Hi Matthias..
Hi Stefan,
the "not showing" is(hopefully was) the biggest bug.
The strange values can be compensated easily if you use
the field evaluator via scripting, but in this case it
needs some additional lines in the manual.
You might be right that its the big part of the bug..
Another big part of it, can be the fact that it seems to flip rapidly from a
positive huge value to a negative huge value, in a not controlled way......!
(this just cant be correct..)
Maybe some of it can be worked around via scripts..
Though it needs a fix, and not a mentioning in the manual of a workaround...
As most are not at all "friendly" with scripting.. including me..=)
Script should be a complement or another way of doing things, only in
extreme case should it be for workarounds..=)
(Workarounds a good when waiting for a fix, or if the workaround is much
better than a fixed feature..!)
I totally agree, that the field evaluator can be a very
helpy tool, IF it's working.
Yep it sure can..
So lets try to make it work.. =)
The FE had always a strange behaviour through the different
versions of RS: Working very well, not working, strange
values, ......
Let's see what v7 will do for us, maybe a working FE again ;-)
Lets hope so, it might only need a proper bugreport...
I have some base now for such, I will try to find some energy to file one..!
FE needs abit more in the manual !
The tutor prjs needs better values for the chors (will see if I can take a
look)
Also maybe a good way of finding out what values needs to be set for the
chors, to get good, smooth and nice working anim.
Maybe a way to "plot" the values that the FE spits out..to easier find the
Min and Max values.. and so on !
(This might already possible to be easily done by a script made by someone
that knows scripting..?=)
I really like the app, and if the animation system will
work as exspected it would be great to do more with RS.
Yep me too. its my only app and will always be.. no matter what =)
Though I agree, the anim sys should work as expected and also more..
If you can grab a developer, whisper something like this to him:
"For character anim the "Skinmanager" is a must have,
the build in weight paint is not so funny."
Again.. if someone would be really close by here, as I am a strong guy I
could grab most of them..=)
I am not sure that would be the best tactics though..=)
On a more serious note..
Carlo talked (long time ago), about RealPuppet he was working on that would
replace the 'Skinmanager'
When he brought that up, Juha asked Carlo to fill him in on this... not sure
if anything was done about it..
Lets hope that RealPuppet can see the light of day... and get implemented in
RS..
That would not only get your weightpainting, but char anim tools as well if
I understand it correctly..
btw. One thing that maybe is one reason for the built in weightpaint in Rs
to not be so funny, is that it has a bug in it..
it only works for the Parent skel and not subskels...
Best regards,
Matthias
Take Care
Best Regards
Stefan Gustafsson ( Beg-inner )
A Proud Owner and User of Real3D and Realsoft3D..
----- Original Message -----
From: "Beg-inner" <[email protected]>
To: <[email protected]>
Sent: Sunday, October 18, 2009 11:09 PM
Subject: Re: V7 release date yet? Field Evalutator..
Hi Matthias..
Sorry for this long post...=), though even if it was a fast test I did..
I
found a lot which cant be written in a few words..=)
After checking your prjs out, I then made some fast and simple tests on
simple setups..
Here is my first fast findings..
'Total frames' = 100
1. Created an analytic Cube and a Field Evaluator.
2. Multiselected them and in 'Prop' win set so they both were at 0, 0, 0
3. Dropped them to a Level.
4. In 'Prop' Win 'Phys' Tab set Spin X field to 90 (using degrees in
prefs)
5. Select 'Spec' Tab and in 'Select 'Win select the 'Field Evaluator'
obj.
6. Play anim and look at the 'Spin' fields...
Things I see there...
A1. The field that is always showing a value is the not the X field, but
the
Y field.
B1. That value is not 90, but a 3.96 times higher value 356.4 (or -356.4)
(356.4/90=3.96).
C1. I see that the other fields changes to larger values that I cant see,
until I (by chance) stop on the correct frame !
(Those values are -> SpinZ = 35640 (Or -35640) (100 times the SpinY
value)
Angular Acc = (+-)7056720)
C2. At times you stop on a frame/time, and the following values show up.
SpinX = +-17820 SpinY = +-356.4 SpinZ = 17820
AngularAccX = +-1764180 AngularAccY = 0 AngularAccZ = +-5292540,000
C3. The high odd values in SpinZ for example show up very frequently
during
frames 25-75 (Total frames 100 and 25 frames/sec set)
D1. Using Motion Blur with samples and Exposure settings, shows up
different
values but odd high ones..
Some of these odd values can only be caught up, when using MB...
Velocity has similar issue...same number comes up here.. 3.96 =)
1. Using same objects as above, but reset the SpinX 90 value to 0 and set
Velocity Y = 1
2. Play anim and look at FE Velocity Y value (correct field=), but it
says
3.96
A1. Using Motionblur can show up values that you only see flash by..like
Velocity X = +-63,936 (instead of 1, or the above odd 3.96 values)
Acceleration Y = +-159056,409
All is also depending on what the 'Total Frames' is set to !
So if some of this is correct, then its not a wonder that FE fails or
gives
odd behavior for animated objects based on values feeded from FE.
I guess it can be the handling of the Fields..that is failing, which
makes
it possible to workaround it somewhat accessing it via scripting.
btw an odd thing, in V6 the FE didnt show up in the view, for newly
created
FEs.. only for loaded prjs with FEs..
In V7 they show up...
I tested also the ''Gauge" prj.. but my tests didnt show what you
experienced (I guess you mean RS was Non responding state), either in V6
or
V7..
A note on the prjs in the Input folder..
I think some of these dont work well, due to that the Input range is not
completely correct (The chors Start and End values..)
Which makes them snap to fast and looks just wrong...
Ex. "Bend based on acceleration" prj
The chor is set to -10 to 10 but the acceleration is changing alot more.
(checking the FE)
Setting the Chor Start -80 and End 80 gives almost a good looking and
working feel.
Hopefully some of this and yours can make this be pinpointed, found and
fixed so the FE will work as it should, as it has tons of nice use for
anims.
btw your flexing cube there looks cool in its motion when working..!
(I wish all those Modifiers, like the 'Bend' Modifiers actually would
animate the Angle attribute and NOT as 'Points' (they now are animated by
keyframing the Bending endpoint... and so on))
I hope I havent made too many errors in the post or in my testing..=)
Take Care
Best Regards
Stefan Gustafsson ( Beg-inner )
A Proud Owner and User of Real3D and Realsoft3D..
Hi Stefan,
here's a "buggy" scene.
As you can see in the "exposure-time-05.r3d"
the field evaluator is giving different values
(the plate is bending totally in some frames)
The "exposure-time-1.r3d" is working as exspected.
You need the motion blur to see it.
The scene in general is my personal workaround
for the "bend based on acceleration" scene from
the "Realsoft3D\tutorprojects\animation\input" folder.
The acceleration is transferred via "poor mans scripting",
because then it's easier for me to scale the
acceleration <-> bend construct.
Some other simple thing to test the field evaluator
is: "Realsoft3D\tutorprojects\animation\fieldeval\gauge.r3d"
Simply keyframe->translate animate the fieldevaluator.
It's giving an "computing till the end of time" on my machine.
I'll look for other "buggy" scenes.
Best regards,
Matthias
----- Original Message -----
From: "Beg-inner" <[email protected]>
To: <[email protected]>
Sent: Friday, October 16, 2009 3:58 PM
Subject: Re: V7 release date yet? Field Evalutator..
Hi Matthias..
Why didnt you even say a single word about the fade issue ?
Where are any of the problematic Field Evalutator prjs (with an
explanation
of what actually is not working and what is buggy) ?
So I can test them....
Just so I know that I dont look into these matters in vain !
Take Care
Best Regards
Stefan Gustafsson ( Beg-inner )
A Proud Owner and User of Real3D and Realsoft3D..
--------------------------------------------------------------------------------
No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 8.5.420 / Virus Database: 270.14.20/2440 - Release Date: 10/16/09
06:32:00
