HCPT/Session 3: Editing
1 How to Edit a Talk
1.1 Before Starting
- Ensure that everybody (particularly the speaker) has been informed well in advance, and that you have fully signed consent forms before the event (c.f. HCPT/Session 1)
- You must have the program Audacity: download from here: http://audacity.sourceforge.net/
- If you are following this tutorial in the context of CUiD / University of Cambridge, you won't need to generate mp3 files: This will be taken care of through the UCS-SMS service. However, if you are following this tutorial elsewhre, you may need an program to export your audacity project MP3. In that case, you can use itunes (http://www.apple.com/itunes/download/ ) or LAME (http://audacity.sourceforge.net/help/faq?s=install&item=lame-mp3). A tutorial on how to do this with iTunes is here Encoding mp3 with iTunes.
- Open Audacity and select
- "Default Sample Rate 44100Htz"
- "Default Sample Format 32-bit float"
- File Formats
- "Make a copy of the file before editing"
- Go to
- Browse to find the recording, click 'Open'
We need to get rid of all the bits you don't want, for CUiD we'll just take top and tail it, for speed and efficiency. When you start removing bits from the middle it gets messy and non-continous.
- Use the selection tool (top left of screen, looks like an 'I')
- Select the bits you want to cut (i.e. beginning of talk, before the introductions, the end and possibly the questions)
- Go to
- Cut (or just Ctrl+X)
This limits the recording to a certain range so it is audible on all computers. However it means that any loud parts may skew the recording, so we need to get rid of these first:
- Go to
- All (or just Ctrl+A)
- Select "Remove any DC offset (center on 0 vertically)"
- Select "Normalize maximum amplitude to -3dB"
In some cases there is a big discrepancy between the loudest and quietest parts of the recording. If you have some very loud parts (e.g. the applause is much louder than the speaker or there is a sudden bang in the middle of the talk) these can distort how the sound might be heard.
The vast majority of talks will NOT need this second step. For most talks, just normalise it (as per instructions above), and then carry on to export the file).
Below are some instructions for reducing the amplification of the loudest parts (those with the longest vertical lines on the recording). DO NOT SELECT SPEECH. Only carry out this second step for sudden loud noises or very loud applause.
Select the loudest parts of the recording (e.g. applause) and go to
- Set the amplification to a negative number and press 'ok'
- Keep doing this until your selection is at a reasonable level with the rest of the recording
- Repeat with all applause/loud noises through the whole recording
N.B. Ensure you listen to the changed parts to check it sounds ok, before moving on to the next stages
Your recording should now be ready to go. Listen to it all the way through to make sure the sound is ok from your computer.
N.B. If you find that your audacity screen consists of two rows of media which aren't the same and you have problems when normalizing, try clicking on the arrow which points downwards, next to the name of the track immediately to the left of the media and select split stereo track. You can then deal with the media separately.
When you import a track into Audacity, it sometimes happens that only one row of media has the recoding on and the row below it just shows a straight line. (This is common if you have been using the Fostex for example). If this occurs, split the stereo track immediately (instructions above) before editing. On the media track with the straight line, you can then press the 'x' button next to the name of the track, directly to the left of the media and delete it so that you are only editing the track with the recording on it.
1.5 Exporting the file
Now you are ready to export the file.
- Export as WAV
- And save it to your desired location
You are now ready to publish the recording online.