Audacity Tutorial (WIN/MAC/Linux Users)

Welcome to the audacity tutorial where everything will be going through audacity. Many people using audacity but don't know how to use this software. Don't worry, I'll try walkthrough as much as I can, so here are some step by steps:

Usage software

Here is an audacity:

Oooo, scary... (also your theme might be different so i'm using custom theme for that)

You may be saying this is complicated, or you could say it's easy but y'know what it is.

Let's start things off one by one of what they can do:

Main bar

I'll go try and explain them as much as i can.

Those are the playbacks and stuff, you can just playback the audio and stopping it. Specifically, some tools also exist, I usually don't use them but here's one you should use:

These two arrows use to slip or moving through audio. Which in case if you want, you can move some audios like this:

Really useful not gonna lie...

The next tool we have is the silence and trimming audio. Those two buttons are also useful to remove some unnecessary noise.

Do note that: Audio trimming only trims part of the audio so, probably don't use that.

Last but not at least the zoom buttons!

It's useful if you don't know any keyboard shortcuts, which we'll get into that later. However, if audios are too small to see or too big, you can use these buttons. Which I'll show you:

Only use plus and minus zooms.

Alright now that covered up let's start off with this side:

Timeline

Ah, waveforms.

So this is, timeline or tracks or... I dunno whatever you want to call but usually I like to call them a timeline, As you can see here things nothing to do here in the timeline but except selecting things and controlling the volume, did I say controlling the volume? yes. on the left side, you have properties and volume, things like if audio is too loud or too quiet. But we'll cover that up later.

Oh, and on the right side, that's your main audio track. Especially down their label track. Which is nice but also neat.

One more last thing to say before we start. Let's follow through with some keyboard shortcuts. I only know a few of them and here are a few:

Keyboard Shortcuts

Spacebar - Plays and stops the audio

CTRL/Command + Mouse Wheel Up/Down - Zooms in and out to the waveform.

CTRL/Command + B - Making a label

And, well I guess that's pretty much it. ON TO THE MAKING DATASET!

Making dataset in Audacity

Alright here we go, you are in audacity.

And right now you probably don't know what to do, remember what I taught you about buttons and stuff? yeah, but right now we haven't covered how to import sound.

Well how can we do it is 3 things:

1. Dragging and dropping audio into audacity.

Excuse the OS language.

2. Opening an audio file in audacity.

3. Importing audio files in audacity.

Working inside audacity

First things first you may notice that this audio is in stereo, which is two tracks. Tacotron needs at least mono, 22050 sample rate and, 16-bit wav. But we'll cover these in a moment.

Let's start with making the stereo into mono.

If your audio is already in mono you can skip this section.

Making audio stereo into mono

Here what you should do, click on that dropdown and select make a stereo into mono:

You'll notice that it's been split into two tracks, that's fine. As long as you wanna mute one of the tracks tho. Because two tracks playing = loud. So keep in mind that. It splits into two tracks of left and rights.

Changing Sample Rate

This next step is not that hard. Do you see that down there? That's the sample rate right here:

That one down there.

You wanna change that to 22050 and you should be good to go!

Labeling dialogs

Okay, this is the part where you wanna label the dialogs. Basically how it works is that you label them and it saves them as a trimmed audio file. No matter how many times you labeled it saves audio files your trimmed labels.

Let's say you labeled 50 dialogs and audacity will save them into 50 audio files.

Pretty neat right? For now, let's start with the fun part:

Here I'm doing an audio dialog with Yin from Yin Yang Yo! and what do I do?

Easy, selecting the audio part you need (hold the left mouse button and select audio):

And pressing CTRL/Command+B to add a label:

This will create a label track and a label. Now you gotta have to do is to type a number. 1 dialog = 1.wav

I know this may be confusing, but, you'll get used to it.

My audio has background noise to it.

If audio has a little background noise then there's a way to do this.

Audacity has an effect called Noise Reduction, this will remove unnecessary quiet background noise.

So what you wanna do is to select the noise part and select effect -> noise reduction:

You will then be greeted by this window:

So here's what you wanna do, click get noise profile, this will grab noise and then select specific dialog that has noise to it, do effect and select repeat noise reduction or as keyboard shortcut is CTRL/Command+R

Once that done, it should be removing some unnecessary background noise.

Otherwise, if that didn't work, vocal remover from colab page may help you with that.

pageVocal Remover Colab Page Guide

Exporting audios

Once you have done gather enough dialogs you can go ahead and export audios now.

Here how we can do this, select File --> Export --> Export Multiple...

Now after that window will pop up:

Well, it shouldn't be that hard but here what you should do. First, select location in the folder section. Doesn't matter where you save.

In format, select WAV (Microsoft) signed 16-bit PCM

And lastly, select Labels.

Once you are done click Export. And you should be done.

Fun fact

Did you know that you can import video into audacity? Yeah, you actually can. Basically, you drag a video into audacity, and then FFmpeg will try to convert the video to audio. Pretty neat yeah?

Finish

If you have done everything correctly, then congratulations but we haven't done it yet, we need to transcribe the audios. So we'll be going to transcribing audio.

pageTranscribing Dialogues