I am wondering if there is a way to export the raw audio from audacity with the label in alongside the raw audio in a timestamped .csv file.
The label could either denote a start and end or just a repetition of the label throughout the whole time period. Any relevant information would be appreciated. The .csv would have the channels as its own columns then a column for the labels.
Labels can be exported as tab separated values. See: https://manual.audacityteam.org/man/importing_and_exporting_labels.html
I see how to export the labels. I meant to ask if there was a way to export the raw timestamped audio to a file with the labels denoting the annotations in a single .csv file in a simple way (few clicks, or documented workflow) from audacity, or in a few steps. Thanks.
What do you mean by “raw timestamped audio”?
The raw audio data has numeric values for each sample. Typically for a stereo track there are 44200 samples per second, so for a 3 minute track there are 15,876,000 numbers. To print that to a csv file would require about 8 bytes per number, so that’ll give you a csv file that’s around 127 MB.
Last time I looked, Microsoft Excel had a limit of 1 million lines for a spreadsheet page. The raw data from a 3 minute stereo song is nearly 16 x greater than that limit.
What are you actually trying to do? What’s the task?
I am trying to build a classifier to classify different types of audio using Audacity’s labelling feature. Specifically, I have 10 wav files of a siren noise that I would like to annotate and upload to a cloud machine learning platform in order to classify the files (ie IBM Watson).