how I turn heads

[neeters_guy]

Okay, Bee-ruce2, here you go. I hope it helps. Be warned, it's still pretty long.

Neeters_guy’s notes on Sinclair’s head turn method
---
PHASE ONE: FULL HEAD FRONT

1. Make a vector layer consisting of a circle (head) and two more circles (eyes). Name the layer "head front"

2. Make a new switch folder, name it "mouth front". Place a vector layer in it, name it "AI". Draw the mouth in the AI position, it will consist of a black circle (for the back of the mouth), tongue, teeth, and skin. Build them in that sequence, ie., black circle is back and skin is front. (Appendix B below.)

3. Once the "AI" layer looks good, copy the layer nine times, one for each of the standard phonemes. Adjust the vectors in the each layer to match the desired phoneme shape.

• a. There are ten phonemes and they are case sensitive for Papagayo: rest - the closed position, etc, E, AI, L, MBP, FV, O, WQ, U.
  
  b. All of these layers have to have the same vector count, or this won't work, so don't delete any vectors, do not add any, do not change the thickness on any of the lines, do not change the colors of any of the objects on each layer or the objects stack sequenc (on every layer, the "skin" is on the top and the "black circle" is on the bottom).
  
  c. As you adjust each layer, turn on interpolation on the switch layer, set the first keyframe on the timeline to "AI" and the next keyframe to whatever phoneme layer you have completed (make sure to space the keyframes a second apart). When you scrub through the time line you'll see the "AI" layer interpolate into the new layer.

4. Put those two layers "head front"(vector layer) and "mouth front" (switch layer) into a group folder, name the group folder "Full Head Front".

5. Make a copy of the "head front" vector layer, pull it out of the "Full Head Front" group folder and name the new vector layer "front switch" (this is not a switch layer).

6. Copy and paste all of the vertices on the "rest" vector layer in the "mouth front" switch layer and paste them onto the "front switch" layer.

7. Check point: If all is correct at this point, you have one vector layer with a full head (head, eyes, mouth) and a group layer with the exact same head (as long as your mouth switch layer is in the "rest" position). Turn the "front switch" layer on and off to make sure that all of the vertices are in the exact same spot; as you turn it on and off (exposing the folder beneath it) you should see no changes in the head.

PHASE TWO: FULL HEAD SIDE

8. Copy the "front switch" layer and name it "side switch". Move the vectors on the "side switch" vector layer so that the head and mouth are in the side view position. Important: The "side switch" and the "front switch" vector layers have the same amount of vectors (just like the mouth switch layer).

9. Put the "Full Head Front" group folder, and the "front switch" and "side switch" vector layers into a new switch layer. Name it "FULL HEAD" (this will be the top level layer). Make sure the switch layer is set to interpolate. Set the first keyframe on the switch layer to "front switch", then scoot down the timeline and make the next keyframe "side switch". When you scrub through the timeline, the head should turn.

10. Set up the group layer "Full Head Side":

• a. Select and copy all of the mouth features on the "side switch" vector layer. Paste them onto a new vector layer, name it "rest" and put that layer into a new switch layer named "mouth side".
  
  b. Select and copy the circle (head) and two small circles (eyes) on the "side switch" vector layer and paste them onto a new vector layer named "head side".
  
  c. Make a new group layer, name it "Full Head Side" and place the "mouth side" layer and "head side" layer in it.

11. At this point your layer order consists of a "FULL HEAD" switch layer which contains:

• a. "front switch" vector layer
  
  b. "side switch" vector layer
  
  c. "Full Head Side" group layer which contains
  • i. "mouth side" switch layer
    ii. "head side" vector layer
  d. "Full Head Front" group layer which contains
  • i. "mouth front " switch layer
    ii. "head front" vector layer

12.The only thing that needs to be done is to complete the "mouth side" switch layer. I find it easier to just copy the "rest" layer multiple times and move the vectors into position rather than copy the "front mouth" switch layer and move each layer into the side position. It can be done either way of course.

PHASE THREE: TESTING

13. To see the head in action, set both mouth switch layers to a .dat file (generated from Papagayo). Because the mouths shapes interpolate, you'll have to add a "rest" keyframe to the .dat file in the spaces where nothing is being said or the mouth will interpolate to the next phoneme over the length of silence.

14. Set the first keyframe on the "FULL HEAD" switch layer to "Full Head Front" group layer. Whenever there is a pause in the characters speech and the "mouth front" is in the rest position, add a keyframe to switch from "Full Head Front" to "front switch" before the character starts talking, then add a keyframe to "side switch" and then another key frame to "Full Head Side".

15. The character will start talking in the front position. When it pauses, the head will turn and begin talking in the side position.

Notes: I generally complicate the file (as if this isn't complicated enough) by adding hair switch layers, nose switch layers, and jaw switch layers (the same way the mouth layer is built).

You can use this technique to switch from up, down, left, and right positions (repeat steps 8 to 12), if you’re willing to build that many variations of the head.

APPENDIX A: LAYER HIERARCHY


APPENDIX B: MOUTH DETAIL

[Bee-ruce2]

Wow......I did'nt realize I was going to make you write such an essay!..........but thanks for sharing, it looks like a detailed and thorough explanation.Will try it out in the next couple days as soon as I have time.

[sinclair]

I'll throw in my 2 cents here: I wonder if its necessary to have more than side, 3/4 and front views for face. I mean are you really going to need a 3.5/4 view? I think just having 3 views in switch layers is more than adequate.

3.5/4, not really necessary but down and up position do add to the mix-

http://www.youtube.com/watch?v=r6n9LXjLmu0

It really comes down to story, (or, in my case, lack there of) and what you want to do with the character-

And (again) very good stuff Neeter, I tried to keep my original explanation as simple as possible, (because the more I added, the more confusing it got) the breakdown on how the mouth is built is very important because I normally build the eye switch layer, jaw switch layer and hair/head/ear switch layer the same way so that I can "play" with all of the models features.

I generally just copy and paste the vectors from one layer from each switch folder (eye, jaw, hair/head/ear, mouth), name each of those layers "rest" in their original switch folders, and paste them onto the actual switch vector layer in the exact stacking order the switch layers are in-

Once the "switch" vector layer is copied and all of the vectors are manually moved to the side or 3/4 view, I double the new "view/switch" layer (for reference) and copy each set of vectors and put them on new layers that will become the new view switch layer.

For instance, I select all of the vectors that make up the eyes, copy them, paste them to a new vector layer, then delete them from the "reference" layer.

I delete them from the reference layer so that it is less confusing when trying to isolate each specific group of vectors (I've tried grouping them, but when you make a copy of the layer the groups never end up copying onto the new layer)-

A shortcut I use when building the mouth is to double the "rest" layer, name the new layer "AI" move all all the vectors into the "AI" position then put them both into a switch layer. Then I copy the "rest" layer (again) the layer will automatically be named rest 2, set "rest 2" for the first keyframe on the switch layer, then set "AI" as the next key frame about 30 frames down in the time line.

Then I select all of the vectors on the "rest 2" layer and scroll through the time line watching the vectors move from the "rest 2" position to the "AI" position, when I see the moving vectors closely resembling another phenome I tap the "translate point" tool on the layer (being sure that when I hit it that I don't actually move anything) to make a key frame. I then copy that keyframe to 0 on the timeline (replacing the original keyframe on the layer)-

With a few minor tweaks "rest 2" becomes "E" rather than moving all of the vectors from the "rest" position into the "E" position manually I let anime studio do the work. It saves alot of time when building multiple view switch layers. (I do the same thing with the eye switch layer (start with "closed" and "wide open")

Again, its a basic concept taken to excess, but I like the way it works-