Announcement

Collapse
No announcement yet.

Mercury

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Mercury

    Im trying to create some mercury forming into a shape (yes very T-1000)

    What would be the best approach to get that mercury look?

    Thinking higher surface tension? but also want it to move freely.


    Thanks

  • #2
    I imagine that if it's a T1000esque situation, then depending on where the liquid is initially, body force is animated
    to increase its attraction to the target geometry along with animated gravity. The surface tension is fairly easy to iterate to get a look that you like.
    I just did a little test...and I don't know PFD very well at all...and it looks ok for a quick setup.
    I'll post it when it finishes, if it's of any use

    It's actually quite odd that there is not really any tutorial out there that covers this...well, there is one but it's part of a paid course, so...
    https://www.behance.net/bartgelin

    Comment


    • #3
      Yup, surface tension with the Droplet Formation option enabled should do it

      Just this morning I remembered how 8-9 years ago when somebody wrote "I can't find ANY tutorials about Phoenix FD!!!!" and I was like "omg how is that possible, there are at least five tutorials online!". Glad this has changed a bit since then.

      Mmercury would make for a great toolbar preset though! I was thinking of pouring mercury, but it could be a simulator with initial fill up that morphs into the shape you've used to create the preset
      Svetlin Nikolov, Ex Phoenix team lead

      Comment


      • #4
        Honestly, all I did was was to steal the logo morph idea (the example setup) and tweak it using animated gravity, with a suitable mercury surface tension look, so
        the info is all there; however a preset could possibly help in the initial stages of doing something like this

        I rendered a rough sequence but way too tired to cut it/upload it now...tomorrow if it helps...
        https://www.behance.net/bartgelin

        Comment


        • #5
          Actually my first test when I looked at it was not right although it looked cool. So I spent a while trying to get something a bit more pleasing.
          I stopped here https://vimeo.com/732671260
          The second half is obviously sped up, as the time to actually resolve to the shape was 600 frames and still not completely done.
          That's body force tweaking I think, which I'm still figuring out the parameters of.

          Anyway, it was not as simple as I first thought and I had to force it to do what it's doing here, in order to get it in any way looking plausible.
          I have to say that the process is much more satisfying than the experience I had in the past, with Realflow years ago...what a pain!

          The problem was mimicking how mercury actually behaves, so its mageticism, the jitteriness (didn't get that done) and the absolute urge it has to become spherical...very awkward.
          Devs can surely point to a better solution but what I did is a reasonable first attempt I think...starting with spheres and with a gravity of 0.1 so they don't go mad initially.
          Then...and this is what I learned that PHX does brilliantly (I didn't know it could) I discovered I can just interactively switch off gravity (or in fact add/delete/tweak other stuff as it sims) at a certain frame to help the liquid get to its target
          more easily...and happily adding some fake magnetism to it.
          This ability is a revelation and a very cool feature indeed

          Something I noted with this test is that it was tricky to assess how many particles and at what resolution would actually end up being enough to complete the shape and have it look smooth enough.
          In my initial test I had an initially filled duplicate as an emitter, which subsequently was re-attracted by body force to its clone.
          This worked ok but the 'mercury' just looked wrong and was the wrong approach for what I imagined was required.
          The only thing that is useful here is that we can know the particle count and ideal resolution from the initial fill, so can subsequently do what I did for the eventual result, with that info as a basis.

          This newer version was trial/error by adding/deleting emitters until it looked like there'd be enough (not sure it was in the end, but close).
          For this the res was 49mil cells though only 105k particles were actually generated.

          Whilst playing with various shapes I also noticed that (maybe 'of course') it'll resolve to a single object volume but has massive difficulty doing so for some e.g. text, so multiple objects.
          This is possibly due to the particular surface tension needed for mercury. That's another problem though, that may or may not have a simple solution...multiple separate interacting sims maybe...
          https://www.behance.net/bartgelin

          Comment


          • #6
            This looks pretty cool
            Thanks for input

            Comment

            Working...
            X