README.md

In pmoracho/loudness: loudness analisis

Package for loudness analysis. It performs a series of analyses from an audio file by first transforming it into a timeline with the loudness values. This package depends on the ffmpeg tool that should be previously installed.

The length of the timeline will be from a tenth of a second to multiple hours, or days. The limit will be on the length of audio file that ffmpeg is able to process, the disk space, for the extraction of the LUFS values, the memory capacity to handle the amount of observations involved. As an example, about 24 hours of recording, are transformed into a lutl of 868899 observations and about 23 MB, processed in about 15 minutes

First, we can load an audio file, to generate a lutl object, which is nothing more than, a series of loudness level observations per time.

library(loudness)

tl <- lutl_from_file('audio/R_MIC_200903-165318.mp3')
#> Get loudness from audio/R_MIC_200903-165318.mp3 using ffmpeg... 
#> Parsinng data... 
#> Process datetime... 
#> Elapsed time: 5.01 
#> Lenght of tl: 00:00:09:54

The lutl objects like any other object in R can be saved and it is recommended to do so:

saveRDS(tl, 'audio/R_MIC_201010-132132.Rds')
tl <- readRDS('audio/R_MIC_201010-132132.Rds')

From the timeline, we can study the distribution of loudness levels:

plot_density(tl)

You can also reproduce some part of the timeline

play(tl)

We can also study loud sounds within a timeline

tl <- readRDS('audio/R_MIC_201010-132132.Rds')

plot_timeline(tl, 99)

pmoracho/loudness documentation built on July 16, 2021, 2:04 p.m.