MatlabHTK: a simple interface for bioacoustic analyses using hidden Markov models

Summary 1.Passive bioacoustic recording devices are now widely available and able to continuously record remotely located sites for extended periods, offering great potential for wildlife monitoring and management. Analysis of the huge datasets generated, in particular for specific biotic sound recognition, remains a critical bottleneck for widespread adoption of these technologies as current methods are labour intensive. 2.Several methods borrowed from speech processing frameworks, such as hidden Markov models, have been successful in analysing bioacoustic data but the software implementations can be expensive and difficult to use for non-specialists involved in wildlife conservation. To remedy this, we present a software interface to a popular speech recognition system making it possible for non-experts to implement hidden Markov models for bioacoustic signal processing. Octave/Matlab functions are used to simplify the set up and the definition of a bioacoustic signal recogniser as well as the analysis of the results. 3.We present the different functions as a workflow. To demonstrate how the package can be used we give the results of an analysis of a bioacoustic monitoring dataset to detect the nocturnal presence and behaviour of a cryptic seabird species, the common diving petrel Pelecanoides urinatrix urinatrix, from Northern New Zealand. 4.We show that the package matlabHTK can be used efficiently to reconstruct the daily patterns of colony activity in the common diving petrel. This article is protected by copyright. All rights reserved.

[1]  Kuntoro Adi,et al.  Acoustic censusing using automatic vocalization classification and identity recognition. , 2010, The Journal of the Acoustical Society of America.

[2]  Michael T. Johnson,et al.  A Framework for Bioacoustic Vocalization Analysis Using Hidden Markov Models , 2009, Algorithms.

[3]  Massimo Venturelli,et al.  Possible Predictors of Involuntary Weight Loss in Patients with Alzheimer’s Disease , 2016, PloS one.

[4]  Otilia Kocsis,et al.  Integration of Temporal Contextual Information for Robust Acoustic Recognition of Bird Species from Real-Field Data , 2013 .

[5]  Lawrence R. Rabiner,et al.  A tutorial on hidden Markov models and selected applications in speech recognition , 1989, Proc. IEEE.

[6]  Klaus Riede,et al.  Automatic bird sound detection in long real-field recordings: Applications and tools , 2014 .

[7]  J A Kogan,et al.  Automated recognition of bird song elements from continuous recordings using dynamic time warping and hidden Markov models: a comparative study. , 1998, The Journal of the Acoustical Society of America.

[8]  M. Hauber,et al.  Analyses of sex and individual differences in vocalizations of Australasian gannets using a dynamic time warping algorithm. , 2012, The Journal of the Acoustical Society of America.

[9]  M. Fry,et al.  Vocal classification of vocalizations of a pair of Asian small-clawed otters to determine stress. , 2015, The Journal of the Acoustical Society of America.

[10]  Amal Punchihewa,et al.  Birdsong Denoising Using Wavelets , 2016, PloS one.

[11]  Charles E Taylor,et al.  Automated species recognition of antbirds in a Mexican rainforest using hidden Markov models. , 2008, The Journal of the Acoustical Society of America.

[12]  Louis Ranjard,et al.  Integration over song classification replicates: song variant analysis in the hihi. , 2015, The Journal of the Acoustical Society of America.

[13]  Panu Somervuo,et al.  Parametric Representations of Bird Sounds for Automatic Species Recognition , 2006, IEEE Transactions on Audio, Speech, and Language Processing.

[14]  L. E. Richdale,et al.  Biology of the birds of Whero Island, New Zealand, with special reference to the Diving Petrel and the White‐faced Storm Petrel , 2010 .

[15]  Manuel Vieira,et al.  Call recognition and individual identification of fish vocalizations based on automatic speech recognition: An example with the Lusitanian toadfish. , 2015, The Journal of the Acoustical Society of America.

[16]  T. S. Brandes,et al.  Feature Vector Selection and Use With Hidden Markov Models to Identify Frequency-Modulated Bioacoustic Signals Amidst Noise , 2008, IEEE Transactions on Audio, Speech, and Language Processing.

[17]  Ricardo B. Machado,et al.  Comparisons Between Autonomous Acoustic Recordings and Avian Point Counts in Open Woodland Savanna , 2015 .

[18]  T. Mitchell Aide,et al.  Real-time bioacoustics monitoring and automated species identification , 2013, PeerJ.