 | | | Mixer question | Mixer question 2005-03-31 - By mark wilson
Back
I'm pretty sure that still won't work correctly once you start using it Brad.
Although its late here now, and my brain has stopped working.
Oscar: I like your last email, I didn't pick up on that idea when I first read
your previous mails, I'll give it a go in the morning. This does still seem to
be something that could be quite easily improved with the mixer though, by just
having a third blending mode which didn't require the rest/Neutral position to
be 0.
Anyway, I'll give it a shot andhopefully I'll be sorted.
Thanks,
Mark
__ ____ ____ ____ ____ ____ ____
From: owner-xsi@(protected) on behalf of Brad Friedman
Sent: Thu 31/03/2005 22:59
To: XSI@(protected)
Subject: Re: Mixer question
then create two clips for each. one full spread. one no spread. set
up and expression on the weight of one to invert the other. proxy param
the free one.
create two clips for each finger's bend. one straight. one bent. set
up and expression on the weight of one to invert the other. proxy param
the free one for each finger.
now you have overall spread on one proxy param and individual curl on
each finger. put it all in a compound clip and blend it against other
animation when you need more control.
-brad
Oscar Juárez wrote:
> yeah but you then can have troubles with updating due to theres no
> information driven
> by the mixer in some channels if you turn off the weight of one of the
> clips.
>
>
> rad Friedman wrote:
>
>> I believe you can turn off channels individually in clips (I could be
>> wrong on this). Therefore, in the situation you have laid out, one
>> clip should have the Y rotation channels turned off. The other clip
>> should have the z rotations turned off. That way, the two clips will
>> not collide and be normalized. Each can be weight animated on its own.
>>
>> -brad
>>
>> mark wilson wrote:
>>
>>> Like I said in my other reply, the blending doesn't work the way I
>>> want it to. I know how to get the stuff into the mixer and linked
>>> up. If you try this with the default rig and blend between the Fist
>>> and the spread, you don't get a tight fist, you get a 50/50 of the
>>> two. What I want is this:
>>>
>>> I set a spread pose where only the global y rotation of the root of
>>> the fingers rotate to close the fingers together
>>> I then set a fist pose where I only rotate the finger joints in the
>>> global z to close up the hand and make a fist, also the thumb would
>>> be rotated into position.
>>>
>>> Then when I blend between these two I want the blending to combine
>>> the two to have 100% of the fist rotations about the global z and to
>>> have 100% of the spread rotations about the global y, not, as it
>>> seems to do, average between them. Taking off Normalise gets close
>>> to what I want but just adding them is not much good unless you take
>>> into account the default pose. I think I'm saying I would like to
>>> work in additive mode but from a default pose thats values were
>>> treated as 0. At the moment if you default pose has some rotation on
>>> it, it will get added to the sum and you won't get the correct pose.
>>>
>>> Is anyone getting what I'm saying? Sorry if I'm making it confusing.
>>> Mark
>>>
>>>
>>> __ ____ ____ ____ ____ ____ ____
>>>
>>> From: owner-xsi@(protected) on behalf of Brad Friedman
>>> Sent: Thu 31/03/2005 18:28
>>> To: XSI@(protected)
>>> Subject: Re: Mixer question
>>>
>>>
>>>
>>> store each of your predefined poses as a clip. Apply them to the
>>> mixer... one per track. Use the parenting options in each clip to make
>>> them always cover the whole timline. Then create proxy parameters from
>>> each clip's weight into a CPset. Make sure the mixer is set to
>>> normalize.
>>>
>>> from this base technique you can expand. For example, you can put all
>>> the hand clips under a compound clip. Then you can blend between that
>>> compound clip, and other hand animated clips to override your automated
>>> CPset when you want to.
>>>
>>> -brad
>>>
>>> mark wilson wrote:
>>>
>>>
>>>
>>>> I don't like the way the default rig does it. Having a slider for
>>>> every bone is just crazy imo. You might as well just animate the
>>>> bones direct. The point of my idea is to have predefined poses for
>>>> the hands that are hooked up to sliders, so you can easily blend
>>>> into hand poses. Then if you need to tweak those you can just
>>>> animate the bones to override it. I just think it would be nicer,
>>>> and its what I'm used to doing in other packages. I guess this
>>>> isn't possible though.
>>>>
>>>> Mark
>>>>
>>>> __ ____ ____ ____ ____ ____ ____
>>>>
>>>> From: owner-xsi@(protected) on behalf of Matt Morris
>>>> Sent: Thu 31/03/2005 17:19
>>>> To: XSI@(protected)
>>>> Subject: RE: Mixer question
>>>>
>>>>
>>>> Is there any specific reason you're going into the mixer for this?
>>>> I find the default xsi rig way of doing it is pretty decent, where
>>>> you have a proxy parameter set of sliders and also a set of saved
>>>> actions on the bones for different poses...
>>>>
>>>> -- --Original Message-- --
>>>> From: mark wilson [mailto:owner-xsi@(protected)]On Behalf Of mark
>>>> wilson
>>>> Sent: 31 March 2005 15:00
>>>> To: XSI@(protected)
>>>> Subject: Mixer question
>>>>
>>>>
>>>>
>>>> Is there any way to get the mixer to do a difference between
>>>> tracks. What I am trying to do is blend between static poses to
>>>> control fingers. This doesn't seem possible with normalized mixing
>>>> or additive. When I say blend between static poses I mean change
>>>> the weight vaue for the track. So I have a track for each finger
>>>> position and set the track weights to get different poses. Does
>>>> that make sense? If theres a better way to do it, I'm all ears, or
>>>> hands, or whatever. ;-)
>>>>
>>>> Mark
>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>> ---
>>> Unsubscribe? Mail Majordomo@(protected) with the following text in
>>> body:
>>> unsubscribe xsi
>>>
>>>
>>>
>>>
>>
>
> ---
> Unsubscribe? Mail Majordomo@(protected) with the following text in
> body:
> unsubscribe xsi
---
Unsubscribe? Mail Majordomo@(protected) with the following text in body:
unsubscribe xsi
|
|
|