Plug-in to speed workflow?

Ok, I am somewhat limited in my capabilities with the software but have managed to digitize my entire tape collection and do some audio recording for the kids. I have searched the plug-ins and these forums, but I am not completely up with the lingo, so this is probably my shortcoming in finding an answer to my question.

I have an audio file that looks like this:
aud.png
Is there a plug-in or a command that will select the next portion of audio without me having to use the mouse every time to select it? In other words, I want to be able to skip the whitespace and have the software select the next portion where audio exists (I don’t know if I am being clear here.) This is one of many files of foreign language words–I have about five thousand words to extract in several files. It would sure speed up the workflow if there is a suitable plug-in!

Thanks for any/all help!

Spencer

There are a couple of Audacity functions that relate to this: the Silence Finder and the Sound Finder (you probably need the Sound Finder).

They are both available from the Analyze menu. See the end of this page in the 1.3 manual: http://manual.audacityteam.org/man/Analyze_Menu

The Sound Finder will ignore the silences an produce a label track wtih the chunks of sound labelled. If you have focus in the label track then you can Tab forwards to the next label (or Shift+Tab to move back to previous label).

You may interested in the silence remover: Effect > Truncate Silence See: http://manual.audacityteam.org/man/Truncate_Silence

WC

Can you elaborate on that. What exactly do you mean by “extract”? Do you want to make lots of very short files with one word in each file?

Steve: yes, I am making very small .mp3s of each word, all separate files. Sorry, don’t know the proper terms here, still learning. I am selecting each portion of the audio, naming it and exporting it as an .mp3. I am using Butler to input all the other keystrokes (with just one keystroke combination) I need except the naming of the file, so my workflow is pretty streamlined except for having to select the audio portions.

WC, I will check out what you suggest, thanks much!

Spencer

As waxcylinder suggested, try the Sound Finder to mark the words with labels, then use Export Multiple to export each labelled region as a separate file.
The process is similar to what is described in the first part of this tutorial: http://wiki.audacityteam.org/wiki/Splitting_recordings_into_separate_tracks

Well, it probably would have helped if I had copied over the entire Audacity folder with all the plug-ins upon installation. A new install of the latest version has the Sound Finder plug-in that Steve references–I did not have that with my previous install.

Works beautifully and turns out that the guy wrote it to do exactly the same thing I am doing now.

Cheers and thanks!

If smallness in memory is important you should have a look at the the “sample rate” and “bit rate” for the mp3 …
http://manual.audacityteam.org/index.php?title=MP3_Export_Options

i.e. use the smallest values for both which give you adequate sound quality if minimizing the file size is important.


Nice to have a contributor from China. :slight_smile:

Trebor, I have fiddled with the export options for the .mp3s and I am averaging ~25k for each file at 128k encode. I couldn’t get rid of the distortion below that, even with the variable options, and am quite happy with that file size. Thanks!

Sumtingwong is the Chinese SNAFU. :smiley:

If the samples are of speech you can get away with a sample rate of 22050Hz, instead of the standard 44100Hz, and half the file size.

[ I often use speech sampled at 22050Hz then encoded to Mp3 at 64Kbps, (mono, not stereo) ].

I sampled at 44.1KHz, will try at half that with the other settings you suggest. I was thinking that the files were going to be much bigger, around 40-100k, so expectations exceeded! Honestly, with the files the size they are now, they are quite reasonable given bandwidth and ubiquitous storage available. Not trying to sound lazy, but want the best audio possible and will give up some file size for the quality.

Thanks!

In that case, keep the sample rate at 44.1 kHz, convert the tracks to mono and keep the MP3 settings at 128 kbps - that should give you excellent sound quality with a mono file (roughly equivalent to 250 kbps for a stereo file).

Steve, when I convert the track to mono using “Tracks → Stereo Track to Mono”, it gives me no file size savings and no choice under the .mp3 export options to save as a mono file. I also tried splitting the track from stereo to mono from the “Audio Track” dropdown and exporting as well as splitting the track into left and right tracks deleting one of them and then exporting, again with very little difference in the file size. I am assuming that the files will be half the size with one one channel, is that correct? Am I doing something wrong here or am I making a bad assumption?

Correct.
When you export a project that is all mono, the exported file is automatically mono.
When exporting as 128 kbps MP3, the resulting file size will be the same, whether you export as mono or stereo, because there will be 128 kb of data for every 1 second of audio. The difference is that if you export a stereo track as 128 kbps MP3, those 128 k bits of data are shared between the left and right audio channels. If you export a mono track as 128 kbps MP3, all 128 k bits of data are used for that one (mono) channel, producing higher sound quality.

128 kbps mono will produce excellent results with very little loss of sound quality.

To make the file smaller, use a lower MP3 setting. 64 kbps will produce sound quality on a mono file that is almost as good as 128 kbps for a stereo file, but only half the size (because there are now only 64 kb of data for each second of audio rather than 128).

For speech, 64 kbps mono is likely to be more than adequate, 128 kbps mono will be near perfect.

Ok, got it. I assumed that when I converted the track to mono or got rid of either the left or right channel that I was getting rid of half of the data. Good to know!

Don’t know if you could answer a question about the Sound Finder plug-in, but when tabbing through the labeled items/sounds/(don’t know the correct term here) on the track, focus is on the label with the cursor at the end of the label. Is there a way to modify the script so that the label is highlighted so that I could just sart typing rather than have to delete the old label?

Currently Sound Finder just adds a number to the label text - to delete that number when the cursor is at the end, use the backspace key, then start typing.
I’m currently working on a new version of Sound Finder that adds a bit more flexibility to the label text. If you keep an eye on this topic Sound Finder / Silence Finder improvements I’ll be posting my proposed updated version soon.

Awesome and thanks! Will keep an eye out.