Bioacoustic signal denoising: a review

Xie, Jie; Colonna, Juan Gabriel; Zhang, Jinglan

doi:10.1007/s10462-020-09932-4

Cited by 42 publications

(31 citation statements)

References 80 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In some cases, classification was derived from band-level indices (e.g., Širović et al, 2015;Oestreich et al, 2020). When the training data are insufficient, additional exemplars can be added by transforming the true signals (and labeling them as false) or adding different types and amounts (Barchiesi et al, 2015;Benetos et al, 2015;Shane et al, 2015;Zabalza et al, 2015;Freeman and Freeman, 2016;Lin et al, 2017a,b;Lin et al, , 2019Lin et al, , 2021Tsao, 2018, 2020;Colonna et al, 2018;Guan et al, 2018;Reis et al, 2018;Seger et al, 2018;Virtanen et al, 2018;Abeßer, 2020;Gatto et al, 2020;Mooney et al, 2020;Shajahan et al, 2020;Xie et al, 2020;Dias et al, 2021;Dimoff et al, 2021;Gabrielli et al, 2021;Ick and McFee, 2021;Ozanich et al, 2021).…”

Section: Analytical Approaches To Soundscape Analysesmentioning

confidence: 99%

“…The literature referenced simply serve as examples, and not an exhaustive or systematic search of the literature. References cited in the table:(Barchiesi et al, 2015;Benetos et al, 2015;Shane et al, 2015;Zabalza et al, 2015;Freeman and Freeman, 2016; Lin et al, 2017a,b;Lin et al, , 2019Lin et al, , 2021 Tsao, 2018, 2020;Colonna et al, 2018;Guan et al, 2018;Reis et al, 2018;Seger et al, 2018;Virtanen et al, 2018;Abeßer, 2020;Gatto et al, 2020;Mooney et al, 2020;Shajahan et al, 2020;Xie et al, 2020;Dias et al, 2021;Dimoff et al, 2021;Gabrielli et al, 2021;Ick and McFee, 2021;Ozanich et al, 2021).…”

mentioning

confidence: 99%

See 1 more Smart Citation

Advancing the Interpretation of Shallow Water Marine Soundscapes

et al. 2021

View full text Add to dashboard Cite

Soundscapes offer rich descriptions of composite acoustic environments. Characterizing marine soundscapes simply through sound levels results in incomplete descriptions, limits the understanding of unique features, and impedes meaningful comparisons. Sources that contribute to sound level metrics shift in time and space with changes in biological patterns, physical forces, and human activity. The presence of a constant or chronic source is often interwoven with episodic sounds. Further, the presence and intensity of sources can influence other sources, calling for a more integrated approach to characterizing soundscapes. Here, we illustrate this complexity using data from a national-scale effort, the Sanctuary Soundscape Monitoring Project (SanctSound), an initiative designed to support collection of biological, environmental, and human use data to compliment the interpretation of sound level measurements. Using nine examples from this diverse dataset we demonstrate the benefit of integrating source identification and site features to interpret sound levels across a diversity of shallow water marine soundscapes (<150 m). Sound levels from sites in high human use areas reflect the anthropogenic influences on the soundscape, especially when measuring broad frequency bands, whereas sites with relatively low human activity and high sound levels reveal biological features of the soundscape. At sites with large tidal changes, sound levels correlated with the magnitude of tidal flow, and sound levels during high tidal flow periods were similar to sound levels at sites near commercial shipping lanes. At sites in relatively close proximity (<30 km), sound levels diverge because of the presence of different proximate sound contributors and propagation features of the site. A review of emerging methodologies for integrated soundscape analysis, including acoustic scenes, provides a framework for interpreting soundscapes across a variety of conditions. With a global growth in monitoring efforts collecting standardized measurements over widely distributed arrays, more integrated methods are needed to advance the utility of soundscapes in marine resource management.

show abstract

Section: Analytical Approaches To Soundscape Analysesmentioning

confidence: 99%

mentioning

confidence: 99%

Advancing the Interpretation of Shallow Water Marine Soundscapes

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Determining what constitutes noise in recordings is non-trivial and impacts what type of noise reduction algorithm can and should be used. In a systematic review of noise reduction methods in bio-acoustics, Xie et al ( 2020 ) outline six classes of noise reduction algorithms used for bio-acoustics: (1) Optimal FIR filter (e.g., Kim et al, 2000 ), (2) spectral subtraction (e.g., Boll, 1979 ; Kiapuchinski et al, 2012 ; Sainburg et al, 2020b ), (3) minimum-mean square error short-time spectral amplitude estimator (MMSE-STSA; e.g., Ephraim and Malah, 1984 ; Alonso et al, 2017 ; Brown et al, 2017 ) (4) wavelet based denoising (e.g., Ren et al, 2008 ; Priyadarshani et al, 2016 ) (5) image processing based noise reduction, and (6) deep learning based noised reduction. These noise reduction algorithms can be broadly divided into two categories: stationary and non-stationary noise reduction ( Figure 1A ).…”

Section: Signal Processing and Denoisingmentioning

confidence: 99%

Toward a Computational Neuroethology of Vocal Communication: From Bioacoustics to Neurophysiology, Emerging Tools and Future Directions

Sainburg

Gentner

2021

Front. Behav. Neurosci.

View full text Add to dashboard Cite

Recently developed methods in computational neuroethology have enabled increasingly detailed and comprehensive quantification of animal movements and behavioral kinematics. Vocal communication behavior is well poised for application of similar large-scale quantification methods in the service of physiological and ethological studies. This review describes emerging techniques that can be applied to acoustic and vocal communication signals with the goal of enabling study beyond a small number of model species. We review a range of modern computational methods for bioacoustics, signal processing, and brain-behavior mapping. Along with a discussion of recent advances and techniques, we include challenges and broader goals in establishing a framework for the computational neuroethology of vocal communication.

show abstract

“…As the amount of noise increases-Signal-Noise Ratio (SNR) decreases, the correlation between two notes increases, eventually giving low NVI scores than expected (Sup materials). For recordings with some amount of background noise, various techniques like denoising (Brown, Garg, & Montgomery, 2018;Xie, Colonna, & Zhang, 2020) can be used before implementing NVI. Though denoising methods can improve the efficacy of NVI, it is recommended to use good quality focal recordings of the target species.…”

Section: Limitations Of Note Variability Indexmentioning

confidence: 99%

Spectrogram cross-correlation can be used to measure the complexity of bird vocalizations

Sawant

Arvind

Joshi

et al. 2021

Preprint

View full text Add to dashboard Cite

Birdsong is an important signal in mate attraction and territorial defense. Quantifying the complexity of these songs can shed light on individual fitness, sexual selection, and behavior. Several techniques have been used to quantify song complexity and be broadly categorized into diversity indices, measures of stationary probabilities, and measures of sequential variations. However, these methods are unable to account for important acoustic features like the frequency bandwidth and the variety in the shape of syllables which are an integral part of these vocal signals. This study proposes a new complexity measure that considers intra-song note variability and calculates a weighted index for birdsongs using spectral cross-correlation. We compared the previously described methods to understand the advantages and limitations based on the factors that would be affecting the complexity of songs. We developed a new method- Note Variability Index (NVI), which incorporates the spectral features of notes while quantifying complexity. This measure alleviates the need for manual annotations of notes that can be error-prone. We used Spectrogram Cross-Correlation (SPCC) to compare notes within a song and used the output values to quantify song complexity. To check for the efficacy of the new method, we generated synthetic songs to caricature extremes in song complexity and compared selected conventional complexity measures along with the NVI. We provide case-specific limitations of these methods. Additionally, to examine the efficacy of this new method in real-world scenarios, we used natural birdsongs from multiple species across the globe with varying song structures to compare conventional methods with NVI. To our knowledge, NVI is the only song complexity method that captures the variation of spectral features of notes in songs where the conventional methods fail to distinguish between similar song structures with different note types. As NVI does not need a manual classification of notes, it can be easily implemented for any type of birdsong with existing sound analysis softwares; it is very quick, avoids the possible biases in note classification, and can possibly be automated for large datasets in the future.

show abstract

Bioacoustic signal denoising: a review

Cited by 42 publications

References 80 publications

Advancing the Interpretation of Shallow Water Marine Soundscapes

Advancing the Interpretation of Shallow Water Marine Soundscapes

Toward a Computational Neuroethology of Vocal Communication: From Bioacoustics to Neurophysiology, Emerging Tools and Future Directions

Spectrogram cross-correlation can be used to measure the complexity of bird vocalizations

Contact Info

Product

Resources

About