Show Us Your Bryce Renders! Part 2

1202123252650

Comments

  • GussNemoGussNemo Posts: 1,855
    edited December 1969

    @Dan: Like what you've done.

    @David: This color on the first two looks much better, though the lighting in the first one is a bit dark. The third image looks bleached, but at least the skin doesn't look plastic. The landscape and the one with the creature both look very good. As usual.

    @Dave: I like the first two light sabers better. The pink one in the second image doesn't look quite right.

    @eireann: That image is a real eye catcher. And I agree with David about the floor. And no, I don't need a new pair of glasses. I like your results. BTW, where did you get the plugin for Wings?

    @Rareth: Both of those images are very nice. One thing I noticed in the second image is where the walk(?) meets the water, on the left side. The edge isn't as clean as on the right. Still, both images look very good.

    @Mark: Ah, ok, I see now. I've not tried using HDRI's so I'd be no help offering suggestions. Still, on the right side of that particular HDRI it's distorted, the wall tiles don't line up. And it seems in the middle the tiles bulge upward. Is that way on your screen as well? If not then something is causing it to look that way. Yeah, finger food comes in many forms.

  • HoroHoro Posts: 4,347
    edited December 1969

    I'm not sure if that is considered sloppy work on the vendors part or this is normal when going to the opposite side from which the HDRI was meant to be viewed? I didn't really care that much because the image I made from it was more or less just playing around.

    I think it would be unfair to say it's a sloppy work on the vendors part. If what we see is the "back" of the HDRI we could conclude that it was acquired using a mirror ball. I started with that method and know the advantages and disadvantages of this method. As far as light is concerned, it doesn't really matter. It is only if it is used as backdrop when distortion and size play an important role.

  • eireann.sgeireann.sg Posts: 0
    edited December 1969

    GussNemo said:
    @Dan: Like what you've done.

    @eireann: That image is a real eye catcher. And I agree with David about the floor. And no, I don't need a new pair of glasses. I like your results. BTW, where did you get the plugin for Wings?

    .

    Hi Guss,
    Happy you liked the image,
    I got the plugin from Manifoldlab
    here the link for downloads of the plugin:
    http://s331378245.onlinehome.us/manifoldlab_releases/
    there ts also a link for daily releases which install Wings plus plugin.
    But the plugins on the URL given are fine. Just download the .tar.gz file and then in WINGS ...File...Install Plugin..

  • Electro-ElvisElectro-Elvis Posts: 258
    edited December 1969

    I developped this new and very useful item for any household. It not only looks good but also has excellent haptic characterics. In the background you recognize a early prototype with wings3d. :)

    In other words: One of my first attempts with Hexagon

    hexagon_thingie.jpg
    804 x 603 - 23K
  • RarethRareth Posts: 1,457
    edited December 1969

    [quote author="GussNemo" date="1353913756
    @Rareth: Both of those images are very nice. One thing I noticed in the second image is where the walk(?) meets the water, on the left side. The edge isn't as clean as on the right. Still, both images look very good.


    Well the Bryce image in the second picture is loaded in the background in Poser, the figure is kneeling on textured primatives, the effect was supposed to be her on a ledge/balcony over looking the water, but there really isn't any DOF to give more of that feeling, what I need to do is generate a depth mask of the Bryce scene then layer with the bryce image in photoshop to simualte dof, then use the resulting image as the backdrop in poser.

  • HoroHoro Posts: 4,347
    edited December 1969

    @electro-elvis - very nice household item. What is it good for? It is very nicely modeled. Never got the hang of Hex.

  • Electro-ElvisElectro-Elvis Posts: 258
    edited December 1969

    Horo said:
    @electro-elvis - very nice household item. What is it good for? It is very nicely modeled. Never got the hang of Hex.

    Thank you very much. I am afraid I have no idea what it's good for. It developped a life of its own. The only thing I know, if the red light is on, it is running.

  • pumecopumeco Posts: 0
    edited December 1969

    lol@electro-elvis

  • HoroHoro Posts: 4,347
    edited December 1969

    The only thing I know, if the red light is on, it is running.

    It is very important to know when it is running - whatever it runs. Perhaps it gives the owner a happy feeling because of its smooth form. If so, it should be on all the time. :-)

  • dwseldwsel Posts: 0
    edited November 2012

    dwsel_ said:
    Yeah, much more natural with shine! Thought I wonder if it fits the toned blue environment...

    Now time to test with different light scheme if it does behave well under different light (maybe boat scene again or some very contrasty non tonemapped hdr?). How well does the skin respond to purely diffuse lighting (flat white environment) and very strong single light without fills.

    I'm wondering mostly about the IOR if it will be the correct one that I won't let too much of light through the character. I took liberty to draw on your render to mark some areas. I really like the look of the ears. I think it's the correct behaviour to get reddish when the light shines through. Also the area marked in violet looks translucent, but I'm wondering if it's caused by the correct light. I'd turn off key light for a small test to try if that area is caused by the rim light or it's the effect of leaking light through the neck. If it was gone then I think that for such closeup Vicki would need some skeleton made from reflective sphere and cylinder placed inside the mesh (that will make the light return to the upper layer of the skin). The other solution might be playing with the IOR to change the direction of the rays. As for the nose and lips - I think that lips look enough reddish but it's not caused by light going through but rather reflection, but for the (inside of the) nose I think the effect is very subtle. I'd make Vicki even more transparent or add some skeleton so that light will return to the surface and light the thinner parts like lips, ears and nostrils.

    The problem I am having, beside the most obvious, is that I don't have a good "absorber" to put inside the model. Ideally I would like a volumetric cloud that followed the geometry of the model and soaked up the rays that got diffused into the core. Instead I've used a shrunken version of her own head. But due to the geometry of the model, it is difficult to get the absorber close enough to the skin while at the same time not intersecting the mesh and becoming visible. I tried using Wings3D to produce a "shrunken" version of herself, but the mesh was unmanageable with my limited skills at Wings.

    The IOR is a bit of a problem because only at refraction 0 is it possible to avoid showing the "seams" up in the mesh model. There is also a conflict between the specular halo setting I want to drive the specular effect and the setting needed for blurring transmissions and reflections. Three different controls would help at this point since the changes needed, while subtle, seem to have a significant impact on the plausibility of the skin.

    Here I've tried to apply the skin for the whole body, setting back a duplicate model away from the camera to act as an absorber and prevent rays travelling right through the figure. The problem being, because the duplicate model is not smaller but just further away, the depth of skin that can be entered by the rays varies and this creates lines where the absorber gets close to the surface. Like in the inside of the arm where a hard shadow has occurred.

    A skeleton is a good idea, but I think it would let too much light through the model. What's needed I reckon is a "fat" skeleton.


    Sorry for late answer, meh... I can't catch up behind literately anything recently.

    I see so you're already using some kind of blocker inside. Speaking about skeleton I meant in fact a 'skeleton' build out of primitives or simplified mesh as you've suggested rather than real skeleton which as you noted would allow too much light to pass through. Building simplified mesh would be the most helpful atm, but it's taking additional time. As for volumetric medium - it's quite close to reality, but inability to set edge softness is not helping in this situation. There's possibility to paint density map so that the volume would be thicker in different parts of body. The dedicated depth control would be easier to control though. To build skin that's good enough (also realistic water) there's at least the need for combining standard material with volumetric, or at least working absorption.

    The main reason why currently rendering is that slow is that often dedicated skin shaders are only an approximation of a single scatter (incoming ray going into medium in random direction, travelling for defined depth, going in 'given' direction phase and leaving the surface or not) , so 3 bounces might be enough for full simulation. Currently the combined shaders are cause of performing lot more redundant calculations with no real impact on the quality. Now, can explain what I meant by 'given' direction: direction of ray going out of the surface is the result of direction of incoming ray, IOR, diffusion strength (we try to simulate this with specular halo), phase controlling backward-forward scattering ratio (random direction with probability to go forward or return in incoming direction, we try to control this with introducing the dark, blocking medium inside - forward scatter, or bright or reflective back scatter). Making these controls joined tightly with the current shader model causes loss of control of the other basic features while trying to simulate the other material behaviour. Your suggestion to create a separate specular halo for reflection and refraction is a very reasonable feature request considering this case + would be useful in some others.

    The seam that you're getting with IOR different from 0 might be caused by reflection being joined with refraction combined with insufficient blurring for refraction after setting the correctly looking specular halo. I believe that IOR for skin is something like 1.33 as the skin is composed out of water.

    Coming back to your skin shader. I drawn the areas that have visible sss (green), areas where I'd more expect them to be (orange) and bonus double-fingers-area (blue). As for the head render (with key light + hdr ibl) I see that the light is not leaking through, but I'm still more more strength of sss (for nostrils and lips).

    Maybe one of the solutions is to render in passes?

    [edit]
    Oh.... and the nice thread about making realistic skin: http://forums.cgsociety.org/showthread.php?f=176&t=664371&page=14&pp=15
    on that page there's an example on which you can see the 'skeleton' I meant
    Warning - thread with lots of skin and nudity

    Post edited by dwsel on
  • David BrinnenDavid Brinnen Posts: 2,852
    edited December 1969

    dwsel_ said:
    dwsel_ said:
    Yeah, much more natural with shine! Thought I wonder if it fits the toned blue environment...

    Now time to test with different light scheme if it does behave well under different light (maybe boat scene again or some very contrasty non tonemapped hdr?). How well does the skin respond to purely diffuse lighting (flat white environment) and very strong single light without fills.

    I'm wondering mostly about the IOR if it will be the correct one that I won't let too much of light through the character. I took liberty to draw on your render to mark some areas. I really like the look of the ears. I think it's the correct behaviour to get reddish when the light shines through. Also the area marked in violet looks translucent, but I'm wondering if it's caused by the correct light. I'd turn off key light for a small test to try if that area is caused by the rim light or it's the effect of leaking light through the neck. If it was gone then I think that for such closeup Vicki would need some skeleton made from reflective sphere and cylinder placed inside the mesh (that will make the light return to the upper layer of the skin). The other solution might be playing with the IOR to change the direction of the rays. As for the nose and lips - I think that lips look enough reddish but it's not caused by light going through but rather reflection, but for the (inside of the) nose I think the effect is very subtle. I'd make Vicki even more transparent or add some skeleton so that light will return to the surface and light the thinner parts like lips, ears and nostrils.

    The problem I am having, beside the most obvious, is that I don't have a good "absorber" to put inside the model. Ideally I would like a volumetric cloud that followed the geometry of the model and soaked up the rays that got diffused into the core. Instead I've used a shrunken version of her own head. But due to the geometry of the model, it is difficult to get the absorber close enough to the skin while at the same time not intersecting the mesh and becoming visible. I tried using Wings3D to produce a "shrunken" version of herself, but the mesh was unmanageable with my limited skills at Wings.

    The IOR is a bit of a problem because only at refraction 0 is it possible to avoid showing the "seams" up in the mesh model. There is also a conflict between the specular halo setting I want to drive the specular effect and the setting needed for blurring transmissions and reflections. Three different controls would help at this point since the changes needed, while subtle, seem to have a significant impact on the plausibility of the skin.

    Here I've tried to apply the skin for the whole body, setting back a duplicate model away from the camera to act as an absorber and prevent rays travelling right through the figure. The problem being, because the duplicate model is not smaller but just further away, the depth of skin that can be entered by the rays varies and this creates lines where the absorber gets close to the surface. Like in the inside of the arm where a hard shadow has occurred.

    A skeleton is a good idea, but I think it would let too much light through the model. What's needed I reckon is a "fat" skeleton.


    Sorry for late answer, meh... I can't catch up behind literately anything recently.

    I see so you're already using some kind of blocker inside. Speaking about skeleton I meant in fact a 'skeleton' build out of primitives or simplified mesh as you've suggested rather than real skeleton which as you noted would allow too much light to pass through. Building simplified mesh would be the most helpful atm, but it's taking additional time. As for volumetric medium - it's quite close to reality, but inability to set edge softness is not helping in this situation. There's possibility to paint density map so that the volume would be thicker in different parts of body. The dedicated depth control would be easier to control though. To build skin that's good enough (also realistic water) there's at least the need for combining standard material with volumetric, or at least working absorption.

    The main reason why currently rendering is that slow is that often dedicated skin shaders are only an approximation of a single scatter (incoming ray going into medium in random direction, travelling for defined depth, going in 'given' direction phase and leaving the surface or not) , so 3 bounces might be enough for full simulation. Currently the combined shaders are cause of performing lot more redundant calculations with no real impact on the quality. Now, can explain what I meant by 'given' direction: direction of ray going out of the surface is the result of direction of incoming ray, IOR, diffusion strength (we try to simulate this with specular halo), phase controlling backward-forward scattering ratio (random direction with probability to go forward or return in incoming direction, we try to control this with introducing the dark, blocking medium inside - forward scatter, or bright or reflective back scatter). Making these controls joined tightly with the current shader model causes loss of control of the other basic features while trying to simulate the other material behaviour. Your suggestion to create a separate specular halo for reflection and refraction is a very reasonable feature request considering this case + would be useful in some others.

    The seam that you're getting with IOR different from 0 might be caused by reflection being joined with refraction combined with insufficient blurring for refraction after setting the correctly looking specular halo. I believe that IOR for skin is something like 1.33 as the skin is composed out of water.

    Coming back to your skin shader. I drawn the areas that have visible sss (green), areas where I'd more expect them to be (orange) and bonus double-fingers-area (blue). As for the head render (with key light + hdr ibl) I see that the light is not leaking through, but I'm still more more strength of sss (for nostrils and lips).

    Maybe one of the solutions is to render in passes?

    [edit]
    Oh.... and the nice thread about making realistic skin: http://forums.cgsociety.org/showthread.php?f=176&t=664371&page=14&pp=15
    on that page there's an example on which you can see the 'skeleton' I meant
    Warning - thread with lots of skin and nudity

    Thank you Dwsel,

    Faking the SSS with an absorber - there is almost an inbuilt function in DS I have discovered - through asking questions - http://www.daz3d.com/forums/discussion/12144/ - that very nearly gives me the absorber I need. Create a shell and offset the mesh inwards. Unfortunately, instead of vanishing, small structures, ears, eyelashes and so forth, turn inside out and grow inconveniently big.

    So, I'm still a bit thwarted. And of course, there is the render time.

    And I'm trying to formulate this into a tutorial. That can be followed. And produces reasonable results in a reasonable time scale.

    Progress is slow.

    I can currently get from the first image to the second image in about one hour. I'd like to cut that down though. The longest tutorial so far is about thirty minutes - and that feels like a long time... Probably moreso for those viewing.

    The second image does not include SSS. But it does include specular and anisotropic response, blurred reflection, two direct lights and some general IBL/TA driven lighting. As well as a HDRI backdrop to provide context.

    The figure's memory footprint is pretty large even with the resolution reduced genesis import. And at this size, it is at the limit as far as the mesh smoothing appearance is concerned. So... still a bit of a pigs ear. But we are around the cusp of a compromise. Render times soar to about 12 hours with the addition of SSS... The hair does not help either... So I don't know. That's the state of play.

    Thanks for the link to contemplate.

    Bridge_to_start1_test1.jpg
    850 x 850 - 633K
    Bridge_to_start1.jpg
    850 x 850 - 527K
  • pumecopumeco Posts: 0
    edited December 1969

    Horo said:
    Perhaps it gives the owner a happy feeling because of its smooth form. If so, it should be on all the time. :-)

    Just so you know, it's not a dildo, wrong shape ;-P

  • David BrinnenDavid Brinnen Posts: 2,852
    edited December 1969
    Bridge_to_start2.jpg
    850 x 850 - 215K
  • edited December 1969

    @electro-elvis: I noticed the red light, but it doesn't seem to be running. Not even walking I think.

  • RarethRareth Posts: 1,457
    edited December 1969

    actually that looks pretty good.

    I found an old article on faking SSS in skin using an incidence ramp which I think is a fancy term for calculating light vectors / angle of polygons

    "The Incident Ramp fakes a sub-surface scattering (sss) effect – probably the most important part of the node setup. Sss is where light enters the skin, and either pops out the other side, or is bounced back a centimeter or two from where it entered. The Incidence Ramp fakes this effect by assuming that as skin becomes parallel to the main light, there will be more light passing through the skin, and that light will be tinted a red colour. So as the skin polygons of the model become parallel to the main light, tint the diffuse texture red. Mathmatically, there is done by taking the dot product of the vector from the main light and the vector of the normal of the polygon. That’s the theory"

    so I started digging into the DTE to see if Bryce had any functions that would work.. I found the Orientation Filter and got the following results after playing around.

    atmosphere off, sun off, 1 single point light just moved to different locations

    bryce-sss-test3.jpg
    1754 x 875 - 417K
    bryce-sss-test2.jpg
    1754 x 875 - 489K
    bryce-sss-test.jpg
    1754 x 875 - 378K
  • Electro-ElvisElectro-Elvis Posts: 258
    edited December 1969

    dyret said:
    @electro-elvis: I noticed the red light, but it doesn't seem to be running. Not even walking I think.

    You can hardly see it, because it is running around its little glass sphere ;-)

  • RarethRareth Posts: 1,457
    edited December 1969

    more playing around with the orientation filter, this is interesting, the red tint is on the transparency,

    matlab1.jpg
    641 x 508 - 101K
    bryce-sss-test-7.jpg
    1754 x 875 - 229K
  • RarethRareth Posts: 1,457
    edited December 1969

    and one more

    bryce-sss-test-8.jpg
    1754 x 875 - 225K
  • RarethRareth Posts: 1,457
    edited December 1969

    last one I swear, not exactly SSS but an interesting effect..

    cylinder-test1.jpg
    1272 x 850 - 58K
  • David BrinnenDavid Brinnen Posts: 2,852
    edited December 1969

    Rareth - interesting yes, these are certainly close relatives of SSS. That you should try orientation filtering is something that had not occurred to me. I'm running tests with curvature filtering at the moment, trying to do away with the need for an "absorber" but the render times are still prohibitive.

    And some more unusual curvature material tests...

    Bridge_to_start1_test2.jpg
    850 x 850 - 251K
    Curvature_response1.jpg
    400 x 225 - 39K
  • David BrinnenDavid Brinnen Posts: 2,852
    edited December 1969
    Bridge_to_start2_TA1.jpg
    850 x 850 - 235K
  • RarethRareth Posts: 1,457
    edited December 1969

    David - wow that one looks alot better, very nice

  • David BrinnenDavid Brinnen Posts: 2,852
    edited December 1969

    Rareth said:
    David - wow that one looks alot better, very nice

    Thank you. Here is another experiment. Which is a combination of renders, one of which features a curvature controlled decaying copper kind of material effect.

    Bit_rusty.jpg
    850 x 850 - 302K
  • RarethRareth Posts: 1,457
    edited December 1969

    Holy Moly call the CDC she's been infected!!

  • RarethRareth Posts: 1,457
    edited December 1969

    ok I think I've reached my limit for now for hammering on SSS, last pic then I go chase another shiney..

    andy-final.jpg
    1754 x 875 - 336K
  • dwseldwsel Posts: 0
    edited December 1969

    Experimenting with different fillers

    13_133_c2.png
    900 x 300 - 397K
    13_133_c3.png
    900 x 300 - 402K
    13_133_c1.png
    900 x 300 - 394K
    13_133_c4.png
    900 x 300 - 436K
  • David BrinnenDavid Brinnen Posts: 2,852
    edited December 1969

    Interesting and also curiously appetising.

    Just for fun I've been playing with more metallic finishes.

    Bridge_to_start2_tiled3.jpg
    850 x 850 - 599K
    Bridge_to_start2_tiled2.jpg
    850 x 850 - 677K
    Bridge_to_start2_tiled1.jpg
    850 x 850 - 492K
  • GussNemoGussNemo Posts: 1,855
    edited December 1969

    @eireann: Thank you for the link. Now to figure out how to use it.

    @electro: That is an interesting object. Since no one knows what it does I believe it could be use for almost anything. As long as the light is lit.

    @David: Your new girlfriend looks much better with the different textures. Without them there appear to be striations on her legs and lower torso. Still, those texture tests like mighty nice.

  • eireann.sgeireann.sg Posts: 0
    edited December 1969

    GussNemo said:
    @eireann: Thank you for the link. Now to figure out how to use it.

    .


    The most interesting seem to be booleaning with the latest .tar.gz plugin
    and rightclick on empty scene...Plugins Incubator...Shape extruder.

  • GussNemoGussNemo Posts: 1,855
    edited December 1969

    @eireann: I'll have to give that one a try. It took these old eyes a moment to realize which choice you were referring to. I didn't think the plugin had installed properly, but it did.

This discussion has been closed.
Rocket Fuel