rigging together a character for all positions...

[rpc9943]

hey im watching the animation on that duck that is walking, checking out that you can change layer order in the timeline that is so cool... okay. so would it be possible to do a head rotate on both sides, rig it up to a switch that interpolates, draw the left, right, and middle face positions, and then im set right? hell i could even make a switch above that with different emotions.. right?

how would it all work? i also want to be able to change each mouth and eye position within these rigs... is all of this possible? i could then just start developing different character heads and throwing in some of that in a template file....

(my cartoons at least right now are just going to be talking floating heads like aqua teen hunger force... but in my own humor, etc)

am i headed towards the right direction??? pls let me know... so excited

RonC

[heyvern]

Uh... it's a lot more work than you think. It all depends on the time you have and any type of production deadline. Trying to cover all the bases before you start is a TON of work... trust me a I know. It's my "Search for the Holy Grail". Trying to create a head or body of a character that can do ALL things so I am not limited during animation.

However that preliminary work up front is intensive. Using switch layers as you describe is also a very complex set up. There will be a lot of key frames on many different switch layers.

I have done exactly what you described; a head switch with left and right turns interpolated... yikes... it's not fun to set up. You have to modify the "front" main view of the head to match those turn views. Plus you have to have mouth positions for each view? That's a lot of work. I was only doing one or two views of the mouth position switch layers. Then the head could turn to one of those positions to talk... but it couldn't talk in ALL positions.

I gave up on that idea a long time ago and went instead with a scripted bone solution and a "2.5D" head. I can turn it anyway I want and create any expression on the face I want at any angle..

... however the set up for that is even worse than the switch layers! 250 bones!

The reason I like it so much is that it is piece of cake to animate, I can rotate the head to any view and lip sync it from any view. I can even lip sync and rotate the head later. Very flexible... BUT A FREAKING PAIN TO SET UP. You would not do this for a short quick fast turnaround project. It takes a lot of up front work. each character is unique.

In a tight deadline situation it might be better to plan ahead of time what angles of the head you want or need and only create those views. You may only need one or two specific views of a characters head with maybe one switch interpolation for quick turns.

You have to remember there is a trade off with this stuff. You can spend a lot of time BEFORE animating to make a versatile character rig (that's what I like to do) OR you can do a simple set up and only animate exactly what you need based on a storyboard or whatever. Many people prefer that. The deciding factor for which way to go is based on the deadline or budget of the project.

-vern

[slowtiger]

One day I really should set up a website with all this basic stuff. In the meantime, here's an example of my typical character setup:
viewtopic.php?t=13052&highlight=character+rig+setup

Remember that you can flip layers easily, which means that you create only one side view and use it for left and right view both. My typical "full" character rig would have these views:

- front
- 3/4 side
- side
- 3/4 back
- back

I don't use a rig which interpolates between those views, which means all turns have to be quite fast.

Remember that you can have different flavours of this setting, taylor-made to your requirements. Let me have one example:

It's possible to put all expressions into one mouth switch layer, like this:
- mouth switch
- - A smile
- - M smile
- - A angry
- - M angry
and so on. (Naming layers like this means you can't use Papagayo for automated lipsync.)

It's also possible to have a setup like this:
- mouth expression switch
- - smile mouth switch
- - - A
- - - M
- - angry mouth switch
- - - A
- - - M

[heyvern]

(Naming layers like this means you can't use Papagayo for automated lipsync.)

Holy cow this is a great idea! You could put all the expressions in one switch and just modify the dat files with a text editor! You append that extra switch layer name to the phoneme name using a text editor and a simple regex expression. What you could do is use Papagayo to lip sync the whole dialog. Then you pick and choose which frame ranges in the dat to change the expressions (Angry, smile, etc). Use your text editor and select just those sections to search and replace for the different switch layers.

I use jEdit. This regex expression would add "Angry" to the end of each phoneme in a papagayo .dat file in one shot. Jedit lets you perform search and replace on just selections in the file. So you can swipe over the lines in the dat file that represent frames 212 to 340 and change that to angry.

Example .dat file code:

Code: Select all

MohoSwitch1
174 etc
180 AI
185 E
187 FV
....

Search for this. It looks for all the upper and lowercase letters at the end of a line ($ is the end of a line) The parenthesis around the brackets is the pattern you want to save for the replace:

Code: Select all

([a-z A-Z])$

replace it with this pattern. The $1 is the first pattern from the seach. Then just add in the new text you want for that swtich:

Code: Select all

$1 Angry

Result after replace all:

Code: Select all

MohoSwitch1
174 etc Angry
180 AI Angry
185 E Angry
187 FV Angry
....

This is COOL!

-vern

[Sage]

heyVern,

what is jEdit? I am not great with code but are you basically saying there would be a faster way then having say 4 mouths switch layers and inserting the Data sound file in each manually?

[madrobot]

If we can get some sort of anme file referencing I think you'd be able to work up some sort of generic rig or rigs (tailored to your own preference of course) which might serve as a basis for setting up your characters.

Wouldn't work all the time obviously, characters won't always be similar enough to be derived from the same rig, but it would definately serve (for me at least) as a great starting point which would save a lot of work.

And with referencing, and updates, tweaks and improvements would ripple through any scenes you have them referenced in. (As would any stuff-ups )

[heyvern]

heyVern,

what is jEdit? I am not great with code but are you basically saying there would be a faster way then having say 4 mouths switch layers and inserting the Data sound file in each manually?

Yes exacctly. You wouldn't need to put the dat file in a whole bunch of different switch layers. However you still have to edit the one dat file... so the extra work is a trade off. You would be changing the name of the switch layer that is used in the dat file. So for instance frames 20-100 would use the normal switch layer phonemes and then 101-160 might use the "Angry" switch layers.

They are all in the same switch layer but each set has a different expression name after it. So when you load the modified dat file it uses that one switch layer for everything instead of loading the same dat into many switch layers and then keying a parent switch over it.

This has the added bonus of allowing interpolation for ALL the switch expressions. You could smooth interpolate from "AI Angry" to "AI Happy" because they are in the same switch layer.

-------
http://www.jedit.org/

Jedit is just a text editor. It's open source (free) and runs on everything, Mac, Win, Linux. It is mainly geared towards programmers and HTML but I use it all the time for editing AS files and other application files that have a text based format. It's also what I use for writing scripts. It has built in syntax highlighting for many languages.

-vern