PDA

View Full Version : NevronMotion Genoma Preset RIg Video.



lino.grandi
08-21-2013, 03:15 PM
There's a new video available:

LightWave 3D NevronMotion - Using the Genoma Preset Rig (http://www.youtube.com/watch?v=PhcU2V9H1UUhttp://)

Be sure to watch it in HD!

;)

Phil
08-21-2013, 03:49 PM
Link is borked. Should be :

http://www.youtube.com/watch?v=PhcU2V9H1UU

UnCommonGrafx
08-21-2013, 04:26 PM
Lino!!!

Excellent job on the captioning!!

Can we get more like that? Captioned, I mean.

lino.grandi
08-21-2013, 04:32 PM
Link is borked. Should be :

http://www.youtube.com/watch?v=PhcU2V9H1UU

Works here!

LW_Will
08-21-2013, 05:17 PM
Hey, Lino... look at this shot from SIGGRAPH 2013!

116538

WOW! Looks like a happy guy who won a NevronMotion and a Kinect!

Wish he actually HAD the NevronMotion... so I, that is "he" could actually try it and give feedback.

I'm just saying... ;-)

stevecullum
08-21-2013, 05:48 PM
Thanks for the video Lino!

allabulle
08-21-2013, 06:38 PM
Thanks for the video.

Burchigb
08-21-2013, 06:46 PM
Once again excellent detail
Thank you

Phil
08-21-2013, 06:46 PM
Works here!

It shows a trailing http:// on the link here, but YouTube drops it.

allabulle
08-21-2013, 10:02 PM
Worked just right, here.

RebelHill
08-22-2013, 12:43 PM
Hang on...

The feet slip because the proportions dont match...

So... Its not TRUE retargeting?? I was under the impression thats what nevron did, a proper, proportionally compensated retarget.

tcoursey
08-22-2013, 01:54 PM
Lino, Is there plans to have a video like this (great job by the way) for using Nevron Motion with the Kinect Skeleton? What steps would be different, maybe nothing...please let us know. Thanks.

lino.grandi
08-22-2013, 02:26 PM
Lino, Is there plans to have a video like this (great job by the way) for using Nevron Motion with the Kinect Skeleton? What steps would be different, maybe nothing...please let us know. Thanks.

The steps are just the same.
The only thing that changes is the source (which is the Kinect, not a Motion Capture file).
I'll make a video showing how to use the Kinect as our retargeting source anyway, so to show how the Kinect Rig works! ;)

tcoursey
08-22-2013, 02:28 PM
The steps are just the same.
The only thing that changes is the source (which is the Kinect, not a Motion Capture file).
I'll make a video showing how to use the Kinect as our retargeting source anyway, so to show how the Kinect Rig works! ;)

Thanks Lino. I know this seems simple enough for you. But many of us have struggled and if nothing else want to see it performed "flawlessly" right in front of our eyes! :) I know many of us have gotten 70% there, but there seems to always be those little qirks that if you run into them on the video you'll walk us through how to fix. Thanks.

lino.grandi
08-22-2013, 02:50 PM
Nothing is simple! And everything can be. ;) I'll do my best to fullfill any doubt you may have.

lino.grandi
08-22-2013, 02:53 PM
Hey, Lino... look at this shot from SIGGRAPH 2013!

116538

WOW! Looks like a happy guy who won a NevronMotion and a Kinect!

Wish he actually HAD the NevronMotion... so I, that is "he" could actually try it and give feedback.

I'm just saying... ;-)

You still have to receive it???? Let me take a look.... ;)

Can you please send me the ID of your user account to [email protected] ?

lino.grandi
08-22-2013, 02:56 PM
Thanks for the video Lino!


Thanks for the video.


Once again excellent detail
Thank you

Thank you all! Nothing would be possible without your support.

bobakabob
08-22-2013, 04:02 PM
Hang on...

The feet slip because the proportions dont match...

So... Its not TRUE retargeting?? I was under the impression thats what nevron did, a proper, proportionally compensated retarget.

Lino addresses this in the video and explains how to fix it...

RebelHill
08-22-2013, 05:35 PM
Lino addresses this in the video and explains how to fix it...

I know, I did watch it.

My point is that I wouldnt have expected it to NEED fixing. I had presumed retargeting in nevron was just like retargeting everywhere else, where character proportions are automatically accounted for (and you have a SOLID retarget in one click, none of this "fix it" extra workload). I also find myself wondering how well this fixing is gonna work out when you really do get VERY different proportions between your target character and the source motion...

Hmmm

bobakabob
08-23-2013, 02:41 AM
It would be interesting to see a Nevron tutorial rigging a toon character with wildly extreme proportions e.g. tiny legs / long body a la Deputy Dawg style designs. Is it possible or would it throw out animated walking and result in foot sliding issues etc? Lino, any more info on the way?

lino.grandi
08-23-2013, 03:41 AM
It would be interesting to see a Nevron tutorial rigging a toon character with wildly extreme proportions e.g. tiny legs / long body a la Deputy Dawg style designs. Is it possible or would it throw out animated walking and result in foot sliding issues etc? Lino, any more info on the way?

I'll make a video showing how there's no problem at all using characters with wildly extreme proportions.
Some days ago I also put in a little change in how the IK solver is used to solve reatargeting that should make the workflow even better. It will be available in the next NevronMotion update. ;)

Spinland
08-23-2013, 06:41 AM
It would be interesting to see a Nevron tutorial rigging a toon character with wildly extreme proportions e.g. tiny legs / long body a la Deputy Dawg style designs. Is it possible or would it throw out animated walking and result in foot sliding issues etc? Lino, any more info on the way?

Lino alluded to some of this in his response to my questions in this thread:

http://forums.newtek.com/showthread.php?137238-Nevron-Genoma-preset-rig-couple-of-questions

RebelHill
08-23-2013, 07:15 AM
Ok... so it seems its operating off of a kind of "space match"... rather than calculating for the proportion adjustment, you're essentially "pre-processing" the mocap by scaling it to match the targets hip>ankle distance. So that'll give you stride scaling, no problem, as well as enable correct matching of hip>ankle movements... What I wonder further then is how this will play out for other parts of the hierarchy when there are major differences. For instance, what if a target character has great big clown feet (whilst obv the source has normal feet). In a situation like that then simply matching up the ankles wont be of much use, as the animation below the ankle becomes important to match for correct foot action and placement.

I also wonder what would happen in the case of arms... If you had some IK style action going on with the arms as well (lets say like a cartwheel motion), where both arms and legs needed to stick properly, then simply stretching out the scale of the source would spoil things Id be imagining.

Hmmm... Im gonna have to do some experiments here.

lino.grandi
08-23-2013, 07:39 AM
Ok... so it seems its operating off of a kind of "space match"... rather than calculating for the proportion adjustment, you're essentially "pre-processing" the mocap by scaling it to match the targets hip>ankle distance. So that'll give you stride scaling, no problem, as well as enable correct matching of hip>ankle movements... What I wonder further then is how this will play out for other parts of the hierarchy when there are major differences. For instance, what if a target character has great big clown feet (whilst obv the source has normal feet). In a situation like that then simply matching up the ankles wont be of much use, as the animation below the ankle becomes important to match for correct foot action and placement.

I also wonder what would happen in the case of arms... If you had some IK style action going on with the arms as well (lets say like a cartwheel motion), where both arms and legs needed to stick properly, then simply stretching out the scale of the source would spoil things Id be imagining.

Hmmm... Im gonna have to do some experiments here.

Thank you. We've been doing all the needed experiments already, and verified that NevronMotion works well with all the situations you're mentioning. ;)

But, you know, "repetita juvant".

RebelHill
08-23-2013, 08:20 AM
you know, "repetita juvant".

That it does... couldnt agree more. But I think the experiments I have in mind are gonna attack this from a different angle... When/if I actually find time to get around to them ofc.

Ryan Roye
08-23-2013, 11:53 AM
Since Nevron runs right in Lightwave couldn't one just create Python/Lscripts to further automate some processes to their preferences? For instance, many batches of mocap typically only use one rig, so a script could make assumptions about how the mocap should be fitted to any given character to save time. Just throwing ideas out there.

stevecullum
08-23-2013, 11:57 AM
I'll make a video showing how there's no problem at all using characters with wildly extreme proportions.

That would be cool. I'd like to see how it might work with a character with elongated legs and tiny arms. Like a tyranasoarus...

LW_Will
08-23-2013, 05:21 PM
You still have to receive it???? Let me take a look.... ;)

Thanks, Lino. ;-)

lino.grandi
08-24-2013, 01:07 AM
Thanks, Lino. ;-)

Sure!

snsmoore
09-16-2013, 09:39 PM
The steps are just the same.
The only thing that changes is the source (which is the Kinect, not a Motion Capture file).
I'll make a video showing how to use the Kinect as our retargeting source anyway, so to show how the Kinect Rig works! ;)

Lino,

Any progress on this 2nd video showing the Kinect as the retargeting source? (or do need to wait for the next release?)

I came close but am getting an unusual deformation. (see this thread: http://forums.newtek.com/showthread.php?137330-Active-Kinect-retargeting-using-the-Nevron-Genoma-preset-95-there )

Would there be any differences with disproportionate characters, like a large body with small arms. (i.e. the cube from the Genoma tutorial in the documentation)?

-shawn

Ryan Roye
12-19-2013, 06:54 PM
Two things I feel should have been mentioned more clearly in the video:

1) It probably should have been explicitly mentioned how the controllers were meant to be controlled. The video implies, but does not really explain that they are primarily meant to be manipulated in pairs (IE: Forearm/ForearmRoll should both be selected to get desired arm movement). If another "introduction to nevron" video is made, be certain to mention this if only briefly. People who animate all the time in Lightwave can figure it out, but less experienced users may become frustrated.

2) It should be explained that the Nevron controllers (magenta/yellow) represent the original position of the character motion after retargeting, which essentially tells user the difference between the motion capture and any corrections/adjustments that have been made. I can't speak for others, but I was initially confused as to why the controllers did not move with the character before realizing the reason for it.