Basic tutorial for creating voice files for fdRPG

First draft by Clint Herron (hanclinto at gmail dot com) 5/2/2005

Step 1) Creating your recording environment

You don't need a special "sound room" to record audio, but there are a few things you can do in a normal house to make sure that you get the best recording possible.

1. Don't record your audio in a room with a lot of echo. Most rooms that have computers have a fair bit of stuff in them, so this won't be an issue.

2. Minimize all background noise. This includes fans, air conditioners, loud TV's, etc. Just try and make the room quiet.

3. You don't need a phenomenal sound card, but try and get one that can record audio without having a buzzing sound in the background.

4. You don't need a great microphone either - most professional mics aren't compatible with your average computer sound card, so a standard microphone will have to do. A microphone that can be held close to your mouth is ideal (as opposed to the kind that is mounted in a monitor or laptop case). Another good kind of microphone would be a gaming headset.

Step 2) Good recording technique

1. When recording your voice, hold the microphone close to your mouth. Somewhere in the range of 6 to 16 centimeters is a good distance (about 2 to 6 inches). Just make sure that you're comfortable, and that you can record at a decent signal level (more about this later).

2. When speaking, do not speak directly into the microphone. If you do this, you will get a lot of breath and wind noises. Instead, hold the microphone at the corner of your mouth and speak *past* the microphone. You will get a much clearer recording this way.

Step 3) Capturing a good recording

1. Once you have good speaking technique, the next most important thing is to make sure you're recording at a decent level. Clipping is defined as audio that goes too high, and sounds cut off at the top. A decent level can be defined as a recording where you can see a good variation in the waveform, but it never maxes out.

2. Record several takes of each of your lines, but be careful, as often your first take will be your best as it's most comfortable and most natural.

3. On long lines especially, sometimes one will record one section perfectly, but will mess up on the others. Save the take anyway! You can splice them together later if you don't wind up being able to speak the whole line in one continuous, good take.

4. Listen to your takes, and if you can, it helps to have a good friend there who will give you his or her honest opinion, and help you to change what you need to change.

5. Give yourself large entrance and exit times to the recording - you don't need to start speaking the instant you click "record", and you shouldn't click "stop" immediately after you think you've finished. Instead, give yourself a second or two of entrance and exit silence. This way, you can be more comfortable, and won't run the risk of cutting off a long "s" sound at the end, or having a pop at the beginning that you can't get rid of because it's on top of your speaking.

6. Record as much as possible, without worrying about cleaning it up. Just take all of your takes and save them, then go back and edit them later. Don't worry about editing each take - just save it and come back to it later. Recording long dialogues can be tiring, and you don't want to have to stay "in character" any longer than you have to.

Step 4) Cleaning up the audio

Once you have a good take and have selected it as your final run, there are three things that need to be done. Removing any glitches, normalization, and trimming.

1. The first step in cleaning up audio is to remove any glitches such as pops in your audio. If you see a spike that shouldn't be there, and it sounds bad, you can select it in Audacity and hit "delete" to cut it out. Play with this, and you should be able to clean up any bad sounds in your audio.

2. If you have 10 different speeches that are going to be played one after the other, they all need to be at roughly the same volume. However, I can almost guarantee you that your takes will not all be the same volume. Therefore, we need to do something called "normalization". Normalization looks at the loudest volume in the selected sample, and then adjusts the volume of the selection so that its peak is at -3 dB. This is the standard that fdRPG uses - all of your audio should be normalized to -3 dB. To normalize in Audacity, select your audio (you can do this with Ctrl-A), then go to your effects menu and click "Normalize". Both check boxes should be checked. Note: You can normalize individual parts of your sample separately, in case one half is loud and the other half is soft (like if they're spliced together from seperate takes), then you can normalize each section individually in order to get a consistent volume throughout the final recording.

3. Once you've done that, then you need to trim out all of the extra silence at the beginning and at the end. The standard for fdRPG is to have between .5 and 1 seconds of silence at the beginning, and no unneeded silence at the end. When trimming audio off of the end, be *very* careful as it's easy to cut the tail off of trailing "s" sounds. Remember, "undo" is your friend! :)

Step 5) Final steps

1. All of your takes should be saved in the Audacity format, but your final sample will need to be exportd to Ogg Vorbis (.ogg).

2. Then add your audio to the game (or ask one of the developers to do it for you), and you're done!