callsync is an R package intended for users that use collar or backpack microphones and want to analyse the resulting recordings. It allows users to create a pipeline from raw recordings until final feature vectors, but specific functions can also be replaced by alternatives.

The main features are:

1. alignment and partitioning of drifting microphones using signal compression and cross correlation
2. call detection using an amplitude envelope
3. fine-scale alignment and call assignment across recordings using cross correlation and energy content
4. fundamental frequency tracing
5. analysis of the resulting traces and wav clips

The package also offers:

• flexible spectrograms
• spectrographic cross correlation
• analysis of amplitude and frequency modulation
• loading selection tables from Raven and Audacity
• calculating performance of the detection and assignment functions

A vignette going trough the main functions can be found here.

R version 4.1.0 or later. Earlier versions might work if you replace the |> function with %>% and load the package tidyverse.

The tested version can be installed from CRAN:

install.packages('callsync')
library(callsync)

To install and load the developmental version run:

install.packages('devtools')
library(devtools)
devtools::install_github('simeonqs/callsync')
library(callsync)

The package should be fully functional, and extra features are actively being developed. Feel free to contact me (simeonqs@hotmail.com) with any questions, suggestions or requests.

If you want a framework that uses all the main functions you can read our preprint and can clone this repository. To cite the package, you can use the following citation:

Smeele, S. Q., Tyndel, S. A., Klump, B. C., Alarcon-Nieto, G. & Aplin, L. M. (2023). callsync: an R package for alignment and analysis of multi-microphone animal recordings. BioRxiv. <https://doi.org/10.1101/2023.02.07.527470>