Audio Processing

En este trabajo se presenta el modelo de un altavoz de doble bobinado. Se revisa el modelo clásico basado en las propiedades físicas de la parte mecánica y eléctrica del parlante, y para salvar las discrepancias observadas entre las mediciones y el modelo matemático se incluye el efecto semi-inductivo causado por las corrientes parásitas, que da lugar a un modelo de orden fraccionario. El principal aporte del trabajo es considerar el efecto semi-inductivo también sobre el acoplamiento mutuo entre los bobinados, efecto que no ha sido observado en la literatura consultada. El modelo teórico se valida mediante resultados experimentales.
//
This paper presents a model for a dual voice coil loudspeaker. The classical model is reviewed based on the physical properties of the mechanical and electrical parts of the speaker. To bridge differences between the measurements and the mathematical model,  the semi-inductive effect caused by eddy currents is included. This results in a fractional-order model. The main contribution of this paper is to consider the semi-inductive effect also on the mutual coupling between the windings, an effect that has not been reported in the literature. The theoretical model is validated by experimental results.

In this article we present an API and a set of Javascript modules for the synchronized scheduling and aligned playback of predetermined sequences of events such as notes, audio segments , and parameter changes as well as media streams (e.g. audio buffers) based on the Web Audio API logical time. The API has been designed to facilitate the development on both ends, the implementation of modules which generate event sequences or media streams as well as the integration of such modules into complex audio applications that require flexible scheduling, playback and synchronization.

En este trabajo se presenta una primera aproximación al control por realimentación de un parlante reproductor de bajas frecuencias (“woofer”) sensando la aceleración del cono, usando controladores discretos. La técnica de diseño se basa en un método lineal algebraico que permite especificar de manera sencilla la respuesta en frecuencia del sistema de lazo cerrado. El diseño se ensaya sobre un procesador digital de señales de punto fijo, y se incluyen resultados de simulación y experimentales. Aunque el enfoque es simple, los resultados obtenidos en el laboratorio son similares a los presentados en la literatura con controladores mucho más complejos.

//

In this paper a first approach to feedback control of a low frequency  speaker ("woofer") sensing the acceleration of the cone, using discrete controllers is presented. The design technique is based on a linear algebraic method which allows to easily specify the frequency response of the closed loop system. The design is tested on a fixed point digital signal processor. Although the approach is simple, the results obtained in the laboratory are similar to those presented in the literature with more complex controllers. Simulations and experimental results are included.

State-of-the-art sound event detection (SED) methods usually employ a series of convolutional neural networks (CNNs) to extract useful features from the input audio signal, and then recurrent neural networks (RNNs) to model longer temporal context in the extracted features. The number of the channels of the CNNs and size of the weight matrices of the RNNs have a direct effect on the total amount of parameters of the SED method, which is to a couple of millions. Additionally, the usually long sequences that are used as an input to an SED method along with the employment of an RNN, introduce implications like increased training time, difficulty at gradient flow, and impeding the parallelization of the SED method. To tackle all these problems, we propose the replacement of the CNNs with depthwise separable convolutions and the replacement of the RNNs with dilated convolutions. We compare the proposed method to a baseline convolutional neural network on a SED task, and achieve a reduction of the amount of parameters by 85% and average training time per epoch by 78%, and an increase the average frame-wise F1 score and reduction of the average error rate by 4.6% and 3.8%, respectively.

The HuComTech project aims at developing a theory of multimodal human-computer interaction linking knowledge about human-human interaction to technological implementation. The purpose is to contribute to a more efficient and human-like human-computer interaction system by defining the main structural elements of communication, identifying their markers and defining their alignment with other markers in a multimodal environment. The novelty of the proposed system is its bidirectionality (both analysis and synthesis). This advantage is utilized in a new multimodal corpus and database.

Emerging technology is facilitating collaboration and peer-production across a variety of industries. As Tapcott and Williams (2006) note, peer-production is particularly effective where a process can be broken down into a series of steps or tasks. It follows therefore, that is has become attractive to the field of music production whether driven by the need to drive down costs, or the desire for a differentiated product.

Rumsey and McCormick (2014) note that in popular music recording, there are at least two distinct stages – ‘track laying’ and ‘mixdown’. It is therefore no surprise that one of the most obvious uses of peer-production in the recording process is the use of a remote mix engineer. In short, a dedicated mix engineer is contracted to mix the recording which has usually been recorded by another engineer. This rise has largely been facilitated by the increase in internet network bandwidth bringing greater ease to the exchange of large audio session files (Bailey 2001). The concept can be seen across all sectors of the music industry from unsigned artists through to global recording artists. Mix engineers such as Chris Lord Alge and Mark ‘Spike’ Stent have a high status and are associated with a variety of globally-successful acts. However, bedroom producers or garage bands may also send their sessions to be mixed by a remote mix engineer with better facilities or expertise.

This paper interrogates the role of remote mix engineer in order to illuminate the reasons behind the rise in popularity.  Specifically, it looks at technological approaches, how the expertise of remote mix engineers differs and becomes attractive to clients, before lastly looking at the mix engineer as a brand whereby their association is seen to be a mark of quality.

The HuComTech project aims at developing a theory of multimodal human-computer interaction linking knowledge about human-human interaction to technological implementation. The purpose is to contribute to a more efficient and human-like human-...

The e-book industry is starting to flourish due, in part, to the availability of affordable and user-friendly e-book readers. As users are increasingly moving from traditional paper books to e-books, there is an opportunity to reinvent and enhance their reading experience, for example, by leveraging the multimedia capabilities of these devices in order to turn the act of reading into a real multimedia experience. In this paper, we focus on the augmentation of the written text with its associated audiobook, so that users can listen to the book they are (currently) reading. We propose an audiobook-to-ebook alignment system by applying a Text-to-Speech(TTS)-based text to audio alignment algorithm, and enhance it with a silence filtering algorithm to cope with the difference on reading style between the TTS output and the speakers in the ebook environment. Experiments done using 12 five-minute excerpts of 6 different audio-books (read by men and women) yield usable word alignment errors below 120ms for 90% of the words. Finally, we also show a user interface implementation in the Ipad for synchronized e-book reading while listening to the associated audiobook.