This page explains how to create audio files for Language Mentor lessons. We discuss acquiring audio (getting your scripts recorded) and editing audio (breaking the recordings into "chunk files" that Language Mentor can use). You can do these steps yourself or hire someone else to do them.
We assume that you have already written your lesson's script and created its lesson XML file. Both of these steps are explained in previous Creating Lessons pages.
First, Acquire the Audio
You can record audio yourself or you can hire professional voice talent. Which approach you take depends on your budget and your target audience.
If you are creating quick personalized lessons for individual students we suggest that you record the audio yourself. You'll save both money and time.
If you are going to make your lessons available to the public, we suggest that you consider hiring professional talent. Recording quality audio isn't easy, and the people who use your lessons will be listening to this audio over and over, and over and over! If the audio quality isn't good they'll be pushed away. Recording a lesson doesn't require much time, so the cost of hiring voice talent doesn't need to be substantial. Also, voice artists charge a wide range of rates. You don't need hire expensive talent to get excellent results. We've found great people at affordable rates by advertising on sites such as Craigslist.
Recordings should be made in a lossless format such as WAV. Record the audio in mono - Language Mentor plays mono, so recording in stereo will only increase file size. The audio should be recorded with a sample rate of 44.1 KHz and a 16 bit "bit depth".
Check the Audio
Once the audio is recorded, check it carefully. It's much easier to re-record audio at this point than to deal with problems later. Check these things:
- The quality of the recorded voice. Do you like the sound of the voice? Is the pace and tone as you'd like it?
- The quality of the recording equipment, especially the microphone. Even a great voice won't sound very good if it's recorded through a cheap microphone.
- The "sound ambiance" of the room where the recording is made. Rooms that have too many hard surfaces - walls, uncovered windows, etc. - can create a subtle echo effect that degrades the quality of the recording.
- Does the audio correspond precisely to your script? Yes, voice artists occasionally make mistakes. If there's an error you can either request that the chunk be re-recorded or, if it makes sense, simply change the script.
- Finally, check for the absense of background noise. Nothing should be audible in the gaps between words and phrases.
Checking for Background Noise
Do this check with the audio at a fairly high volume - not loud enough to hurt your ears, but almost that loud. If the audio is recorded at a low level you may need to do an initial "amplify" edit in order get sufficient volume.
Background noise falls into two categories:
- Environmental noise: This is noise that comes from the environment - the sound of a refrigerator, cars passing by, etc. Professionals avoid this by using sound-proof rooms - others use strategies such as recording at 3:00 AM, recording multiple takes, etc.
- Equipment noise: This is "hiss" and/or "hum" created by recording equipment. Equipment noise can be difficult to eliminate - sometimes the only solution is to use different equipment. If you hire professional voice talent they should know how to avoid this problem. If you are doing the recording yourself we suggest that you start with your microphone. Try borrowing a quality microphone, and see if the problem goes away. It may also help to use a preamplifier to amplify the microphone's signal.
Most audio editing software has a "noise reduction" feature. This feature may be helpful in some cases, but be aware that it also degrades the quality of the audio. This can be a worthwhile tradeoff if you only need to eliminate low level background noise. If you have a larger problem on your hands, noise reduction usually just replaces one problem (background noise) with another (audio that sounds strange). The best and easiest way to deal with this problem is to record audio without background noise in the first place.
- Audio editing software: If you don't have this, we suggest that you try Audacity. It's fairly good, and free.
- Basic audio editing skills: If you want to learn, we suggest that you look at Audacity's help and tutorials (free) and/or Lynda.com's audio tututorials (paid).
Editing audio consists of taking the raw audio that you have acquired and a) breaking it up into "chunk files", and b) "polishing" the audio in these files.
Creating Chunk Files
If your lesson is a single language lesson you'll create one file for each of your lesson's chunks. This will, of course, be target language audio.
If your lesson is a dual language lesson you'll create two files for each chunk - one for the target language and one for the lesson's native language.
In general, the process consists of opening raw audio within your audio editing software, then repeatedly selecting a single chunk's audio and saving that portion as a separate file. But that's a bit of an oversimplification. Let's look at some details.
Polishing the Audio
First, if your audio has any "pops" or "clicks" like this:
you can reduce them by zooming in, selecting the area of the "pop", then reducing it's volume, like this:
We use the word "reduce" but this often works so well that the sound ceases to be noticable.
Next, ensure that all of your chunks are at the same volume relative to each other. This can be a bit more complicated than it sounds. Different voices have different characteristics that give them different subjective volumes. If you adjust all voices so that they have the same objective volume (i.e. the same decibel level), some voices will sound louder than others.
We recommend that you create a "volume sample" audio file that you can compare your files to, and that you adjust each file's volume "by ear" so that it matches the volume of that file.
While you are ensuring that your audio files have the same volume relative to each other, you may also want to make the volume fairly consistent within each file. It's up to you whether or not you wish to do this step, but we find that chunks with relatively even volume, like this:
are are easier to hear clearly than chunks with more variation in volume, like this:
How to Name Chunk Files
We touched on this subject briefly when discussing the fileNameRoot element. We'll now go into more detail.
Audio files are named using this format: [chunkFileNameRoot].[languageCode].mp3, where:
- "chunkfileNameRoot" is the contents of the chunk's fileNameRoot element in the lesson's XML file.
- "languageCode" is the ISO 639-3 code for the file's language.
- If you're creating a dual-language lesson that teaches Mandarin Chinese to English speakers...
- And the lesson XML's chunks element looks like this:
Then the audio file names for these four chunks should be:
Audio Formats & Bitrates
Your ultimate goal is to create audio files in the MP3 format, but we suggest that you initially create chunk files in the WAV format or some other lossless format.
The MP3 format is a lossy format, which means that every time you re-encode an audio file as an MP3 you lose a little audio quality. Doing this once isn't a problem, but if you repeatedly edit and save in the MP3 format the loss of quality can be significant.
It's a good practice to keep a set of "source" WAV format chunk files. This allows you to edit your files repeatedly without lowering their quality. Then, when you are ready to publish your lesson, use an audio batch conversion program to create copies in the MP3 format. (We use Switch by NCH.)
WAV Format: Use 44.1 KHz 16 bit format.
MP3 Format: Use a bitrate of at least 40 Kbps. A higher bitrate will improve the audio quality slightly, but bear in mind that there's a direct relationship between bitrate and file size. A lesson with 80 Kbps audio will take twice as long to download as one with 40 Kbps audio, will take up twice as much storage space on the user's device, etc.
Where To Put The Files
We suggest that you create a "wav" subfolder in the lesson folder and store the WAV format files you create there. Then, when you make MP3 format versions of these files, put them into the base level of the lesson folder, like this:
If you have questions about the lesson creation process, please contact us on our Creating Lessons forum. We're here to help.
You've now created all of your lesson's files. Next we'll show you how to set up a lesson library on the Internet, and how to publish your lesson to a library. The first step in this process is setting up your library folders.