https://doi.org/10.5061/dryad.9w0vt4br6

Description of the data and file structure

We analyzed each propagated and recorded song using the R package barulho (Araya-Salas 2020), where each reference signal was compared to each recorded signal at different distances and at two equal heights. Barulho added acoustic markers at the beginning and end of our propagation stimuli to time-sync re-recorded signals, which facilitates the quantification of degradation variables. The degradation variables we measured included blur ratio, “the degradation of sound as a function of the change in signal energy in the time domain,” as described by Dabelsteen et al. (1993). Blur ratio is calculated as Ex / Ey, where Ey is the energy of the re-recorded signal. A high blur ratio value indicates greater degradation. Signal-to-noise ratio measures signal amplitude in relation to ambient noise (Dabelsteen et al., 1993). Signal-to-noise ratio is calculated as 10 log((Ey—En) / En), where ‘n’ equals background noise. A low value of signal-to-nois,e ratio indicates greater degradation. Excess attenuation measures additional loss of amplitude over what is expected due to spherical spreading (Wiley and Richards 1978). Excess attenuation is calculated as –20 log(k)–A, where A equals attenuation and ratio k is the ratio between the energy of the reference and re-recorded signal (Dabelsteen et al. 1993). A greater value of excess attenuation signifies greater degradation. Envelope correlation measures the similarity of the amplitude envelopes between two signals in the time domain. A low value signifies greater degradation. Tail-to-signal ratio measures the energy ratio of the signal's tail to the energy in the signal itself (Balsby et al. 2003). Tail-to-signal ratio is calculated as 10 log ((Etail − En) / (Ey − En)), where the tail end of the signal represents Etail and is described as the successive fading of the signal’s tail (Holland et al. 2001). Greater values of tail-to-signal ratio signify greater degradation.

Feature measurem,ents of common and rare song types from our maser sound fileweres measured using the bioacoustic computer program Luscinia., Using the program Luscinia (Lachlan 2007) we measured the spectral and temporal features of whole songs, of introductory notes (either a pure tone or buzz eleGaussiand of the syllable portion of the song. We used a 221 Window Length, 80% window overlap, dynamic range = 50 dB, frame length = 5 ms, time step = 1 ms, and a gaussian window function for our spectrogram seFigure. For each song or song syllable, we measured spectral-based properties that included peak frequency, frequency bandwidth, and maximum and minimum frequency (figure 2). For the temporal-based properties of each song or song syllable, we measured the syllable gap and duration (Figure 2). Spectral and temporal-based measurements were made using an amplitude threshold, which Luscinia provides.

Files and variables

File: IntroNote_measurements.csv

Description: Contains acoustic measurements of introductory notes from common and rare song types.

Variables

• Song.name: Unique identifier for each recorded song.
• Length: Duration of the introductory note (milliseconds).
• Peak.freq: Frequency containing the greatest amount of acoustic energy within the introductory note (Hz).
• Min.freq: Minimum frequency of the introductory note measured using the amplitude threshold (Hz).
• Max.freq: Maximum frequency of the introductory note measured using the amplitude threshold (Hz).
• Freq.bandwidth: Difference between the maximum and minimum frequencies of the introductory note (Hz).
• song.type: Classification of the song as either common or rare.

File: Syllable_measurements.csv

Description: Contains acoustic measurements of individual syllables from common and rare song types.

Variables

• Song.name: Unique identifier for each recorded song.
• Syllable.number: Sequential number identifying the syllable measured within the song.
• Length: Duration of the syllable (milliseconds).
• Peak.frequency: Frequency containing the greatest amount of acoustic energy within the syllable (Hz).
• Min.freq: Minimum frequency of the syllable measured using the amplitude threshold (Hz).
• Max.freq: Maximum frequency of the syllable measured using the amplitude threshold (Hz).
• Between.syllable.gap: Silent interval between the measured syllable and the following syllable (milliseconds).
• Freq.bandwidth: Difference between the maximum and minimum frequencies of the syllable (Hz).
• song.type: Classification of the song as either common or rare. 

File: WholeSong_measurements.csv

Description: Contains acoustic measurements made at the whole-song level for common and rare song types.

Variables

• Song name: Unique identifier for each recorded song.
• Element.Number: Sequential number of the vocal element
• Length: Total duration of the song from the beginning of the introductory note to the end of the final syllable (milliseconds).
• Peak.Freq: Frequency containing the greatest amount of acoustic energy within the entire song (Hz).
• Maximum.Freq: Highest frequency measured across the entire song using the amplitude threshold (Hz).
• Minimum.Freq: Lowest frequency measured across the entire song using the amplitude threshold (Hz).
• Freq.Bandwidth: Difference between the maximum and minimum frequencies of the entire song (Hz).
• Song.Type: Classification of the song as either common or rare.

File: Degradation.csv

Description: Contains acoustic degradation measurements used in analyses of song propagation. It also contains other related measurements used in this study. 

Variables

• Sound.files: Name of the propagated sound file analyzed.
• Start.time: Start time of the annotation from the propagated song (seconds).
• End.time: End time of the annotation from the propagated song (seconds).
• Min.freq: Lower frequency limit of song from the propagation sound file (Hz).
• Max.freq: Upper frequency limit of song from the propagation sound file (Hz).
• Site: Name of the site where the propagation experiment was conducted.
• Height: Speaker and microphone height used during sound propagation (categorized as either short or tall).
• Julian.date: Julian day on which the propagation trial was conducted.
• Wind.Speed: Wind speed during the propagation trial (m/s).
• Tree.Density.Hectare: Measure of tree density at the propagation site (meters squared per hectare).
• Distance: Distance between the loudspeaker and recording microphone (meters).
• Signal.Type: Classification of the propagated song as either common or rare.
• Reference: Reference recording used to calculate degradation metrics by comparison with the rerecorded signal.

• Blur.Ratio: A Degradation variable, calculated as Ex/Ey, where ‘Ey’ is the energy of the re-recorded signal, and ‘Ex’ is the energy of the reference signal. A high blur ratio value indicates greater degradation. Propagated songs in the data frame at a distance of 1.5m contain NAs because baRulho uses propagated songs recorded at 1.5m as reference signals. Since these recordings serve as the reference against which recordings at greater distances are compared, blur ratio cannot be calculated for the 1.5 m recordings themselves.
• Amp.Env.Corr: Amplitude envelope correlation. A degradation variable that measures the similarity of the amplitude envelopes between the reference signal and the propagated signal in the time domain. Amplitude envelope correlation measures degradation similarly to blur ratio; however, it includes the effect of the overall energy attenuation (Brown and Handford 1996). Propagated songs in the data frame at a distance of 1.5m contain NAs because baRulho uses propagated songs recorded at 1.5m as reference signals. Since these recordings serve as the reference against which recordings at greater distances are compared, amplitude envelope correlation cannot be calculated for the 1.5 m recordings themselves.
• Signal.to.noise.Ratio: A degradation variable that measures signal amplitude in relation to ambient noise (Dabelsteen et al. 1993). Signal-to-noise ratio is calculated as 10 log((Ey – En)/En), where ‘n’ equals background noise. A low value of signal-to-noise ratio indicates greater degradation. 
• Tail.to.signal.Ratio: A degradation variable that measures the energy ratio of the signal’s tail to the energy in the signal itself (Balsby et al. 2003). Tail-to-signal ratio is calculated as 10 log ((Etail − En)/(Ey − En)), where the tail end of the signal represents Etail and is described as the successive fading of the signal’s tail (Holland et al. 2001). Higher values of tail-to-signal ratio indicate greater degradation.

• Excess.Attenuation: A degradation variable that measures additional loss of amplitude over what is expected due to spherical spreading (Wiley and Richards 1978). Excess attenuation is calculated as -20 log (k)-A, where A equals attenuation that is due to spherical spreading and ratio k is the ratio between the energy of the reference and re-recorded signal (Dabelsteen et al. 1993). A greater value of excess attenuation indicates greater degradation. Propagated songs in the data frame at a distance of 1.5m contain NAs because baRulho uses propagated songs recorded at 1.5m as reference signals. Since these recordings serve as the reference against which recordings at greater distances are compared, excess attenuation cannot be calculated for the 1.5 m recordings themselves.

File: ESM-Manuscript.Bioacoustics_.docx

Description: Electronic Supplementary Material accompanying this manuscript uploaded to Zenodo.