Lukas Treybig

The paper presents an alternative approach for estimating reverberation time from measurements in real rooms when the requirements of the standard DIN EN ISO 3382-1/2 for the characteristics of the sound source, receiver, and measurement positions cannot be met. The main goal is to minimize the variance of the calculated reverberation times when using a directional source and receiver, or source-receiver relative positions with very small distances. For this purpose, the energy decay curve for individual octave bands is sampled in time. The estimation starts 2 ms after the direct sound. This is followed by several estimates of the RT over a 20 dB drop, starting 1 dB later with each iteration. The best fit mean of these values gives the estimated reverberation time. A comparison with the standard reverberation time estimation shows a variance reduction of 10% to 30% for binaural room impulse responses (BRIRs). The proposed method finds its application in situations where measurements can only be made at a few positions in the room and/or only in a few areas of the room. Furthermore, the method should be better suitable for measurements with receivers located near or at the head of a person.

This paper shows various approaches for determining acoustic (dis-)similarity based on room acoustic parameter values derived from real measurements. The similarity is calculated across different room configurations and/or between different microphone-loudspeaker positions within the same room configuration. We compare supervised (LDA, Random Forrest) and unsupervised techniques (PCA, SPPA) and pre-selected visualizations in terms of their ability to exhibit inter- and intra-room (dis-)similarities. The data set generated comprises spatially high-resolution room impulse responses obtained from multiple source-receiver positions within a room configuration. The room acoustics are varied by introducing active walls and geometries accounting for specific room configurations. The results show that the separation of room configurations primarily relies on specific acoustic parameters, with the reverberation time playing an important role. Within a given room configuration, the acoustic parameters excluding the reverberation time mainly capture the orientation and distance between the source and receiver.

To achieve the goal of a perceptual fusion between the auralization of virtual audio objects in the room acoustics of a real listening room, an adequate adaptation of the virtual acoustics to the real room acoustics is necessary. The challenges are to describe the acoustics of different rooms by suitable parameters, to classify different rooms, and to evoke a similar auditory perception between acoustically similar rooms. An approach is presented to classify rooms based on measured BRIRs using statistical methods and to select best match BRIRs from the dataset to auralize audio objects in a new room. The results show that it is possible to separate rooms based on their room acoustic properties, that the separation also corresponds to a large extent to the perceptual distance between rooms, and that a selection of best match BRIRs is possible.

This contribution gives a field report for the use of network-based array microphones in a television news broadcast studio. Presenters are captured with multiple beam patterns. A network based signal processing unit used to generate this multiple directivity. The test setup has been integrated into the production environment with the IEEE AVB/TSN network standard suite using the existing network infrastructure. The main part of this contribution refers to the question in how far the deployment of array microphones with digital signal processing is suitable for television studios. For this purpose, the sound quality of the microphone array beams is compared with signals captured with conventional clip-on microphones. Moreover, resulting acoustic problems and solutions are discussed.

Mikrofone sind ein wichtiges Element von Sprachkommunikationssystemen. In vielen Anwendungen werden dabei Freisprecheinrichtungen verwendet. Dies führt dazu, dass neben dem eigentlichen Sprechersignal auch Störgeräusche aufgenommen werden, welche die Qualität des Sprechersignals beeinträchtigen können. Digitale Signalverarbeitungsverfahren ermöglichen es Störungen zu unterdrücken und damit das Sprechersignal zu verbessern. Dabei werden unteranderem Mikrofonarrays, bestehend aus mindestens zwei Mikrofonen, eingesetzt. Ein Prinzip ist dabei das Beamforming, bei welchem die einzelnen Laufzeiten der Mikrofonsignale bezüglich des Sprechersignals synchronisiert werden, damit diese sich bei der anschließenden Signalmittelung unverzerrt überlagern. Gleichzeitig werden Störquellen aus anderen Richtungen und ungerichtete Störungen gedämpft. Im Rahmen dieser Arbeit wird eine dynamische Laufzeitsteuerung für zwei Mikrofonsignale entwickelt, welche die Laufzeitdifferenz zwischen den Signalen mittels Analyseverfahren bestimmt und ausgleicht. Die Grundlagen bilden die theoretische Betrachtung der Mikrofonsignale und verschiedener Laufzeitanalyseverfahren. Des Weiteren werden die Evaluation der Laufzeitanalyseverfahren, sowie die Implementierung einer Laufzeitsteuerung in ein eingebettetes System beschrieben. Als Zielplattform dient dabei das ADSP-21364 EZ-KIT Lite Evaluationsboard der Firma Analog Devices. Der Digitale Signalprozessor bietet dabei die Möglichkeit einer flexiblen und latenzarmen Verarbeitung der Audiosignale. In der Auswertung werden die Funktionalität der Laufzeitsteuerung und deren Grenzen aufgezeigt. Zudem werden zu lösende Probleme und mögliche Ansätze zur Weiterentwicklung dargelegt.