Integrating Technology in the Foreign Language Classroom

1. Capture sound [File/New]

• Avoid background noise.
• Use a good microphone. Talk across your mic, don't blow air into the mic.
• Use 16-bit samples (32-bit yields better quality for editing your working copy, though you can reduce the final copy to 16).
• Select sample rate (22050 is ok for voice depending on desired quality; 44100 yields CD quality for music).
• Higher sample rate creates larger file size and slower downloads but better quality.
• Adjust volume to maximize the sound but avoid clipping (if a wave is too loud, the top of the wave is clipped off).
• Leave at least a second of silence in the recording to aid in future noise reduction.

If you have trouble getting your own sound file for the following steps, you may download a sample sound and practice editing that : soundsample.mp3. All of the manipulations of digital sound are really mathematical transformations, remember graphing functions in algebra class?

2. Center Wave (aka DC bias adjust, DC offset) [Transform/Amplitude/Amplify]

VERY important step after many recordings! Your sound might not be centered for amplitude (see below for a wave that is not centered).

Centering should be done FIRST. In Audacity, DC Bias can be found under the Normalize effect. If you have additional effects to apply, you might prefer to center the wave first and Normalize last.

Make sure that the wave is centered before applying any other transformation.

3. Noise reduction  (FFT 8192, precision 7)  [Transform/Noise Reduction]  But only do this if you think it needs it.

In Audacity, the noise removal effect is fairly straightforward. Select a few seconds of noise, then get the noise profile. Next, go back to your sound, and select the whole thing. You can preview before actually removing the noise. If you don't like the result, you can try removing more or less of the noise until you are satisfied. Without going into the details, noise reduction transforms the part of the sound that it considers "noise" into something softer. Too much noise reduction creates unwanted distortion, so it is not a magic bullet. It is better to record clean audio in the beginning than to try to fix it later.

4. Trim [Edit/Trim]

  

Audacity When trimming, use the Zoom tool to find the precise segments that you wish to select. Listen to your selection to be sure it is what you want. Select area of the sound you want to keep then select Edit / Trim.

 

5. Listen carefully and clean up anything you don't like. To avoid breaking the rhythm of your sound sample, reduce the volume (Effect / Amplify) of problem noise instead of cutting it out. Using a negative dB reduces the volume.

6. Normalize  [Transform/Amplitude/Normalize]

Cool Edit / Audition

• Use this to ensure that all of your sound files are at the same level.
• Set to 95% then calculate. 
• Click OK to make the change.
• Whereas Noise Reduction, Noise Gating, and Compression may not be needed, Normalization is generally required to make your project sound good. You can do the DC Bias Adjust (Center Wave) in the Normalization window to save a step (only if normalization is the only modification you are making to the sound).

Audacity

In Audacity, Normalize and DC Offset (aka: Bias, center wave) are found in the same Effect. Normalization is set to -3 dB or about 75% of full volume for the highest peak of your sound. All sounds in a project should be set to the same level, but you might prefer a higher volume than 75%.

 

It is possible to use the Amplify Effect in Audacity to Normalize by setting the same Peak Amplitude for each sound file in a project. -0.0dB sets the highest sound peak at the maximum sound level. Some people prefer to avoid using the maximum sound level in order to be sure to prevent clipping, so you might use -0.1dB or -1.0dB. As long as you use the same setting for all sounds.

10. Save backup copy

• Keep a backup at the best sampling rate possible for future use.

• For use on the Web, it can be necessary to reduce sample rate for faster file transfer.
• Software such as RealAudio may require a particular sample rate.
• In general, just leave this alone.

12. For use on the Web most sound files should be saved as .mp3. It usually is not useful to use stereo for most voice-only applications.

The most important steps in sound editing are : Capture, Center Wave, Trim, Normalize.

In Audacity, the View / History feature is very nice and will allow you to undo any of the edits that you have made to a sound.

Cool Edit used to be purchased from Syntrillium Software http://www.syntrillium.com/cooledit/
But they were bought by Adobe.  The program is now called Adobe Audition. Ask about academic pricing / educational discounts for teachers. It remains quite expensive. Adobe also has a less expensive sound editor called Soundbooth (should be available for less than $80 with academic discount). Other high quality, and expensive, tools are Pro Tools, Sound Forge, and Wavelab.

For this course, we are using a free program called Audacity. It works quite well and can do everything you will need:  http://audacity.sourceforge.net/ (The current version is 1.3.9, and this requires an additional Lame plug-in to export mp3.)

A very inexpensive solution that also seems to do a good job with sound is either MIXCRAFT Recording Studio, about $65, from Acoustica http://www.acoustica.com/
AVS Audio Tools has an unlimited evaluation period, but also very inexpensive.
Goldwave is great in it's price range, about $49.

Roger Nichols Recording Guide
Audacity Tutorials from Sourceforge
Audacity Tutorial(Adventures in Multimedia)
Digital Domain - excellent resource
Editing Sound Files (mp3handbook)
Digital Audio Tutorials and Resource
Audio software listing from Sonic Spot