Dawn 2.0 Underway

[Miss B]

Ohhhh, Pugsley. I WANT Pugsley!!

 

[kobaltkween]

Hum, would you prefer to see more teeth morph options? Perhaps we leave the base teeth straight as general?

I'll look them over when I get to the teeth and gums. I'm planning to rework them on the new Dawn last, before UV work and morphs.

IMHO, the _huge_ thing with the teeth is to get them right for shading. I want to say that DS/Iray would be a good test for this, but teeth shaders are so difficult and my DS/Iray knowledge is decent but basic. I know that Blender/Cycles with Random Walk SSS would be good for testing, but I'm betting that's outside of your team's wheelhouse.

Lots of teeth are made _way_ too thick. Which is immediately evident if you have a good shader and renderer that should be showing translucence where they're thin and doesn't. Proper SSS is generally the difference between teeth looking real or like dentures. If the teeth are too thick, or even just too thick in the wrong areas, you can only get teeth that have way too much scattering or way too little.

I don't know if I know of any figure that's gotten teeth definitively right. I do know that when I was messing with Dawn's teeth materials, I found it really helpful to morph her incisors to be thinner. That could just be me, and could just be the fact that I wasn't using Random Walk SSS. I would suggest, however, that teeth, and maybe ears, get tested with scattering shaders _before_ their mesh is finalized, because you cannot get them to shade properly if they're the wrong thickness.

 

[Ken1171]

I know that Blender/Cycles with Random Walk SSS would be good for testing

You can adjust how deep SSS light scatters using the Physical Surface root node in Poser by setting the desired real-world distance in millimeters on the red, green and blue parameters - the last 3 at the bottom. For example, if you set SSS red to "1", light will penetrate 1 mm into the surface and scatter on the red channel up to that distance.

 

[kobaltkween]

You can adjust how deep SSS light scatters using the Physical Surface root node in Poser by setting the desired real-world distance in millimeters on the red, green and blue parameters - the last 3 at the bottom. For example, if you set SSS red to "1", light will penetrate 1 mm into the surface and scatter on the red channel up to that distance.

Technically, that's supposed to be true. I haven't found it actually works accurately in either renderer. But regardless, in the version of Cycles that is Superfly, how SSS gets calculated just isn't full on accurate. It involves the _entire_ mesh, even if your matzone with SSS is small, or underneath a transparent area. I reported this as a bug, and Brecht, the main developer of Cycles and its original creator, told me to just switch to random walk.

But even if it did calculate the right area, it's still just not entirely accurate. Look up the videos about random walk SSS in 2.8. They demonstrate the difference, which is most visible where meshes get thin. Mind, I find most renderers perform about like the less accurate method, so it's not like the lack of accuracy makes it relatively low quality. It's that the random walk calculations produce very high quality, at least to go by the tests I've seen.

 

[Ken1171]

Technically, that's supposed to be true.

My problem with SSS in Superfly is a different one. It adds TONS of noise to the render, of which it does a very poor job cleaning up. This has [probably] been already fixed in Cycles, but Superfly hasn't been updated ever since, or ever. I rarely use SSS in SF because of that, and when I do, it takes forever and I still end up with a noisy render. So yes, there are issues.

 

[kobaltkween]

My problem with SSS in Superfly is a different one. It adds TONS of noise to the render, of which it does a very poor job cleaning up. This has [probably] been already fixed in Cycles, but Superfly hasn't been updated ever since, or ever. I rarely use SSS in SF because of that, and when I do, it takes forever and I still end up with a noisy render. So yes, there are issues.

I think I've found that's mostly an issue of settings and lighting. All my images use Superfly with SSS. SSS hasn't been a big time or noise sink for me, unlike multi-layered, complex, transmapped hair.

Cycles got noise reduction some time back. IMHO, Filmic color management was a more important addition. I find proper color space absolutely necessary for realism, especially in skin.

 

[kobaltkween]

OK, testing now, and to be fair the performance hit isn't so much transparency as translucency. Which is like SSS in effect, but not in calculation.

 

[Satira Capriccio]

PUGS!!!!

 

[Ken1171]

OK, testing now, and to be fair the performance hit isn't so much transparency as translucency. Which is like SSS in effect, but not in calculation.

Let's leave this for another thread.

 

[Ken1171]

I can tell you that Paul is chomping at the bit to get his hands on this new Dawn 2.0. He is really digging what he's seeing. I too can't wait to get it to him to continue with his magic.

So here is just a small taste of the Unimesh concept with our Dawn 2.0. I've added in our Gorilla morph to her new mesh. It's still a bit sloppy and in need of refinement, but this is just a test. Check out those oblique muscles. Couldn't get that with our current Gorilla.

Oh I am eager too. Say, I was wondering - is there something specific in Modo that allows a morph to be transferred from the current Dawn mesh to the new one? If I asked 3DS MAX to do that, it would call me crazy.

 

[Chris]

Oh I am eager too. Say, I was wondering - is there something specific in Modo that allows a morph to be transferred from the current Dawn mesh to the new one? If I asked 3DS MAX to do that, it would call me crazy.

I had to add all figures and morphs that share the same mesh into one humongous scene file. It's quite cumbersome and slow.

 

[Chris]

Some arm work. Still need to do some thinning, and a few more triangles to work out of the mesh on the arm.

 

[Ken1171]

I had to add all figures and morphs that share the same mesh into one humongous scene file. It's quite cumbersome and slow.

So there is a way to do it in Modo when the meshes are different, like from DawnSE to Dawn 2.0?

Some arm work. Still need to do some thinning, and a few more triangles to work out of the mesh on the arm.

It's funny that I've got so used to recognizing Dawn by her thick arms, wide shoulders, thick neck and strong chin, and those are now finally being toned down. I remember when Dawn was just released, the very first critique was that her shoulders and chin looked large and masculine. I like this new shape much better. You're doing a great job, Chris! ^^

The suggestion of making the lashes a separate conforming figure seems odd at first, but that can be a huge time-saver when creating head morphs. First because we never want to deal with lashes when sculpting the face, and second because conforming items can automatically refit to the new shape. Even if the refit is not perfect, it is still better than doing it from scratch. On that aspect, it's a relief to know that the double lashes are gone. I don't think they were ever used, or at least not used enough to justify the cons. As for compatibility with CC3, it doesn't care for geometry - it only cares for the shape, and lashes don't change much from one figure to another. Conforming lashes are seamless in DS, but in Poser we would have to manually copy morphs to them. But the idea is still interesting nonetheless. What do you think?

 

[RAMWolff]

Chris, are we sticking with the T Pose as default with the info that was shared about reducing stretching for mapping using the A pose?

 

[Art_of_Mind]

I've just seen this today. I had lost my motivation to create but after seeing this I'm interested in making new content for sale again. I started learning and crafting when Dawn was new. When I seen this thread and was reading through it I remembered a thread from way back when; a comment where people judge a figure based off their first impressions of what they see when they first load it out of the box.
First impressions of Dawn?.....V4 is safe. - Page 7

I highly advise you to have her default form load looking ultra fine, fitness, dont hide details that need to be loaded, have her load all guns blazing.
Legs. Would love to see Dawn 2.0 step out of the box and walk all over all the other figures. She's got to have envious legs. I'm going to upload images of a current wrestler/bodybuilder that has mighty fine sculpted legs; great definition without the hulk effect. The knees. Man you have the geometry in place, you have mad skills, please sculpt us the most realistic looking knees any figure has ever had. Knees front and back.
Lastly, asymmetry/asymmetrical morphs for realisim seem to be getting popular.
For packaging I recommend the base version be a light version, cheap or free. By light version I mean you get the base shape and maybe a few alternate shapes; something like the base shape is fitness, with alternates of thin, thick, backward compatible original Dawn. Quick edit, another thought on a cheap or free light base figure, leave the texture off too. Going light on the base morphs will also speed up clothing creation, less morphs to add to the clothing.

This is all just friendly advice and suggestions, you are the Boss and your final decision is always respected.
I look foward to creating content for Dawn 2.0

 

[Ken1171]

Well, the DAZ forums are not exactly unbiased about competing figures, but some of the critiques there match my own, starting with the thigh joint's pivot being incorrectly placed, which led to most of the posing issues. However, they made the comparison against a JCM-heavy figure, taking that for granted as if it meant nothing. I wrote about that earlier in this thread, where there is a price to pay for the extra JCMs - they don't come for free. If they want to compare figures, it has to be on the same grounds.

They made some valid points on the default shape, though. Dawn and Genesis had more or less the same body proportions, but Genesis was the slender and more feminine type. They have noticed Dawn's thick arms and broad shoulders, which led some people to claim she looked manly. That is no longer an issue with Dawn 2.0, where Chris might make the arms even thinner, so that won't be an issue anymore.

However, it made no sense to compare Dawn to Gia and V6 because those are not figures, but character morphs. I suppose DAZ calls those as "figures" nowadays, but they are not. Maybe they just wanted to emphasize how "manly" Dawn may look when comparing to supermodel morphs for Genesis.

There was a lot of comparisons with V4, simply because she holds the title for the most supported figure ever - and that happens because she's the easiest figure to create contents for. Above they have dismissed JCMs completely when comparing G2F to Dawn, while V4 is the most supported because she doesn't require any JCM adjustments whatsoever. Good part of the backslash on the original Dawn was because of the default shape, the default textures (fixed on DawnSE) and the rigging issues. Shapes and textures are easy to change, but the rigging... is cast on stone. The thigh pivots are not at the right place, and no JCM can fix that. Even if I moved the pivots to the right place with animatable origins, the bones will no longer match the weight maps.

To me, the new thighs rigging will make a world of difference on this new Dawn. That's a high point for me. Ball joints like shoulders and thighs are the toughest things to get right, and I am reasonably happy with the shoulders - just not the thighs. This is the chance to get it right, which makes me excited about it. I don't take this for granted because I know how hard it is to rig a figure's thighs. But we only get 1 chance to get it right before it gets forever cast on stone. This time I have my hopes high.

 

[Chris]

Chris, are we sticking with the T Pose as default with the info that was shared about reducing stretching for mapping using the A pose?

Clothing conversion will be easier with the T-pose. So yeah, we'll stick with the T.

 

[Chris]

Showing a bit more work on the upper arm and shoulder. Comparisons with Dawn 2.0 and Dawn S.E. Look closely and you'll notice the refined and thinner shoulder and upper arm of Dawn 2.0.

Dawn 2.0

Dawn S.E.

Dawn 2.0 Notice the reworked edge loop flow around the elbow.

Will move to lower arm now, then the hand.

 

[Miss B]

Looking good Chris.

 

[Ken1171]

Dawn 2.0 Notice the reworked edge loop flow around the elbow.

I remember having posing issues with some of my body morphs on the back of the upper arms. That area is rather dense with polys on Dawn, and when I made the arms thinner, the dense edges started to overlap when posed because they were being pushed around by the weight maps. The result were frequent creases on the back of the upper arms when bent and twisted. That was rather tough to fix in morphs because making the arms thinner made the edges get too close together, creating the undesired creases when posed. I wonder if that could be helped?