Thread: Software for 3 Axis Skulls- VSA, Parlay, etc

Tweet

ok-
I've got my 3 axis skull made, and IT'S ALIVE!!! MUHAAHAAHAHAHAHA!!!!

I have a number of questions about the SOFTWARE for programming it. Unless you have a clean wave file, the WaveMotion Analysis really sucks as it moves the servos to every noise in the file.

What's the easiest way to program the talking motion of a song/sound file?

I am using a Saitek joystick- which I love for the head motions- but, might not be good for the jaw movements. It seems like you just can't move it fast enough...
ideas?

Probably the easiest thing would be to write a vu meter-like program and map the extremes of the high and low values to the high and low values of the skull's servo and save the mapped values to a file you can import into VSA.

ummm, sure... what was that? Connect the doohickie to the thingamabob and gyrate that podular modulating sansquat? 220, 221, whatever it takes...

LMAO! Bob- I'd prefer to use your program! Parlay looks awesome- but been having trouble connecting it to the SSC32 and plugging in the servo min/max values when set to mini-ssc mode. Any ideas on either?

The other thing that kinda sucks with the joystick- this one 'centers' pretty strongly at the center point. So when moving from left to right, it kinda of pauses in the middle- or speeds up to the middle. It's got a strong spring in it. It's probably something I have to just deal with and get used to- but FYI for anyone else.

Here's how I did it.

I made a CD of the song I wanted the skull (or Skulls) to sing.

I set up my computer with a microphone to record my voice.

I put on headphones and played the CD I made of the song in a walkman into my headphones.

I sang along and recorded my voice singing just the lyrics that the particular skull I was working on would sing.

Then I made a wav file of the track with just my voice. In multi track recording software, (I used Adobe Audition) I matched up (synced up) my voice track with the voice on the real recording.

Then I saved the adjusted file making that a wav file. I had to make sure that it was exactly the same length as the original file and that my voice started at exactly the same number of seconds in as the real track

Then you load up the new track with just your voice and run wavform analysis on it.

After your done with that, save that routine. Then delete the audio track and insert the original.

The jaw now only reacts to the voice part and ignores all the music and other noises, but you hear evertything and if you practice a few times, you will stay in sync the whole way through.

You don't have to be on key or even know every word. As long as you make a sound at the right times at the right volume it will be fine.

I hope that clears it up for you.

There is no other way that is as effective as this.

Thanks Bob- that's exactly what I thought. I was hoping there was an easier way- but- ehhh, maybe this isn't all that bad.
Thanks!

Bob,

Glad to hear that works. I've had mixed luck with both Parlay and VSA motion capture. I've thought about going that route, but I held out until I got my Lynxmotion board today (I've been using parallax, which won't work with Parlay.)

I didn't had any better results with the Lynxmotion board. In fact, the movement is a lot more random and jerky than with my parallax board. I'm wondering if I need a bigger power supply. I've got one 5.2V, 400ma supply powering all four servos and the Lynxmotion board itself. With the parallax, the supply only has to power the servos. The board is powered by the USB port.

Has anyone else used both of these boards and notice any difference in performance?

Shoot, I just moved my mouse in sync to the recording played back through VSA to capture the jaw servo motion. Took a couple of "takes" but it was QUICK and EASY. Yes, you have to know exactly where the vocals are and it takes a little practice but it works.

ok, I'm doing this- but when I bring in my modified audio track of just my voice- it barely picks up on the voice. When it does, it mainly 'chatters' the jaw, instead of opens/closes it fully. Do I need to amplify the audio track here?

Abunai.. I'm currently(no pun intended 'mate) runnin' a 1 amp supply fer me board an' three skulls in testin' mode, bu' have a 230w Micro ATX power supply from Frys Electronics fer $20, which supplies 5vdc @ 20amps(current limit) along with the standard 12vdc's too. I figured with the 18 or so skulls I'm goin' ta run, 20amps would be more than enough to keep everythin' fed..

Spectre, after setting me servo limits, I imported th' song file, then ran Waveform analysis fer each of th' jaw channels (3 in Pirates o' th' Black Tide video). After the waveform is finished loadin', I went in an removed a lot o' th' jaw events, so it doesnt look like the jaw is chatterin' so much, but openin' an' closin' more an' wider. It doesnt have to move exactly like the audio file drives it, just open/close somewhat in time to the audio(words). I lengthened some of th' events an' shortened some others.. main thin' is I be lettin' th' wavform do the hard work, then I cleaned it up. Fer parts of th' song, where one skull sings an th' others dont, I started with th tracks of th' guys not singin'... I zoomed in a little, marked each of th' blocks that dont get sung by th' guy an' removed them. Did the same fer music that triggered jaw movements.. jus removed them. Dint really take long... A Pirates Life fer Me took lil over an hour ta clean up fer me three singers... Pirates o' th' Black Tide took 'bout alomst 3 hrs ta clean up, bu' tha' included using VSA's capture option an' me mouse to load head movements fer th' left two skulls.. so you can do it whateva' way werks fer ya 'mate...