sebkaine Posted May 7, 2015 Share Posted May 7, 2015 (edited) Hi Guys, Sorry for this noob question, but i haven't find a good enough way to do this. I have a simple cache of animated point generated in SOP on 100 frames. they have : - @v - @trigger - @viscosity I just want to use this object as emitter in my FLIP. And i want this : - when $F >= @trigger the point is release in the flip sim with initial velocity @v and viscosity @v I have try the following without success - particle field initial data doesn't look to allow animated point sequence - if i plug a POP source in post solve , it works but the point doesn't look to have FLIP properties they behave like simple POP - if i use a sop solver in post solve with an object merge it also work but i have the same problem, particle doesn't have a FLIP behavior So i would be curious to know what is the recommanded workflow to use a particle sequence with custom attribute as a FLIP source. How to be sure that i am not making some sort of weird hijacking, and that i am injecting a proper FLIP point in my sim ? Thanks for your input ! Cheers E Edited May 7, 2015 by sebkaine Quote Link to comment Share on other sites More sharing options...
loudsubs Posted May 7, 2015 Share Posted May 7, 2015 In SOPs, use a blast node, set to points and put @trigger>=$F in the group parm. In Dops, use a particle fluid emitter to source from these points into FLIP. 2 Quote Link to comment Share on other sites More sharing options...
sebkaine Posted May 7, 2015 Author Share Posted May 7, 2015 Thanks a lot loudsubs ! this particle fluid emitter was exactly the one i need ! Quote Link to comment Share on other sites More sharing options...
sebkaine Posted May 7, 2015 Author Share Posted May 7, 2015 (edited) i would also have another question ... Would you recommand to also select the sop point in flip object initial data , or does particle fluid emitter alone is good enough ? Basically does the fact of seting a initial data is a better practice or it doesn't matter at all ... I love the approach of preping stuff in sop , so if using particle fluid emitter is giving as good result than using object initial data, i think i'm gonna use this 99% of the time now. Edited May 7, 2015 by sebkaine Quote Link to comment Share on other sites More sharing options...
pbarua Posted May 8, 2015 Share Posted May 8, 2015 particle fluid emitter is based on old POP, use new POPSource DOP instead which is faster and give you same result. You need not to feed sop points in flip object initial data. That's for one time sourcing like tanks. Quote Link to comment Share on other sites More sharing options...
sebkaine Posted May 8, 2015 Author Share Posted May 8, 2015 (edited) Thanks for your help Pradeep ! So you mean using the standard POP source node as FLIP emitter ? That's actually what i did at first . But i have the feeling that using this node is not a good thing. I compare using initial data and POP source as initial emitter, and the 2 gives very different result, Particle emit with POP source look like they were missing FLIP attributes, (like the dumb particle in realflow). So maybe i am totally wrong and i did a mistake somewhere else. But could you confirm me that the most accurate / elegant / recommanded way to use POP/SOP sequence as flip emitter, is by using the standard POP Source node ? Edited May 8, 2015 by sebkaine Quote Link to comment Share on other sites More sharing options...
pbarua Posted May 8, 2015 Share Posted May 8, 2015 For fluid attributes you need to unlock POP Source DOP and then attribtransfer from flipobject. Quote Link to comment Share on other sites More sharing options...
sebkaine Posted May 8, 2015 Author Share Posted May 8, 2015 (edited) Thanks again for your help Pradeep ! So there is no real official workflow if you want to use a pop/sop sequence as a flip emitter. You have to hack the source POP to transfer @. Well that's kind of a geeky approach ... It is sometime really exhausting to re-program the matrix everytime , especially when you just want some few point emitting into a fluid and that you are under short deadline ! i think i'm gonna stick with loudsubs tricks and use the particle fluid emitter even if it's old school cause i can't imagine a fluid emitter that doesn't create fluid attributes ... Cheers E EDIT : i've got good enough result with particle fluid emitter, i've attach an exemple scene. If you have a better / faster method to get this , i would be happy to know flip_sop_emit.hip Edited May 8, 2015 by sebkaine Quote Link to comment Share on other sites More sharing options...
DLCool Posted May 8, 2015 Share Posted May 8, 2015 Hi Emmanuel. I tried this way. I used your point sequence as an input to the fluid source SOP node. And I added point attributes you wrote in your first post. Then I used the source volume node as the flip emitter in the DOP network. I guess this is not the simplest and the fastest way to add particles to flip. But it is more natural for flip in my opinion. Because the fluid source node in the SOP generates particles with the right separation and the source volume in the DOP doesn't allow to add additional unneeded particles killing them inside the surface field. flip_sop_emit1.hipnc Quote Link to comment Share on other sites More sharing options...
sebkaine Posted May 9, 2015 Author Share Posted May 9, 2015 (edited) Thanks a lot for your help Dmitry. I think you're absolutey right, using the source volume is the best way to ensure that you are not injecting some trash in your flip sims. Actually all exemple and shelf presets i've found that use emission inside FLIP grid use this method. The thing is that i am using cloth to draw some liquid pattern and i generate some ptc with custom @. I will try to convert those ptc to volume ... but initially i find more straightforward to use the exact ptc with known @ptnum and custom @. I am not sure injecting pop inside FLIP directly is a good practice ... cause in fact you are right you violate particle separation ppties. well i now have 2 methods that works ... that 's cool i will compare both ... if someone has an other tricks to share for injecting SOP/POP ptc sequence inside FLIP , by keeping the FLUID properties intact feel free to share ! Thanks again for your hip ! cool stuff inside ! Edited May 9, 2015 by sebkaine Quote Link to comment Share on other sites More sharing options...
DLCool Posted May 9, 2015 Share Posted May 9, 2015 I have just tried a sop solver and a particle source. All stuff works. Regards. Look inside: flip_sop_emit2.hipnc Quote Link to comment Share on other sites More sharing options...
sebkaine Posted May 9, 2015 Author Share Posted May 9, 2015 (edited) That's cool now we have 4 different methods ... I've made a scene to compare each one ... SOP solver / Fluid Emitter / POP source give very very similar result - Like Pradeep say previously it looks that Fuid Emitter is replacable by POP source ... - In each of those 3 methods i guess that separation is not perfectly respected Volume source give similar behavior but we can definitly see that overlaps beetween particles are avoided ... Well beetween all those 4 methods i would not be able to know what would be the most efficient. One problem with Volume Source s that it's not very accurate when you want to follow precisly a mesh. So i would say that if you want -> big amount of water in your flip and you need speed, volume source would be my candidate. -> draw a precise form from a mesh, point emission and POP source / SOP solver source would be my candidate -> Fluid Emitter which work with old pop inside looks more like a legacy node of the SPH workflow ( pradeep is right ! ) sometime houdini remind me the overlook hotel with all those obscur room inside ! emission_compare.hipnc Edited May 9, 2015 by sebkaine 1 Quote Link to comment Share on other sites More sharing options...
anim Posted May 9, 2015 Share Posted May 9, 2015 (edited) ... One problem with Volume Source s that it's not very accurate when you want to follow precisly a mesh. ... don't forget that Source Volume uses particles from your source geometry directly and doesn't need volume from source (vel nor surface), so if you point it to /obj/EMIT_PTC/PTC_OUT, it will be as precise as others, and still will remove all particles being sourced within FLIP's surface SDF but no matter which of the mentioned methods you use all of them are just SOP Solver DOP anyways for FLIP I would stick to Source Volume DOP and if you care about particle separation being enforced, you can try Flip Solver/Particle Motion/Separation tab Edited May 9, 2015 by anim 2 Quote Link to comment Share on other sites More sharing options...
DLCool Posted May 10, 2015 Share Posted May 10, 2015 yeah you are right. Thank you. Quote Link to comment Share on other sites More sharing options...
sebkaine Posted May 10, 2015 Author Share Posted May 10, 2015 Thanks Tomas very useful info here ! Quote Link to comment Share on other sites More sharing options...
dpap Posted May 13, 2015 Share Posted May 13, 2015 I think there is one more way. If you use the "Particle Field" option in the flip object but in the creation tab you put $FF instead of 1. I added an example in your file I hope it helps. emission_compare2.hipnc Quote Link to comment Share on other sites More sharing options...
anim Posted May 13, 2015 Share Posted May 13, 2015 I think there is one more way. If you use the "Particle Field" option in the flip object but in the creation tab you put $FF instead of 1. I added an example in your file I hope it helps. that's super dangerous what that does is, it will create new flip object per each frame, with brand new vel, surface, and other fields and initial particles from that frame you will end up with hundreds of independent small flip sims (not interacting in any way), super slow and super high memory usage, you would probably never want to do this Quote Link to comment Share on other sites More sharing options...
dpap Posted May 19, 2015 Share Posted May 19, 2015 Thank you Tomas for pointing out that it's not the right way to do it. I didn't know the things said and the fact that the first frames simulated fine, made me think that it was a valid solution. I did another test with the same technique in another file and although it didn't eat up more ram it produced far more particles and overall was a lot slower. So I guess you are right. that's super dangerous Hehe, I hope i didn't blow up anyone's computer ! Imagine what will happen when I start meshing with the pyro solver Sorry If I misled anyone, I was just trying to help... Quote Link to comment Share on other sites More sharing options...
anthonymcgrath Posted August 11, 2016 Share Posted August 11, 2016 On 19/05/2015 at 1:02 AM, dpap said: Hehe, I hope i didn't blow up anyone's computer ! Imagine what will happen when I start meshing with the pyro solver Sorry If I misled anyone, I was just trying to help... glad i read this as i literally just typed $FF in then fortunately read through this thread before things got too heavy! Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.