View Full Version : RayFire Tool


Pages : 1 2 [3] 4 5 6 7 8 9 10

PsychoSilence
07 July 2008, 02:37 PM
Is love to see the pieces of that window slowley entering the reactor sim as the teapot pushes through, rather than all the bits falling the frame of impact... hehe! I only say that cos im trying to do that effect at the moment. BTW ray fire is pretty cool!!! just playing around with the demo now!

actually this is using the PhysX solver which enables you to have veeery tiny pieces...reactor would quit when pieces get too small! No real size limitation with PhysX unless its only one face or so...

@ Steve: i tested PWrapper etc with the Blood feature/SuperSpray in RayFire. Works like a charm :) We might get rid of the Blood option for now though because no one ever seemed to use it. PLUS SuperSpray can be SLOOOOW when u have 100s of impacts...We might add it in again with some proper solution just like the FumeFX and Afterburn features that are dropped for now too. We donīt wanna offer half baked features... :arteest:

Steve Green
07 July 2008, 02:56 PM
"We donīt wanna offer half baked features..."

No, that's Autodesk's job... ;)

Seriously, though - I think the most general useful thing would be to have automation of particles spawning from general demolition.

I'm not sure what's the best solution, whether to use position object and selected faces on the interiors of the fragments?

I'm not sure how well that would work on a slow demolition though - it's a case of finding whats the best way to trigger particle birth at a per fragment level, so as ia frag moves, you get a chunk of particles falling off, possibly linking the birth rate to velocity initially?

So you'd get a trickle when it moves slowly, then a bit more as it accelerates and you get the initial demolition.

I guess the birth rate should then be capped or slowly die off, but maybe spikes when a fragments velocity increases dramatically (so when a frag hits the ground it kicks up a lot more debris.)

I guess all this particle stuff would only need to collide with the floor and maybe some proxy geometry (like other buildings) and maybe use some wind turbulence to get some variety going.

Just rambling now... :)

- Steve

mmoses00
07 July 2008, 03:57 PM
What is the timeframe for this new release?

I use a Quadro FX 4600... DO I get any advantage/disadvantage using this Physx engine with this card?

THx....

PsychoSilence
07 July 2008, 04:25 PM
Good news first: Blood stays in for now :)

the current built for testers uses a PhysX engine that is not yet released and was given to us to test stuff in advance so basicly we have to wait until NVidia releases that built officially and then we go public with RF1.4

For now u can check out the downloadable built of PhysX at: http://sourceforge.net/projects/physxplugin there you should find any kinds of specifications.

ACiD80
07 July 2008, 06:23 PM
For now u can check out the downloadable built of PhysX at: http://sourceforge.net/projects/physxplugin there you should find any kinds of specifications.

No x64 version :cry:
or did i overlook something?

mir-vadim
07 July 2008, 06:52 PM
You right.
For now only 32 bit version.

Steve Green
07 July 2008, 08:03 PM
That's a pain, when you consider heavy duty particle/physics scenes would be an ideal candidate for the extra memory of 64bit.

- Steve

SoLiTuDe
07 July 2008, 08:13 PM
^ Yup... ageia / nvidia is working on it i'm sure... but given when they started there's really no reason it was 64 bit from the get go.

JohnnyRandom
07 July 2008, 08:53 PM
What is the timeframe for this new release?

I use a Quadro FX 4600... DO I get any advantage/disadvantage using this Physx engine with this card?

THx....

AFAIK No, 260, 280, 9800 gtx, maybe that will change, but that is what nvidia specs say.

Glacierise
07 July 2008, 07:10 AM
Hey Steve, on your question about the slower demolition - there's actually a much easier solution that I use all the time. The trick is to position particles, and then spawn by travel distance from them. So they don't spawn at all when they don't move, and they start spawning when it starts crumbling. Try that, it works great for me. Otherwise - I would like to have just two particle systems - one for rayfire, that just creates a pflow event whenever a projectile hits. I can easily build all kinds of effects afterwards. And a second system for fragmentations, that places particles on the inside faces.

Steve Green
07 July 2008, 06:33 PM
Hi,

that works really well - simple but effective! Thanks for the tip.

Cheers

Steve

circusboy
07 July 2008, 07:58 PM
FumeFX and Afterburn features that are dropped for now too. We donīt wanna offer half baked features... :arteest:
Hmm-I was just getting excited about the FumeFX intergration the overview demo suggested was there. So is it out completely?
I was also curious about the muzzle flash (FumeFX) possibilities in the mix-if/when thats production ready.

floopyb
07 July 2008, 02:16 AM
Hi, We just purchased two copies of rayfire for a project we are working on at the moment (well the boss is pressing the purchase button now!). Great tool! but Im using it in demo mode now and im trying to use the Instant action feature on a simple champher box and I can do the first bullet hit (Reactor Debris - predefined, sparks, debris, smoke, bullets and impact flash) but when I choose the scond one I get a "Unable to convert undefined to type: Point3".
I can sometimes cancel and reload and get a second hit in but its pretty buggy. The other thing is that when I do an instant action hit and then delete the layer the original box is not restored. It stays there with the hole in it.
I think this is an important feature to get right as directing timing and location of the hits is VERY important in production.
Cheers,

Jordan Walsh

PsychoSilence
07 July 2008, 02:39 AM
I would like to have just two particle systems - one for rayfire, that just creates a pflow event whenever a projectile hits. I can easily build all kinds of effects afterwards. And a second system for fragmentations, that places particles on the inside faces.

Problem might be if u have like a hundred hits to deal with because u have to simulate shotgun shooting with a faily low shooting frequency but like 50 small hits at once...50 pflow systems is nothing i would want to deal with...especially when its all unique systems so that pflow creates 50 particle engines...that just must go wrong. meybe u could automatily delete every particle engine after the first but i dont know...just sonds all messy.

for the inside thing thats an easy one to achieve yourself :) just declare a pure white/black material in slot 2 of a multi sub as your fracturing/surface material and use a position object by grayscale using submat 2 or N.Should be slow with 2000 fractures but defintiely should work :)

Glacierise
07 July 2008, 02:59 AM
Hi Anselm, thanks for that tip ;)

On the pflow thingie - I didn't mean create a new pflow for every hit, that'd be quite useless. I meant - one particle per hit, or more like 2 particles - one on the surface, and one bounced off. I think all effects can be derived from that.

edit: Another and maybe more practical approach could be a creation of a pflow with the bullets in it - we can easily collide them with the geometry manually, and create spawns from that.

PsychoSilence
07 July 2008, 04:45 AM
The particle request would go too much under the hood imo...RayFire is ment to be an easy to use application by adding objects and using the built in tools without any external knownledge about PhysX,Reactor,Particles & Fracturing. Thatīs what parts it from other products on the market where itīs almost impossible to create something that quick without deeper knownledge of whatīs happening "under the hood" and/or scripting background or particle knownledge.

The quickest and easiest workaround for shooting would be to set up a particle system that "hits" the obstacle. Then scrub thru your animation and use the Instant Action feature on every frame your particle is supposed to put a hole in the wall or what ever. Quickly tested it and it works just fine that way :) That would be the quickest way to combine both worlds as we donīt plan to get everybody trained on Particle Flow or TP here :D

See my 5 minutes solution to your request:

I simply aligned and linked the Find Target to the Gun Target and the Particle Flow Icon to the Gun Plane. So eveything follows nicely to what ever u animated in RayFire with a little more tweaking...

http://www.3delicious.de/spielwiese/PFlow_As_Gun.gif

We might add that scene as a workaround example to the installation or the website support once 1.4 is out ;)

EDIT: You could even set up a particle system of your choice and grab it as Custom Object to shot with and now every time a shot hits the obstacle you have that single spawned particle u wanted ;)

http://www.3delicious.de/spielwiese/PFlow_As_Custom.gif

with kindest regards,
Anselm

Glacierise
07 July 2008, 07:57 AM
Awesome and comprehensive, thanks Anselm!

mir-vadim
07 July 2008, 09:57 AM
Hi, We just purchased two copies of rayfire for a project we are working on at the moment (well the boss is pressing the purchase button now!). Great tool! but Im using it in demo mode now and im trying to use the Instant action feature on a simple champher box and I can do the first bullet hit (Reactor Debris - predefined, sparks, debris, smoke, bullets and impact flash) but when I choose the scond one I get a "Unable to convert undefined to type: Point3".
I can sometimes cancel and reload and get a second hit in but its pretty buggy. The other thing is that when I do an instant action hit and then delete the layer the original box is not restored. It stays there with the hole in it.
I think this is an important feature to get right as directing timing and location of the hits is VERY important in production.
Cheers,

Jordan Walsh

Hi, floopyb. Thanks for pointing me, in last updates I mainly improved fragmentation and dynamic simulation features and completely forgot about shooting stuff and Instant Action.
I just tried Instanct Action and all works fine, so problem might be in Your scene.
Please send it to me for tests, and I will try to find this problem and fix it.

Mills
07 July 2008, 10:45 AM
Hi
Sorry if this has been posted. What is the time line for the release of RF1.4.?
Would like to use the physics on a upcoming job and I don't want to propose the concept without knowing when its going to be available.
Thanks
M

mir-vadim
07 July 2008, 11:49 AM
Approximately 2 weeks.

JohnnyRandom
07 July 2008, 03:57 PM
Glacierise, Jay Haywood was building a PFlow based shoot'um-script called WinchesterFX (http://forums.cgsociety.org/showthread.php?f=183&t=451701&highlight=winchester) a while back, the problems encountered using Pflow generally had to do with the creation of the flow and the wiring, basically you set up the scene run the script and you had to manually debug the flow, by rewiring events, ect. but it works really pretty well. He stopped development:sad:

Way cool script though, I suppose you could integrate it with Rayfire. :)

http://forums.cgsociety.org/showthread.php?f=183&t=451701&highlight=winchester

floopyb
07 July 2008, 07:21 AM
First of all let me say how good the fracture options are!!
I would love to see the ability to use a bunch of selected objects as the fracture objects, as I am trying to do some wood and need long thin splinters rather than chunks.
Also Fragment relative to mouse doesnt seem to work for me. I add a box to the impact objects, go to RaXplosion tab, choose Relative to mouse and hit and press 'Fragment Impact Objects' but it just does a uniform style fragmentation.
Cheers!

depleteD
07 July 2008, 08:22 AM
Man I gotta say, I jsut started using ray fire. This thing is saveing my ass man. These fragmentation tools are ****ing incredible.

Gonna really start taking these things through the ringer. Its so awesome for TP set ups.

mir-vadim
07 July 2008, 08:34 AM
I would love to see the ability to use a bunch of selected objects as the fracture objects, as I am trying to do some wood and need long thin splinters rather than chunks.
Yeah, this feature already in my feature list. But for now I spend all time for PhysX integration in 1.4.

Also Fragment relative to mouse doesnt seem to work for me. I add a box to the impact objects, go to RaXplosion tab, choose Relative to mouse and hit and press 'Fragment Impact Objects' but it just does a uniform style fragmentation.
Cheers!

RayFire fragment The BIGGEST object from ALL fragments UNDER cursor. But if You will move mouse cursor far from fragments it will start fragment the bigggest object from ALL fragments, so You can quickly fragment object where You need and then move it to some distance just to let RayFire to fragment the biggest objects.


Thanks, depleteD. :)

floopyb
07 July 2008, 09:09 AM
Yeah, this feature already in my feature list. But for now I spend all time for PhysX integration in 1.4.

Great!! Looking forward to the PhysX stuff!



RayFire fragment The BIGGEST object from ALL fragments UNDER cursor. But if You will move mouse cursor far from fragments it will start fragment the bigggest object from ALL fragments, so You can quickly fragment object where You need and then move it to some distance just to let RayFire to fragment the biggest objects.


I have seen the videos of how it should work, but i leave the mouse over the lower right corner of the box the whole time and it makes no difference!

mir-vadim
07 July 2008, 09:20 AM
How much iterations You use?
You have to fragment objects a little bit before RayFire will able to start fragment objects under cursor.
Also try to fragment some thin objects, just to be sure that this fragmentation type works.
If it will not help You, send me scene, please, will test it to find the problem.

floopyb
07 July 2008, 03:51 AM
im using about 60-120 iterations.
I was testing it by opening max, creating a box say 100x100x3 then trying to fragment. No mouse fragmentation!

PsychoSilence
07 July 2008, 05:19 AM
im using about 60-120 iterations.
I was testing it by opening max, creating a box say 100x100x3 then trying to fragment. No mouse fragmentation!

it looks like rayfire finishes the fragmentation faster then the wanted manual operation...for fragmentation by mouse u should crank up both amount spinners to the max and quit the operation when ever u reach the desired detail level. try 2000 and 300 and let it run and/or quit when ever you feel like.

floopyb
07 July 2008, 07:15 AM
hmmm, i tried that too, was sitting there for a min or so with it still doing nothing!

mir-vadim
07 July 2008, 11:51 AM
hm, ok, I will test fragmentation feature precisely.

Also I created few video tutorials dedicated to PhysX engine in RayFire:

1. PhysX - Rigid Body support. 12 Mb. (http://www.mirvadim.com/Video/tut/RayFire14_PhysX_RigidBody.wmv)
2. PhysX - Influence rollout. 45 Mb. (http://www.mirvadim.com/Video/tut/RayFire14_PhysX_Influence.wmv)
3. PhysX - RayFire Bombs. 28 Mb. (http://www.mirvadim.com/Video/tut/RayFire14_PhysX_RFBombs.wmv)

Will try to create more tutorials soon. With something more complicated.

wreath
07 July 2008, 12:05 PM
Just amazing! especially that wind controlled realtime fracturing blowed me away now Max have über powerful dynamics as Houdini. :bounce:

Steve Green
07 July 2008, 12:27 PM
Very cool - I saw you changing the gravity settings in realtime - are they keyable?

I could imagine some great effects where gravity is suddenly restored and things crash back to earth.

Regarding the splinters post someone made - I think what would be good is if you could bias the fragmentation depending on the dimensions of the object's faces - so you could optionally have a fairly flat box fragmented along the length (like plaster on a wall)

I've no idea if there's any way to detect this kind of thing and instruct ProBoolean/cutter to work like this - I guess simple boxes are easy, but a more complex object would be tough.

Anyway, just a thought - still marveling at the PhysX implementation.

Cheers

Steve

mir-vadim
07 July 2008, 12:35 PM
RayFire Ui gravity property is not keyable, but You can use animated wind or gravity space warp instead, though I noticed that PhysX own gravity works better.
So I will add later special object (like RF_Bomb) with some PhysX properties which You will be able to animate.

Steve Green
07 July 2008, 01:44 PM
Cool - thanks Mir!

Cheers,

Steve

Steve Green
07 July 2008, 04:22 PM
On a related topic, does the drag spacewarp work with PhysX currently?

Also is there anything in PhysX similar to the water collection?

I've not tried it but presumably using Reactor you could have an object crash into water and the fragments bob about (or slowly sink depending on the settings)

I was thinking if the first works, it would go some way to simulating at least heavy stuff sinking into water, but meeting more resistance.

Thanks,

Steve

mir-vadim
07 July 2008, 04:37 PM
No, there is no water in PhysX at this moment.
And I did not add support for drag force yet, but it will be added later.
Also I want to add support for vortex space warp.

mir-vadim
07 July 2008, 05:32 PM
heh, just quickly added support for drag force.
Only for unlimited range and Linerar damping.
I hope it will be enough for a first time.

Steve Green
07 July 2008, 07:20 PM
Cool, thanks - yeah I think a falloff drag might be useful if you wanted pieces to just drag when they hit the water... pretty specific request I know.

Cheers

Steve

PsychoSilence
07 July 2008, 03:39 AM
Mir never sleeps and works at the speed of light! I just wonder which studio is going to grab em first :)

mir-vadim
07 July 2008, 08:46 AM
I hope it will be studio Private, or Vivid at least. :drool:
I already sent them my demoreel. :)

Glacierise
07 July 2008, 09:28 AM
You'll be shattering the ladies then? Nice plan! That industry will really suck you dry though :D

Steve Green
07 July 2008, 04:33 PM
Mumbles something about soft and rigid body dynamics... and fluids obviously.

- Steve

JohnnyRandom
07 July 2008, 06:49 PM
I hope it will be studio Private, or Vivid at least. :drool:
I already sent them my demoreel. :)

Just don't leave us hanging:D

Whoever it is will be stoked!

Really excited to try that PhysX implementation :)

ddustin
07 July 2008, 08:34 PM
OK we just purchased the Rayfire plugin.

1. The install file does not install to any version except for 2008.
2. We tried to run the Buy now thingy and used paypal, but there was no place to enter the generated code.
3. How do we install this for max 2009?
4. There is really no 64 bit version, yuk.

I sent an email, hopefully they will get back to me soon (anxious to play)..

David

PsychoSilence
07 July 2008, 08:56 PM
you have to point to where ever your max is installed manually then if youīre running another max and or it is instaled to a different location. i installed it to for max9 and it works just fine here.

the reason for non-64 bit atm is mainly because there is no PhysX SDK for 64bit yet! So nVidia is the right place to ask for 64 bit support atm. Not sure about the plugin tho.

cheers
Anselm

ddustin
07 July 2008, 09:13 PM
Thanks for the tip.

I had tried that but mistakenly tried to install it in the 64 bit directory (duh).

Hopefully mir-vadim can get the code to me soon.

It seems to generate a different code for each version (max 2008 and 2009).
Does that mean we need to register it for each version?

Thanks,

David

JohnnyRandom
07 July 2008, 10:50 PM
Did I miss something?

Version 1.34 works in 64 bit, 9, 2008, and 2009.

Ansi you talking about 1.4 right?

mir-vadim
07 July 2008, 11:00 PM
Yes, he talked about 1.4.
1.34 works on 32 and 64 bits.

PexElroy
07 July 2008, 12:19 AM
PhysX implementation is fantastic, and a :wise: move, it is far more robust than reactor. Good use of wind, bombs and soon vortex. Looking forward to 1.4

ddustin
07 July 2008, 02:07 AM
OK so if we create it in 32 bit, can we render in a 64 bit environment (we would need to be able to open it in 64 bit).

My render farm is all 64 bit.

What is the latest estimate as to the arrival of V1.4?

Thanks,
David

mir-vadim
07 July 2008, 08:37 AM
Yes, You can.

About 1.4 release. The Tool is ready, all I need for now is released for public Nvidia PhysX plugin 1.0.4, which RayFire uses for now, and EULA for PhysX in RayFire, which NVidia will give me soon.

ddustin
07 July 2008, 04:42 PM
Well I am very impressed.

This is an excellent tool (very addicting I might add).
The tutorials you have made are very helpfull as well.

David

ddustin
08 August 2008, 01:04 AM
I am having a problem with the Rayfire plugin.

The example in the video posted by Mir shows the introductory setup of a scene, where there are boxes and debris objects used.
All is good except my debris is scattered at impact, then hangs in the air, failing to fall to the ground.

I have looked over any setting I can think of and can't determine what the problem is.
Here is an image of the debris hanging in the air.

Thanks for any thoughts.
David
http://www.dustinproductions.com/wip/objects.jpg

PsychoSilence
08 August 2008, 01:38 AM
my first guess is a wrong scene scale. maybe post the file, max version, rayfire version for us :)

ddustin
08 August 2008, 02:13 AM
Max 2009, Build 13402.
File is here http://www.dustinproductions.com/wip/RF03.zip

Thanks for looking.
David

ddustin
08 August 2008, 01:07 PM
Is anyone using rayfire with a render farm?

Do you install just the latest version from the website or is there a different build download with the latest build?

When you use the software, you can update from inside it.
How do render nodes update?

Thanks,
David Dustin

mir-vadim
08 August 2008, 01:42 PM
You don't have to update RayFire on render nodes at all.
In this update I did not modify RF_Bombs class object anyhow, so, don't wary, just install it.

Also I tested Your scene and descried that Your reactor debris are very small for reactor engine.
It just can not simulate such small objects, actually reactor very constrained engine.
I created reactor debris in case users will want to simulate big debris from holes, for small debris better to use particle debris.
But PhysX engine in 1.4 doesn't care about objects size at all.
So in 1.4 this problem will be fixed.

ddustin
08 August 2008, 02:03 PM
Thanks for the quick reply.


David

floopyb
08 August 2008, 04:40 AM
hm, ok, I will test fragmentation feature precisely.



I just found out that it is one of our scripts' callbacks that is interupting the mouse calls in rayFire, so when I disabled our script the mouse fragmenting worked fine! Son it was not a RF problem!! :)

wreath
08 August 2008, 07:29 PM
I'm trying to explode 1 impact object and nothing happens, but when i fracture same object then it works, is that any way to explode unfractured raw objects?

wreath
08 August 2008, 07:37 PM
I'm trying to explode 1 impact object and nothing happens, but when i fracture same object then it works, is that any way to explode unfractured raw objects?

Hmm some weird things happening, i added a sphere to the impact object list and explosion worked but when i remove that sphere, explosion not working with box object is that some kind of bug?

mir-vadim
08 August 2008, 08:13 PM
No, this is not a bug. Problem is that You just can't explode one object. You have to define more then one impact object. After fragmentation RayFire automatically adds fragments in Impact object list, that is why explosion works after fragmentation.
Anyway. this is an obsolete restriction, thanks for pointing me on it. I will remove it righ now.

wreath
08 August 2008, 08:26 PM
Cool! RF is always getting better, i want to ask another thing, can't set explosions start time before frame # 6 is that any specific reason for that?

mir-vadim
08 August 2008, 09:17 PM
Heh, this is what I call fresh eye. Removed. :)

ddustin
08 August 2008, 09:40 PM
Try as I may, I can't get my project to work like your rayfire demolition example.

My guess is there is something different in the scale, but without something to go by, we don't know what the correct scale(s) are.

Might I suggest posting example projects that are setup to where we can run them and see how the settings are?

(still love the product!!)

David

wreath
08 August 2008, 02:24 AM
Try as I may, I can't get my project to work like your rayfire demolition example.

My guess is there is something different in the scale, but without something to go by, we don't know what the correct scale(s) are.

Might I suggest posting example projects that are setup to where we can run them and see how the settings are?

(still love the product!!)

David

David if you mean many intersections / strange looking movements, yes it's probably because of object scales as you can see we have limited value in the reactor tab on RF panel it's 0.5 cm so if you playing objects that have smaller dimensions than this value they may start penetrating each other, i really don't know is that any simple way to solve this without scaling up but i hope.

edit:
You can check these files (http://www.yeatfx.com/small_big_max_2008.rar)

mir-vadim
08 August 2008, 08:46 AM
Thanks wreath, :)
Yes, reactor hates small objects,
It just can't simulate correctly objects less then about 5 cm (2 inches).
So, just go to Customize\Unit Setups, activate Metric Display Unit scale and set sentimeters there. Then create simple box near Your simulation object to compare their sizes. Also You can use taper of course.

Another problem which I notified is that reactor can't simulate correctly small objects with mass less then about 1 kilo, in this case object might freeze in the air and reactor says something like "Object is too small. Removing from simulation". So if You use "Set mass by material" RayFire will apply mass to objects accordingly their material and volume. But since it is useless to apply mass less then 1 kilo, RayFire will apply default 1.4 kilo for all small objects, and because of this You can get really small objects (1*1*1 cm) with gigantic mass for theit volume. This is another reason not to use small objects.

Ok, What is the size of Your simulation objects?

ddustin
08 August 2008, 12:54 PM
Thanks wreath, :)
Another problem which I notified is that reactor can't simulate correctly small objects with mass less then about 1 kilo, in this case object might freeze in the air and reactor says something like "Object is too small. Removing from simulation". So if You use "Set mass by material" RayFire will apply mass to objects accordingly their material and volume. But since it is useless to apply mass less then 1 kilo, RayFire will apply default 1.4 kilo for all small objects, and because of this You can get really small objects (1*1*1 cm) with gigantic mass for theit volume. This is another reason not to use small objects.

Ok, What is the size of Your simulation objects?

1. My simulation objects are 1000 cm square.
2. Project file has been emailed to you
3. My floor box has been set to as dense as I can make it in max (200 divisions x 200 divisions) yet my objects fall through (collision tollerance has been set to .5).
4. I converted my project to metric. Not to "americanize" this disscusion, but we work a lot with accident reconstructionists who work almost exclusively in Decimal (sometimes fractional) inches. Is there a good guideline for setting up objects?

Best regards,
David

mir-vadim
08 August 2008, 01:42 PM
Well, first of all.
1. Reactor does not care about Your ground object mesh density, in case You did not use Concave simulation geometry. It cooks it's own optimized geometry for simulation.
2. Your Gravity increased in 5 times. This is might be the reason why fragments fall through the ground. I decreased it to default 1.0 value.
3. 0.5 cm collision tolerance too low for 1000 cm falling objects, You can easily increase it up to 5 cm.
4. Scale factor 10 is too big value, keep it in range between 0.5 to 3.

I emailed scene with fixed properties and created demolition to You.
As You will see all works fine.

ddustin
08 August 2008, 02:00 PM
Mir,

Yes it works fine now. Thanks for the tips and fixing the file.

David

JohnnyRandom
08 August 2008, 02:24 AM
Dude, you tease, 1.4 help file! :D

I take it you found the page template?

PsychoSilence
08 August 2008, 11:51 AM
Dude, you tease, 1.4 help file! :D

I take it you found the page template?

we so gonna deliver next week :beer: itīs Siggi Zig Zag Siggraph time, bwwwwoooooy!

mir-vadim
08 August 2008, 02:34 PM
Yes, I finally got EULA from Nvidia and now free to release new version.
RayFire Tool 1.4 will be released on Monday (11 August).
Nvidia PhysX engine support will be only in 32bit version.
There will be Demo version with few restrictions for public.
It will cost 255$.
Also there will be charge in 45$ for 1.31, 1.32, and 1.33 customers.
1.34 customers will get it for free, because they did not get yet any free new versions as other customers, I think it will be fair.

Also I want to mention that this is only beggining.
I added just the basic PhysX Rigid Body support in RayFire.
Later there will be more cool stuff, like cloth, soft bodies and metals.

JohnnyRandom
08 August 2008, 03:21 PM
Absolutely awesome news.

Fracturing non-deformables is fun, but are you going a step further?

Soft body/metals are you saying this can be deformable, like when say an oil can explodes and banana peels the metal? Or am I just getting overly excited?

mir-vadim
08 August 2008, 03:33 PM
Heh, for now You are over exited. :)
I can't say anything for sure right now, but I have a lot of ideas,
which I will try to implement step by step in near future.

Glacierise
08 August 2008, 04:09 PM
There is a recent demo of the metal cloth feature here:

http://www.youtube.com/watch?v=yvUSaNiiYo0

Check it out ;)

JohnnyRandom
08 August 2008, 04:42 PM
Hah, Vadim, still can't wait to try it out:)

recent demo That is just sweet.

More improved version of what I saw on the TP3 deformable video. Wondering what they used? if it was some in-house stuff...didn't see anything but the vids on there sight.

Glacierise
08 August 2008, 05:35 PM
That one is from a game engine, but it's the physx solver, so it's coming to max at some point.

LordHar
08 August 2008, 05:51 AM
Quote:
"Nvidia PhysX engine support will be only in 32bit version."

Will there be support for 64bit systems in the near future?

Glacierise
08 August 2008, 06:26 AM
There is no 64bit version because there's no 64bit PhysX for now. NVIDIA should make the 64bit at some point, but they haven't disclosed when.

gongjinhu
08 August 2008, 07:26 AM
hello mir ,sorry for my poor english! where can i find rayfire tool V1.4 .purchase
i only find "Purchase RayFire Tool 1.34 " one your website! i need 1.4

PsychoSilence
08 August 2008, 07:37 AM
hello mir ,sorry for my poor english! where can i find rayfire tool V1.4 .purchase
i only find "Purchase RayFire Tool 1.34 " one your website! i need 1.4

Dear gongjinhu,

RayFire Tool 1.4 will be released on Monday (11 August).
Nvidia PhysX engine support will be only in 32bit version.
There will be Demo version with few restrictions for public.
It will cost 255$.
Also there will be charge in 45$ for 1.31, 1.32, and 1.33 customers.
1.34 customers will get it for free, because they did not get yet any free new versions as other customers. (quoted from above)

mir-vadim
08 August 2008, 11:16 AM
Ok, I updated website, so You can take a look whats new in 1.4
Also You can purchase 1.4, but You will get it from monday on only!
.

mmoses00
08 August 2008, 03:48 PM
Good stuff Mir!

How will you be handling the upgrade customers' purchase?
On the site, I only see the full license available to purchase.

thx,

mir-vadim
08 August 2008, 04:14 PM
Heh,
Matt, You was first RF customer and as I guess You will upgrade Your license first. :)
Check out http://mirvadim.com/buy.html again.

ddustin
08 August 2008, 05:46 PM
So those of us with 1.34 will be given a download link or do we need to download the demo?

Thanks,
David

mir-vadim
08 August 2008, 05:52 PM
You send me request email and I send You full version.

PexElroy
08 August 2008, 07:13 PM
Looking forward to monday;

ps - could you (in time) allow the useful Demolition tools, and other Tabs, to support PhysX ? :wip:

mir-vadim
08 August 2008, 09:04 PM
Sure, for now PhysX support added only for Rigid Body simulation.
Also You can use PhysX for demolitions, but first animation pass (until first impact) in any case will create reactor to store collision info, because PhysX plugin unable to give such info for now.

Also I didn't add PhysX support for shooting, because I am planning to rewrite all this part to make it much more flexible for PhysX dynamic simulation. Adding support in current shooting state will be time wasting.

mmoses00
08 August 2008, 02:35 AM
THx Mir.... I hope to use this very often.


I have superheroes that need to break buildings and things!

I'll try and work with the clients to get rights for you to showcase the stuff.

JohnnyRandom
08 August 2008, 09:15 PM
Dude, its Monday in Kaliningrad!

Sorry couldn't resist...

PsychoSilence
08 August 2008, 10:01 PM
same in germany :D

23:52 right now actually so close to monday...canīt sleep :(

JohnnyRandom
08 August 2008, 10:07 PM
Got to used to staying up with everyone visiting?:)

wreath
08 August 2008, 10:59 PM
Hey i'm wondering you x64 guys have a plan to use physX stuff ? like simulate rbd's on x86 than merge in x64 scene workflow, any ideas?

ddustin
08 August 2008, 12:02 AM
Yeah that is what I plan on doing.
Mir says it should be seamless to take it into x64 and he's NEVER wrong..

David

wreath
08 August 2008, 12:54 AM
Wow 1.4 is online! :bounce:

mir-vadim
08 August 2008, 01:16 AM
Yeah, it is released finally. :)
Amazing, 40 days ago I didn't even dreamed about such abilities in RayFire, which it provides now.
So, all 1.34 customers, just send me request email, and I will send You full 1.4 version with new serial ID.

wreath
08 August 2008, 02:50 AM
I just started using 1.4, new physX solver integration is A-MA-ZING!!! :bowdown: you da man Mir!

JohnnyRandom
08 August 2008, 02:58 AM
That is almost an understatement:) PhysX is the Bomb:D

Glacierise
08 August 2008, 07:27 AM
What I absolutely love about RF&Physx is that I am able to adjust parameters in real time WHILE THE SIM IS PLAYING and it reacts, so no resimming, it's such a breeze!

ddustin
08 August 2008, 10:25 AM
Any idea when on Monday we will get the update to 1.4?

(Sorry, very anxious).
David

PexElroy
08 August 2008, 11:05 AM
This is great to have the PhysX for rigids so far. useful and fun :cool: to just mess with ideas fast and quickly, in real-time.

Cool stuff, but noticed I get smooth animation of the sim in the viewports, but when time to bake what I like, the sim generates many detailed frames which are not being baked per frame [skipped], and small motions are lost; how could we go about slow motion baking?

mir-vadim
08 August 2008, 11:20 AM
It is because You use substeps value more then 1.
Substeps property define how many substep PhysX will do between two frames.
If You will turn off "Update viewports" in Option menu, You will decrease simulation time, but max will update viewports only on each frame.
That is why it looks smooth when You create it, this is just animation between frames.

I am planning to add some kind of "Time warp" property in next build tho.

Glacierise
08 August 2008, 11:31 AM
Dude if you just can make a timewarp curve, like the one in Bobo's maxtrix, or in CAT, that will be AWESOME. That's a spline in the curve editor, and it's slope defines the speed at that point (hello differential calculus :D ) By the way, I tried scaling the keys on baked animation these days, and it slows down very nicely. It's quite hard to do gradual slowdowns/speedups this way though :D

PexElroy
08 August 2008, 12:29 PM
Yep some Curve maybe to control time scale, or a field, and then bakes each substep or half substep etc

ddustin
08 August 2008, 02:15 PM
Anybody install the update on a x64 machine?

David

mir-vadim
08 August 2008, 02:49 PM
Yes, Glacierise Installed it on 64 bit system and all works fine.
Do You have any problems?

Glacierise
08 August 2008, 03:06 PM
Hey Dave, and anybody else that has doubts - I run a win XP64, and I have installed RF 1.4 on a 32 bit max 2008, no problems whatsoever.

ddustin
08 August 2008, 03:10 PM
Did you install it to the Program Files directory or Program Files (x86)?

I get an error about being able to write one of the dll's.

(I am using max 2009 btw)

David

Edit: I had to install to the 64 bit directory first, then the 32 bit (x86) to get it to work.
On the 2nd install to 32 bit, I just opted not to re-install the Physix portion

wreath
08 August 2008, 08:42 PM
Dustin try install on program files (x86) / autodesk / 3d studio max 2009

PexElroy
08 August 2008, 09:48 PM
- the control with real-time PhysX influence is great ;)

suprun
08 August 2008, 06:30 AM
Приветствую, всех, если кто прочтет, еще меньше кто поймет!

Обращаюсь к Вам MirVadim как соотечественнику, и просто хорошему человеку! Я прекрасно осознаю что Вы крайне занятой человек, но всё же обращусь к вам с предложением. Вы как человек прекрасно разбирающийся в реакторе и написании скриптов, наверняка можете разобраться в этом деле! Поверьте перешарил весь интернет вдоль и поперек, но не нашел урока по нормальной настройке анимации танка в реакторе, особенно гусениц. Craft director не достаточно реалистично просчитывает динамику для танка! Сможете ли Вы справиться с такой проблемой при помощи реактора. Хотябы направить на путь правильный!

ddustin
08 August 2008, 04:15 PM
Приветствую, всех, если кто прочтет, еще меньше кто поймет!

Обращаюсь к Вам MirVadim как соотечественнику, и просто хорошему человеку! Я прекрасно осознаю что Вы крайне занятой человек, но всё же обращусь к вам с предложением. Вы как человек прекрасно разбирающийся в реакторе и написании скриптов, наверняка можете разобраться в этом деле! Поверьте перешарил весь интернет вдоль и поперек, но не нашел урока по нормальной настройке анимации танка в реакторе, особенно гусениц. Craft director не достаточно реалистично просчитывает динамику для танка! Сможете ли Вы справиться с такой проблемой при помощи реактора. Хотябы направить на путь правильный!

Well I'm not sure I agree with you and I'm not sure I disagree with you (don't understand).
Kidding of course......

David

Glacierise
08 August 2008, 04:23 PM
Sorry brat, this forum's language is English :) Also, this is not the place to ask such a question - use the search, or post in the main 3ds max forum - you'll get help there.

fireknght2
08 August 2008, 02:26 AM
Translation: Russian ŧ EnglishПриветствую, всех, если кто прочтет, еще меньше кто поймет! Обращаюсь к Вам MirVadim как соотечественнику, и просто хорошему человеку! Я прекрасно осознаю что Вы крайне занятой человек, но всё же обращусь к вам с предложением. Вы как человек прекрасно разбирающийся в реакторе и написании скриптов, наверняка можете разобраться в этом деле! Поверьте перешарил весь интернет вдоль и поперек, но не нашел урока по нормальной настройке анимации танка в реакторе, особенно гусениц. Craft director не достаточно реалистично просчитывает динамику для танка! Сможете ли Вы справиться с такой проблемой при помощи реактора. Хотябы направить на путь правильный!_SetupBidi('source'); Greetings, all, if someone will read, even fewer who understand!

I am writing to you as MirVadim compatriot, and just a good man! I am well aware that you are extremely busy man, but still turn to you with the proposal. You as a person well versed in the reactor and writing scripts, can certainly look into this matter! Believe peresharil all along and across the Internet, but has not found a lesson on the normal setting animation in the reactor tank, especially caterpillars. Craft director proschityvaet not realistic enough momentum to the tank! Can you cope with this problem through the reactor. Hotyaby directed to the correct path!

PsychoSilence
08 August 2008, 10:42 AM
although this in not really the right place for tank thread reactor setups im always happy if i can help :)

Craft Crawler Extension Plugin (https://secure.craftanimations.com/index.cfm?objectid=387B2AC1-AB18-8901-4C343FDB95F7ADAD&cat=AA1EAF6D-ECF1-1742-8D115BEFF3775801)
and
more about what it does (https://secure.craftanimations.com/index.cfm?cat=AA1E79DB-0A6F-9FA5-A171CD98FD810180&objectid=3885F9DA-C92E-AA3B-E94A0FD46CE906D9)

http://www.youtube.com/watch?v=He7b98XURt8 < in use

hope i could help

back to RF because it rocks so hard!

Glacierise
08 August 2008, 10:55 AM
Awesome thing I just found about RF (sorry, can't show yet :D ) - When you do a physx sim, until frame x, then bake the animation, then fracture, and then explode - the initial animation stays intact, the effect is of the initial animation, and then the fracture/explode animation with no hassle whatsoever. Mighty cool!

mir-vadim
08 August 2008, 01:15 PM
suprun'у: к сожаление одним реактором тут не обойтись, тем более если опыта работы с ним не много. Тут нужно более комплексное решение. Советую сходить по ссылкам которые дал PsychoSilence в первом посте. А вообще по таким специфичным вопросам лучше создавать тему отдельно, быстрее помогут.

Actually, Yes. Fragmentation able to keep animation if Impact object has it.
You can turn On\Off baking using Bake Animation checkbox in Advanced fragmentation rollout.

Glacierise
08 August 2008, 05:51 PM
Here's a tiny demo:

http://cg.glacierise.com/2-stage_rayfire.avi

That's just the thing I was describing - first animate, then record animation, break the object, make new animation! Another plus point for RF! I hope somebody is keeping notes ;)

JohnnyRandom
08 August 2008, 06:07 PM
heh... sweet, taking notes...:arteest:

ddustin
08 August 2008, 09:14 PM
Here is a link to an animation simulating a laser.
http://www.laser-engrave-it.com/

It is for a new business I am starting for my wife to run.

Thanks,
David

EDIT: Here is a link with motion blur http://www.laser-engrave-it.com/LEI-Mo-blur.html

Steve Green
08 August 2008, 02:52 PM
Has anyone tested Rayfire with hardware PhysX yet?

I installed the new drivers, thinking it might work with my 8600GTS (I thought all 8 series cards would support PhysX), but computer says no...

Thanks,

Steve

jigu
08 August 2008, 02:57 PM
I have got 8800GT which supports cuda and physx. As 8800GT is latest model and it uses G92 GPU processor which supports physx.

Does demo version of rayfire 1.4 have physx? ... gonna have to try it.

videofxuniverse
08 August 2008, 04:22 PM
I have found this tool to be a very time removing blessing. All the messing around getting all the reactor bits, and particle systems can really slow down your work flow, but with a few clicks you have it all done for you. The only 1 gripe i have with it is that when i use the bomb simulation on an object pieces always seems to fall through the ground, no matter what i do. Is there anyway to make sure the ground stops any objects falling through?

mir-vadim
08 August 2008, 05:02 PM
I tested PhysX plugin with 8800 GT, and didn't notice any difference.

to videofxuniverse: try to set Concave simulation geometry for Unyielding objects.
Send me scene if this will not help.

Good news, I just added Time Scale feature for PhysX engine.
Now You able to interactively slow down or speed up You simulation.
Check out this video:
http://www.mirvadim.com/Video/RayFireTool_1.4_TimeScale.wmv

Glacierise
08 August 2008, 05:09 PM
Do you use reactor or physx? Have you added the ground in the unyielding objects? If you are using physx, try with a box instead of a plane for the ground, I've found it to work best.

Glacierise
08 August 2008, 05:13 PM
That time scale feature rocks so much! Another great RF update in no time, respect dude!

mmoses00
08 August 2008, 05:21 PM
Animating Time scale !

Cool. Slomo ramps on destruction never were easier..... you cannot do anything like that in PFLOW or Fume or whatnot as easy as this.

Bummer that I use a Quadro 4600... and I cannot get hardware Physx support. :(

JohnnyRandom
08 August 2008, 07:40 PM
Time scale = cool:cool:



Bummer that I use a Quadro 4600... and I cannot get hardware Physx support. :(

If your feeling spunky check out the latest Beta driver ( http://developer.nvidia.com/object/opengl_3_driver.html ) Supports quadro 4600, PhysX, Cuda, and OGL 3.0.

Steve Green
08 August 2008, 07:50 PM
The time scale thing is cool, but I'm guessing that it only works within the Rayfire sim?

If it somehow manages to work with the rest of the scene, including PFlow, materials, Fume etc. then it's fantastic.

If it's just within Rayfire then wouldn't it be pretty limited in what situations it could be useful in?

Sorry to point it out, it is very cool...

- Steve

Glacierise
08 August 2008, 08:08 PM
Making a retimer for max as a whole is far beyond the scope of rayfire :D Nevertheless, that time curve will be very usefull, since these shots are carefully planned and directed, and the other elements can be retimed too, that usually are particles and fumefx/afterburn.

Steve Green
08 August 2008, 08:28 PM
I just thought it would be easier to go the brute force approach and render every frame at the slowest speed and do it in post, rather than try to sync Rayfire and the rest.

or...

would there be a way of using Rayfire to visualise the timing, and send that data to Max to render what frames are needed for the final shot?

So the whole animation (n the rest of Max - fume/particles) is calculated at the slowest speed, but only the frames that are needed to mimic the rayfire visualisation are rendered?

You're still simming more info, but I don't see any other way round it.

- Steve

mir-vadim
08 August 2008, 08:47 PM
As for FumeFX, You can use the same animated controller for FumeFX's time scale property.
Just simulated fire with animated time scale, all works fine

JohnnyRandom
08 August 2008, 09:09 PM
Steve, well for one it is not that simple to tweek Fume's speed by simply adjusting the time scale. below 1.0 your basic setup starts to go to sh!t linearly. You have to start adjusting more and more parameters as the time scale drops to retain a "proper" look.

Steve Green
08 August 2008, 06:48 AM
Hi,

yeah, that's why I think you'd just have to bite the bullet and animate and sim everything at the maximum frame rate and then do any time-ramping in post.

A typical scene could contain point-cached characters and cloth, hair (no idea how you could time ramp hair), Pflow, Atmospheric Effects, PhysX or Reactor.

Just seems much less of a headache to accept the extra sim/render time than attempt to retime everything individually.

Cheers,

Steve

videofxuniverse
08 August 2008, 08:03 AM
here is a short video of my rayfire renders


http://www.youtube.com/watch?v=le4n4JP_1Co

can you see what i mean by some of the debris falling through the ground, specifically the statue and chrome ball. i am also sure i had the settings made as you said

mir-vadim
08 August 2008, 08:55 AM
Which reactor solver You used?
RayFire uses solver defined in reactor's About rollout.
Make sure that it is Havok 1.
Also try to lower Your fragments mass,
sometimes big mass cause interpenetration through ground.

videofxuniverse
08 August 2008, 09:02 AM
thank you for your help i will try it, and also thank yuo very much for bringing out an absolutly awesome reactor tool. To set up all those things manually especially with the gun, would take 40 mins, but you have helped save a lot of time

JohnnyRandom
08 August 2008, 05:31 PM
Hi,

yeah, that's why I think you'd just have to bite the bullet and animate and sim everything at the maximum frame rate and then do any time-ramping in post.

A typical scene could contain point-cached characters and cloth, hair (no idea how you could time ramp hair), Pflow, Atmospheric Effects, PhysX or Reactor.

Just seems much less of a headache to accept the extra sim/render time than attempt to retime everything individually.

Cheers,

Steve

Dude, were you on the Debate Team?:D

Most cases that is true, since everything typically gets comped anyway.

None-the-less it is a feature, which personally, could come in handy, and since it is a feature you don't have to use it.

What about motion blur, a high frame rate animation ramped down to a matrix style speed never shows mblur at its slowest point, but you see it at velocity, depending on your angles there is a good chance you won't be able to add it in post.

Specifically, Fume doesn't respect FPS (one of the most requested features for update), it is a frame-by-frame simulator, sure you can tweek the timescale, and as I said it just doesn't come out the same.

Steve Green
08 August 2008, 06:21 PM
Hah, no, I wasn't...

I jsut wanted some clarification, Mir said it worked fine, but you said it didn't with regards to Fume.


Cheers

Steve

JohnnyRandom
08 August 2008, 06:29 PM
LOL, just checking :D

I am sure itdoes work fine in regards to the linking the timescales. ramping-them-wise

It is just the way fume behaves at lower time scales. Here is a swf of some tests I did, and you'll notice the difference in the sims.

Example 8.5 mb SWF (http://www.4rand.com/TEST/FumeFX/Smoke/TimeScale.swf)

Which is not rayfire's fault but Fumes propagation in the voxels.

Steve Green
08 August 2008, 07:31 PM
Going a little OT regarding time ramping and motion blur...

I use RSB for post motionblur with a velocity channel - the obvious way to do it (to me anyway) is for Twixtor by the same company to take frame rate changes into account when RSMB is using the velocity channel to do the motion blur.

Not sure if it does that already, since I don't have Twixtor, but it seems worthy of a suggestion to them.

Cheers

Steve

JohnnyRandom
08 August 2008, 08:43 PM
Sorry Vadim for all the semi-OT

Those test are ages old but at least you see what happens in the grid when you mess with the timescale on the low end. Yeah, you would need something to say hey vel. data needs to differ according to frame rate change. Hmm

PsychoSilence
08 August 2008, 09:08 PM
Example 8.5 mb SWF (http://www.4rand.com/TEST/FumeFX/Smoke/TimeScale.swf)


sorry fot getting back to semi-OT but these tests are nice :) must have missed them on other threads. i never had to do animated timescales in fume with collision objects...only free flames where u animate certain parameters til it looks right...

mir-vadim
08 August 2008, 08:48 PM
It is time to update Your tools. :)
Build 1.40.03 released.

Here is changelog:

1. New Time Scale property for dynamic simulation. This property interactive for PhysX dynamic simulation.

2. New RF_PhysX object in Helpers\RayFire Tool category. Allows You to animate physical properties for PhysX engine. You should add this object in PhysX Influence list to make it active.

3. RF_Bomb object moved to Helpers\RayFire Tool category. IMPORTANT: Old Geometry RF_Bomb object not supported anymore.

Glacierise
08 August 2008, 09:01 PM
Yohoho, and a bottle of whatever! I must say I made my first Rayfire bullet time in like 5 min :D The new helpers open interesting oppurtunities, too, even subsamples is animatable! Another great release, cheers Mir! And, I think these features deserve a higher version increment, at least 1.45? :)

P.S.: here's the aforementioned 5-min-bullet time :D Not gonna win at the Oscars, but paving the way for the coolness to come :D

http://cg.glacierise.com/rf_5_min_bullet_time.avi

JohnnyRandom
08 August 2008, 12:07 AM
Was the RF goe bomb not working correctly? I haven't used it much but I was getting strange behaviour in regards to detonation time, ie the physx sim set to set at frame 10 and the bomb set to detonate at frame 20, the det starts at frame 10?

floopyb
08 August 2008, 06:20 AM
Hi, Im currently using rayfire 1.34 in production for machine gun fire which is working very nicely. A few features that I would like to see in there would be a way to have the fire rate randomized a bit (or with a randomize % spinner) rather than just every 4 frames (or what ever you have the Shooting rate set to)
Also being able to specify what material ID for the Holes left by the booleaning and the interior of the fragmented hole chunks explicitly would be very handy.
Overall a great plugin and very useful!

mir-vadim
08 August 2008, 08:14 AM
to floopyb: for now You can't add variation in shooting rate, but I am planning to create special RF_Gun object with more properties and presets in next versions. Material ID You can set in Advanced Fragmentation Rollout.

to JohnnyRandom: How You defined objects which You want to explode? As Impact or as Inactive? If as Impact then try to check on "Deactivate Static Impact objects" and check on "Activate by force" in PhysX influence rollout.

mir-vadim
08 August 2008, 07:37 AM
Here little rendered demo for Time Scale feature.
http://www.mirvadim.com/Video/RayFireTool_1.4_timescale_rendered.mov 2.5 MB

PexElroy
08 August 2008, 07:24 PM
great addition, useful in many shots, glad you found this solution to time scaling, and it looks smooth - :cool:

Steve Green
08 August 2008, 06:57 PM
Hi,

one thing I haven't seen much discussion of is the best method of building geometry for RayXplosion to work with.

For example on a building, I started with everything as a single mesh and elements within that (so the walls were basically a box with a shell modifier, and lots of separate boxes for floors within the shell of the building) - This didn't work very well with the frag booleans.

So, I'm now looking at 4 separate objects for the walls and separate objects for the floors as well - is this the best method to get a successful demolition? I really wanted to get some interior in the buildings + add some particle debris of the fabric of the building to give it some volume rather than it just being one solid lump.

Anyone got any tips/comments on this - am I going broadly in the right direction?

Thanks,

Steve

Glacierise
08 August 2008, 06:38 PM
Hi guys, I'm back! I was very sick for a while, shit happens :)

@Steve: Check out the house demolition demo at the blastcode site, also these old nuke tests on youtube where they blast a house. There was also a great article a while ago on how they made the houses blow up in a cool game called "Company of heroes", but I can't find it. It all boils down to the idea that to make that realistic, you need to play with all the real elements of a construction - the outer cement/wood/plastered wall, the internal woodbeam/reinforced concrete framing, the masonry, the inner cement layer. All these break differently ad produce the complex effect of a structure crumbling. I might make something of the like, once I finish my current assignments :)

mir-vadim
08 August 2008, 07:28 PM
As for "real elements".
For now You can create dynamic simulation using only one material type for Impact, Unyielding and Inactive objects. In next version I will add "Material" property in Custom Fragmentation properties.
So, if object will have this material property, RayFire will assign it's own materail properties for it.
This will make rigid body setup more precise and deep.
And final dynamic simulation more real.

Steve Green
08 August 2008, 08:08 PM
Hi,

yeah it was more a question of what makes procutter/reactor/physx happy.

For example it looks like an object with separate elements is not such a good idea. Better to have many separate contiguous meshes.

It was more that, rather than how the house is constructed (you could do the same structure in two different ways - attach everything and try to run the frag programme, or leave them all as separate objects.

I'm guessing that fragging a continuous wall would be OK, but having a floor and ceiling within the same object might cause problems - I think it's just the poly within a poly, typical of a box with a shell modifier that gives it headaches.

Cheers

Steve

Glacierise
08 August 2008, 08:22 PM
@Mir - that sounds interesting ;)

@Steve - applying the same fragmentation on the different materials won't cut it, since they shouldn't fragment the same way. Even the dynamics for the different components shouldn't be the same - it it is, the structure would fall like a lego broken lego castle - everything at once. Separate components should be simmed separately, with the heaviest being simmed first, so their animation can be baked and the lighter pieces can bounce off them.

Steve Green
08 August 2008, 08:40 PM
Hi,

yeah, I wasn't really thinking of anything of the detail of the blastcode demos, more that there's just something inside (floors for example) so it's not a completely hollow shell.

(Think wide shot of a skyscraper along the lines of the original demo vid from Mir)

So I think in that case it would be maybe 3 layers, a concrete exterior, and a lighter weight interior, and then particles to add volume.

Cheers,

Steve

Glacierise
09 September 2008, 11:21 AM
A new thing with RF : http://cg.glacierise.com/reel_opener.avi - my new reel opener ;)

mir-vadim
09 September 2008, 07:00 PM
Hey.
Good news. :)
Thanks to Nvidia PhysX plugin developer now I am able to implemet Interactive Demolition feature.
Actually it is already implemented on about 40%.

Here first results.

http://www.mirvadim.com/Images/Ad/RF141_Demol_1.jpg
http://www.mirvadim.com/Video/rend/PxDemolition_1.mov (http://www.mirvadim.com/videos/rend/PxDemolition_1.mov)

And here 2 previews from another test.

Substeps: 20 Minimum break force: 50
http://www.mirvadim.com/Video/prev/PxDemol_2.avi (http://www.mirvadim.com/videos/prev/PxDemol_2.avi)

Substeps: 30 Minimum break force: 25
http://www.mirvadim.com/Video/prev/PxDemol_3.avi (http://www.mirvadim.com/videos/prev/PxDemol_3.avi)

First I will describe how current 1.4 Reactor Demolition works.
Since RF can't affect on animation while reactor creates it, RF have to create first animation pass to maintain all collision info. Then using this info RF fragments collided objects and creates second animation pass where it breaks all prefragmented objects.
As You expect this method can work only for first 2 -5 collisions. But there is no way to break fragments further, except create third animation pass to maintain new collision info and so on...

So, using PhysX plugin RF can start animation with solid objects, simulate animation frame by frame, and fragment objects right after their collision with something, add new fragments into simulation and continue simulation. When new fragment will collide with something again it will be fragmented further and so on.
Theoretically You can turn some object into a dust, but this will take a lot of time. :)

mmoses00
09 September 2008, 07:31 PM
Mir,


If I am understanding you correctly, Physx Demolition will keep breaking the bits if enough force? Or can you add an iteration value to keep it from breaking further... say breaking 3 times then just fall normally?

Also... anything in Physx that can bend the fragments... like a bullet thru metal plate effect?

PsychoSilence
09 September 2008, 07:47 PM
ur right about the velocity/ force. as long as it is more then ur Breaking velocity it will shatter. for now thereīs no way to limit the shattering to like 3 times. wouldnt be natural ;)

for the bending we have to wait on a proper version fix for the cloth featire in the physx plugin for max.

mir-vadim
09 September 2008, 07:23 AM
Rendered demo of one of the previews.
http://www.mirvadim.com/Images/Ad/RF141_Demol_2.jpg
http://www.mirvadim.com/Video/rend/PxDemolition_2.mov

mir-vadim
09 September 2008, 04:40 PM
If I am understanding you correctly, Physx Demolition will keep breaking the bits if enough force? Or can you add an iteration value to keep it from breaking further... say breaking 3 times then just fall normally?

Also... anything in Physx that can bend the fragments... like a bullet thru metal plate effect?

Yes, Physx Demolition will keep breaking the fragments further if enough force.
But there are few ways to limit demolition.
First as You can see in previews with falling cylinder, RayFire didn't demolish the big cylinder, dark plates, and big dark boxes.
This is because I selected them and added Custom Fragmentation Options.
After that I set their Material (new Custom fragmentation property in 1.41) to Lead (You can set any metal material from list). So, in new version You will be able to set material for each objects separately and if it is any metal, RayFire will not demolish it no matter how big is collision force. Unfortunately for now there is no way in RF to bend fragments.

And another way is as You suggested and I already implemented is sub demolition level.
So You can set it to 1 and can be sure that RF will demolish only actual Impact objects but not their fragments.
Or set it to 2 and RayFire will demolish only Impact objects,
and their fragments (if collision force enough), but not the fragments of fragments. :)

JohnnyRandom
09 September 2008, 08:46 PM
That's cool, so you can... say keep breaking fragments into smaller pieces? So if I had a cylinder rolling back and forth the fragments would keeps getting smaller?

mir-vadim
09 September 2008, 11:50 AM
Yes, you can. And Yes, they would.
Also I already added different material solidity for demolition.
I will share some screen grabs later, so You will see how this works.
But I already can say that it looks and works awesome. :)

Edit: Saying "material" I mean material from RayFire's material preset list.
Not material from material editor.

PexElroy
09 September 2008, 01:16 AM
nice additions ;)

thiagopaim
09 September 2008, 01:19 PM
Hi! Iīve just downloaded 1.34 demo version and the "Add/Clear/Menu" buttons in the "Impact Objects" window are not working. I canīt get any impact object...

Is it a MAX 9 bug? Is there any way to fix it?

Thanks a lot!

Debneyink
09 September 2008, 01:16 PM
...when letting fragments fall to the ground many of them appear to slip through the net(or box)
how do i gt them all to land on the ground and stay there?

PsychoSilence
09 September 2008, 02:05 PM
...when letting fragments fall to the ground many of them appear to slip through the net(or box)
how do i gt them all to land on the ground and stay there?

u should have a look at the help about stepsize and global scale.

mir-vadim
09 September 2008, 02:19 PM
...when letting fragments fall to the ground many of them appear to slip through the net(or box)
how do i gt them all to land on the ground and stay there?

In case You use reactor: Increae a litle bit substeps, change Unyileding Simulation geometry to Box, decrease a little bit gravity and/or set Impact objects material to something lighter, like Brick or Glass.
In case You use PhysX: There is only 1 way to avoid such problem with PhysX. Increase substeps property.


Hi! Iīve just downloaded 1.34 demo version and the "Add/Clear/Menu" buttons in the "Impact Objects" window are not working. I canīt get any impact object...
Is it a MAX 9 bug? Is there any way to fix it?

Well, this is definitely not a max 9 bug, because You are first who told me about such problem. Can You PM how exactly they are not working? Do You get any error messages? Can You see how buttons change their brightness when You move mouse cursor over them and when You click them?

Debneyink
09 September 2008, 04:03 PM
im wondering about peoples simualtion times im using ray for the first time on a job right now and my meshes are going to be fairly complex, whats been others experiences? im on max 9 64bit

each time i simulate my keyframe tab freaks out then I find myt run out of patience, how long should i expect to wait? is there a table of times v poly pieces?

thiagopaim
09 September 2008, 04:30 PM
Well, this is definitely not a max 9 bug, because You are first who told me about such problem. Can You PM how exactly they are not working? Do You get any error messages? Can You see how buttons change their brightness when You move mouse cursor over them and when You click them?

The same problem occurs with 1.4 demo version...the "Add/Clear/Menu" buttons in the "impact object" window doesn`t move, not even change colors with the mouse over (as do all the others buttons). I tried to click on every pixel over the buttons and around them, but nothing happens...any ideas!

Thanks again!

Erka2
09 September 2008, 08:44 PM
im wondering about peoples simualtion times im using ray for the first time on a job right now and my meshes are going to be fairly complex, whats been others experiences? im on max 9 64bit

each time i simulate my keyframe tab freaks out then I find myt run out of patience, how long should i expect to wait? is there a table of times v poly pieces?
If you have big number of objects you can try to set Mass by option to Random: range. I have recently come to situation where calculating of Mass by material density was taking some good time every time at beginning of simulation, changing of this option to Random greatly speeded up whole process.
I'm sure Vadim is already working on solving of this issue

mir-vadim
09 September 2008, 11:15 AM
I'm sure Vadim is already working on solving of this issue
Yes, I already fixed this bug. Now calculating objects mass accordingly to it's material and volume for 700 objects takes 5 seconds.

to thiagopaim: Yeah, seems that You have some problem with ImgTag controller. Which Max version You use?

Debneyink
09 September 2008, 03:05 PM
so im having problems...



basically all i want to do is have a bunch of buildings in my scene that have crumbled to the ground as set dressing if you will. each time i explode one it seems to go out of control. I just want the pieces to fall and land.
I've tried bomb strength settings to minimum but its still goes way too powerful. all i want to do is destroy...any one give me advice as to a simple way to do this would be most welcome.

mir-vadim
09 September 2008, 06:01 PM
Hi, Christian.
First of all tell me which engine You use? Reactor or PhysX?
In case You use PhysX check off "Force strength by mass" in PhysX influence rollout.
when this checkbox is On You might get weird results creating explosion, I will fix that in next version.
If it will not help You, please, save scene right before You want to create explosion and send it to me. So, I will better understand what actually You want to create and where might be the problem.

Debneyink
09 September 2008, 07:24 PM
thanks I'm going to have a look into that im using max 9 64bit o i dont think i can use physix?

ill keep having a play if i get no joy then ill send over the file

thanks again

thiagopaim
09 September 2008, 10:40 PM
to thiagopaim: Yeah, seems that You have some problem with ImgTag controller. Which Max version You use?

MAX 9,0 commercial 32-bit + windows xp profesisonal sp2...

Same problem with MAX 2009 32-bit.

Thanks!

mir-vadim
09 September 2008, 04:17 PM
Hey, guys.
I am planning to release new 1.41 version on next few days.
So, here short technical screen grab.

http://www.mirvadim.com/videos/screen/StickToMouseTest.wmv

More detailed and informative tutorials with sound will be created on next few days.

to thiagopaim: I made few changes with add\Clear\menu buttons in 1.41.
Cant ssure You that this will help, but try it when 1.41 will be released.

mmoses00
09 September 2008, 05:46 PM
wow.... that is crazy-cool!


sometimes, you just want to throw stuff around.. and get a certain effect quickly...
NICE!

PexElroy
09 September 2008, 07:57 PM
cool interaction mir; useful and fun :arteest:

Debneyink
09 September 2008, 12:43 PM
http://www.youtube.com/watch?v=qjk6cFgHr8o


so im blowing up a house bit by bit (slow progress) im getting artifacts left behind in mid air in my solves so im having to go to 0 delete the keyframes and solve all the individual bits again which is clowing me down.... what can i do to solve this??

mir-vadim
09 September 2008, 01:01 PM
This happens because reactor can stops simulation for some objects in case it will decided that they are too small\light.
In Physics Tab / Impact objects (in case these objects defined as Impact objects), go to Impact objects rollout and set "Mass by" to "By volume range", and then set some range using mass spinners, in this case mass will be calculated not by material density and volume, but by mass range and volume. Smalles objects will get lowest mass spinner value, the biggest will get maximum mass range value.
This should fix this propblem.
In 1.41 I added few functions which should fix such problem automatically.
Anyway, write me how this will help You.

Debneyink
09 September 2008, 01:21 PM
great thanks ill check it out

...too much to learn not enough lifespan...

Debneyink
09 September 2008, 01:27 PM
worked a treat thanks

Debneyink
09 September 2008, 03:47 PM
im finding myself having to delete fragments underneath the fragments that i want.

say i frag an object then frag it again the first frag piece is still there im sure its somthing im not clicking but not sure what it is?? at the moment im going throgh my pieces and deleting by hand!!

Debneyink
09 September 2008, 06:31 PM
another thing....im getting geomtry is too small so it only gets to a few frames of solving...how do i solve this problem? (deadline looming)

mir-vadim
09 September 2008, 06:45 PM
im finding myself having to delete fragments underneath the fragments that i want.

say i frag an object then frag it again the first frag piece is still there im sure its somthing im not clicking but not sure what it is?? at the moment im going throgh my pieces and deleting by hand!


When You frag object RayFire create copy of this object, hide original and frag this copy.
In advenced Fragmentation Options You can check ON "Do not store original objects"
and in this case You will not able to delete fragments and restore original objects via Layer manager.
Also, maybe You checked ON "Animate impacts/fragments visibility" in Adv. frag options.
In this case RayFire will animate visibility for original objects and for fragments relative to You current frame. You can use this for objects with glass like material. Check this checkbox and turn it OFF if it is ON.

What is Your minimum Mass range spinner in Impact objects properties?

Debneyink
09 September 2008, 07:05 PM
think you answered that one already!

mir-vadim
09 September 2008, 07:12 PM
So, did this help You?

Glacierise
09 September 2008, 02:21 PM
Cool new release! Very cool, in fact, due to the PhysX demolition! You guys should check out my Rayfire free video tutorial course over here: http://forums.cgsociety.org/showthread.php?f=183&t=678366

mir-vadim
09 September 2008, 03:09 PM
Well, 1.41 released. :)
http://forums.cgsociety.org/showthread.php?f=59&t=679063

PexElroy
09 September 2008, 04:27 PM
great, released before weekend :drool:

about the two new fields under Demolition options, depth level and material solidty, when you could go into more detail on those.

Glacierise
09 September 2008, 04:57 PM
Depth level is the number of times fragmentation is performed when a demolition event occurs. Material solidity is a global multiplier for the material solidity of all materials - increasing it will make materials more solid, and decreasing it - more fragile. This stuff will be explained and demoed soon.

TwiiK
09 September 2008, 06:03 PM
I've tested it some now and it looks awesome. The speed and ease of the PhysX stuff is amazing.

I'm a bit dissapointed in the very limited state of the demo though, I thought 20 days was more or less the only limitations, but I was barely able to follow Hristo's intro tutorial because I can't fragment that much nor simulate that long.

I was going to rant a little here now on how so many software developers have horrible demos who end up working against their intended purpose, but instead I'm just going to ask if you there is another demo or test version available for those who ask? :)

The time range limitation is basically what's keeping me from having fun with the plugin now. If it was something like 300 frames I wouldn't even notice it because I'm to impatient to simulate or render longer than that anyway. :p

mir-vadim
09 September 2008, 10:27 PM
great, released before weekend
about the two new fields under Demolition options, depth level and material solidty, when you could go into more detail on those.

Tomorrow I will create video tutorial where I will show how to use new features.
Was very tired to create them today. :)

And I will make demo restriction less constrained. But not tomorrow, on next week.

PexElroy
09 September 2008, 11:51 PM
great to hear mir, maybe take a nap now ;)

simple demolish example (http://www.areagrey.com/other/rfpass01w.mov),

TwiiK
09 September 2008, 07:35 AM
Sounds great, Mir.

About the PhysX stuff. Is this available in 3ds max outside of Rayfire now that I installed the plugin? Will it or can it replace reactor? It just seems so fast and intuitive compared to reactor.

TwiiK
09 September 2008, 07:44 AM
Another thing. Is there no help included with the plugin?

If I go to Help->Additional help->Ray Fire tool nothing happens. And I don't see any in the help folder.

Is the help on your page the only one?

Glacierise
09 September 2008, 07:50 AM
There is the physx plugin for max that comes from NVIDIA, you just need an interface to it, which Rayfire does. There could be another interfaces too, that use the same Physx plugin, so it's a very open situation. On PhysX replacing reactor - The reactor that is in Max is very old, and especially in the new light of dynamics getting paralelized and taking advantage of new CPU/GPU hardware, I think the old reactor is just gonna take its place in the honorable max history book (geez, I have the time for long posts with Fume simming at high res :D ). PhysX is not yet feature complete, have that in mind. NVidia are still working on it, and when it's complete, it will have liquids, metal cloth (a special cloth solver to help with making bendy/mangly metal surfaces) and who knows what more. So we need to wait on NVidia to build it up, and somebody to make a cool interface system for it - Rayfire is already a great thing, for example. And I guess that will happen, because I don't see anything coming for Max that is comparable to PhysX. There could be something AME does, it's all secret, but I think they won't be wasting effort with another solver when NVIDIA is giving PhysX for free. Maybe write an interface for it, buy box#2 from Oleg or something. Anyway, there's cool stuff to come! (fumefx done simming, end of obnoxious posting)

Glacierise
09 September 2008, 08:13 AM
@Twiik: Open up rayfire, click "about" in the main menu, then "online help".

mir-vadim
09 September 2008, 03:24 PM
I just released few tutorials about new features.
http://www.mirvadim.com/index.php?id=tutorials

TwiiK
09 September 2008, 05:09 PM
I'm looking at them now, but I seem to be unable to get the interactive demolition to work. I'm following what you do in the 'basic' video, but no matter what material I use for my impact object it doesn't break.

mir-vadim
09 September 2008, 05:17 PM
There should be tutorial scene in Max Scenes folder.
Open it and start demolition simulation, it should work.
Also, did You installed new Physx plugin version?
In case You already had old version installed, what You clicked when it asked You about another version of physx plugin installed?
Here intallation instruction:
http://www.mirvadim.com/forum/viewtopic.php?f=3&t=21
Write here how this will help You.

TwiiK
09 September 2008, 05:24 PM
It worked with the included scene, thanks.

Guess I have to see what the differences between that scene and a fresh scene is then. :)

Also, regarding the demo version. Have you thought about having a more or less fully functional, but time limited demo similar to how VRay do it?

I've already used 10 of my 20 tries and I haven't touched 5% of the features yet. And the short frame range makes it hard to test out things properly. :p

TwiiK
09 September 2008, 11:43 PM
I was messing around with the shooting simulation, and decided to render out what I had before I go to bed.

I haven't used Afterburn in a looong time so I just threw something on just before I rendered and hoped it came out ok. It almost did. :)

Here: http://www.twiik.net/files/videos/RF_Wall_AB.mov

The speed, intuitiveness(?) and interactivity of Rayfire and PhysX blows me away. It's sooo fun to work with.

Airflow
09 September 2008, 12:32 PM
When is the max2009-64bit with PhysX support demo gonna be realeased?

mir-vadim
09 September 2008, 01:17 PM
For now Physx Plugin supports only 3ds Max 32 bit.
This mean that You still can use it on 64 bit PC, but only with 32 bit Max.
And unfortunately I can't say for sure when 64 bit version will be available.

Abdel
09 September 2008, 02:22 PM
And I will make demo restriction less constrained. But not tomorrow, on next week.

Thats would be great, the demo version is kinda limited, especially the 'time range and literations' (for lots of fragments) options.. Is it also not possible to use the 'stick to mouse' feature in the demo version? I followed the video and it looks like the 'move timeslider' is greyed out, is it because of that?

mir-vadim
09 September 2008, 02:27 PM
To use Stick selected to Mouse feature You have to set "Stick selected to mouse strength" to some value more then 0. Because it is 0 by default. You can find this spinner in PhysX Influence rollout. And this feature doesn't depend on Move timeslider checkbox.

Abdel
09 September 2008, 03:20 PM
Thanks for the fast reply, figured it out. The 'Interactive demolition' doesnt work in the demo?
When i hit the simulate button, the option unchecks itself...

mir-vadim
09 September 2008, 04:08 PM
Interactive Demolition Feature works only in Bake Animation mode.
It checks OFF when You launch Preview mode.
Check out new tutorials, I noticed about this in them.

Abdel
09 September 2008, 05:31 PM
Im getting a '--unknown property: "getNumContacts" in PX' error when trying to bake animation. BTW im also having troubles when applying a glass shader to a object, you can see the fragmented pieces, what is the correct approach?

mir-vadim
09 September 2008, 05:48 PM
Im getting a '--unknown property: "getNumContacts" in PX' error when trying to bake animation.
I guess You had RyFire 1.4 installed and when You installed 1.41 You didn't install new PhysX plugin version.
Install 1.41 again, but first read this instruction carefully.
http://www.mirvadim.com/forum/viewtopic.php?f=3&t=21

And can You describe Your second problem in detail?

Abdel
09 September 2008, 06:36 PM
Yea i did, i reinstalled it and know it works correct.
Having problems with materials, here's a example...
Thx,

(http://img371.imageshack.us/my.php?image=glasqg7.jpg)http://img371.imageshack.us/my.php?image=glasqg7.jpg (http://img371.imageshack.us/my.php?image=glasqg7.jpg)

mir-vadim
09 September 2008, 06:58 PM
I see.
Download ant watch this tutorial:
http://www.mirvadim.com/videos/tut/RF141_Physx_Dml_2_Glass.mov
I added special demolition way for objects with refracted (glass) material .
Shortly, before create demolition You should set material in RayFire UI material presets list to Glass. In this case RayFire will create copy of original object and will animate visibility for it and for created fragments. So, before impact You will render solid object and after You will get fragments on You rendered image.

depleteD
09 September 2008, 07:08 PM
Amaaaaazing work Mir, Absolutley awesome

Glacierise
09 September 2008, 07:33 PM
Only bummer is that mental ray doesn't give a crap about the visibility track :)

P.S.: Tipped by Mir, I must correct myself - mr doesn't respect the itnermediate values between 0 and 1, but it does respect 0 and 1, and that's all RF needs, so - no problem!

PexElroy
09 September 2008, 11:03 PM
Wine glass (http://www.areagrey.com/other/rfpass06c.mov)

Cool stuff, but their is a problem with PhysX Glass/Ice materials when the demolition Depth level is 2 or greater.

sample of issue; max 2008 (http://www.areagrey.com/other/rf034.rar). a box spins, falls and breaks. It is set to Depth level 1, so sim at 1 first with PhysX and render frame 1, it looks good and renders right. Now Delete sim, change Depth level to 2, sim again and render the first few frames before it breaks.

mir-vadim
09 September 2008, 06:08 AM
Ok, I see the problem. Will fix it today.
New update will be available via Automatic Update system. (About menu \Check for updates)
Thanks for report.

PexElroy
09 September 2008, 03:07 PM
thanks mir; will revisit after ;)

Any tips in how to avoid fragments acting like popcorn popping? Some fragments might sit still ok, while others may touch other bits and pop around (PhysX).

Abdel
09 September 2008, 05:00 PM
I see.
Download ant watch this tutorial:
http://www.mirvadim.com/videos/tut/RF141_Physx_Dml_2_Glass.mov
I added special demolition way for objects with refracted (glass) material .
Shortly, before create demolition You should set material in RayFire UI material presets list to Glass. In this case RayFire will create copy of original object and will animate visibility for it and for created fragments. So, before impact You will render solid object and after You will get fragments on You rendered image.

But what if for example i want to fill a glass with fluids (with Glu3D), the settled particles are simulated in the glass (glass filled) but then when i animated the glass being fractured the fluids stays in the copied version of the glass that RF makes. I tried different ways how to achieve a proper method bit with no luck so far.

Thanks for the help, it is mutch appreciated..

mir-vadim
09 September 2008, 05:04 PM
Ok, just fixed bug with visible glass fragments.
You can get new build via About menu / Check for updates.
Demo version with new build will be release later, at the end of this week.

Any tips in how to avoid fragments acting like popcorn popping? Some fragments might sit still ok, while others may touch other bits and pop around (PhysX).Try to lower substeps, sometimes high value might cause such behaviour.

But what if for example i want to fill a glass with fluids (with Glu3D), the settled particles are simulated in the glass (glass filled) but then when i animated the glass being fractured the fluids stay in the copied version of the glass that RF makes. I tried different ways how to achieve a proper method bit with no luck so far.Try to simulate demolition first and then use fragments as collision objects for fluids.
Or maybe simulate fluids until glass explosion, then add fragments as collision objects and continue simulation.
Here few demos with exploded glasses with liquids.
http://www.youtube.com/watch?v=jV5dC10fCqg
http://www.youtube.com/watch?v=fJqOLKRq4Xo

Glacierise
09 September 2008, 10:23 PM
As far as I've experienced, interpenetration is the leading cause for popping behavior - you could increase col. tolerance too.

Also, the next, third Rayfire tutorial in my course is up: http://forums.cgsociety.org/showthread.php?f=183&t=678366 - this one is about the advanced fragmentation options.

PexElroy
10 October 2008, 04:41 AM
ah thanks for tips, the bits may pop or jiggle around for a few reasons, like if touching or rest on top of each other, or a pile of many fragments.

trying to find some work arounds, but it maybe linked to the physx sim not able to use the concave mesh yet.

Glass cuts nice now :wise: - wine glass II (http://www.areagrey.com/other/rfpass07.mov)

JohnnyRandom
10 October 2008, 04:06 PM
Glass cuts nice now :wise: - wine glass II (http://www.areagrey.com/other/rfpass07.mov)

Nice one :thumbsup:

PexElroy
10 October 2008, 11:02 PM
Thanks. now trying to test a bullet effect thru the glass. Going through one side and out other; not sure it will work tho, in how it randomly fragments it.

TwiiK
10 October 2008, 07:56 AM
What do you mean randomly fragments it?

In what I've learned through my weekend's experience with RayFire I would think there's 2 very viable ways of doing a bullet through glass:

- The first one is to use interactive demolition with an animated bullet going through the glass and fragmentation set to 'relative to impact point'.

- And the other would be to not use interactive demolition, but instead prefrag it using 2 dummies positioned where the bullet would intersect the glass and fragment the glass relative to those 2 dummies.

There are probably more ways of doing it, but wouldn't those 2 methods work and give a realistic effect instead of a random one? :)

mir-vadim
10 October 2008, 10:00 AM
- The first one is to use interactive demolition with an animated bullet going through the glass and fragmentation set to 'relative to impact point'.

I have to note that there might be some problem with animated bullet in case You will define it as Unyileding object.
Problem is that animated Unyielding objects changes it's position only at each frame.
But Physx simulate animation accordingly to substeps value. It means that in dynamic simulation this bullet will chage it's position only at each frame and will not take into account substeps, while Impact and Inactive objects will change their position accordingly to substeps value and animation will be more qualitative. I am lookig for ways to solve this problem, but for now the only way to avoid it is to animate bullet with low velocity or define it as Impact object.

PexElroy
10 October 2008, 02:46 PM
yeah the unyielding objects do not move each per-substep that RF is set to, but rather frame by frame. we cannot yet do a fracture on glass and use inactive objects, we'd see the fractured cut edges before motion.

the NIVIDA physX plugin for max doesn't yet have concave mesh, the space of a cup or a torus hole. Convex meshes are easier to compute (spheres are the easiest, boxes come next and concave meshes afterwards). but still testing different things ;)

Glacierise
10 October 2008, 03:00 PM
I think the animated unyielding and the substeps isn't so hard to solve problem - the motion should be interpolatable, right? So if object is at 0,0,5 at frame 5, and at 0,0,6 at frame 6, it is interpolatable that the object is at 0,0,5.3 at the 3rd of 10 substeps between frames 5 and 6. It's not so trivial with deforming objects, but it's in the same spirit, just for verts. That is an issue NVIDIA should work with though.

On the glass shatter - that's a more complex effect, and requires a more complex approach. I'd do a draw fracture for the spiderweb break, then overlay a rayfire point of impact layer, and do a multi-stage simulation. I'll show that in one of the later RF tutorials I do. My point is - just because it's not immediately doable by the push of a button does not mean it's not doable. Yes, it will be more complex and yes, it will be an approximation, but that's what art is! Cheers ;)

Glacierise
10 October 2008, 08:04 PM
Fourth tut is online : http://forums.cgsociety.org/showthread.php?f=183&t=678366 - Dynamics part one.

PexElroy
10 October 2008, 07:25 PM
Cool 4th tutorial. ;) I agree that some work can be an approximation or art, for us and tests, but a client may not accept that in some results.

ice-boy
10 October 2008, 06:30 PM
rayfire is very expensive for me so i need a lot of info about this plug in.
on their official site i dont see a special link for 3ds max 9. so its for all versions the same?

mir-vadim
10 October 2008, 06:50 PM
Yes, it is the same for all max versions.

ice-boy
10 October 2008, 08:45 PM
i can not buy the 1.34 version anymore right? its just 1.41.
because i read somewhere that the ones who had 1.34 just sended yo uan email and you send them teh uppdate right?

p.s. it looks like a really good plug in. i watched the 4 videos and its really amazing.

Debneyink
10 October 2008, 11:13 PM
does rayfire work for max 7 pretty odd question i know but i was hoping to teach it in a masters class,

mir-vadim
10 October 2008, 11:32 PM
Yes, 1.34 is not available for purchasing anymore and all 1.34 customers can get new 1.41 as free update.

No, unfortunately RayFire doesn't support Max 7. Only 9, 2008, and 2009 for now.

By the way, tomorrow I will release build 1.41.02 with few fixes and improvements.
It will be available via automatic update system.

mir-vadim
10 October 2008, 03:56 PM
Just released build 1.41.02.
It is available via About menu \ Check for updates.
I fixed few bugs in this build and added one neat property for physx dynamic simulation:
Collision Damping.

Also, check out these 2 new video tutorials:
http://mirvadim.com/videos/tut/RayFire141_PhysX_Objects_Activation.wmv
http://mirvadim.com/videos/tut/RayFire141_PhysX_Animated_TimeScale.wmv

PsychoSilence
10 October 2008, 04:35 PM
Great! downloading :)

PexElroy
10 October 2008, 04:37 PM
Downloaded; thanks for fixes and videos mir - :cool:

Glacierise
10 October 2008, 05:22 PM
That collision damping is such a cool feature. Awesome work dude!

mmoses00
10 October 2008, 05:43 PM
Mir,


I have been busy with other things...... but I really want to try out some of these new features.... Your updates keep fixing the exact things I needed to be fixed.... I think that this latest update has brought this above anything else out there for the speed/quality/control of destuction. The timescale and damping were the perfect additions....

I hope to get some kind of cool sample posted soon.
THX! Great work!

mir-vadim
10 October 2008, 02:13 PM
Hey,
Build 1.41.03 just released:
I made few improvements and separated update viewport toggle for dynamic simulation and fragmentaion.
But main feature is Groups support. With groups support You can simulate a group of objects as one solid object. Moreover, Physx will simulate a bunch of objects in one group as Concave geometry, each object in group will be treated as convex, but whole group will behave like Concave.
Here short tutor:
http://mirvadim.com/videos/tut/RF141_Physx_Groups.mov

PexElroy
10 October 2008, 02:27 PM
great additions - thanks mir for fixes and new grouping feature, trying it out today ;)


EDIT - when I try to toggle update viewport-Fragmentation, it is does not show the cutting process if solo picked, it does show the fragmentation if both are toggled on.

PsychoSilence
10 October 2008, 04:04 PM
mir "the mashine" vadim :)

this is the best addition since like forever :D

LeifKE
10 October 2008, 06:46 PM
mir-vadim, this is great work, i have been waiting for someting like this, support for groups.
To bad we cant breake the groups yet, but anyway great work. And thanks again for making this possible.

Leif.

mir-vadim
10 October 2008, 07:56 PM
To bad we cant breake the groups yet, but anyway great work. And thanks again for making this possible.
Haha, actually I already add such feature. It will be in next minor build 1.41.04
I added "Demolish groups" check box in Interactive demolition options.
Now You able to define whether You want to demolish only geometry objects, only groups, or the both. Also group demolition based on Interactive demolition system.
So, for instance, You can throw or hit some group, then it will be demolished after collision and then each object in this group might be demolished further, depends on it's own material properties. Looks very cool for now. :)

Before release it, I want to implement some kind of connectivity feature like in reactor fracture helper also.
So, when group will collide with something, it will not be demolished totally.
Only collided parts will be excluded from group while all other parts will stay in it.

MartinRomero
10 October 2008, 03:40 AM
Hello Mir,

I just wanted to let you know that Rayfire is the best plugin for max that I ever tried. Looking forward on getting version 1.4.

Later

Glacierise
10 October 2008, 07:15 AM
Yay, another cool new version! Just keep them coming dude :D Also, a new tut is up:

Dynamics part 2: http://mirvadim.com/videos/hristo/rf_tut05.mov

Check it out, that one's kinda cool.


1