental parameters.
This paper is based on the work undertaken by the Institute of Applied and Computational Mathematics (IACM) of the Foundation for Research and Technology-Hellas in the framework of the QUIETMED project funded by the Directorate General of the Environment of the European Commission. In this respect, the IACM group designed, developed and tested autonomous underwater listeners which were installed in two types of mobile stations for short term and medium term measurements. This paper describes the fourth pilot experiment undertaken by IACM in the framework of the QUIETMED project. The paper includes a description of the experiment and some examples of recorded data and processing results.
Acoustic and Measurements
Since the first levitation of an object with sound, almost a century ago, acoustic levitation is still a rather active field of research; mainly because there are no constraints regarding the levitated object’s material and because of the manipulation precision. These two aspects are enabling several applications both for lab experiments and for human-computer interaction apparatus. The two broadly used techniques involve standing waves and focusing the acoustic energy in specific locations, with one or more arrays of phased ultrasonic transducers. However, by putting the first one into practice the levitation points are in predefined positions and with the second one each acoustic trap is getting weaker as a new one is being introduced. Here we introduce a method combining the above techniques by adopting one transducer or a single phased array and a grid of reflecting surfaces with various heights. Our method is using numerical optimization techniques to optimize the reflecting surfaces’ heights to form multiple levitation points at given locations. By combining the emitted with the reflected sound pressures the acoustic field gets stronger and more capable of levitating multiple objects.
Sound power is an important and widely used acoustic quantity. It can be determined by sound pressure or sound intensity measurements. Despite the applicability of a number of different sound power determination methods, the results deviate from one another, mainly due to the different theoretical assumptions. The establishment of traceability of unit watt in airborne sound aims at overcoming the up-to-date deficiencies. The study explains the measurement chain that is required for establishing the traceability. The potential improvements of the current sound power determination methods are discussed. A research was performed within a funded metrology programme. The performed investigation included theoretical modelling along with measurements in both calibration conditions and in situ, results of which are presented. Apart from measurements and the required corrections, the related uncertainty was also determined in a transparent budget. This means each uncertainty factor was individually estimated and its contribution to the combined uncertainty revealed.
We present in this work a methodology for the design of idiophones, indicatively a vibrating beam or an acoustic tube. The methodology is based on optimization algorithms utilizing numerical models of the finite element method and their respective sensitivity indexes. Our target is the determination of the shape in order to obtain desired ratio of harmonics (eigenfrequencies) in respect to the fundamental one. An evident application of this current methodology is the implementation for the design of parts of musical instruments as distinctively are the keys of metallophones, xylophones and these of marimba, among others. The methodology can be easily dapted for the case of other similar problems as is the design of material distribution for appropriate stiffness and inertial properties arrangement. Furthermore, the chance of automization of the current method, makes it a potential useful tool for technicians and instrument makers. In this current work an attempt to construct a beam of shape that arise of this approach has been accomplished and experimental modal analysis as been conducted in order to compare with theoretical results.
In the present work, three-dimensional structures which can be considered as Phononic Metamaterials and exhibit large and complete Phononic Band Gaps during the propagation of acoustic waves were studied, theoretically and experimentally. These structures are created from air to solid material through 3D printing. The results showed complete and large Phononic Band Gaps for all structures studied. At the same time, the case in which an imperfection is placed in the structure was also studied. The existence of such structures enables the development of many applications such as acoustic filters, acoustic insulation, sensors as well as sound sources with strong directionality.
In the present study a convenient depiction of the diffraction field around a rigid wedge is proposed called polarity plots. The polarity plots show the shadow boundaries, the curve of zero-valued diffraction, the polarity of the impulse response and the number of geometrical acoustics contributions for any combination of the angular locations of source and receiver around a given wedge. The analysis is based on an approximate time domain solution for diffraction around wedges that is applicable for all types of simple incident radiation (plane, cylindrical and spherical incident signals). Furthermore, non-dimensional forms of the polarity plots are introduced which can be constructed for any wedge by only 4 basic pairs of source and receiver locations. Finally, an experimental procedure is proposed based on the polarity plots, where a series of measurements of the acoustic field around a wedge can determine the wedge angle.
In this paper, two methods for calculating the acoustic field around an acoustically hard rectangular plate are presented. The methods are used to calculate the acoustic field as superposition of the geometrical acoustic field and the diffraction fields from the four edges of the plate. The solution for the diffraction field used in the first method is analytical, thus the method is called the analytical method. The solution used in the second method is approximate hence the method is called the approximate method. Results of the application of the new methods show that they are in good agreement with similar existing methods, with other methods that treat the plate as a single body, and with experimental data. Finally, it is demonstrated that the two new methods are computationally faster by orders of magnitude compared to existing methods, while the approximate method is faster than the analytical method by two orders of magnitude.
In recent years, new methods have been developed for the generation of extremely high-frequency acoustic waves in materials using optoacoustic conversion of ultrashort laser pulses into mechanical waves. The generated acoustic waves have frequencies from tens of GHz up to a few THz, with wavelengths in the range of μm to a few nm respectively, providing excellent spatial resolution when they are used in nanoscopic structural and elasto-mechanical characterization of materials. This paper presents applications of this method developed at the Institute of Plasma Physics and Lasers in collaboration with the Laboratory of Physics Acoustics and Optoacoustics, Department of Music Technology and Acoustics, Hellenic Mediterranean University.
In this paper two solutions for the diffraction of a spherical wave around an acoustically hard edge are presented. The first solution is an integral along the edge. In comparison with other integral solutions, the presented solution: i) is easy to compute, ii) reduces the number of the independent variables of the finite diffraction problem from 8 to just 2, thus allowing better physical understanding of the problem, and iii) can be used for the definition and the a priori calculation of the critical length after which the edge can be considered to behave as an infinite edge. The second solution is an analytical approximation of the first solution and does not require integration along the edge, which makes it faster to calculate. Finally, the approximate solution does not yield considerable relative error compared to the analytical solution. In calculations of insertion loss the difference between the two solutions does not exceed 0.12dB.
Environmental Acoustics - Soundscapes
Urban areas are those in which high levels of environmental noise and low levels of air quality are observed. Properly informed citizens become much more aware and active in participating in actions towards limiting such problems. The present research is based on the assumption that there is a correlation between air and noise pollution, up to an extent which is still under investigation. The aim of this work is twofold: on the one hand, to identify sources of air pollution given their environmental audio footprint and, on the other hand, to prototype a crowdsourcing application which will allow citizens to contribute environmental audio data to a machine learning system.
The study focuses on sounds that are recorded in memoirs of fighters and historical records. The sounds are classified in three categories: 1. Low SPL sources, that propagate tens or hundreds of meters away. 2. Medium-range sounds, that propagate over distances of a few kilometers. 3. Long distance sounds, that spread, according to the historic sources, over distances of tens or even hundreds of kilometers. Since the sound propagation in many cases seems, to us today, exaggerated, it is examined if the historic sources are true or, under which conditions they could be true. A possible explanation is the low environmental noise. In general, it can be argued that the people of that time relied quite a lot on sound to communicate with each other, overcoming the absence of telecommunication means that we have today.
This paper describes the design of historically informed soundscapes to be incorporated in modern travel-guide applications. The work stems from the research project TRACCE (TRavelogue with Augmented Cultural & Contemporary Experience). Apart from the creative process and technical details, this paper discusses the design challenges, which mainly stem from a) the limited data available which would allow for an accurate and convincing reconstruction of the acoustic environments, b) the need for diverse auditory displays which would grasp the users’ attention, and c) the difficulty in designing soundscapes which would be interesting, appropriate, and informative for a wide audience of various age groups, educational backgrounds, and sensory abilities. The paper concludes with preliminary results from a user evaluation study of the designed soundscapes.
The aim of this research was to create an easy-to-use platform for highlighting sound preferences related to a hypothetical public sound environment. Sound recordings were conducted in the categories of anthropophony, biophony and geophony using a digital recorder. Utilizing the capabilities of the Audacity software, the opinion of the residents of an Athenian neighborhood were studied in regard of the combined content of the recordings. The participants' preferences were used in order to extract a new sound composition, creating a unique sound environment and thus soundscape for each participant. These new compositions were analyzed in order to extract the acoustic biodiversity indicators and later studied in terms of acoustic complexity and diversity.
This dissertation aims to create a podcast consisting of soundscapes as an alternative way to live the museum experience. It is based on the museum of contemporary arts of MOMus (Metropolitan Organization of Museums of Visual Arts of Thessaloniki). In more detail, wanting to modernize the museum’s guided tour, so guests can have a more augmented and immersive experience through the usage of current media trends and digital technologies, soundscapes are produced. The latter guide guests through the museum’s permanent exhibition, provide basic information about the selected showpieces, and transfuse to the visitors the feelings behind them. Additionally to the production of soundscapes, it reviews the theory behind podcasts and soundscapes, and it investigates similar projects.
The increasing population movement towards urban centers has brought about a number of environmental pressures, including environmental noise, with implications for the quality of life and the quality of the overall environment. The purpose of this research is to highlight circular economy methods to reduce the noise footprint. The main tools of the research were noise modelling and mapping. Noise management scenarios structured in a circular economy framework were then formed and modelled. More specifically, the cases of green facades and electric vehicles were studied. The results showed a decrease of 4 dB(A) following the implementation of the circular economy actions.
The public gardens in the Mediterranean cities are places of relaxation, recreation and social interaction, especially during the extended summer period. Their design requires the utilization of bioclimatic parameters that constitute objects of study and result from the collaboration among multiple technical and geotechnical specialties. Sound protection from environmental noises has emerged as a key evaluation indicator, a parameter that contributes to ensuring spaces of tranquility within the urban area.
The proposed announcement is part of an ongoing research that evaluates various urban green spaces, with the objective of studying and understanding the bioclimatic conditions, in a sample of twenty (20) gardens in Cyprus and Xanthi.
Acoustics in Historical Landmarks and Sound Heritage
This paper presents the results of a series of experiments, designing and interacting between architects and music composers concerning the architectural acoustic qualities of the historical monuments of Thessaloniki (Rotonda, Achieropoiitos and Panagia Chalkeon). The acoustic measurements are included in the music composition as a series of data necessary for the creation of digital as well as real sound spaces. The impact of this interaction, in addition to the creation of sound installations in specific locations (site-specific sound installations), contributes to the expansion of the sound experience of each musical work and an even integrated design.
The announcement attempts to record the current state of eleven (11) ancient theaters, as a continuation of a long-term observatory survey (since 2004) which is repeated at regular intervals (in another 20 ancient theaters), but was interrupted during the last three years, due to the pandemic
The aim of the research is the acoustic evaluation of the existing or potential re-use of the ancient theaters, with parameters such as the type of sources and the levels of noise. The evaluation of the building condition of the theaters refers to the orchestra, the stage building, but also the modern operational infrastructures of the space. As can be seen, in the current period the restoration works of the monuments have been intensified, while the intense urban pressures (traffic & urban noise) in their immediate environment continue.
The decision for designing the first Municipal Theatre in Nicosia was taken during the 50s. The theatre has officially opened in 1967. In June 2008 soon after the refurbishment works and the addition of new lighting and electro-acoustical equipment the roof of the auditorium collapsed completely and remained in this situation for almost 8 years. Due to the above it has been decided to preserve only the external building shell and reconstruct in full the entire building.
The aim of the redevelopment was the creation of a contemporary Concert Hall for classical music & symphonic orchestras. Further performing art events were also included in the schedule of uses such as, ballet as well as theatre performances.
Given the above context, the present article delineates the intervention of the architectural design, the measurement results and the key acoustic highlights that contributed in the enhancement of the spatial envelopment.
Key Words: Nicosia, Concert Hall, Classical Music, Architectural redesign, Spatial impression, Collapsed
The impact of noise within large urban agglomerations, is increasing, concerning the well-being of mankind and the entire planet. A seemingly easy solution to this problem is to create or find and “re-use” quiet areas in the urban fabric. This paper, attempts -by using a wide range of tools and methodologies-, to evaluate the sound characteristics of quiet areas of Thessaloniki’s (Greece) Historic Centre, which are at a different level from that of the current road network and to highlight whether this elevation difference contributes substantially to the formation of the sound identity of a potentially quiet area in the urban fabric.
Studies have shown that noise levels in the centre of Athens are above the noise levels set by WHO. This study is based on the hypothesis that in the north suburbs of Athens – where no measurements have yet been conducted – these levels would follow the example of the rest of Athens. Initially, the acoustic environment is recorded and analysed within the theoretical framework of the soundscape approach. Sound measurements were carried out along with soundwalks, at different times of the day, applying the microscopic and macroscopic methods, allowing soundscape understanding of a specific place but also a wider environment. This identified areas with impact of high-level or negative sounds.
The study reveals that the residents of these areas experience noise levels above 50 dB, as determined by WHO. However, this study also shows that exposure to high level sounds or negative sounds is not constant and is associated with commuting. These areas, characterised by large avenues but also large spaces where positive sounds dominate, are suitable for living.
According to the believers, art historians, and those involved with the Devine affairs in general, in the doctrine of the Orthodox faith, iconography plays the role of a bridge between man and the Divine. In other words, it is one of the key factors in uniting believers with their faith.
Following the period of the iconoclasm, during the 7th Ecumenical council of 787 AD,. the "meaning" of the "icons", was restored, creating the well-known tradition of the iconography of temples. This tradition was extended, with the help of its "historical" and "theological" origins, and dominated, "philosophically" but also spatially in Byzantine temples.
Considering the "expansion" and "dominance" of the iconography of the temples during the Byzantine era and taking into account the lack of "active", surfaces, in the interior, due to this "dominance", the recorded parameters of the acoustic field, are expected to be, inevitably, at prohibitive levels under the contemporary view. However, in cruciform temples, this usually is not the case. The present essay, through "differential" reasoning, presents a solution for improving the acoustic field, in large spaces, without the use of absorbent means, with the widespread use of the phenomenon, as it seems, in the properties of cruciform temples.
The work describes the acoustic properties of one of the two ancient theatres that existed in the city of Larissa, in Thessaly in central Greece. This theatre was initially built in the first half of the 3rd century B.C. It consists of the orchestra having a diameter of 25.5m (compared to 20m for Epidaurus), the cavea is divided in 11 sectors each with 25 tiers of seats and a partially preserved stage building is also in existence.
The work presents a first record of the acoustics of this monument with in-situ measurements based on: (a) an omnidirectional and (b) with a binaural dummy head. The results of these measurements are also compared to other ancient Theatres.
This paper aims at recording and prototyping the dedicated procedures for the conversion of an existing building shell into an integrated area for music rehearsal and sound recordings. Specifically, an existing urban ground floor building, with total area of 290 m2 and height of 4.70 m, is being analyzed and redesigned. Initially, a review of the literature on building acoustics, computer based modern design methods and construction detailing is carried out. In the next step the existing layout of the building is being imprinted and reviewed, to identify constraints to be considered during the design phase. Internal layout alternatives are then analyzed and compared, according to wave theory, proportional dimensions of rooms, parallel activities and functional architectural design. The acoustic performance of the interior is being carried out through room simulation models. The required sound insulation capacity of all rooms and the building shell is next calculated. The outmost target is to extract and test best practices that can be used in similar cases.
This paper presents a novel dynamic directional amplification (DDA) mechanism suitable for low-frequency noise mitigation. The DDA mechanism increases the Inertia of the panel and shifts the resonance frequency to the very low frequency regime. Τhe implementation of such a mechanism is very simple and is comprised of flexible supports and a rigid link, without the need of any additional masses or damping elements. Initially the mathematical model of the amplifier is described followed by its experimental validation. Next, an example of a panel with and without the DDA mechanism is presented through analytical analysis.
This paper presents the architectural and acoustic design methodology, the results of acoustic measurements carried out in the "Miltiadis Evert" amphitheater according to ISO 3382-2 as well as the comparison with the CATT-Acoustic acoustic simulation model. The calculation of various acoustic parameters such as Reverberation Time RT60, Early Decay Time EDT, Clarity C-80, Definition D-50 etc. gives important data for the acoustic behaviour and evaluation of the auditorium. The design complexity of the space was based on a unique acoustic ceiling, with an arrangement of prisms which consists of 564 pieces, combining diffusive, reflective and absorbing zones. Detailed analyzes of acoustic parameters based on relative source and listener positions also provide important insights into the room's impulse response and time-frequency relationship. During the measurement process, a "Talk Box" noise source was used which was calibrated according to the frequency response of the human voice in order to measure and evaluate the STI & STIPA indicators, while the indicators for the Public Address System were also measured ) according to IEC 60268-16:2020.
Finally, the variation of the results is examined by comparing different simulations of the space, based on the complexity of architectural details, the choice of diffusion coefficients in the prismatic roof.
Το πρόβλημα που καλείτε να επιλύσει η σύγχρονη αρχιτεκτονική είναι η ηχομόνωση των κατασκευών ώστε να ικανοποιούνται οι απαιτήσεις των ενοίκων. Το πρόβλημα είναι εντονότερο σε κτήρια συνάθροισης κοινού όπως ξενοδοχεία, κτήρια γραφείων κλπ. Διαχωριστικές τοιχοποιίες μεταξύ γραφείων ή δωματίων αλλά και πολυσύχναστων διαδρόμων καλούνται να έχουν την απαραίτητη ηχομείωση, ώστε να εξασφαλίζεται η «ησυχία» στα δωμάτια. Εξάλλου τα τελευταία χρόνια απαιτήσεις ηχομόνωσης ορίζουν και οι μεγάλοι τουριστικοί πράκτορες αφού προδιαγράφουν την ηχομόνωση ως ένα πλεονέκτημα του καταλύματος.
Στο πλαίσιο αυτό σχεδιάστηκε ένα πρόγραμμα πειραματικών μετρήσεων σε διαχωριστικές τοιχοποιίες. Το πρόγραμμα είχε ως απαίτηση τις ελάχιστες απαιτήσεις του DIN 4109 που αποτελεί το πιο διαδεδομένο πανευρωπαϊκά κείμενο απαιτήσεων ηχομόνωσης κατασκευών και αποτελούνταν περισσότερες από 50 μετρήσεις. Αναλυτικότερα, μελετήθηκαν συστήματα με μια ή δύο γυψοσανίδες εκατέρωθέν καθώς επίσης και ένας ή δυο σκελετοί. Παράλληλα μελετήθηκαν διάφορα πάχη σκελετών, όπως αυτά διατίθενται στην αγορά. Στα συστήματα ως ηχομονωτικό υλικό χρησιμοποιείται ο πετροβάμβακας σε δυο πυκνότητες 50 και 100kg/m3 και σε διάφορα πάχη. Τέλος, μελετήθηκε η συμπεριφορά ηχομονωτικών και απλών κυτίων σε ηχομονωτικές λύσεις.
Από τα αποτελέσματα των μετρήσεων προέκυψαν χρήσιμα συμπεράσματα για την συμπεριφορά των συστημάτων ανάλογα των επιμέρους στοιχείων του, καθώς και του τύπου των κυτίων, ώστε να επιλέξουμε το κατάλληλο σύστημα για την κάθε κατασκευή.
The Acoustics Research & Simulation Laboratory, NTUA School of architecture, dept IV was founded in 1954 by late Professor Dr. V. Papathanasopoulos and initiated its research activities under the title “Laboratory of Sound Technology”. During its almost 70 years of operation it has always been the spearhead of the acoustics research for the major Hellenic projects. Since 2018 the ex “Laboratory of Sound Technology” starts a new era by expanding its research capabilities at the areas of acoustical simulations, research on the physical reconstruction of auditory scenes and soundscapes, as well as the support of research related to the enhancement of the acoustic performance of Hellenic building materials. The above-mentioned breakthrough has also been accompanied by the composition of a totally new internal operations regulation, the initiation of the registration process to the National Accreditation System as a notified body and the renaming of the laboratory as “Acoustics Research and Simulation Laboratory NTUA”. One of the pivotal highlights of the above mentioned plan was -amongst other subjects- the research and development of a flexible modular diaphragm, by the Lab’s research team (2&3). The aim of the new modular partition in matter was to speed up the test element mounting times, reduce the cost of construction especially when having multiple test elements of different sizes and at the same time investigate in practical terms the measured R’max of the laboratory, for supplementing the already excellent performance of the lab facilities. Further to the presentation of practical examples of the diaphragm’s application and its contribution in the qualitative measurements of high insulation elements’, the present article also engages the results of the measured Rw in a wider discourse about its contribution in the research and development of the production chain of the Hellenic sound insulation and soundproofing materials and building arrays.
Key Words: NTUA Acoustics Research & Simulations Laboratory, R’max, Architecture, Sound insulation, Airborne Noise, Hellenic building industry, Certification
This paper tries to address the aims and objectives of the building acoustics study and its contents.
The parameters of acoustic comfort are difficult concepts for non-experts or even those familiar with acoustics. The acoustic study has to translate them into drawings and constructional details and describe a subjective perception of everyday noises. The study team of the building project (architect, civil engineer, HVAC engineer, etc.) must make clear decisions and give specific instructions for the geometry and construction of all structural elements of the project.
The conversion of acoustic requirements into construction details must be proved by calculations by recognized standards and with the assistance of acoustic measurements. This theoretical and practical proof should allow an acoustic consultant to "control" the study at any state.
The above steps, impose the limitation of the "philological" texts of the study and the increase in the acoustic designs and the construction details, both of which, are based on computer models indicated by the accepted standards of the industry.
The subject of this work is the adaptation of an underground storage space in the buildings of the Thessaloniki Concert Hall (TCH) into a rehearsal room for the Megaro Symphonic Youth Orchestra (MOYSA). The challenge of the project was to create a room with an acoustic behavior suitable for its use, addressing the problem of its short hight, in relation to its other dimensions. The suspension of curved reflectors on the ceiling, as well as the application, on separate sections of the walls, of inclined surfaces and sound diffusers, resulted in the achievement of the goal.
After all these years, the expression of its meticulous modern Greek architecture is perfectly amalgamated with the city’s context, while its state-of-the-art acoustics fulfill the vision of the numerous people that aspired this majestic project through the entire development. CCCC encapsulates a top-notch main auditorium (750 seats), a smaller chamber dedicated to classical music events (250 seats), two theatrical venues and other satellite spaces with primary functions such as seminar rooms, a conservatory and various musicians’ warm up areas.
The main auditorium is designed in order to support classical music concerts. The acoustic design also allows the dynamic adjustment of the hall in order to also fully support drama opera, folk music, ballet and contemporary dance performances. Moreover, high-end electro-acoustical equipment has been integrated for allowing the delivery of lectures, presentations, conferences as well as the projection of cinematic movies.
In the current lecture the subjects that will be discussed include but are not limited to:
a- The main acoustical features that have been incorporated into the architectural design and the implementation of the dynamic acoustic components that realize the hall’s adjustability.
c- The pivotal contribution of the first Hellenic physical acoustic simulator that has been developed for over 7 years by the authors. The simulator was employed for the first time during the design of CCCC for the spatial reconstruction of the main auditorium’s acoustic environment. The purpose was to analyze the psychoacoustic perception of the hall during the final design phase -without any signal post-processing corrections and investigate for any potential final adjustments, in a «design-through-hearing» manner.
Key Words: Crete, Concert Hall, Classical Music, Spatial impression, DEPANAL, CCCC
The different acoustic comfort requirements in speech and music spaces, after the classical era, returned during the Renaissance, receiving a prevailing, but incomplete, solution during the Baroque era. For the layout and form of these rooms, specific stereotypes were chosen that resulted in the rectangular, uni-axial layout (shoebox type), a baroque form that dominates until nowadays.
The emergence of an alternative version with the vineyard layout appeared at the Berlin Philharmonic (1963), a format that goes back to theatrical design proposals for open stage or arena. In the modern era (20th - 21st century) concert halls are places equipped with advanced technological means and continue to stand as symbols of the wealth and development of a place.
Sound and Music
The aim of the paper is to investigate the effect of three different spatial sound reproduction systems (mono, stereo and surround) on the perceptual parameter of envelopment, overall listening experience and emotional arousal of listeners. A listening experiment was conducted that asked participants to evaluate these parameters while listening to binaurally recorded music pieces. Τhe results indicate that the envelopment of binaurally recorded musical signals can be adequately modelled adequately by the linear combination of binaural features. A linear model of the Overall Listening Experience (OLE) is also implemented that examines the extent to which perceptual, cognitive and affective factors influence the overall preference of listeners. It was found that the preference for music content dominates on OLE formation, while spatial sound quality affects OLE to a lesser extent. Emotional arousal was not found to be linearly related with OLE.
This paper presents a preamplifier circuit based on Metal Oxide Semiconductor Field Effect Transistor (MOSFET) technology, which emulates the quality of the sound produced by vacuum tube amplifiers used for electric guitars. The preamplifier output signals are analyzed and compared to output signals coming from a popular combo tube guitar amplifier under the same signal driving conditions. The comparison revealed noticeable similarities as far as overdriven or distorted sound quality is concerned between the MOSFET emulator and the original tube amplifier. The listening tests conducted confirmed the results of the analysis and comparison of the output signals.
Drum pattern generation is a part of the wider research field of algorithmic composition which focuses on the music element of the rhythm. Related research offers various methodologies, however there is a lack of approaches relying on audio data to generate rhythmic drum sequences. The current paper addresses this issue by introducing a system which receives a raw musical excerpt and generates symbolic drum patterns which are rhythmically compatible with it. Core pillar of the proposed system is the retrieval of the so-called Onset Strength Signal (OSS) and its processing by a Seq2Seq model. The system evaluation was carried out by conducting an online survey in which participants were asked to listen to and rate qualitative characteristics of certain audio samples. Statistical analysis of the results validated the effectiveness of the system.
The present research aims to record and classify as much information as possible related to music listening habits in Greece after gathering and processing 5252 samples. As it emerged after a thorough literature research, this issue had not yet received a comprehensive academic study.
This research has been carried out through the creation and distribution of a highly expanded original questionnaire, in order to capture as many parameters as possible, related to both the musical material itself and the media and of course also the listener himself.
Noise - Vibrations
Noise pollution from food stores (restaurants, pubs, bars, cafés) during their function and provision of services when using music, is a frequent phenomenon disturbing the peace and quiet of individuals all around the area and also causing problems to social coexistence. As a phenomenon of a dual nature, it is faced by our legal order through both private and public law. The interpretation and application of this legislation is provided by the jurisprudence (case law). Thus, this work includes the whole existing legislation, as well as the relevant jurisprudence, and ends with conclusions on that legislation and jurisprudence.
A new approximate analytical solution is presented for the prediction of the diffracted field around a rigid wedge. Results for receiver locations in the shadow zone behind the wedge are in reasonably good agreement with the exact Oberhettinger solution. However, the analytical solutions for the diffraction field around a wedge are mathematically complex. Based on the new solution the insertion loss in the shadow zone behind a wedge is investigated and the following are proposed for fast and easy calculations: (i) a prediction chart as a family of curves, and (ii) a semi-empirical relation containing simple mathematical functions. The proposed semi-empirical relation is compared with regards to its accuracy and its computational cost with existing semi-empirical models.
This research deals with the determination of sound power levels for each one of multiple sound sources based on sound pressure measurements taken during normal operation of complex industrial facilities. The numerical method presented correlates the data from sound pressure measurements at various points with the unknown sound power levels and calculates the latter through numerical optimization. It is thus possible to determine the sound pressure levels at any point in the examined space as well as the sound exposures of operators in order to propose the optimal noise control measures. The effectiveness of the method is demonstrated through a real case study realized for an industrial production line.
This work concerns a technical study prepared on behalf of the Operator of the Thessaloniki Sewage Treatment Plant (SWTP), with the aim of suppressing the noise emitted by the plant's centrifugal air compressor complex. The study of the application of a muffler and sound trap in the air intake section for each of the turbochargers of the assembly is presented, and the results of the emitted noise measurements, before and after the application, are evaluated.
The purpose of this paper is the development of a simplified and reliable methodology for the evaluation of the noise produced by large wind turbines. It is also proposed the adaptation of the evaluation process of the relevant Greek legislation to the specific characteristics of noise produced by large wind turbines. The assessment refers to the technique of measuring the noise with the appropriate precesion methods-instrumentation and not the assessment of the produced noise with anthropocentric methods (such as questionnaires). The development of the methodology was based on the British Standard (BS) 4142: 2014, ETSUR97(1996) and the simplification of the IEC61400: Part 11 technique. Additional evaluation criteria of the generated noise are introduced, applying level penalties in the process in case the noise characterized by tonality
In present work an Active Noise Control System (ANCS) mounted on the headrest of an aircraft’s seat is proposed. It is based on the non-linear FsLMS algorithm alongside with a linear extrapolation technique used to estimate the acoustic pressure around the passenger’s ears. In order to check system’s effectiveness, real-time experiments took place inside an aircraft’s cabin mock-up for an acoustic disturbance measured during the flight of a tilt-rotor aircraft. The results have shown a more than 10 dB attenuation of different BPF harmonics and a significant SPL reduction that varies from 8 dB to 11 dB around the passenger’s ears.
This paper presents a case study comparing and analyzing structure borne noise (vibrations) that were transmitted in a first-floor apartment due to the operation of a natural gas compression unit in an adjacent gas station in Attica, Greece. Following the initial measurement of the vibration transmission the proposal of gradual vibration transmission reduction actions in four different stages was carried out. The effectiveness of each stage was with successive series of vibration measurements. The aim of this paper is to investigate the vibration levels and to evaluate the interventions carried out. The presentation of the results is comparative between the phases of the anti-vibration improvement actions. The aim was to collect adequate data in order to quantify the reduction achieved in each phase.
In later years, the awareness on noise reduction in working places is rising. There are rules and laws from the government [1,2] or, in the best cases, the employer is willing to provide a better working place for the employees. In Greece there are laws derived from EU recommendations . In this paper, a case study in a production line of a factory is discussed. The measurements, the model development and the study of the solution are described. The utilization of commercially available acoustics simulation software  and the way it can support the study is highlighted. Details on the model setup and the simulations are discussed. Finally, the results look promising and are presented along with the proposed solution.
The process of time reversal has the forward and the backward stage and can be used to focus sound coming from one or more distant sources. The method has been applied to the areas of underwater acoustics, biomedicine and a few other. The present work deals with the numerical implementation of the time-reversibility process in a vibro-acoustic environment. We study a twodimensional resonant medium in conjunction with a finite number of beam elements. Although the topic of time reversibility in a reverberant environment has been investigated in the past, the present work also includes the interaction with elastic bodies that are inside the acoustic medium, an element that constitutes the innovation of this work. In the numerical experiments presented, the elastoacoustic medium is excited with the appropriate shock pulse while the waves are recorded at points located within the domain of propagation, at the boundaries of the domain and on the elastic bodies. There is a finite number of recorders and the duration of the simulation is variable, while the finite element method is used for the experiments. In the backward stage the signals are inverted and emitted into the elastoacoustic medium, creating at a specific time a strongly focused virtual source. Using various indicators the quality of the focus will be assessed, depending on which and how many signals are used in the backward stage, by comparing for example the primary maximum of the sound pressure at the time of the maximum focus with the secondary maximum. By comparing two scenarios when there are elastic beams in the domain with the case where they are not, it is shown that the presence of elastic bodies improves the focusing indices. Possible applications of the findings will also be discussed.