dforce Hair. It's risen after simulation.

Sci Fi Funk

Hi,

I have applied dforce to the non-dforce hair - Zofia Hair. I split the wieght map into a skull cap, and rest of hair. The skull is to stay in place and the rest of it is to blow in the wind.

I'm happy with the simulation, in that it flutters around in the gentle breeze.

However it is raised up now as a whole. I include screen shots of before and after running the dforce simulation, plus the settings I used including the wieght map.

Does anyone know why it has raised up?

Thanks

[EDIT] - I should add that this is G3 Hair on a G8 character. It auto-fts fine, but does dforce bypass / ignore auto-fitting?

Richard Haseltine

I would imagine that the Collision Offset value is causing the mesh layers to "lift and separate".

Sci Fi Funk

Richard Haseltine said:

I would imagine that the Collision Offset value is causing the mesh layers to "lift and separate".

Ok thanks very much - I'll make some adjustments and try again tomorrow.

Cheers!

Sci Fi Funk

@Richard Haseltine Well the collision offset was part of the solution, but for anyone reading this thread, I think the other part of the issue was my simplistic approach to separating the hair into scalp and the rest of it.

This might work for some hair in a no wind situation, but when wind is applied, with this carefully shaped hair, it just won't cut it (pun intended).

One of the issues is that I've lost the parting. So I'm going to have to seperate that and I believe the sides around the ears into seperate surfaces - otherwise I've lost control of the shape:- I.e. If I drop the collision offset value, the sides are OK but the crown is flat, and if I keep it at default, the crown is fine, but the sides balloon out.

 

Sci Fi Funk

I ran a dforce simulation on the hair without wind, no other edting - it's still the simplistic approach of scalp (not reacting to deforce), and rest of hair (progressively reacting to defoce down to the bottom tip with max reaction).

At first glance, the hair appeared to hold the original shape (default collision offsets), so I concluded it's how the hair reacts to the wind.

Then I tried a version of the scene with wind that is out of range. When I looked closely i could see that although it's similar to the original shape, the original hair is stylized to curve to the right (facing the camera) and when a dfoce weight map is applied it logically straightens the hair out.

 

Sci Fi Funk

Adding to my previous comment.

Now I'm trying to figure out - how do you apply a weight map that doesn't destroy stylized hair - i.e. hair that is sprayed in real life to defy gravity? In this case the ends of the hair swerve to the right, otherwise it seems pretty logical (in terms of gravity).

I'm looking for a way to figure that out, before I get back to tackling the main issue, preventing the wind node puffing out the hair (when it's in range of the wind).

Anyone faced this in the past?

Sci Fi Funk

Well I've spent all day going through each parameter and I've had to reach a compromise. In the end @Richard Haseltine's suggestion of changing the collision offset made the most difference.

I don't understand why changing it in the opposite direction of what would have made sense to me, actually worked the best (still with compromise on shape), but it did. By changing it from the default of 0.2 to to make the hair more prone to stand off each other, it kept shape. I thought I had to decrease it to bring the hair down (that was the first thing I spotted that seemed wrong), but making the hair more stand offish kept shape better. Pehaps it wasn't that high in the first place, perhaps it was more of a shape issue. It seemed higher though.

The final parameters then that differ from the defaults are:-

Friction 0.8 (probably made the least difference)

Collision offset 0.5 (kept shape the best at the bottom and top of the hair, at a cost of slightly puffing it out on one of the sides). - Dont forget I divided the hair into scalp vs rest and put a smoothed increase on the lower part of the hair.

Dynamics strength 0.45 (but this can stay at default if no wind, and perhaps If I had moved the wind around better it could have still stayed at default).

Density GSM - 20 - I had it at 21. I understand hair blows up <20 but I've not tried this.

I may return to this in the future, but I have deadlines to make so this is the best compromise I can get right now.

Please feel free to add any other suggestions. It's Zofia hair which needs to be made into dforce hair first.

Lastly, I tried a few different wieght maps, but the simplistic one worked best after all.

benniewoodell

I'm going to have to keep working on trying to add dForce to a non-dForce hair, once in a blue moon I can get it to work where the hair doesn't just explode immediately or fall off the head within two frames. I have the weight map set up too where the top of the hair is gray. I can get it work in Blender just fine weight painting so it should be similar, but I don't know. This is definitely giving me a little bit of a different insight so thank you for giving the updates on what was working and what wasn't! 

Sci Fi Funk

@benniewoodell I hope it helps, but an update is that it seems to be rendering out the same as it always does. So I'm back to square one.

Ugh. I'm pulling my own hair out on this one. :(

 

benniewoodell

Oh no! Keep us posted on if you figure out the culprit this time. I'm hoping for the best for ya. 

Sci Fi Funk

Cheers man. Will do.

charles

If you want absolute full control of your hair you can't rely on dforce, you have to sculpt it in blender (excluding strand based dforce hair). For animation I've come up with a new trick (for me anyways), make a series of scultped morphs and transition the morpsh during the animation sequence.

Sci Fi Funk

@charles OK thanks - at least I know it's not me going mad. I don't get on well with Blender, but I do have the Dforce to Morph script which I haven't used yet (just getting used to dforce first) - so perhaps I can use that.

UPDATE  - If I render my animation with a script in the viewport - it's as expected (so near to what I want) - If I render normally it's the risen inflated hair! This is not the first example I've come across where viewport rendering is correct and the main render is not.

So at the moment my conclusion is you take each piece of hair on it's own merit to see if it will play nicely. I don't like unpredictabliity - that means a risk of a missed deadline.