The biggest problem we get is sound quality. People can put up with the video not looking perfect, but if they can’t hear what’s going on then they’ll give up watching. So wherever possible we need to try to improve sound quality.
Most of the time the sound on videos people send us is too quiet. We need to do the following to fix that:
- Extract the audio from the video, so it’s a separate file we can work on
- Increase the volume of the audio, cleaning up any nasty bits
- Put the original video and edited audio files into OpenShot, lining them up to the sound is in sync with the video, and reducing the volume of the original video so we just get the edited sound
Easy! Here’s how.
Extracting audio from video
Open VLC, and in the “Media” menu select “Convert/Save” This will show a window like this:
Drag the video you want to extract the audio from into the “You can select local file…” box, or use the “Add” button to select it. Then click “Convert/save” and you’ll see this:
Make sure you select “Convert” and “Profile = Audio – MP3”. Click “Browse” and choose the folder which contains the video you’re working on – it’s a good idea to keep everything in one folder. Give the file a sensible name then click “Save”:
Press “Start” and VLC will extract the audio from the video and save it as an MP3 file. Great.
If you add multiple files to convert at once in VLC (by dragging and dropping, or selecting incividually) then you don’t need to specify where to save them to – VLC will automatically put the MP3s in the same folder as the video. This is a big time-saver!
Now we can edit the MP3 audio. Open Audacity and drag the MP3 file into it. You’ll see something like this:
Those blue squiggles are the audio waveform. Welcome to the wonderful world of audio editing!
First, we zoom out to look at the whole file – you can use the plus/minus magnifying glass buttons, or hold down “Ctrl” and use your mouse wheel.
Step 1: Normalise
The first thing we do is normalise the file. This makes the volume as loud as it can be (without distorting) by stretching out the squiggles so they are taller.
Double-click the waveform to select all of it (or go to the Select menu and click ‘All’), then select the “Effects” menu and click “Normalise” and make sure the settings look like this:
Step 2: Remove spikes
No we need to look for “spikes”, where there’s a sudden loud noise. Perhaps the person speaking dropped something, or a door was slammed! Here’s what spikes look like:
The reality is that for speech you will see MANY spikes all the time. This is normal, but too many spiky spikes means that the overall volume of the person talking is too low. We need to flatten the spikes a bit.
The easiest way to to this is to select all the wave form (press Ctrl-A, or “Select” > “All”) then go to the “Effects” menu and click “Limiter”:
Make the settings look like this:
Click “OK” and Audacity will flatten the spikes, so they look less … spiky:
Step 3: Normalise again
Once you’ve got rid of most spikes you can normalise the file again. Follow the instructions for step 1.
We want to peak amplitude (the posh word for volume) to be -1.0dB. Yes, that’s minus 1. Decibels are weird. That volume will make the speakers voice a nice volume on the video, but not be distorted.
People’s voices have natural variations which mean the sound goes up and down in levels all the time. That’s fine, by editing the audio we want to make sure that we don’t have a really quiet bit of the service. We want everyone to sound roughly at the same volume.
Here’s what a waveform will look like before and after normalising and limiting a couple of times – you can see here how much difference this makes to the volume of the person speaking:
Exporting your audio
Once you’ve normalised your audio and had a listen to it to make sure there are no nasty sounds in it, you can click “File” > “Export” > “Export as MP3” and save it in the same folder. It will ask you if you want to overwrite the file, which of course you do.
That’s the audio and video sorted, now we can put them into OpenShot.