Search

Over the past decade, mobile news production has had a growing prevalence and has been established as a new type by modern journalism industry. Journalists understand content capturing and sharing as parts of their role in newsrooms. Mobile journalism (mojo) is an evolving form of reporting in which where people use only a smartphone to create and file stories, and it has been gaining ground during the last decade. This paper aims to examine the difficulties, issues, and challenges in real-world mojo scenarios, analyzing the efficacy of prototype machine-assisted reporting services (MoJo-MATE). A usability evaluation is conducted in quantitative and qualitative terms, paying attention to the media literacy support provided through implemented tools and the proposed collaborations. Students of the School of Journalism and Mass Communications, along with postgraduate-level researchers and professional journalists, form the sample for this investigation, which has a two-folded target: To guide the rapid prototyping process for system development and to validate specific hypotheses by answering the corresponding research questions. The results indicate the impact of mobile/on-demand support and training on journalistic practices and the attitudes of future journalists towards specialized technology in the era of constantly evolving digital journalism.

Speaker diarization consists of answering the question of "who spoke when" in audio recordings. In meeting scenarios, the task of labeling audio with the corresponding speaker identities can be further assisted by the exploitation of spatial features. This work proposes a framework designed to assess the effectiveness of combining speaker embeddings with Time Difference of Arrival (TDOA) values from available microphone sensor arrays in meetings. We extract speaker embeddings using two popular and robust pre-trained models, ECAPA-TDNN and X-vectors, and calculate the TDOA values via the Generalized Cross-Correlation (GCC) method with Phase Transform (PHAT) weighting. Although ECAPA-TDNN outperforms the Xvectors model, we utilize both speaker embedding models to explore the potential of employing a computationally lighter model when spatial information is exploited. Various techniques for combining the spatial–temporal information are examined in order to determine the best clustering method. The proposed framework is evaluated on two multichannel datasets: the AVLab Speaker Localization dataset and a multichannel dataset (SpeaD-M3C) enriched in the context of the present work with supplementary information from smartphone recordings. Our results strongly indicate that the integration of spatial information can significantly improve the performance of state-of-the-art deep learning diarization models, presenting a 2–3% reduction in DER compared to the baseline approach on the evaluated datasets. [ABSTRACT FROM AUTHOR]

This study presents a novel audio compression technique, tailored for environmental monitoring within multi-modal data processing pipelines. Considering the crucial role that audio data play in environmental evaluations, particularly in contexts with extreme resource limitations, our strategy substantially decreases bit rates to facilitate efficient data transfer and storage. This is accomplished without undermining the accuracy necessary for trustworthy air pollution analysis while simultaneously minimizing processing expenses. More specifically, our approach fuses a Deep-Learning-based model, optimized for edge devices, along with a conventional coding schema for audio compression. Once transmitted to the cloud, the compressed data undergo a decoding process, leveraging vast cloud computing resources for accurate reconstruction and classification. The experimental results indicate that our approach leads to a relatively minor decrease in accuracy, even at notably low bit rates, and demonstrates strong robustness in identifying data from labels not included in our training dataset. [ABSTRACT FROM AUTHOR]

MoJo refers to the emerging framework of covering news reporting workflows using smart mobile devices with dedicated software or hardware. However, a fully integrated and dedicated computing environment that can address the individual needs of both professional and citizen journalists is still missing. We introduce MoJo--MATE, (Mobile Journalism Machine AssisTEd reporting) a platform that integrates tools concerning all workflows at all stages of news reporting (content creation, editing, fact-checking, content management and publishing), thus providing interconnection of the newsroom collaborators and the public. We distinguish the client-/terminal-side and the server-side as two discrete components of the platform. The so-called jReporter mobile application includes tools and automations to support journalists in audiovisual capturing and editing along with semantic metadata extraction by making use of the smart devices. This is essential, since, in MOJO, the reporter has full responsibility of correct camera and microphone handling, without the assistance of a professional operator. Dedicated editing tools enhance portability, allowing at the same time the collection of log files with editing history, for authentication purposes. The created content is uploaded from the mobile devices to the server-side repositories, where Graphical User Interfaces are provided for accessing and managing media assets. A database containing information about the registered contributors and their respective types of engagement is maintained, controlling content rights and functionalities depending on the discrete users' roles (e.g. reporter, technical stuff, editor, journalist, and contributor). The metadata created in client-side is utilized for intelligent Multimedia Asset Management, efficient and comprehensible content publishing, including cross-validation and fact-checking aspects. [ABSTRACT FROM AUTHOR]

This paper investigates the applicability of machine-driven Speech Emotion Recognition (SER) towards the augmentation of theatrical performances and interactions (e.g., controlling stage color/light, stimulating active audience engagement, actors' interactive training, etc.). For the needs of the classification experiments, the Acted Emotional Speech Dynamic Database (AESDD) is developed, containing spoken utterances by five actors in five emotions. Several audio features and various classification techniques are implemented and evaluated, based on their performance with the AESDD, while also comparing to the well-known SAVEE database. The trained classifier is integrated in a novel application that performs live SER, fitting the needs of actors' training, while simultaneously augmenting the AESDD repository. [ABSTRACT FROM AUTHOR]