PDA

View Full Version : Cell Mitosis with Pflow help


Lsdinc
02-03-2011, 04:14 PM
Hi all,

I'm working on a way to animate Cell Mitosis using pflow.

So far I have looked at Pete Drapers Tutorial (can be found here: http://www.computerarts.co.uk/about_us/latest_issue/stop_press if your looking for it.) (http://www.computerarts.co.uk/about_us/latest_issue/stop_press)
It is very good but is just not what I'm trying to do.

I have animated a sphere that splits in two and with Pete Drapers pflow I can get it to spawn in different ways but it is not effect am looking for, it is too random.
This is effect I want:
http://www.youtube.com/watch?v=QrMoyOfim2s&feature=related
and hopefully lead to this:
http://www.youtube.com/watch?v=m73i1Zk8EA0

I want a particle to spawn into 2 particles and then from the nucleus of both particles I want to spawn 2 more again and again and again and so on.
I have tried a couple of ways but am just getting a bit confused and bewildered.
If anybody has any suggestions please help my poor brain.

Thanks
Les

JohnnyRandom
02-04-2011, 09:07 PM
You can loop through a couple of spawn events.

http://vimeo.com/19580668

Flow:

Lsdinc
02-07-2011, 08:55 AM
Brilliant!!
I have been working on the same lines with Pete Drapers Pflow but I like the results from yours so I'm gonna have a play with them thanks and will be back.
I have attached my pflow.

Lsdinc
02-07-2011, 09:29 AM
You have a Data operator in there, is that doing anything? Unfortunately we only have one licence for box3 and somebody else is working with it at the moment.
Also your keep apart is set to selection, what events are you selecting to keep apart?

thanks
Les

JohnnyRandom
02-07-2011, 06:24 PM
The Data Op is simply to lock the particles z-axis to world 0.0. Scalar 0.0 -> Output Standard, Position+Z Component. You could just as easy do it with a script too, although it may take a pretty big speed hit near the end when you start to get a huge particle count, you would have to test it to be sure.

For a top-down you maybe able to just get away without it. As long as there are no shadows in the render.

Does box#3 demo mode not read pre-built ops?

The Keep Apart is just set to Selected System PF-Source 01. Probably just me but it seems to work better that way.

Oh, and the spawns I use Delete Parent + Offspring count is 2, this keeps particles from staying in the spawn test event.

Lsdinc
02-07-2011, 08:56 PM
Hey,
It took me all morning and bit of afternoon but I worked all that stuff out :D The "Delete Parent + Offspring count is 2" took me a bit tho :D
The particle instance I want to use an animated but for some reason the instance will not animate, is there any reason in the flow that it should not?

Thanks

Les

JohnnyRandom
02-07-2011, 10:32 PM
If you use an animated shape your need to enable it in the Shape Instance operator, also you will now have the Animation Offset group enabled, select the appropriate value, in most cases it is Particle Age.

So when a particle is first born it will play the keyed animation from the beginning. If you were to choose something like Absolute Time, a particle born on frame 15f would already be 15 frames into the animation.

Lsdinc
02-07-2011, 11:01 PM
Yeah I did all that but the partcle is born at the end of the animation for some reason. I tried a very simple animation to test it and the same thing happened.
I just thought that maybe there was some reason but if not will just keep trying.

Thanks

Les

JohnnyRandom
02-08-2011, 12:21 AM
Here is an example. You have to be somewhat cautious with your animation cycle duration and your particle ages/age tests or it will not look correct. You either need to different animated shapes, one with a birth cycle and another with a death cycle (as I have included in the example) or a single shape that returns to the original shape at the end of its animation cycle.

Lsdinc
02-08-2011, 09:13 AM
OK, I see I have to get it all in the right order in the stack too.
I'm going to have to rethink it though as I cant have the particles moving about so much.
I might need to confine them to a grid or something like that, I would prob need to keep the parent aswell as the nature of the cell is that it replicates itself but does not die.
Like in this example: http://www.youtube.com/watch?v=m73i1Zk8EA0
The cell just divides and then there twice as many but my director wants to start from one cell to two to four to eight etc and all not moving away but close to each other.
I have to learn box 3, pflow its self is not enough, they have promised me a licence :)

Lsdinc
02-08-2011, 10:38 AM
This is exactly the sort of thing was thinking of
http://www.turbosquid.com/FullPreview/Index.cfm/ID/316140

I dont know if they used a pflow for it or just did it by hand tho.

JohnnyRandom
02-09-2011, 05:14 AM
You can instead of deleting the parent, just send the parent particle to the next event with another age test. I simply did it that way because of simplicity, least amount of operators.

No idea how that person did it, would need to see the animation to at least get an idea.

Maybe add a particle mesher (blobmesh or pwrapper) to the particles volume select the top of the blob, delete the selection and cap the holes, turbosmooth, subdivide, and relax. Something like that anyway.

No idea how to texture it though not really my field, maybe some type of xray shader.

Lsdinc
02-09-2011, 08:59 AM
Hey Johnny,

Yeah that was a way I was doing it alright it is just that the spacing and were the particle moves to are a bit random so I need to look into that.
The animation for the turbosquid thing can be found here: http://www.turbosquid.com/Download/Index.cfm?ID=316140_1223058

For the animation of the splitting of the cell I'm using vol select with relax etc and a bit of animated noise. Have been testing with blobmesh but the effect is not very good and is hard to texture and control. When the bits get to close to each other they just join up again.
I was thinking of a system where they come apart and then sort of squish up against each other again. I need to find a way for sort of creating a surface tension for the cell membrane.
My director reckons it would be a job for realflow but I think thatís a bit of over engineering.
Am gonna experiment with some reverse Skin wrapping and see what i can get from that.
As for shading that is not a prob, that is more my field :D.
I have attached what the vol select and relax system looks like with a bit of noise over it to blob it up a bit.
Thanks for keeping with me on this, your the only person to have gotten back to me.
Best
Les

JohnnyRandom
02-12-2011, 07:13 AM
When the bits get to close to each other they just join up again.


Yep I noticed that too, among the texturing issue.

As for realflow unless you get dead funky with it (way beyond my knowledge) you will run into the same mesh convergence issue. It will just handle a shitton more particles and greater mesh capability.

That clip looks pretty good :)

Lsdinc
02-14-2011, 08:57 AM
**Update**

Have had pretty good success with you pflow and a couple of data operators and I think have pretty much cracked it but alas all the particles are intersecting each other when it gets critical.
I have to find a way now to get them to not intersect!! It never ends!
Here is video and screen grab of pflow.
The data operator is just a small one to lock rotation after each spawn.
Any ideas on how to avoid intersecting particles?
Best
Les

JohnnyRandom
02-14-2011, 04:44 PM
yeh intersections can be a problem, without using Box#2, you have to be very creative with your timing, speeds, and the Keep Apart operator, that is about all I can suggest.

CGTalk Moderation
02-14-2011, 04:44 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.