Show Us Your Bryce Renders! Part 3
This discussion has been closed.
Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
Cris333, I meant to say I like your pyramidal render. Although remember dof does tend to make big things look small.
Back at the ranch...
And my bad puns about being cool with poles...
Spherical maps by their very nature do tend to map well onto spheres. However, they don't always look so good on flat surfaces, in the last image, I used one of the mapping options to improve tiling.
The previous crude experiments with arrows suggest however there are patterns which can map onto infinite flat surfaces and spheres. I don't know how this subset is mathematically defined, but I know how to spot them. That next.
At this point, due to the last image I got diverted into reflection maps... but that's straying into deep waters and I figured that will have to wait for it's own tutorial.
Then I got to pondering what Graham Dresch showed us the other day, that is that displacement will work most of the time if the priority is set to low. So how does displacement respond to spherically mapped images? On a sphere...
There's good news and bad news.
Good - it works.
Bad - sort of.
First image looked promising, but on closer inspection something was amiss.
Second image shows just the displacement.
Third image shows the displacement with the displacement spherical map mirrored - this seemed to look like a better match for the pattern in the first.
Fourth image is the image from the first and the displacement from the third. This mirroring issue with spherical maps is not uncommon since all the methods I can thing of for getting a spherical map of something require you to put the camera inside the object being mapped. So as with the globes shown before, the earth map gets mirrored when you generate the spherical map. Either by compositing squares or using a handy purpose made tool Spherical Mapper
And aside from Graham, what got me pondering this? The last image shows a pattern which can be successfully applied to both flat surfaces and spheres and cubes and goodness knows what else, without discontinuities appearing.
But this again got me thinking about bump. Bump is after all much faster than displacement.
Image one shows how they compare - it also shows that spherical mapping displacement onto a cube causes the pattern to go out of phase with the displacement. So that's not very good. But beside the point. The bump mapping is also wayward. Some bits which should be out are in and the other way around.
Another of my mapping bug bears. So my final two images, to test different lighting models to see if heightmap mapped bump is any better or worse.
First obscure light. Hmn... no.
Then HDRI, both no for the inversion issue and also due to only accepting 8 bit height maps instead of 16 bit there is also some ugly stepping in the bump response. What is needed?
Normal maps and some lovely parallax mapping - to act as a halfway house option between bump and true displacement.
OK so not the most delightful selection of renders, but I hope it gets you thinking and experimenting for yourselves.
Cheers,
David.
Wow David is at it again, nice renders.
oh fyi I found another fractal toy to play with
http://www.incendia.net/
Look at this artists gallery, some good incendia there - mostly all rendered in C4D.
http://nic022.deviantart.com/
incendia object rendered in Bryce..
2nd Incendia object rendred in Bryce
Very nice, David. And Rareth, looks super cool what you're doing with the space stuff!
Thanks I used to do DKBTrace 2.x ray-tracing of glass items on an Amiga A500 back about 1988. I made a Christmas tree with a blob, and then added glass and metal spheres to it. Maybe I'll have enough time to do that in Bryce this year.
If I'm not mistaken, image size doesn't matter, but file size does. 399,999,999 Bytes or 381.4 MB. Nice picture, by the way.
@GeroblueJim - interesting picture.
Thanks !
And the lesson to be learned here is; Don't pick your lizard up by it's tail! :cheese:
Those are great objects, I can see loads of uses for the fire one.
oh there are loads of uses for the fire one,
Thanks for the heads-up Rareth. This looks great!!
@Rareth: The last image is very interesting. Can't decide if it's a creature best left alone or an art object. Love the look.
As usual all the renders are great.
David all your experiments are very encouraging. I love the underwater scenes something to look forward to when I get to that tutorial.
Rareth – The Shadows render is awesome. I’m a bit confused, the render took 9 hours and there are still white spots, I think David refers to them as flies, was this a premium TA render.
Still learning from Bryce-info, at a snail pace.
I combined David’s skyscraper and volumetric cloud tutorials (Bryce-tutorials.info). The height map for the building was done in Photoshop as I did not succeed in the DTE.
@Rareth - very good idea to put this - whatever it is - into a torus.
@mermaid010 - nicely done.
those white spots are from the Celestial and Stars turned on in the sky lab.
skyscrapers in bryce take a decidedly different approach from doing terrains, if going with the DTE,
I find its a bit easier to just use primatives and textures if making them in Bryce,
here is a quick 5 minute thrown together building.
A new image... just a bit of something to keep moving forward. A little bit of photoshop, but everything here is bryce, including the atmospheric bits. The render time was a short 6 hours, with two light sources. One of the "suns" had a squared falloff, which probably complicated the render time.
@Rareth - for such geometric buildings, I agree with you that "boolean-ing" them together is the simpler means.
@M1chaelFrank - good to see you back. Another outstanding render from you. In fact, it looks rather like a painting and not like computer graphics - and this is excellent.
Thanks very much, Horo, I really appreciate it. I'm still quite interested in taking the "shine and polish" off of my images; retaining a painterly quality. Of course, painters frequently employ a great deal of techniques to physically handle or manipulate their canvas, and I'm always keeping that in the back of my mind when making my digital images.
@David: Really like your experimental results with mapping. I especially like how the use of TA optimized radial lights made the last image look as though the objects were sitting on a quilted cloth of some kind. Nice effect.
@mermaid: That looks really nice.
@Rareth: Like the look of your building, though I had to chuckle at the 5 minute time you stated. It'd be 5 minutes once you know what you're doing.
@M1chael: That's a very interesting image, plenty of things to study. Until Horo mentioned it, it does look like a very good painting. Nice work.
not if you approach like building blocks, its all cylinders and blocks, no booleans used in this one, the longest time was spend tweaking the window texture (it was a preset, came with Bryce) so it fit properly. build time was right around 5 minutes give or take 30 seconds :)
Thanks very much!
something different from the space images.. the outcome of randomly clicking on things with structure synth.
Ooh, look. I take time to go away and play in Bryce isntead of constantly monitoring forums, and M1chael sneaks another new image in. And great as normal.
Beautiful work Michael, somehow it looks like something being displayed on an old CRT, like something out of Quatermass and the Pit. Which means I like it a great deal.
OK, well here right about as far away from fine art as it is possible to get, I've been grubbing around in reflection mapping and trying to make some sense of things with varying degrees of success. I will have to make a video at some point, but here are the results of my findings for the curious.
First image v0 - shows from left to right, spherical, parametric, reflection mapping and finally true reflection. Clearly there is some promising distortion offered by reflection mapping but... it's clearly not reflection.
v1 - On the left reflective wings cube and sphere, on the right reflection mapping. What is obviously an issue is that the reflection map mapping mode does not take account of the incidence angle between the surface normal and the camera - that's one issue. The other is that if a ground is included that's not part of the map.
v2 - here's a closer look at the sphere. So what I need is for the reflection map to look like this reflection. Now I can't capture a reflection directly from inside the sphere. If I move the camera inside the sphere it will be dark. If I make the sphere transparent. I won't be able to see the reflection and anything inbetween will be just internal reflections. So... how to get this incidence angle accounted for?
v3 - this is v2 object front mapped onto a sphere. This is not a reflection, this is a map. Which means, I can now move the camera and the area's will stay put. Indeed, I need to move the camera inside. (the variation between the v2 and v3 is accounted for by perspective, this could be somewhat corrected for by moving the camera further away and narrowing the FOV - but in the greater scheme of things, since this is all going to be a matter of approximation, we will just press on OK).
Final image in this set of five is a spherical map made from inside v3. I used my trusty Spherical Mapper but if you don't want to part with any money, you can achieve the same results by assembling six cube faces. Horo has a tutorial on his website. It's a bit fiddly that way, but the results are slightly more accurate. Then again, re my comment above. Ease of use wins out for me.
As always, a pleasure to hear from you... Many thanks!
This image has to be mirrored and offset to compensate for a) being a map of inside the of sphere and not outside of the sphere and b) mapping rotation convention within Bryce. This gives v5. Which looks more like reflection I think than v1. So... now we have a method, time to test it in a more serious way.
v6 - is the equivalent of v2 and going through the process, this gives rise to
M_on_map_sm2_mirror_offset - our incidence angle corrected spherical map reflection of one of Horo's fine HDRI's.
v7 with a slightly reflective ground shows up further disadvantages of this method. Reflections of reflection maps do not look right (something that's going to be hard to get around since this introduces a new incidence angle) and that half the ball is blank (a problem that is far less obvious on the more complex geometry).
M_on_map_sm3_mirror_offset - a work around that just mirrors the front half of the ball onto the back - this is the spherical map for that.
This gives rise to v9 which doesn't look too bad, but the underside of the ball is still a bit suspiciously blank. The idea being, If I can get the ball looking right, the cube will follow.
What about capturing the other side of the ball as seen from the floor? Now that's an interesting question is it? I mean, in real life, that would be impossibly, but thankfully, we are in Bryce land where the physics is iffy - which can be bad, but it can also mean it can be exploited. So here is an exploit - a one way mirror. I swapped the camera around to look at the opposite face of the mirror ball, made the infinite plane into a perfect one way mirror (more details on one way geometry can be found in the Bryce Mentoring DVD ) and rendered the far side. v10
This was then combined into a new spherical map. M_on_map_sm4_mirror_offset
And mapped into the objects on the right. v12
The final test. A more complex scene using this map, v14 - the cubes are not reflective - if they were the render takes 4x longer!
So that's what I've been pondering.