Maya nHair - how hard to repro?

[Kerro Perro]

Hy guys i rarely look back at Maya but am quite impressed by the presentation of their new hair solver:



Now of course as usual it might not work so hot in practice but there are a two things i really llike:

1- the ability to generate hair from geo

2- the way it seems to hold it shape during simulation

Now i've played with Melena a little and especially 1 is nor possible afaik. This stuff might be do-able with Kristinka but as much as i think Anto's a great guy i find his stuff near impossible to comprehend...

EDIT: hmm embedded video isn't showing up in my browser: http://www.youtube.com/watch?feature=pl ... eWisUaz3bE

Moderator edit: Yes it does, if you use the tags correctly... ;;) - HB

Moderator edit, continued: Hovering over the tag name in the editor gives a short explanation,
more info can always be found in The tag FAQ - HB also, obviously

kp edit: whoops yeah i get now, sorry HB!

[Rez007]

The Maya nHair does look pretty impressive.

[Bellsey]

Essentially its now the Nucleus solver that's driving nHair so it improves things abit.

I'm waiting to see some of this stuff though ;)
http://www.3dworldmag.com/2011/08/09/au ... echnology/

[Kerro Perro]

Essentially its now the Nucleus solver that's driving nHair so it improves things abit.

Hmm yes but i really like the way the geometry seemed to generate a volume of hair - been thinking about it and this must be do-able in ICE: if you can fill a volume with particles why not with strands? I should look into this even though i have no need for it at all atm ;)

I'm waiting to see some of this stuff though ;)
http://www.3dworldmag.com/2011/08/09/au ... echnology/

That sounds even more interesting!

[mattmos]

If you check out this guy's demo reel, he uses geometry extrusions and point indexing to generate a strand volume. Very cool stuff.

http://www.youtube.com/watch?v=579TxZhgapQ

[Mathaeus]

Hmm yes but i really like the way the geometry seemed to generate a volume of hair - been thinking about it and this must be do-able in ICE: if you can fill a volume with particles why not with strands? I should look into this even though i have no need for it at all atm ;)

Since ICE modeling in SI 2012, as well as SDK improvements in... 2011.5 I think, UI widgets an so... it should be *more* doable than before. I had something like, done even only with ICE in 7.01, but it was a whole small system, a lot copies of this or that just in order to work. Not that much for 'comprehend'
Definitively not just a fill in volume, decision about direction and distribution should be done, too. And this decision could affect the final look, dramatically. Achieving the acceptable final result, could take a lot of time, re-work of entire concept, so on.

Now someone *just* need to do that. I wouldn't, I don't need this...

Simulation engine looks interesting, don't know is it possible to do something like, using Syflex on curves, Momentum or so..

[origin]

mesh based hair gen is cool, but if you worked with hairfarm for example, you know how annoying it can be...

[Kerro Perro]

@Mathaeus: Yeah as you told me before i should just get more familiar with "basic" ICE, it's just hard sometimes to study something so general - i tend to learn by crisis if you know what i mean ;)

Like the fact today i needed to make dynamic cables so i did some googling and found syflex curves > deformed the cable geo by curve > tada!

Perhaps i'll take on the strands volume someday but as said i dont actualy need it either, I'd probably try to tackle the "ICE greeble" thing first - been wanting to do that one for a long time!

@Origin: Really? Why?

[Rez007]

Essentially its now the Nucleus solver that's driving nHair so it improves things abit.

I'm waiting to see some of this stuff though ;)
http://www.3dworldmag.com/2011/08/09/au ... echnology/

That does sound great! Bellsey, I don't know if you have contact with the decision makers higher up on what new technology goes to what platform, but if you want to mention to those people, that if Autodesk wanted to really sell a lot of Suites, then it would be the smartest move to put that XGen into Softimage first - that would sell the most Suite seats -fast. If they put it into Maya first, that would do little for the sales of Suites.

[Bellsey]

I do have contact with people in the product teams, but whether they listen to anything say, well who knows, lol.
However, in that article Disney do say that Maya is at the core of the pipline.

[ace63]

Well Softimage is in dire need of a new hair system so why not take this stuff?

[Rez007]

I do have contact with people in the product teams, but whether they listen to anything say, well who knows, lol.
However, in that article Disney do say that Maya is at the core of the pipline.

Yeah, that is true. I guess it is wishful thinking on my end, but my logic, I think does work the best if Autodesk wants to sell more Suites. I am just saying for the first year or so, then bring it to the other 2. I highly doubt that putting this in Maya first is going to sell more Suites, if that is what Autodesk is aiming to do. Anyway, I guess with this article and the recent departure of most of the Softimage team to the newly created Maya VFX team, I think one can put the pieces together on what they might be working on next. I just wish it was for Softimage...

[Pancho]

Well Softimage is in dire need of a new hair system so why not take this stuff?

I don't think so. Since the arrival of ICE and strands it isn't too hard to come up with your own hair solution. I'm currently on my way to do so and I must say it comes along quite nice. So I'm not a computer nerd.

Though you are right, as not everyone wants to/has the ability to look into ICE, a basic hair solution in ICE would be nice. But Melena and Kristinka are there. Why don't you go for them?

[milanvasek]

also, there is another hair system on the way:
http://vimeo.com/41125869

but i really like that idea of putting the cool stuff into Softimage first to make all those Maya and Max users buy premium or ultimate suites it makes perfect sense from autodesk (greedy) point of view

[ace63]

Well Softimage is in dire need of a new hair system so why not take this stuff?

I don't think so. Since the arrival of ICE and strands it isn't too hard to come up with your own hair solution. I'm currently on my way to do so and I must say it comes along quite nice. So I'm not a computer nerd.

Though you are right, as not everyone wants to/has the ability to look into ICE, a basic hair solution in ICE would be nice. But Melena and Kristinka are there. Why don't you go for them?

I am pretty familiar with ICE but building a hair system with it seems pretty cumbersome to me.
First - you don't have any proper styling tools. I like the styling tools the original Softimage Hair has - this is one of the biggest pros of it.
Furthermore, you cannot generate strands at rendertime with ICE - you have to generate and simulate all of your hair in the scene, which
results in massive performance loss.

[Mathaeus]

First - you don't have any proper styling tools. I like the styling tools the original Softimage Hair has - this is one of the biggest pros of it.
Furthermore, you cannot generate strands at rendertime with ICE - you have to generate and simulate all of your hair in the scene, which
results in massive performance loss.

well, if 'render time' means entire strand generated by renderer, Let's say, something like King Konger for 3delight, nothing in out-of-the-box SI can do that. It's only about viewport display, and ICE equivalent of 'render XSI hair in viewport' is 'particle display' property, automatically attached to any ICE point cloud.

Of course, would be great to have XGen implementation in Softimage, but.... definition of 'artists friendly' by Disney... does it means the same from forum life, 'I want it simple, even on the price of poor final result, even I can't show anything well done with that simple', I don't think so. I'm pretty sure, Maya users will have chance to enjoy the complexity.

Regarding Kristinka Hair, as it is already mentioned in first post, usually I have some free time during June or July. So , for this year, general plan is:
1: refining of existing stuff, simpler and more efficient interface here and there, a few new nodes. Let's say, version 3.1 or so ( current is 3.0).
2: tutorial. Nothing spectacular, more like all exact steps to get some particular hairdo.

Cheers