Audio segmentation of broadcast news in the Albayzin-2010 evaluation: overview, results, and discussion

Abstract

Recently, audio segmentation has attracted research interest because of its usefulness in several applications like audio indexing and retrieval, subtitling, monitoring of acoustic scenes, etc. Moreover, a previous audio segmentation stage may be useful to improve the robustness of speech technologies like automatic speech recognition and speaker diarization. In this article, we present the evaluation of broadcast news audio segmentation systems carried out in the context of the Albayzín-2010 evaluation campaign. That evaluation consisted of segmenting audio from the 3/24 Catalan TV channel into five acoustic classes: music, speech, speech over music, speech over noise, and the other. The evaluation results displayed the difficulty of this segmentation task. In this article, after presenting the database and metric, as well as the feature extraction methods and segmentation techniques used by the submitted systems, the experimental results are analyzed and compared, with the aim of gaining an insight into the proposed solutions, and looking for directions which are promising.

Albayzín-2014 evaluation: audio segmentation and classification in broadcast news domains

“…Using this database an audio segmentation task was proposed, where the systems were required to identify the presence of speech, music and/or noise, either isolated or overlapped. The Albayzín-2014 Audio Segmentation Evaluation contributed to the evolution of the audio segmentation technology in broadcast news domains by providing a more general and realistic database, compared to those used in the Albayzín-2010 and -2012 Audio Segmentation Evaluations [10,30]. The main features of the approaches and the results attained by seven segmentation systems from four different research groups have been presented and briefly analyzed.…”

“…However, some classes are better described by the statistics computed over longer periods of time (from 0.5 to 5 s long). These characteristics are referred in the literature as segment-based features [29,30]. For example, in [31], a content-based speech discrimination algorithm is designed to exploit the long-term information inherent in the modulation spectrum; and in [32], authors propose two segment-based features: the variance of the spectrum flux (VSF) and the variance of the zero crossing rate (VZCR).…”

Audio segmentation-by-classification approach based on factor analysis in broadcast news domain

“…However, there is an important amount of long segments (longer than 60 s). More details about the database and the labeling process can be found in [19].…”

“…A complete description of the Albayzin 2010 audio segmentation and classification evaluation can be found in [19] where the participant's approaches and the results are presented. We describe the database and the metric used in the evaluation in the next subsections.…”

“…However, some classes are better described by the statistics computed over consecutive frames (from 0.5 to 5 s long). These characteristics are referred in the literature as segment-based features [18,19]. For example, in [20], a content-based speech discrimination algorithm is designed to exploit the long-term information inherent in the modulation spectrum.…”