Knowledge graph centered on Spectrographic song analysis with 51 nodes and 155 connections. Top connected: Rocky Mountain Biological Laboratory, Gothic, Crested Butte, Pekania pennanti, Colorado East River Valley.
Method synopsis
Detailed measurement of acoustic parameters from spectrograms to quantify song characteristics and calculate vocal individuality metrics.
Synthesized from method descriptions across 3 papers using this protocol.
Procedure from a recent peer-reviewed implementation
Analyze recorded songs for syllable identification
Create spectrograms using 256 point FFTs in a standardized window of 0-5.5 kHz and 0-8 seconds. Apply 100-contrast level and brightness level of 70. Create syllable identification key to aid in consistent syllable recognition across recordings. Identify and catalog each syllable in recordings longer than 1 minute containing at least 50 syllables.
Quantities: 220 distinct syllables identified across all individuals, recordings >1 minute with ≥50 syllables used for analysisDuration: Not specifiedConditions: Excluded songs recorded during rainy or windy conditions
Equipment: Raven Pro 1.3 software, Krein et al. 2009 reference
Create data sets for method comparison
Generate three data sets of different sample sizes from individuals with sufficient recordings. Large data set: first 50 syllables from 16 individuals. Medium data set: first 100 syllables from 14 individuals. Small data set: first 150 syllables from 5 individuals. Each data set uses the same recordings to enable direct method comparison.
Quantities: Large (L): 16 individuals × 50 syllables, Medium (M): 14 individuals × 100 syllables, Small (S): 5 individuals × 150 syllablesDuration: Not specifiedConditions: Used same recordings across all methods for accurate comparison