Common Terminology and Acoustic Measures for Human …

DownloaTerms o

JSLHR

Research Note

aDepartmentbDepartmentof Arizona, T

Robin A. Sam

Corresponden

Editor-in-ChiEditor: Miche

Received MayRevision receAccepted Aughttps://doi.org

ded From: httf Use: https://p

Common Terminology and AcousticMeasures for Human Voice and Birdsong
Areen Badwal,a JoHanna Poertner,b Robin A. Samlan,b and Julie E. Millera,b
Purpose: The zebra finch is used as a model to study theneural circuitry of auditory-guided human vocal production.The terminology of birdsong production and acousticanalysis, however, differs from human voice production,making it difficult for voice researchers of either speciesto navigate the literature from the other. The purpose ofthis research note is to identify common terminology andmeasures to better compare information across species.Method: Terminology used in the birdsong literature willbe mapped onto terminology used in the human voiceproduction literature. Measures typically used to quantifythe percepts of pitch, loudness, and quality will bedescribed. Measures common to the literature in bothspecies will be made from the songs of 3 middle-agebirds using Praat and Song Analysis Pro. Two measures,cepstral peak prominence (CPP) and Wiener entropy

of Neuroscience, University of Arizona, Tucsonof Speech, Language, and Hearing Sciences, Universityucson

lan and Julie E. Miller contributed equally to this article.

ce to Julie E. Miller: [email protected]

ef: Julie Lisslle Ciucci

30, 2018ived July 20, 2018ust 13, 2018/10.1044/2018_JSLHR-S-18-0218

Journal of Speech, Language, and Hearing Researc

This work is licensed under a Creative Commops://jslhr.pubs.asha.org/ on 12/12/2018ubs.asha.org/ss/rights_and_permissions.aspx

(WE), will be compared to determine if they provide similarinformation.Results: Similarities and differences in terminology andacoustic analyses are presented. A core set of measuresincluding frequency, frequency variability within a syllable,intensity, CPP, and WE are proposed for future studies.CPP and WE are related yet provide unique informationabout the syllable structure.Conclusions: Using a core set of measures familiar toboth human voice and birdsong researchers, along withboth CPP and WE, will allow characterization of similaritiesand differences among birds. Standard terminology andmeasures will improve accessibility of the birdsong literatureto human voice researchers and vice versa.Supplemental Material: https://doi.org/10.23641/asha.7438964

S ongbirds have long been utilized as models for study-ing the neural circuitry for auditory-guided humanvocal learning and production. Two songbird species,

the zebra finch and Bengalese finch, are well suited to thesepurposes due to their easy breeding and adaptability to cap-tivity, preference for social housing, and the abundance ofliterature on central brain mechanisms for vocal motor con-trol. Similarities and differences in central and peripheralmechanisms of finch and human sound production arereviewed in the following paragraphs, followed by a com-parison of their production mechanisms. It remains unclearhow common acoustic measures of birdsong relate to humanvoice. The purpose of this research note is to propose a“translational dictionary” to facilitate sharing results andknowledge across the human voice and birdsong literature

including demonstrating the feasibility of applying acousticmeasurements made in human voice to structural elementswithin the birdsong.

In the zebra finch species, the males sing and thefemales do not, a sexually dimorphic behavior establishedearly on with the development and growth of the song con-trol system in males but not in females. The finch is anexcellent model for studying neural circuits for humanvocal motor control (Brainard & Doupe, 2013). In finches,there are identifiable song-dedicated brain nuclei locatedwithin cortical and basal ganglia tissue. Evidence fromanatomical and gene expression studies show a high degreeof similarity between finch song control regions RA, LMAN,and Area X (abbreviations used as proper names) withhuman brain regions involved in speech motor productionand planning. Specifically, RA is correlated with the pri-mary motor cortex, LMAN with Broca’s area, and Area Xwith the striatum (Pfenning et al., 2014). Recent electro-physiological evidence has identified a direct connectionbetween the human primary motor cortex and the larynxin the control of pitch during speaking and singing, a find-ing that further supports the use of the finch model tostudy voice control (Dichter, Breshears, Leonard, & Chang,2018).

Disclosure: The authors have declared that no competing interests existed at the timeof publication.

h • 1–10 • Copyright © 2018 The Authors

ns Attribution 4.0 International License.

1

https://doi.org/10.23641/asha.7438964

https://doi.org/10.23641/asha.7438964

https://doi.org/10.1044/2018_JSLHR-S-18-0218

DownloaTerms o

Finches and humans also share a cortico-basal ganglia-thalamo-cortical loop that is under neuromodulatory controlby dopamine (Simonyan, Horwitz, & Jarvis, 2012). In bothspecies, primary and supplementary motor cortices providedescending excitatory glutamatergic input to basal gangliastructures. The basal ganglia provide GABAergic inhibitionof the thalamus, which relays input back to higher corticalstructures. In finches, song-dedicated cortical nucleus RA isthe output nucleus of the song control circuit sending descend-ing drive to the hypoglossal cranial motor neurons in thebrainstem and then on to the syrinx, the main vocal organ(vs. the larynx in humans).

Experimental studies in finches have provided neededinsight into genetic and neuromodulatory control of bird-song as a model for human voice research. Notably, onewidely used example of shared genetic function betweenfinches and humans is the speech gene, FOXP2, which affectsvocal learning and production in both species (Haesleret al., 2007; Heston & White, 2015; Lai, Fisher, Hurst,Vargha-Khadem, & Monaco, 2001; Vargha-Khadem et al.,1998; Watkins, Gadian, & Vargha-Khadem, 1999). Inhumans, abnormal expression of the FOXP2 protein inhuman cortical and basal ganglia regions is associated witharticulatory deficits including impaired sequencing of oro-facial movements; whether FOXP2 contributes to voicedeficits (e.g., loudness, pitch, or quality) has not been exam-ined. Intriguingly, virally driven genetic manipulation ofFOXP2 levels in Area X of adult male zebra finches altersone aspect of voice, pitch control, which is acoustically mea-sured as changes in fundamental frequency oscillation (fo),of harmonic elements in both finch song and in human voice(Murugan, Harward, Scharff, & Mooney, 2013; Stemple,Roy, & Klaben, 2014). Changes in fo are also detectedwith experimental manipulation of dopamine levels in finchArea X during auditory-driven vocal learning tasks in adultmales and dependent on whether the male is singing aloneor to a female (Hoffmann, Saravanan, Wood, He, & Sober,2016; Leblois & Perkel, 2012; Leblois, Wendel, & Perkel,2010). However, fo measurements made in birdsong arerestricted to a small subset of harmonic elements withinthe bird’s song, thereby requiring a large number of birdsto achieve a sufficient sample size. To facilitate future com-parisons between birdsong and voice research studies, it iscritical to identify acoustic voice measurements that can bereliably made from inharmonic aspects of the bird’s songand have the potential to change with age or intervention.

Central mechanisms drive the peripheral mechanisms(e.g., song control nuclei) to produce the acoustic signal.Both central and peripheral mechanisms operate to deter-mine fo (pitch) and intensity (loudness), features that varyas the finch sings (Brumm & Slater, 2006; Kao & Brainard,2006; Kao, Doupe, & Brainard, 2005; Sober, Wohlgemuth,& Brainard, 2008). Humans and zebra finches share severalperipheral sound generation mechanisms and present withsome key differences. As reviewed by Riede and Goller(2010), similarities and differences involve the respiratorysystem, oscillating masses, and supraglottal structures. Bothspecies phonate primarily on exhalation. Muscular control

2 Journal of Speech, Language, and Hearing Research • 1–10

ded From: https://jslhr.pubs.asha.org/ on 12/12/2018f Use: https://pubs.asha.org/ss/rights_and_permissions.aspx

regulates airflow to and from the lungs in both, thoughrecoil forces do not appear to contribute to driving pressuresin avian voice production (Riede & Goller, 2010). Bothspecies have oscillating tissue masses that abduct and adduct,though the avian syrinx has two independently controlledsound generators, the medial and lateral labia. The mostcommon vibratory modes present in human vocal fold vibra-tion are also thought to occur with avian labial vibration.In songbirds, the medial-to-lateral vibratory mode has beenobserved and the rotational mode is hypothesized, thoughlittle is known about the histological composition of thelabia, and songbirds might lack the layered structure of thelamina propria that underlies the rotational mode (Riede &Goller, 2010). The rate of vibration determines fo. The supra-glottal vocal tract contributes to tuning resonances in bothspecies. Muscular control of the upper vocal tract in song-birds, consisting of the tracheal tube and the oropharyngeal–esophageal cavity, is used to enhance select componentsof the acoustic signal, much as humans use the pharynx,mouth, and nose to enhance specific components of the acous-tic signal. In general, the songbird voice production systemsappear to be adapted for high-speed vocal output and controlof timing, and human laryngeal muscles are adapted forfine control of tension (Riede & Goller, 2010).

Research studies using finches and other songbirdsrely upon a standard set of acoustic measurements to describebirdsong. Several of these measures are often unfamiliar toresearchers of human voice, and vice versa, limiting sharingof information across species. In order to interpret birdsongdata in the context of human voice metrics, there needs tobe a common “language” between birdsong and humanvoice. Therefore, we had two aims in the current study:(a) establish a “translational dictionary” of birdsong andhuman voice terminology and measures, (b) determine thefeasibility of using a set of measures familiar to humanvoice researchers that will characterize key features of bird-song, and (c) identify the extent to which Wiener entropy (WE)relates to cepstral peak prominence (CPP). WE and CPPboth provide information about harmonic and noise energyin the sound (Hillenbrand & Houde, 1996; Tchernichovski,Nottebohm, Ho, Pesaran, & Mitra, 2000). Importantly, wehypothesize that WE or CPP can be used to provide similarinformation about the birdsong. To our knowledge, thesetwo aims have not been pursued previously in birdsongresearch and should facilitate future investigations usingthe songbird model to investigate the impact of aging andneurodegenerative diseases on vocal communication.

Aim 1: Establishing a Translational DictionaryThere are two components to the translational dic-

tionary: (a) perceptual terminology related to the soundsbirds and humans produce, with emphasis on voice, and(b) acoustic measures used to define voice quality, loudness,and pitch. Acoustic analysis of human voice and birdsong isnot standardized in the birdsong or human voice literature.We therefore reviewed several examples of acoustic mea-surements made in adult birdsong with relevance to voice.

DownloaTerms o

First, it is necessary to provide a description of the bird-song structure in terms familiar to human voice researchers.Both birdsong and human speech include time-varyingacoustic products consisting of alternating harmonic andnoisy units. The comparison is not intended to suggest simi-larity in meaning (e.g., that a particular unit represents thesmallest unit of meaning). We refer the reader to alreadyexisting literature that draws parallels between phono-logical and syntactical features of birdsong and human lan-guage structure during development and adulthood (seeChapters 9–12 in Berwick & Chomsky, 2013; Lipkind et al.,2013).

The basic unit of birdsong is a sequence of repeatedsyllables known as a motif. Motifs are encoded by spe-cific patterns of neuronal firing in song control regionsand are separated in time by silent intervals (Hahnloser,Kozhevnikov, & Fee, 2002). The motif can vary overmultiple renditions by the order of the syllables (syntax)or insertion of a new syllable type. The audio signals andnarrow-band spectrograms for sample birdsong from threezebra finches are shown in Figures 1A–1C. The basic motifdiffers across birds, and the most complex motif for thesethree birds is in Panel C. Figure 1A shows two motifs. Forcomparison, Figure 1D shows a sentence spoken twice(“Shhh, finches perch in trees”) by a middle-aged humanfemale to illustrate similarities/differences in the spectralstructure of elements, human words, vowels, and consonantsto birdsong. The motifs and sentence show a generallysimilar pattern in that they are both composed of severalcomponents and that the energy in some, but not all, com-ponents has an fo and harmonics.

In birdsong, the motif is a series of “syllables” definedby their spectral profile and labeled in Figure 1 with capitalletters. Each syllable in the birdsong consists of one (e.g.,Syllables A, B, C, and E in Figure 1A) or more notes (e.g.,two notes: Syllable D, Figure 1A). All acoustic analysesin this research note were completed on the syllable level.Birdsong syllables are referred to as “noisy” or “harmonic,”and the designation determines the type of analysis com-pleted on the syllable. A well-defined harmonic syllable(e.g., Syllables “B & E” in Figure 1A and “B” in Figure 1B)has an fo and many harmonics, similar to a vowel such asthe “ee” in “trees.” Note that the fo for the human sentence(see Figure 1D) is generally lower than for the bird motif.The associated human harmonics tend to fade by approxi-mately 5000 Hz, whereas they continue to 10,000 Hz in thebird’s harmonic syllables. A noisy syllable (e.g., Syllable Ain Figures 1A–1B) has poorly defined harmonics with soundenergy visible between them. This type of birdsong syllableis more similar to a human fricative or affricate (such as“sh” or “ch”) than a vowel, yet the birdsong noisy syllablehas more harmonic structure than a fricative or affricate.Some syllables are composed of a harmonic and a noisynote (e.g., Figure 1A, Syllable “D”; Figure 1C, Syllable“I”). In this research note, those syllables are considered“mixed.” In our human sentence, the word finches and treesappear to resemble the mixed type of birdsong syllables, inthat there are harmonic and noisy elements within the words.


Perceptual Voice TerminologyBirdsong and human voice are both discussed using

the perceptual terms loudness, pitch, and quality. Workingdefinitions are summarized in Column 2 of Table 1 andtheir acoustic correlates in Column 3. The meanings ofpitch and loudness are similar in the birdsong and humanvoice literature and are perceptually judged using termssuch as louder or quieter (loudness) and higher or lower(pitch). Quality in birdsong is determined based on howsimilar one performance of a syllable or motif is to the nextrendition; in early song development, quality is defined by thedegree of song similarity (e.g., imitation) between the juve-nile finch and his adult tutor (Tchernichovski & Nottebohm,1998; Tchernichovski et al., 2000). Highly accurate andsimilar performances (scores) in adult males and reducedvariability in fo from one song rendition to the next indicatehigher quality song, and in social contexts, the higher qual-ity song is preferred by a female finch (Kao et al., 2005;Woolley & Doupe, 2008). The accuracy and similarity ofpitch and loudness from one rendition to the next in birdsongis enfolded into the concept of quality. In contrast, humanvoice “quality” is separated from pitch and loudness inhuman voice research. However, quality is difficult to defineand typically described using a series of terms such as breathy,strained, or rough (Bartholomew, 1934; Kempster, Gerratt,Abbott, Barkmeier-Kraemer, & Hillman, 2009). Together,the three aspects of voice provide considerable informationabout speaker identity, mood, physical health, and vitality.

Acoustic AnalysisWhen comparing standard acoustic measurements

of birdsong to a human voice, several similarities and dif-ferences emerge. In the birdsong field, fo and aggregateacoustic measures, known as similarity scores, are usedto describe the acoustic match of the juvenile pupil’s songto that of its adult tutor and to compare the effects of pre-experimental versus postexperimental treatments in finchsong at different ages (Haesler et al., 2007; Heston & White,2015; Miller, Hilliard, & White, 2010; Tchernichovski, Mitra,Lints, & Nottebohm, 2001). Similarity scores are calculatedat the motif and syllable level using WE, frequency modu-lation, pitch, pitch goodness, and amplitude modulation(Tchernichovski et al., 2000). WE is a common measureused to describe the effects of experimental treatment onbirdsong syllables. WE is defined as a measure of the widthand uniformity of the power spectrum and is measured ona logarithmic scale where zero is white noise and negativeinfinity is complete order (Tchernichovski et al., 2000).Thus, syllables with harmonic structure have more negative(e.g., lower) WE scores than noisy syllables.

Birdsong analysis contrasts with human voice acousticmeasurement, where several measures of intensity, fo, andspectral correlates of quality are common together withcomposite measures of quality (e.g., cepstral spectral index ofdysphonia; Awan, Roy, & Cohen, 2014). Current recommen-dations for standard acoustic evaluation include speaking fo,

Badwal et al.: Comparison of Birdsong and Human Voice 3

Figure 1. Birdsong motifs versus human sentence. Audio waveform (top) and spectrogram (bottom). Spectrogram is depicted as time (in seconds,x-axis) versus frequency (in Hertz, y-axis). Color in the spectrogram represents the amplitude of the spectral energy at that frequency, with redbeing the most intense followed by orange, yellow, and blue. (A) Two motifs from Bird R1156 shown as a sequence of repeated syllables (A–E)separated by a pause (red line). Syllables are assigned unique letters based on their visual appearance and confirmed by acoustic measurements.(B) Motif exemplar for Bird W35 composed of Syllables A–E. (C) Motif exemplar for Bird R1157, consisting of Syllables A–I. (D) Humansentence “SHHH, FINCHES PERCH IN TREES,” repeated twice.

DownloaTerms o

standard deviation of fo, maximum phonational frequencyrange, speaking intensity, maximum intensity range, andCPP (Patel et al., 2018). CPP is a robust and frequentlyused measure of overall voice quality that provides informa-tion about acoustic waveform periodicity (Fraile & Godino-



Llorente, 2014; Hillenbrand, Cleveland, & Erickson, 1994;Samlan, Story, & Bunton, 2013). Many other measures ofquality have been used over the past several decades, includ-ing long- and short-term perturbation, ratios of harmonicto noise components, and various measures of spectral slope

Table 1. Definition in human voice and birdsong.

Perceptualcomponentof voice Meaning Acoustic correlates in human and finch

Loudness The percept of the amount or intensity of the sound;volume. Influenced by intensity, fo, and spectral profile

Intensity level or sound pressure level; reported in dB SPL

Pitch The percept of how high or low a sound is. Related inpart to fo (of human vocal folds, finch syringeal labialfolds, or acoustic waveform) and also to intensity andspectral profile

fo, in Hertz

Quality Human voice: the interaction of the acoustic signal withthe listener. Perceived as an overall pattern, related inpart to harmonics and spectral slope, inharmonic energy,time-varying frequency and intensity, and characteristicsof the vocal tract

Birdsong: a similarity score that indicates goodness ofmatch of the song motif across multiple renditions inan adult bird or how similar the juvenile finch’s songis to his adult tutor’s song

Human voice: CPP, in dB, is the currently recommended measureof overall voice quality. CPP describes the prominence ofharmonic energy in the acoustic waveform.

Birdsong: composite acoustic measures of similarity and accuracy(scale of 0–100) that include WE, pitch, frequency modulation,and spectral continuity taken from 50- (similarity) or 7-ms(accuracy) sampling windows at either the motif or syllable level

Birdsong WE: a measure of the periodic versus aperiodic energyin a birdsong syllable. Measured on a logarithmic scale fromzero to minus infinity. White noise log (1) = 0 and completeorder log (0) = minus infinity

Note. Definitions are compiled from selected references (Baken & Orlikoff, 2000; Hillenbrand & Houde, 1996; Kreiman & Gerratt 2000; Kreiman,Gerratt, Garellek, Samlan, & Zhang, 2014; Patel et al., 2018; Tchernichovski et al., 2000). Human measurements in voice research are obtainedusing either sustained vowel phonation and/or connected speech during a standard reading passage. fo = fundamental frequency oscillation;CPP = cepstral peak prominence; WE = Wiener entropy.

DownloaTerms o

(Buder, 2000; Kreiman, Gerratt, & Berke, 1994). To sum-marize, frequency and amplitude measures are commonto analyses of both species, in spite of being used differently(i.e., as part of a composite score in birdsong and as individ-ual measures in humans). Based on these similarities, wedetermined that mean fo, the variability of fo within a sylla-ble, and intensity will serve as part of a core set of measuresfamiliar to researchers of both species. Measures of qualityare defined differently for birdsong and human voice, thougha measure of harmonic or noise structure is common to both(i.e., WE for birdsong and CPP for human voice). Thesetwo measures complete the core set used in Aim 2.

Aim 2: MethodsThere are two subaims: (a) determine the feasibility of

using a set of measures familiar to human voice researchersthat will characterize key features of birdsong and (b) identifythe extent to which WE relates to CPP.

SubjectsAll animal use was approved by the Institutional

Animal Care and Use Committee at the University ofArizona. For the song analyses, three adult male zebrafinches at the midpoint of their life span (~865–898 dayspost-hatch) were used with the expectation that measure-ments made in these three birds would be feasible in youngand elder adult finches. The finches were raised in differentnest boxes within an aviary in which male and femalefinches can select their mates; therefore, the adult tutor wasnot identified but is likely different for each bird and mayrepresent the influence of several tutors based on the distinctsongs sung (see Figure 1). Bird identification codes refer


to the leg band color (R = red and W = white and number:R1156, W35, R1157). Finches were moved to individualsound attenuation chambers (Eckel Noise Control Technol-ogies) and acclimated for 2 days under a 13:11 h of light:dark cycle before beginning recordings.

Song RecordingsMethods followed those of Miller et al. (2008). Songs

were recorded from males housed alone in sound attenua-tion chambers. Isolation of an individual male finch inthese sound chambers is a routine experimental paradigmemployed in songbird research. It enables recording andanalysis of a single song in response to an experimentalmanipulation of the neural circuitry without the need tofilter out competing noise from a female, groups of otherfinches, or human presence (selected references: Jarvis,Scharff, Grossman, Ramos, & Nottebohm, 1998; Milleret al., 2008). Next, we opted to collect songs from malessinging in a solo context, also known as undirected song(UD). UD song characteristics are advantageous for thecurrent study because they include a greater degree ofvariability in acoustic features such as fo compared withfemale-directed song performance (Kao et al., 2005). Thisis similar to human voice research, where it is common toevaluate the subject reading or imitating text and sustainedvowels (Patel et al., 2018). Conversational speech is lessfrequently used as stimuli for laboratory measurement.

Two hours of UD song were collected from lights onin the morning for all birds using Shure 93 lavalier condenseromnidirectional microphones connected to an audiobox(Audiobox: 44.1-kHz sampling rate/24-bit depth). Whensinging the UD song, male finches tend to stay stationary


DownloaTerms o

in their cage; our previous observations noted only a 1-dBchange in sound intensity detected by the microphone ifthe bird was at the far regions of the cage. Sounds wererecorded and digitized using preset parameters for captur-ing zebra finch song in Sound Analysis Pro (SAP: http://soundanalysispro.com/; Tchernichovski et al., 2000), a freelyavailable software platform. Birdsong researchers use SAPand custom written code in MATLAB or in R for acousticanalysis (Burkett, Day, Penagarikano, Geschwind, & White,2015; Miller et al., 2010).

Song AnalysesMotifs were identified as a sequence of repeated sylla-

bles separated by periods of silence. Acoustic features wereanalyzed for 25 renditions of each syllable within the bird’smotif immediately following lights on in the morning. Noappreciable increase in power has been previously observedin any statistical test when conducted on an n ≥ 25 syllablesin a given behavioral condition sung in the 2-hr recordingperiod based on power calculations (Miller et al., 2010).Introductory notes and unlearned calls were excluded fromthe acoustic analyses. Syllables were identified as soundenvelopes that could be separated from other syllables bylocal minima and repeated across the 25 motifs. The motifsand syllables were segmented in Praat and then analyzedin SAP (WE) and Praat (Boersma & Van Heuven, 2001).

Syllable Analysis in PraatAcoustic measurements of birdsong syllables that are

correlates of loudness, pitch, and quality were selected foranalysis (see Table 1). Spectrogram settings typically usedfor viewing speech and voice did not allow visualizationof enough birdsong harmonics to make decisions about seg-mentation and confirm syllable type within the motifs.Settings were manipulated through trial and error so thatthe harmonic structure was clearly revealed. The followingspectrogram settings were used: view range of 0 to 10000 Hz,window length of 0.025 s, and dynamic range of 70 dB.Mean intensity and smoothed CPP were computed for allanalyzed syllables. Mean intensity was computed using the“get intensity” command (Maryn, 2017). The CPP wascalculated using standard methodology as described in thePraat instruction manual included with the program. First,a power cepstrogram was generated using a pitch floorsetting of 300 Hz, a maximum frequency of 20,000, a timestep of 0.002, and preemphasis of 50. All settings are thePraat default settings except the pitch floor and maximumfrequency, which were modified in order to accommodatethe higher fo of the birds. The “get CPPS” command wasthen used with a peak search range of 300 to 1500 Hz (alsomodified to accommodate the bird fo), time window of0.0001, frequency averaging window of 0.00005, toleranceof 0.05, parabolic interpolation, tilt line frequency of 0.001to 0.0, exponential decay, and a robust fit method.

We followed the convention in birdsong analysiswhereby fo is only measured for harmonic syllables where



it is relatively stable, that is, syllable types composed ofonly one note in the form of flat harmonic stacks (Kaoet al., 2005). The fo was not calculated for noisy or mixedbirdsong syllables because, by definition, they lack a consis-tent fo. For frequency analysis, we specified a 75- to 1600-Hzrange for the setting “pitch range” and selected the cross-correlation analysis option. Mean fo, standard deviation,and minimum and maximum fo, were measured using the“voice report” function in Praat for each harmonic syllable(Praat manual). The fo range was computed in an Excelspreadsheet as the difference between maximum and mini-mum fo, in Hertz.

StatisticsTwo undergraduate research assistants made all mea-

surements and rated one bird in common to determineinterrater reliability. Descriptive statistics include meanand standard deviation for each acoustic measure.

Aim 2: ResultsThe Pearson correlation coefficient was used to calcu-

late interrater reliability. Correlations were .87 or higher foreach measure (r ≥ .87), confirming good reliability.

The mean and standard deviation of each acousticmeasure are reported for all syllables within each bird (seeTable 2). The full dataset of mean and standard deviationfor each of the 25 copies of every syllable can be found in theonline supplemental materials (Supplemental Material S1).

As presented in Figure 2, CPP and WE are shownfor three consecutive renditions of a harmonic syllable(top panel, Syllable E) versus a noisy Syllable C (bottompanel) from the bird’s motif in Figure 1A. In this example,the harmonic syllables have low WE (closer to negativeinfinity) compared with noisy syllables that approach zero,closer to white noise (see Figure 2). Both syllable typesare present in a song motif, but there are fewer harmonicsyllables produced making sample sizes small.

Figure 3 shows scatter plots of CPP versus WE foreach bird. In Figure 3a (R1156), the harmonic Syllable Eshows high CPP and low WE (e.g., more negative). Incontrast, noisy Syllable A has lower CPP and higher WE(e.g., less negative). Interestingly, CPP values for harmonicSyllable B are close to noisy Syllable A, but WE is similarto harmonic Syllable E. Syllable D (mixed), composedof two notes, a harmonic and a noisy, can be comparedwith the syllables with one harmonic or one noisy note.In comparison to harmonic Syllables B and E, Syllable Dhas a lower CPP score and higher mean WE (see Table 2).CPP and WE for Syllable D are both lower than for thetwo noisy syllables, A and C.

In Figure 3b (W35), harmonic Syllable B has highCPP and low WE, whereas noisy Syllables A and E havelower CPP and higher WE (see also mean scores, Table 2).Syllables C and D are composed of mixed notes so theirCPP scores are lower than harmonic Syllable B but greaterthan pure noisy Syllables A and E. Mixed Syllable C consists

http://soundanalysispro.com/

http://soundanalysispro.com/

Table 2. Mean and standard deviation scores for acoustic measurements of birdsong.

Bird andsyllable ID

Intensity (dB SPL) CPP (dB) WE fo (Hz) fo range (Hz)

M SD M SD M SD M SD M SD

W35 Syll A 61.82 4.89 9.68 0.21 –2.45 0.18 — — — —W35 Syll B 56.39 1.70 19.70 0.60 –3.08 0.14 644.02 5.30 118.43 15.56W35 Syll C 65.25 1.81 15.20 0.37 –2.88 0.22 — — — —W35 Syll D 57.43 1.12 16.20 0.88 –1.55 0.15 — — — —W35 Syll E 51.26 0.84 10.10 0.41 –2.75 0.20 — — — —R1156 Syll A 62.89 1.19 16.66 0.92 –1.17 0.08 — — — —R1156 Syll B 56.57 1.37 16.40 1.35 –2.97 0.18 615.54 7.81 55.10 13.68R1156 Syll C 60.22 0.83 15.32 0.68 –1.91 0.12 — — — —R1156 Syll D 61.26 1.46 12.00 0.80 –2.73 0.13 — — — —R1156 Syll E 54.38 0.95 20.77 0.90 –2.87 0.20 612.50 6.78 53.23 12.99R1157 Syll A 57.30 3.71 11.42 1.25 –2.86 0.19 — — — —R1157 Syll B 64.80 2.26 10.72 0.45 –2.06 0.10 — — — —R1157 Syll C 65.86 0.78 9.52 0.37 –1.92 0.32 — — — —R1157 Syll D 66.04 0.88 9.30 0.40 –3.34 0.26 — — — —R1157 Syll E 66.22 1.28 9.67 0.52 –2.47 0.18 — — — —R1157 Syll F 69.51 0.79 11.90 0.31 –2.14 0.12 — — — —R1157 Syll G 64.41 0.68 10.43 0.47 –2.31 0.13 — — — —R1157 Syll H 66.42 1.15 19.30 0.99 –2.19 0.14 — — — —R1157 Syll I 71.26 1.63 16.98 0.65 –2.88 0.15 — — — —

Note. Mean and standard deviation are reported for each syllable across three birds. The fo measurements were only made for syllables witha clear one-note harmonic structure. The fo range represents the difference between the lowest and highest fo during a sustained phonation.Units are reported for the mean scores except for WE, which is a pure number and unitless (Tchernichovski et al., 2000). Em dashes indicatedata not applicable. ID = identification; CPP = cepstral peak prominence; WE = Wiener entropy; fo = fundamental frequency oscillation.

DownloaTerms o

of mostly harmonic notes, and therefore, its WE score islower, similar to purely harmonic Syllable B. By contrast,mixed Syllable D has a noisy note with harmonic note,which may explain its high WE score in Figure 3b.

In Figure 3c (R1157), Syllable H has the highest CPPvalue compared with mixed syllables, B, D, E, F, and I andnoisy Syllables A, C, and G. Syllable H has a harmonicappearance, but its fo varies over the length of the syllable,and therefore, its fo was not analyzed. The WE scores arenot consistent among the syllables. Mixed Syllable D hasthe lowest WE, whereas noisy Syllable C has the highestscores.

DiscussionThe current study establishes a translational dictionary

of birdsong and human voice terminology/measures that canbe used as a tool to interpret how manipulations of neuralcircuitry in birdsong can be applied to a better understand-ing of voice disorders. Furthermore, we identified a set ofacoustic measures common to evaluation of human voiceand birdsong. In Praat, measurements of birdsong werefeasible following some adjustments to the settings typicallyused for human voice analysis. Praat presents a platform thatis more familiar to human voice researchers and may facili-tate collaborations between voice and songbird researchers.

We hypothesized that the human voice measureCPP and the birdsong measure WE would provide the sameinformation about the harmonic and noise componentsof the birdsong syllable. Contrary to our hypothesis, therelationship between CPP and WE is complex. For syllables


that were assigned to clear categories of harmonic or noisy,their CPP and WE scores showed a fairly consistent inverserelationship. Harmonic syllables have high CPP values andlow WE scores compared with noisy syllables, suggestingthat both metrics provide similar information. However, inthe case of mixed bird syllables consisting of harmonic andnoisy elements, the relationship was not preserved becauseWE scores were too variable. For mixed syllables, CPPscores appear to be a more reliable measure. The CPP wasgenerally midrange when the mixed syllable had a longharmonic component (i.e., Syllables C and D in Figure 3Band I in Figure 3C) and low when the harmonic compo-nent was short (i.e., Syllable D in Figure 3A and Syllables B,D, E, and F in Figure 3C). Because calculation of CPP doesnot rely on previous determination of fo (Hillenbrand et al.,1994), it is one of the measures commonly used in humanvoice analysis that is reliable when the acoustic signal isnoisy (Awan, Roy, & Dromey, 2009). This is important tothe assessment of disordered voice and also when the assess-ment stimuli (e.g., sustained sounds, words, phrases) containfricatives or affricates, an acoustic situation that mirrorsthe noisy and mixed syllables of birdsong. To our knowl-edge, limitations of WE for noisy and mixed syllables arenot reported in the birdsong literature.

The CPP was not always the more reliable determi-nant of noisy versus harmonic energy, however. There werealso some instances where CPP was similar for noisy andharmonic syllables (e.g., Figure 3A: Syllables A vs. B), andthe WE score differentiated the syllables. In this particularcase, the noisy syllables had some low frequency harmonicsthat appear reflected in the CPP value. Thus, WE could


Figure 3. Scatter plots of cepstral peak prominence (CPP) versus Wiener entropy (WE) scores for all syllables. CPP (x-axis) versus WE (y-axis)scores are plotted for 25 renditions (copies) of each syllable within a bird. Filled circles indicate harmonic syllables, whereas open circlesrepresent noisy syllables. Mixed syllables that consist of two or more notes are depicted as plus signs. Colors are used to represent eachsyllable. R1156 (A) and W35 (B) have five syllables each in their motifs, whereas R1157 (C) has nine syllables. The “?” signifies that the syllabletype is not clear but contains harmonic elements.

Figure 2. Harmonic versus noisy syllable exemplars, cepstral peak prominence (CPP), and Wiener entropy (WE). Three consecutiverenditions of R1156 harmonic Syllable “E” (top) versus Syllable “C” (bottom) are shown. CPP and WE scores are reported below each syllable.CPP is higher for harmonic Syllable E compared with noisy Syllable C. WE entropy scores are lower (more negative) for Syllable E than forSyllable C.


Downloaded From: https://jslhr.pubs.asha.org/ on 12/12/2018Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx

DownloaTerms o

also occasionally distinguish fine detail between two sylla-bles better than CPP. Our results suggest that the measuresprovide complementary information, and utilization ofboth CPP and WE scores is warranted to reliably character-ize fine differences in acoustic properties between syllables.

Given the ability to make successful comparisonsbetween birdsong and human voice measures, our analysescan be extended in future work to include more harmonicsyllables in birdsong and to make comparisons at the motiflevel with human speech measurements including speechand articulation rate. Here, we only studied males singinga UD (e.g., vocally practicing alone), as a model for humanvoice research, but it would be informative for future studiesto analyze female-directed song as a model for conversa-tional speech.

One limitation of using zebra finches as a model forhuman voice is that the males sing and the females do not.Therefore, future studies will need to incorporate othersongbird species (e.g., Northern cardinals, house wrens) inwhich both the females and males sing (Odom & Benedict,2018) in order to make comparisons of pitch, loudness,and vocal quality with human male and female subjects.For example, adult human males have a lower fo comparedwith females (Baken & Orlikoff, 2000); whether that holdstrue for songbird species in which both the females andmales sing requires determination. In addition to differencesbased on gender, changes in human voice are impactedby age—children have a higher fo than adults (Baken &Orlikoff, 2000), men’s fo might increase as they age, andwomen’s fo might decrease (Dehqan, Scherer, Dashti, Ansari-Moghaddam, & Fanaie, 2012; Goy, Fernandes, Pichora-Fuller, & van Lieshout, 2013; Hollien & Shipp, 1972).Whether young male zebra finches learning their song havea higher fo compared with adult finches is not known. In asmall study conducted in a related species, Bengalese finch,reductions in syllable pitch and intensity were detected inadulthood as the birds aged (Cooper et al., 2012).

Because male zebra finches are known as close-endedlearners, meaning they retain a similar motif structure fromdevelopment into adulthood, it would be useful to conductvocal analyses on canaries (open-ended learners) that changetheir songs each breeding season (Nottebohm, Nottebohm,& Crane, 1986). By applying our human voice analyses tocanary songs, we can obtain additional insight into theneural plasticity mechanisms that drive vocal quality basedon environmental needs.

The further development of comparative analyses be-tween voice/speech measures in normal human populationswith birdsong will facilitate future comparisons examiningage and neurodegenerative disease–related changes on vocaloutput.

AcknowledgmentsData collection was supported by University of Arizona

startup funds to Julie E. Miller and financial support to AreenBadwal from the University of Arizona Undergraduate Biologi-cal Research Program. We would like to thank Stephanie Munger


(University of Arizona) for the birdsong collection, Nancy Day (Uni-versity of California, Los Angeles) for the MATLAB code, andUniversity of Arizona Animal Care.

ReferencesAwan, S. N., Roy, N., & Cohen, S. M. (2014). Exploring the

relationship between spectral and cepstral measures ofvoice and the Voice Handicap Index (VHI). Journal ofVoice, 28(4), 430–439. https://doi.org/10.1016/j.jvoice.2013.12.008

Awan, S. N., Roy, N., & Dromey, C. (2009). Estimating dyspho-nia severity in continuous speech: Application of a multi-parameter spectral/cepstral model. Clinical Linguistics &Phonetics, 23(11), 825–841.

Baken, R. J., & Orlikoff, R. F. (2000). Clinical measurement ofspeech and voice. Independence, KY: Cengage Learning.

Bartholomew, W. T. (1934). A physical definition of “good voice‐quality” in the male voice. The Journal of the Acoustical Societyof America, 6(1), 25–33.

Berwick, R. C., & Chomsky, N. (2013). Birdsong, speech, and lan-guage: Exploring the evolution of mind and brain. Cambridge,MA: MIT Press.

Boersma, P., & Van Heuven, V. (2001). Speak and unspeak withPraat. Glot International, 5(9/10), 341–347.

Brainard, M. S., & Doupe, A. J. (2013). Translating birdsong:Songbirds as a model for basic and applied medical research.Annual Review of Neuroscience, 36, 489–517. https://doi.org/10.1146/annurev-neuro-060909-152826

Brumm, H., & Slater, P. J. (2006). Animals can vary signal ampli-tude with receiver distance: Evidence from zebra finch song.Animal Behaviour, 72(3), 699–705.

Buder, E. H. (2000). Acoustic analysis of voice quality: A tabula-tion of algorithms 1902–1990. In R. D. Kent & M. J. Ball(Eds.), Voice quality measurement (pp. 119–244). San Diego,CA: Singular.

Burkett, Z. D., Day, N. F., Penagarikano, O., Geschwind, D. H.,& White, S. A. (2015). VoICE: A semi-automated pipeline forstandardizing vocal analysis across models. Scientific Reports,5, 10237. https://doi.org/10.1038/srep10237

Cooper, B. G., Méndez, J. M., Saar, S., Whetstone, A. G., Meyers, R.,& Goller, F. (2012). Age-related changes in the Bengalesefinch song motor program. Neurobiology of Aging, 33(3),564–568.

Dehqan, A., Scherer, R. C., Dashti, G., Ansari-Moghaddam, A., &Fanaie, S. (2012). The effects of aging on acoustic parametersof voice. Folia Phoniatrica et Logopaedica, 64(6), 265–270.

Dichter, B. K., Breshears, J. D., Leonard, M. K., & Chang, E. F.(2018). The control of vocal pitch in human laryngeal motorcortex. Cell, 174, 21–31.

Fraile, R., & Godino-Llorente, J. I. (2014). Cepstral peak promi-nence: A comprehensive analysis. Biomedical Signal Processingand Control, 14, 42–54.

Goy, H., Fernandes, D. N., Pichora-Fuller, M. K., & van Lieshout,P. (2013). Normative voice data for younger and older adults.Journal of Voice, 27(5), 545–555.

Haesler, S., Rochefort, C., Georgi, B., Licznerski, P., Osten, P., &Scharff, C. (2007). Incomplete and inaccurate vocal imitationafter knockdown of FoxP2 in songbird basal ganglia nucleusArea X. PLoS Biology, 5(12), e321.

Hahnloser, R. H. R., Kozhevnikov, A. A., & Fee, M. S. (2002). Anultra-sparse code underlies the generation of neural sequencesin a songbird. Nature, 419(6902), 65–70.


https://doi.org/10.1016/j.jvoice.2013.12.008

https://doi.org/10.1016/j.jvoice.2013.12.008

https://doi.org/10.1146/annurev-neuro-060909-152826

https://doi.org/10.1146/annurev-neuro-060909-152826

https://doi.org/10.1038/srep10237

DownloaTerms o

Heston, J. B., & White, S. A. (2015). Behavior-linked FoxP2 regula-tion enables zebra finch vocal learning. Journal of Neuroscience,35(7), 2885–2894. https://doi.org/10.1523/JNEUROSCI.3715-14.2015

Hillenbrand, J., Cleveland, R. A., & Erickson, R. L. (1994). Acous-tic correlates of breathy vocal quality. Journal of Speech andHearing Research, 37(4), 769–778.

Hillenbrand, J., & Houde, R. A. (1996). Acoustic correlates ofbreathy vocal quality: Dysphonic voices and continuous speech.Journal of Speech and Hearing Research, 39(2), 311–321.

Hoffmann, L. A., Saravanan, V., Wood, A. N., He, L., & Sober,S. J. (2016). Dopaminergic contributions to vocal learning.Journal of Neuroscience, 36(7), 2176–2189. https://doi.org/10.1523/JNEUROSCI.3883-15.2016

Hollien, H., & Shipp, T. (1972). Speaking fundamental frequencyand chronologic age in males. Journal of Speech and HearingDisorders, 15(1), 155–159.

Jarvis, E. D., Scharff, C., Grossman, M. R., Ramos, J. A., &Nottebohm, F. (1998). For whom the bird sings: Context-dependent gene expression. Neuron, 21(4), 775–788.

Kao, M. H., & Brainard, M. S. (2006). Lesions of an avian basalganglia circuit prevent context-dependent changes to song vari-ability. Journal of Neurophysiology, 96(3), 1441–1455.

Kao, M. H., Doupe, A. J., & Brainard, M. S. (2005). Contributionsof an avian basal ganglia-forebrain circuit to real-time modu-lation of song. Nature, 433(7026), 638–643.

Kempster, G. B., Gerratt, B. R., Abbott, K. V., Barkmeier-Kraemer,J., & Hillman, R. E. (2009). Consensus auditory-perceptual eval-uation of voice: Development of a standardized clinical protocol.American Journal of Speech-Language Pathology, 18(2), 124–132.

Kreiman, J., & Gerratt, B. R. (2000). Measuring voice quality. InR. D. Kent & M. J. Ball (Eds.), Voice quality measurement(pp. 73–101). San Diego, CA: Singular.

Kreiman, J., Gerratt, B. R., & Berke, G. S. (1994). The multi-dimensional nature of pathologic vocal quality. The Journalof the Acoustical Society of America, 96(3), 1291–1302.

Kreiman, J., Gerratt, B. R., Garellek, M., Samlan, R., & Zhang, Z.(2014). Toward a unified theory of voice production and percep-tion. Loquens, 1(1), e009.

Lai, C. S. L., Fisher, S. E., Hurst, J. A., Vargha-Khadem, F., &Monaco, A. P. (2001). A forkhead-domain gene is mutated in asevere speech and language disorder. Nature, 413(6855), 519–523.

Leblois, A., & Perkel, D. J. (2012). Striatal dopamine modulatessong spectral but not temporal features through D1 receptors.European Journal of Neuroscience, 35, 1771–1781.

Leblois, A., Wendel, B. J., & Perkel, D. J. (2010). Striatal dopaminemodulates basal ganglia output and regulates social context-dependent behavioral variability through D1 receptors. Journalof Neuroscience, 30(16), 5730–5743. https://doi.org/10.1523/JNEUROSCI.5974-09.2010

Lipkind, D., Marcus, G. F., Bemis, D. K., Sasahara, K., Jacoby, N.,Takahasi, M., . . . Okanoya, K. (2013). Stepwise acquisition ofvocal combinatorial capacity in songbirds and human infants.Nature, 498(7452), 104–108.

Maryn, Y. (2017). Practical acoustics in clinical voice assessment:A Praat primer. Perspectives of the ASHA Special InterestGroups, 2(3), 14–32.

Miller, J. E., Hilliard, A. T., & White, S. A. (2010). Song practicepromotes acute vocal variability at a key stage of sensorimotorlearning. PLoS One, 5(1), e8592.

Miller, J. E., Spiteri, E., Condro, M. C., Dosumu-Johnson, R. T.,Geschwind, D. H., & White, S. A. (2008). Birdsong decreases



protein levels of FoxP2, a molecule required for human speech.Journal of Neurophysiology, 100(4), 2015–2025.

Murugan, M., Harward, S., Scharff, C., & Mooney, R. (2013).Diminished FoxP2 levels affect dopaminergic modulation ofcorticostriatal signaling important to song variability. Neuron,80(6), 1464–1476. https://doi.org/10.1016/j.neuron.2013.09.021

Nottebohm, F., Nottebohm, M. E., & Crane, L. (1986). Develop-mental and seasonal changes in canary song and their relationto changes in the anatomy of song-control nuclei. Behavioraland Neural Biology, 46(3), 445–471.

Odom, K. J., & Benedict, L. (2018). A call to document femalebird songs: Applications for diverse fields. The Auk, 135(2),314–325.

Patel, R. R., Awan, S. N., Barkmeier-Kraemer, J., Courey, M.,Deliyski, D., Eadie, T., . . . Hillman, R. (2018). Recommendedprotocols for instrumental assessment of voice: American Speech-Language-Hearing Association expert panel to develop a pro-tocol for instrumental assessment of vocal function. AmericanJournal of Speech-Language Pathology, 27, 887–905.

Pfenning, A. R., Hara, E., Whitney, O., Rivas, M. V., Wang, R.,Roulhac, P. L., . . . Jarvis, E. D. (2014). Convergent transcrip-tional specializations in the brains of humans and song-learningbirds. Science, 346(6215), 1256846. https://doi.org/10.1126/science.1256846

Riede, T., & Goller, F. (2010). Peripheral mechanisms for vocalproduction in birds-differences and similarities to human speechand singing. Brain and Language, 115(1), 69–80.

Samlan, R. A., Story, B. H., & Bunton, K. (2013). Relation of per-ceived breathiness to laryngeal kinematics and acoustic mea-sures based on computational modeling. Journal of Speech,Language, and Hearing Research, 56(4), 1209–1223. https://doi.org/10.1044/1092-4388(2012/12-0194)

Simonyan, K., Horwitz, B., & Jarvis, E. D. (2012). Dopamineregulation of human speech and bird song: A critical review.Brain and Language, 122(3), 142–150.

Sober, S. J., Wohlgemuth, M. J., & Brainard, M. S. (2008). Centralcontributions to acoustic variation in birdsong. Journal ofNeuroscience, 28(41), 10370–10379.

Stemple, J. C., Roy, N., & Klaben, B. K. (2014). Clinical voicepathology: Theory and management. San Diego, CA: Plural.

Tchernichovski, O., Mitra, P. P., Lints, T., & Nottebohm, F. (2001).Dynamics of the vocal imitation process: How a zebra finchlearns its song. Science, 291(5513), 2564–2569.

Tchernichovski, O., & Nottebohm, F. (1998). Social inhibition ofsong imitation among sibling male zebra finches. Proceedingsof the National Academy of Sciences of the United States, 95,8951–8956.

Tchernichovski, O., Nottebohm, F., Ho, C. E., Pesaran, B., &Mitra, P. P. (2000). A procedure for an automated measure-ment of song similarity. Animal Behaviour, 59, 1167–1176.

Vargha-Khadem, F., Watkins, K. E., Price, C. J., Ashburner, J.,Alcock, K. J., Connelly, A., . . . Passingham, R. E. (1998).Neural basis of an inherited speech and language disorder.Proceedings of the National Academy of Sciences of the UnitedStates, 95(21), 12695–12700.

Watkins, K. E., Gadian, D. G., & Vargha-Khadem, F. (1999).Functional and structural brain abnormalities associated witha genetic disorder of speech and language. American Journalof Human Genetics, 65(5), 1215–1221.

Woolley, S. C., & Doupe, A. J. (2008). Social context-inducedsong variation affects female behavior and gene expression.PLoS Biology, 6(3), e62.

https://doi.org/10.1523/JNEUROSCI.3715-14.2015






https://doi.org/10.1016/j.neuron.2013.09.021

https://doi.org/10.1126/science.1256846

https://doi.org/10.1126/science.1256846

https://doi.org/10.1044/1092-4388(2012/12-0194)

https://doi.org/10.1044/1092-4388(2012/12-0194)

Online supplemental material, Badwal et al., “Common Terminology and Acoustic Measures for Human Voice and Birdsong,” JSLHR, https://doi.org/10.1044/2018_JSLHR-S-18-0218

Supplemental Material S1. The full dataset of mean and standard deviation for each of the 25

copies of every syllable.

Bird ID R1156: Syllable A.

Syllable Copy# Intensity CPP WE

A 1 59.6 14.8 –1.25

A 3 63.5 18 –1.09

A 5 62 15.8 –1.17

A 7 61.8 17 –1.23

A 9 62.6 15.8 –1.24

A 11 61.9 15.6 –1.14

A 13 62.3 16.7 –1.10

A 15 63.8 16.8 –1.03

A 17 61.9 17.2 –1.10

A 19 63.4 17 –1.28

A 21 60.2 16.4 –1.16

A 23 64.3 16.6 –1.15

A 25 64.2 16.5 –1.17

A 27 64.3 17.2 –1.12

A 29 62.2 16.5 –1.25

A 31 63.2 16 –1.20

A 33 63.8 17.5 –1.07

A 35 63.6 16.5 –1.10

A 37 63.5 17.5 –1.10

A 39 63.6 19.1 –1.33

A 41 63.4 15.7 –1.20

A 43 63.8 17.2 –1.22

A 45 62.9 17.4 –1.16

A 47 63.4 16.3 –1.23

A 49 63.1 15.3 –1.08

M 62.89 16.66 –1.17

SD 1.19 0.92 0.08

Bird ID R1156: Syllable B.

Syllable Copy# Intensity CPP WE F0.mean F0.range

B 1 52.8 14.2 –3.03 613.6 67.2

B 3 54.7 14.7 –3.03 610 51.3

B 5 55.8 15.6 –2.79 612 65.5

B 7 55.4 16 –3.02 609.1 68.4

B 9 54.8 14.8 –2.89 600.5 60.2

B 11 58.3 18.3 –2.87 626.4 35.5

B 13 57.5 18.8 –3.05 615 47.3



B 15 57.5 16.5 –2.97 614.2 64.2

B 17 55.6 15 –3.22 619 52.6

B 19 56.9 15.8 –3.10 608.8 47.6

B 21 54.8 18 –3.21 626 18.2

B 23 57.4 17.1 –2.61 614.7 70.2

B 25 56.5 14.8 –3.05 612.8 45.3

B 27 56.1 16.6 –2.86 606.3 71.4

B 29 55.6 15.6 –2.88 614.9 44.7

B 31 56.2 15.8 –3.02 610.1 63.5

B 33 57.6 17.9 –2.79 611 53.4

B 35 57.6 16.2 –2.92 616.5 71.7

B 37 56.9 16.4 –2.74 610.1 68.2

B 39 56.7 19.6 –2.88 611.1 51.6

B 41 57.9 15.4 –3.29 629.7 66.8

B 43 57.8 16.8 –2.85 620.6 44.7

B 45 57.7 17 –2.82 620 43.9

B 47 58 16.3 –3.35 635.5 37.1

B 49 58.1 16.7 –2.89 620.5 66.9

M 56.57 16.40 –2.97 615.54 55.10

SD 1.37 1.35 0.18 7.81 13.68

Bird ID R1156: Syllable C.


C 2 57.8 15.7 –1.65

C 4 60.2 15.7 –1.86

C 6 60 15.5 –1.91

C 8 59.9 15.9 –2.04

C 10 59.5 15.3 –1.76

C 12 60.7 14.8 –2.02

C 14 59.9 16.1 –1.90

C 16 59.6 15.5 –1.71

C 18 60.4 15.1 –2.17

C 20 60.3 14.2 –2.08

C 22 59 16.2 –1.73

C 24 60.7 14.5 –1.90

C 26 61.5 15.9 –2.04

C 28 61.9 15.3 –2.00

C 30 60 16.1 –1.95

C 32 59.9 15.6 –1.91

C 34 60.5 14.9 –1.96

C 36 60.6 15.9 –1.93

C 38 59.8 15.6 –1.89

C 40 60.3 13.7 –1.94



C 42 61.1 14.8 –1.97

C 44 60.8 15.1 –1.85

C 46 60.6 14.2 –1.82

C 48 61 16.3 –2.03

C 50 59.5 15 –1.79

M 60.22 15.32 –1.91

SD 0.83 0.68 0.12

Bird ID R1156: Syllable D.


D 1 57.3 12.7 –2.32

D 2 59.3 11.6 –2.50

D 3 59.1 12.2 –2.60

D 4 60.1 11.3 –2.88

D 5 59.8 11.6 –2.70

D 6 61.2 12.6 –2.63

D 7 61.5 12.1 –2.82

D 8 61.8 11.8 –2.66

D 9 61.3 11.5 –2.80

D 10 61.9 12.2 –2.82

D 11 59.9 12.5 –2.64

D 12 61.9 12.1 –2.83

D 13 61.6 12 –2.69

D 14 61.1 11.4 –2.83

D 15 61.9 10.7 –2.87

D 16 62.9 12.1 –2.72

D 17 62.9 11.2 –2.72

D 18 62.8 11.4 –2.77

D 19 62.5 12 –2.79

D 20 62.6 11.6 –2.72

D 21 63.6 11.9 –2.94

D 22 61.5 12.6 –2.74

D 23 61.5 12.5 –2.75

D 24 59.5 15 –2.77

D 25 61.9 11.4 –2.64

M 61.26 12.0 –2.73

SD 1.46 0.80 0.13



Bird ID R1156: Syllable E.


E 2 54.4 21.5 –2.89 616.5 34.8

E 4 53.6 20.5 –2.95 608 59.9

E 6 53.6 20 –2.63 602.9 57

E 8 53.3 19.8 –3.00 607.2 63.1

E 10 54.5 20.3 –3.07 612.7 53.3

E 12 54.8 19.9 –2.82 615.9 65

E 14 55.1 20.7 –2.83 611.5 87.1

E 16 53.6 19.6 –2.58 603.3 78.6

E 18 54.8 20.6 –3.17 614.6 49.3

E 20 54.1 19 –3.20 606.9 33.6

E 22 56.1 22.5 –3.15 622.6 34.6

E 24 55.2 20.7 –2.83 611.9 63.4

E 26 54.7 21.6 –2.80 608.5 43.6

E 28 53.9 20.9 –2.88 610.2 54.1

E 30 52.3 19.2 –2.77 613.2 53.4

E 32 54 22.3 –2.63 601.7 45.5

E 34 53.7 20.5 –2.48 613.7 60.3

E 36 55.5 20.7 –3.09 619.5 44.9

E 38 54.9 21.6 –2.96 612.8 40.7

E 40 53 20.4 –2.71 608.9 40.7

E 42 53.2 21 –2.56 601.3 64.7

E 44 55.6 21.7 –3.04 621.3 52.7

E 46 55.6 21.9 –3.04 620.8 52.1

E 48 55.2 21.3 –2.90 625.1 46.2

E 50 54.8 21 –2.76 621.4 52.1

M 54.38 20.77 –2.87 612.50 53.23

SD 0.95 0.90 0.20 6.78 12.99



Bird ID W35: Syllable A.


A 1 49.1 9.8 –2.14

A 2 58 9.7 –2.42

A 3 54.2 9.8 –2.10

A 4 62.9 10.2 –2.63

A 5 55.1 9.2 –2.42

A 6 63.2 9.4 –2.67

A 7 57.9 9.7 –2.19

A 8 58.6 9.6 –2.28

A 9 61.8 9.8 –2.54

A 10 68.1 9.5 –2.64

A 11 65 9.8 –2.27

A 12 65.2 10.1 –2.16

A 13 67.2 9.8 –2.39

A 14 68.3 9.8 –2.40

A 15 63.7 9.7 –2.45

A 16 65.2 9.6 –2.65

A 17 60 9.6 –2.51

A 18 65.2 9.7 –2.60

A 19 66.4 9.6 –2.48

A 20 65.4 9.4 –2.76

A 21 65.4 9.6 –2.59

A 22 54.8 9.7 –2.35

A 23 63.2 9.6 –2.53

A 24 62.4 9.6 –2.56

A 25 59.2 9.8 –2.48

M 61.82 9.68 –2.45

SD 4.89 0.21 0.18

Bird ID W35: Syllable B.


B 1 54.2 19.4 –3.17 653.8 109.6

B 2 54.9 18.9 –2.93 641.3 113.8

B 3 53.4 18.7 –2.70 630.8 144.3

B 4 55.2 19.1 –3.09 640.1 127.8

B 5 54.7 20 –2.91 646.5 104.3

B 6 51.4 18.8 –2.95 642.1 110.4

B 7 55.8 19.2 –3.06 639.4 118.6

B 8 55.9 18.9 –3.20 644.6 129.3

B 9 56.3 19.5 –3.24 640.3 89.1



B 10 57.1 20.8 –3.11 649.8 113.3

B 11 57.3 19.9 –2.93 644.7 118.5

B 12 58.2 20.5 –2.97 647.8 124.4

B 13 57.7 20 –3.14 649 117.7

B 14 56.9 19.8 –2.98 641.8 108.4

B 15 58.5 20.1 –2.95 643.7 122

B 16 58.3 20.7 –3.23 653.6 166.3

B 17 58.4 19.6 –3.18 644.8 136.2

B 18 57.2 20.1 –3.02 640.6 106.4

B 19 58 20.4 –3.07 643.5 100.9

B 20 56.9 19 –3.21 638.2 129.5

B 21 56.6 19.5 –3.16 640.1 126.7

B 22 56.3 19.9 –3.14 650.1 107

B 23 57.4 20 –3.22 649.7 107.9

B 24 56.4 19.4 –3.22 638.5 115.9

B 25 56.7 20.2 –3.20 645.8 112.5

M 56.39 19.70 –3.08 644.02 118.43

SD 1.70 0.60 0.14 5.30 15.56

Bird ID W35: Syllable C.


C 1 62.3 14.8 –2.76

C 2 63.1 15 –2.77

C 3 62.4 14.6 –2.49

C 4 64.5 14.8 –2.83

C 5 62.5 15.1 –2.32

C 6 63.3 15.1 –2.57

C 7 63.7 14.4 –2.54

C 8 64.2 14.9 –2.90

C 9 64.3 14.9 –2.97

C 10 66.2 15.4 –3.06

C 11 66.5 15.3 –2.81

C 12 67.1 15.6 –2.75

C 13 67.8 15.4 –3.04

C 14 66.2 15.4 –2.98

C 15 67 15.1 –3.05

C 16 67.5 16 –3.20

C 17 68 15.1 –3.12

C 18 66.6 15.9 –2.99

C 19 67.1 15.6 –2.91

C 20 66.2 15.4 –3.06

C 21 66.6 15.3 –3.10

C 22 63.4 15 –2.72



C 23 65 15.2 –3.00

C 24 64.7 15.3 –2.99

C 25 65.1 15.4 –2.99

M 65.25 15.20 –2.88

SD 1.81 0.37 0.22

Bird ID W35: Syllable D.


D 1 56.5 15.2 –1.58

D 2 56.7 15.2 –1.55

D 3 58.1 15.9 –1.72

D 4 57 15 –1.77

D 5 56 17.1 –1.56

D 6 56 15.2 –1.52

D 7 57.6 14.8 –1.75

D 8 55.4 15.5 –1.33

D 9 57.8 17.6 –1.49

D 10 58.4 17.9 –1.68

D 11 59.8 17 –1.59

D 12 58.4 16.3 –1.27

D 13 58.5 15.5 –1.42

D 14 59.5 16.7 –1.53

D 15 58.3 15.9 –1.80

D 16 58.1 17.2 –1.73

D 17 57.8 16.6 –1.54

D 18 57.8 16.7 –1.47

D 19 56.8 16.5 –1.66

D 20 57 17 –1.51

D 21 57.2 15.6 –1.66

D 22 56.1 15 –1.32

D 23 55.9 16.5 –1.45

D 24 57.1 16.9 –1.39

D 25 58 16.1 –1.38

M 57.43 16.20 –1.55

SD 1.12 0.88 0.15

Bird ID W35: Syllable E.


E 1 50.4 9.9 –2.94

E 2 50.5 10.3 –2.93



E 3 51.4 9.5 –3.08

E 4 51 9.9 –3.15

E 5 51.2 10.2 –2.84

E 6 50.7 10.9 –3.02

E 7 50.8 9.9 –2.80

E 8 50.2 9.9 –2.55

E 9 48.7 9 –2.83

E 10 51.4 10.4 –2.53

E 11 52.1 10.1 –2.71

E 12 51.9 10 –2.55

E 13 52.5 10.2 –2.61

E 14 52.3 10.5 –2.58

E 15 52.3 10.4 –2.61

E 16 52 9.8 –2.85

E 17 51.6 10.2 –2.41

E 18 51.3 10.2 –2.61

E 19 50.9 10.8 –2.75

E 20 51.5 9.4 –2.69

E 21 50.9 10.1 –2.85

E 22 51.6 10 –2.67

E 23 51 10.2 –2.65

E 24 50.8 10.1 –3.08

E 25 52.4 10.5 –2.46

M 51.26 10.10 –2.75

SD 0.84 0.41 0.20



Bird ID R1157: Syllable A.


A 1 53.1 10.3 –2.86

A 2 58.2 11.5 –2.56

A 3 57.9 11.5 –2.65

A 4 55.3 11.1 –2.82

A 5 57.9 11.8 –2.68

A 6 53.4 11 –2.87

A 7 59.3 11.5 –3.11

A 8 58.8 11.8 –3.16

A 9 54.7 11.1 –2.91

A 10 58.9 11.8 –2.96

A 11 57.4 11.4 –2.77

A 12 53.4 10.9 –2.81

A 13 57.2 11.1 –2.65

A 14 57.6 10.8 –2.87

A 15 56.8 11 –2.74

A 16 72.4 17.1 –2.93

A 17 53.4 10.8 –2.57

A 18 56.8 11.4 –2.77

A 19 55 10.6 –3.32

A 20 58.3 11 –2.93

A 21 57.3 10.9 –2.85

A 22 54.9 12 –2.91

A 23 59.6 10.8 –3.03

A 24 56.9 11.1 –2.64

A 25 58 11.3 –3.05

M 57.30 11.42 –2.86

SD 3.71 1.25 0.19

Bird ID R1157: Syllable B.


B 1 64.7 10.7 –2.08

B 3 65.8 10.4 –1.94

B 5 65.7 10.6 –1.92

B 7 64.6 10.4 –2.04

B 9 65.4 10.5 –1.97

B 11 64.7 10.4 –2.22

B 13 65.7 11.9 –2.19

B 15 66.2 10.9 –2.26

B 17 65 10.9 –1.89

B 19 65.7 10.4 –2.17



B 21 66.2 10.3 –2.02

B 23 62.7 10.3 –2.11

B 25 64.7 10.2 –1.95

B 27 64.8 10.3 –2.10

B 29 64.4 10.9 –2.06

B 31 54.7 11.1 –2.05

B 33 64.1 11.1 –2.00

B 35 66 10.6 –2.00

B 37 65.6 10.3 –2.09

B 39 66.2 10.4 –2.06

B 41 64.3 10.9 –1.99

B 43 65.6 11.5 –2.15

B 45 66.2 11.3 –1.98

B 47 65.7 10.3 –2.13

B 49 65.4 11.4 –2.10

M 64.80 10.72 –2.06

SD 2.26 0.45 0.10

Bird ID R1157: Syllable C.


C 2 64.6 9.6 –1.48

C 4 66.6 8.8 –1.73

C 6 66.6 9.1 –2.20

C 8 65.3 9.4 –1.84

C 10 66.5 9.4 –1.89

C 12 66.2 9.3 –1.82

C 14 66.2 9.7 –1.95

C 16 66 9.9 –1.97

C 18 65.7 9.6 –1.92

C 20 65.4 9.6 –1.72

C 22 65.6 10.4 –1.87

C 24 65.1 9.5 –1.89

C 26 66.1 9.2 –2.07

C 28 65.1 9.1 –1.96

C 30 67 9.1 –2.07

C 32 63.8 10.1 –1.66

C 34 65.5 9.1 –3.23

C 36 66.3 9.4 –1.53

C 38 67 10.1 –1.77

C 40 66.5 9.3 –1.94

C 42 67 9.5 –1.86

C 44 65.6 9.6 –1.77

C 46 66.1 9.6 –1.86



C 48 65.7 9.8 –1.89

C 50 65.1 9.9 –2.05

M 65.86 9.52 –1.92

SD 0.78 0.37 0.32

Bird ID R1157: Syllable D.


D 1 64.8 9.1 –3.01

D 2 65.8 9.3 –3.41

D 3 65.9 9.4 –3.44

D 4 65.3 8.8 –3.36

D 5 65.9 9.6 –3.46

D 6 66.8 9.1 –3.61

D 7 66.9 9 –3.56

D 8 66.6 9.1 –3.57

D 9 66.6 8.9 –3.52

D 10 67.4 8.8 –3.59

D 11 66 9.1 –3.59

D 12 64.4 9.3 –3.22

D 13 65.2 8.5 –3.10

D 14 65.1 9.6 –3.49

D 15 64.6 9 –3.21

D 16 66.2 9.9 –3.20

D 17 65.9 9.6 –2.35

D 18 66.3 9.4 –3.39

D 19 67.7 10 –3.47

D 20 67.6 9.5 –3.40

D 21 66.2 9.3 –3.36

D 22 66.4 9.8 –3.41

D 23 66.2 10.2 –3.28

D 24 65.9 9.1 –3.28

D 25 65.3 9.2 –3.32

M 66.04 9.30 –3.34

SD 0.88 0.40 0.26

Bird ID R1157: Syllable E.


E 1 64.9 9.2 –2.12

E 2 65.1 9.4 –2.35

E 3 64.7 9.7 –2.40

E 4 64.7 10 –2.21



E 5 65 9.6 –2.41

E 6 67.6 9.6 –2.68

E 7 66.9 9.5 –2.60

E 8 67.1 9.8 –2.64

E 9 67.3 9.7 –2.67

E 10 66.2 9.7 –2.58

E 11 67.1 9.3 –2.67

E 12 64.8 9.4 –2.53

E 13 64.9 9.6 –2.48

E 14 65.4 9.6 –2.41

E 15 65 9.3 –2.56

E 16 65 9.2 –2.42

E 17 69.7 11.9 –1.91

E 18 67.3 9.7 –2.44

E 19 67 9.6 –2.48

E 20 65.8 9.8 –2.44

E 21 66.3 9.9 –2.40

E 22 68 9.6 –2.59

E 23 66.9 9.2 –2.50

E 24 66.7 9.9 –2.56

E 25 66.1 9.5 –2.66

M 66.22 9.67 –2.47

SD 1.28 0.52 0.18

Bird ID R1157: Syllable F.


F 1 70.1 12 –2.06

F 2 70.3 11.8 –2.17

F 3 69.4 11.6 –1.93

F 4 69.1 11.4 –1.90

F 5 69.8 12.2 –2.18

F 6 70.3 12.2 –2.10

F 7 70.1 12.4 –2.14

F 8 68.9 12 –2.04

F 9 68 11.8 –2.29

F 10 69.2 12.1 –2.07

F 11 68.3 12.1 –2.20

F 12 69.2 11.5 –2.21

F 13 68.7 11.3 –2.16

F 14 68.1 11.8 –2.07

F 15 69.3 12 –2.19

F 16 69.8 11.7 –2.20

F 17 69.4 11.8 –1.98



F 18 70.2 12.2 –2.06

F 19 69.1 11.9 –2.03

F 20 68.7 12.4 –2.07

F 21 70.2 11.3 –2.15

F 22 70.9 11.8 –2.30

F 23 70.6 12.3 –2.33

F 24 70.2 12 –2.32

F 25 69.9 11.9 –2.35

M 69.51 11.90 –2.14

SD 0.79 0.31 0.12

Bird ID R1157: Syllable G.


G 1 65.6 10.2 –2.41

G 2 65.9 11.9 –2.36

G 3 64.2 11.1 –2.06

G 4 63.7 10.7 –1.94

G 5 65.4 10 –2.18

G 6 65.3 10.3 –2.38

G 7 64.3 10.1 –2.38

G 8 64.2 10.3 –2.36

G 9 63.2 10.1 –2.29

G 10 64.3 10.8 –2.36

G 11 64.1 10.4 –2.46

G 12 64.4 9.7 –2.40

G 13 64.2 10 –2.39

G 14 63.4 10.5 –2.36

G 15 64 10.1 –2.46

G 16 64.5 9.9 –2.41

G 17 64.3 10.6 –2.11

G 18 64 10.5 –2.21

G 19 64.8 10.9 –2.38

G 20 64 10.2 –2.33

G 21 63.7 10.3 –2.32

G 22 65.5 10.6 –2.25

G 23 64.6 10.2 –2.30

G 24 64.3 11.1 –2.29

G 25 64.3 10.3 –2.45

M 64.41 10.43 –2.31

SD 0.68 0.47 0.13



Bird ID R1157: Syllable H.


H 1 67.7 18.8 –2.36

H 2 67.1 21.1 –2.14

H 3 66 19.1 –2.15

H 4 66.6 17.9 –2.08

H 5 67.4 19.8 –2.19

H 6 66.8 19.9 –2.20

H 7 65.8 19.8 –2.38

H 8 66.7 20.4 –2.15

H 9 64.6 18.1 –2.22

H 10 66.3 18.9 –2.17

H 11 66.9 18.7 –2.31

H 12 64.6 17 –2.12

H 13 66.2 18.5 –2.18

H 14 63.1 17.7 –1.86

H 15 65.4 19 –2.26

H 16 66.2 18.6 –2.09

H 17 67.6 19.8 –2.17

H 18 66.3 20.2 –2.01

H 19 66.4 19.6 –2.11

H 20 67.8 20.3 –2.19

H 21 65.8 20.4 –2.12

H 22 68.4 20.4 –2.49

H 23 67.4 19 –2.48

H 24 66.6 19.9 –2.16

H 25 66.9 19.5 –2.19

M 66.42 19.30 –2.19

SD 1.15 0.99 0.14

Bird ID R1157: Syllable I.


I 1 71.3 16.1 –2.73

I 2 70.8 17.7 –2.69

I 3 70.7 16.5 –2.99

I 4 71.4 17.2 –2.79

I 5 71 16.6 –2.81

I 6 72.9 16.9 –3.03

I 7 71.9 16.9 –2.98

I 8 72.1 17.7 –2.94

I 9 72.4 16.9 –2.99

I 10 71.9 18.3 –3.05



I 11 72.7 17 –2.95

I 12 70.7 15.6 –2.77

I 13 71.6 16.1 –2.91

I 14 69.2 16.5 –2.61

I 15 65.2 18.1 –2.95

I 16 72.9 17.3 –2.74

I 17 72.1 16.6 –2.85

I 18 69 17 –2.46

I 19 72 17.4 –2.92

I 20 71.7 16.3 –2.93

I 21 70.1 16.9 –2.83

I 22 72.8 17.9 –3.07

I 23 72.4 17 –3.13

I 24 71.3 16.8 –2.94

I 25 71.3 17.3 –2.84

M 71.26 16.98 –2.88

SD 1.63 0.65 0.15


Date post:	02-Dec-2021
Category:	Documents
Upload:	others
View:	4 times
Download:	0 times

Common Terminology and Acoustic Measures for Human …

Documents