Pilot Packs Part 1 : Recording, Editing, and Encoding

A lot of /m/en have been asking how to make pilot packs as of late. In the interest of getting more user-created content into our games, I've thrown together a quick and duurty tutorial on how to record, clean up, and edit pilot voice samples. Part II will deal with getting the game to recognize the voice files.

Before getting started, I suggest that you avoid samples with full-volume background music; they're hard to clean up and sound retarded. Fortunately most shows ramp down music volume when characters are speaking, so in some cases we can get away with it. Background noise isn't nearly as big a problem, it's easier to clean up, and tends to blend in with the mayhem of the game anyhow.

What you'll need

• Audacity - for recording, trimming, and cleaning up voice samples
• >>http://audacity.sourceforge.net/
• WinLame - for batch encoding
• >>http://winlame.sourceforge.net/

Recording Voice Samples

- Fire up audacity, and select Wave/MP3 as your recording source. This will allow you to record whatever is being outputted to your speakers.



- Open your media player of choice and cue up to a couple seconds before the sound you want to capture plays. In this example I'm pulling sound from a Gundam 00 clip on youtube, of Setsuna saying "I am Gundam" (this will never stop being funny).

- Hit record in audacity, play your clip, then hit stop. Simple as that. I find for noise reduction purposes it's best to have a couple seconds extra before and after the clip.

Noise Reduction

The sample I recorded has a distinct hum on a GN Drive in the background; others I've worked with have had explosions, gunfire, or faint music playing. So far the noise reduction tool has worked fairly well for me. If you have a fairly clean sounding sample you can skip this step.

- Highlight a portion of the sample where just the background noise is playing, make sure that your character's voice isn't included at this point.



- Click Effect -> Noise Removal -> Get Noise Profile

- Now select-all, and click Effect -> Noise Removal and move the slider all the way to the left and click preview. The farther right you move the slider, the more 'hollow' your character's voice is going to sound, chances are you'll never need to move it more than 2 clicks right.

- Adjust the slider it till it sounds good, then click Remove Noise.

- You'll probably want to amplify the clip a little at this point, just click effect -> amplify. Audacity will automagically determine how much it can amp the sound without distorting it, so just use the suggested settings.

Trimming Sample, and using Fade in/out

- Select the portion of the sample you wish to keep and click Edit -> Trim



- Now highlight from the start of your sample, to where your pilot starts speaking (or screaming in hot blood). Click edit -> fade in



- Repeat the last step except at the end of the sample, and use fade out

Batch Encoding your Samples

- Once you have edited your voice sample, export it as a .WAV file. Repeat the above steps until you have a bunch of good sounding WAV's for your pilot.

- Open winlame, select your files, drag and drop them into the big empty white box and click next.

- Choose Ogg Vorbis Encoder as your output module, and the location you want to save to

- At the next screen move the slider all the way to worst, as we want the files to be as small as possible. Chances are the drop in quality won't be audible. Click next.



- Click play button to start encoding your files. Simple as that.

Here are my "I am Gundam" samples;

Before; >>http://bootfighter.googlepages.com/i_am_gundam.wav

After; >>http://bootfighter.googlepages.com/i_am_gundam.ogg


It seems like a lot of steps, but the whole process only takes a minute or two per sample; it's been my experience that actually locating the source material is by far the longest part of this process