During a recent troop with members of the 501st, one of our members showed me a Jawa voice box she bought online. It sounded great and I thought it would be a quick and easy thing to build. As it turns out, I already had all the parts I needed, so the total project build time ended up being less than 6 hours.
I visited YouTube and found several others have built one of these already. I commented on the video and begged for the sound files, but no one responded. So, I set out to find my own. There are many free sites on the Internet that offer .MP3 and .WAV sound clips. However, I was disappointed to see a limited selection of Jawa voice clips. The ones I did find were pulled straight from the movies and had background noise (music, explosions, etc.)
Then, I came across this amazing Disney site that has lots of clean sound files. They offered about 20 Jawa sounds that all sounded great. No background noise – just pure speech. Long story-short, I couldn’t figure out a way to download the sound files, so I just patched my headphone jack to my Mic jack and used i-Sound Recorder software to record the audio to .WAV file format. I then went back and edited each .WAV file using Audacity to get rid of the extra silence recorded before and after the sound clip, and also normalize the volume.
I’ve used the Adafruit Audio FX Mini Sound Board on several other projects, so felt confident that it would work well for this project. I loaded the .WAV files onto the board and tested. The audio volume was very weak. Luckly, I had an Adafruit Stereo 3.7W Class-D Audio Amplifier board (also left over from another project). Once I connected that up, the volume was much better.
It was important for the device to be highly modular so it could be put on/taken off easily while in costume. So, I used a CAT5 cable with an RJ45 connector for the IO pins on the sound board, a barrel plug for the power, and a mono 1/8″ sound jack for the speaker. I cut holes in a small project box to expose the wires, and screwed it all together.
I opted for a 3 x AA battery pack with an on/off switch which was a good compromise on size, weight and power capacity. Ideally, I’d like to get the power pack even smaller and lighter, but didn’t have a LiPo pack available at the time.
I got the idea from watching some YouTube videos – I could use a palm-sized perf board and solder some small momentary buttons on in two rows of 4. An elastic strap helps keep it in place. Curling the 4 fingers allows you to trigger one of 8 sounds. I cut a piece of adhesive foam-rubber for the underside of the pad to make it more comfortable to wear and avoid any pokes from the soldered leads underneath.
I mounted an 2 inch speaker inside the mask. This was a nice speaker I found in an Electronics surplus store that had screw terminals that let me easily and securely connect the wires without any solder. I used E-6000 to affix the speaker to the inside of the fencing helmet. See this earlier blog for how I made the helmet.
Although I had a stereo output from the amp, I elected to use a single speaker to reduce the number of wires. It turned out to be fine for volume.
- The buttons triggered by the pinkie finger are hard to press
- Move some of the most common sounds to the middle so they are easier
- 8 sounds is not a lot – choose wisely.
- Adjust the lanyard so the boxis positioned underneath the bandoleer so the bulge is not noticable when wearing the costume.
- Need one or two more elastic bands to support the wire along the arm
- Routing the CAT5 up the arm, around the neck and down the chest offered some strain relief on the wire.
- 3 AAA batteries provided more than enough power capacity for a typical troop (4 hours)
- I used zip ties to help with wire strain relief inside the electronics box as well as on the hand pad.
- The sound board will not playback when connected to USB port on a PC. I had to use external power to get the playback to work. Not sure why.
- Audio playback seems loud when I test in my workshop, but when I get to a troop and there are lots of other loud noises, they always seem quiet. So, I’ve trained myself to always make the audio playback a bit louder than I think is necessary to accommodate.