how to get Bump textures to work?

[godsgifttoearth]

i'm interested in this program as its incredibly fast, I however cannot figure out how to get a bump texture to work. the material nodes are a maze!

i have rendered the below design using sketchup + indigo render. indigo render is incredibly easy to use and assign materials but not past this level of detail. it is also very slow.

can someone please tell me how i would apply a bump texture in the form of a png to the OBJ mesh, as well as a basic texture image.

also how can i adjust the camera properties to get an HDR image rendered?

thanks and sorry for the ignorant questions.

[FooZe]

Hi,

If you haven't already gone though them there are a bunch of good tutorials found here - they should get you pointed in the right direction.
http://www.refractivesoftware.com/demotutorial.html

Let me know how it goes.

Cheers
Chris.

[roeland]

Another forum member, radiant has also made a good series of tutorials, you can find it in our resources and sharing forum: http://refractivesoftware.com/forum/vie ... =21&t=7566

--
Roeland

[godsgifttoearth]

ok its taking some effort to get my brain working in spider diagrams but i'm slowly getting there.

this software is actually incredibly impressive, in terms of speed and the quality of the render you can quite quickly start producing. takes a lot longer to create the item you wish to render, but i easily eat that time up in rendering as small changes discovered later in the rendering can be corrected without wasting several hours of rendering.

i shall wait a bit longer to try out a few more revisions but i definitely will be purchasing this. great work!

how does one 'undo' linking a node or action they have made cntl+z type action? i keep having to delete a node and start again?

one question, in terms of GPU usage. do my gpu's need to be
1) matched
2) bridged via crossfire/sli link

i currently have a pair of nvidia GPU's, i also have a spare slot for another, instead of getting a 3rd of the same card i can pick up one of the quicker new gen models. will the program happily use all 3 un-bridged, all running at different speeds?

[mib2berlin]

It not needed to have matched cards but it is better to mix cards with Compute Capability > 2.0.

http://developer.nvidia.com/cuda-gpus

You can not use all features in octane with cards with Compute Capability < 2.0 especially in Linux.
SLI has to be disabled for Octane, so not necessary.

Cheers, mib.

[godsgifttoearth]

ok, i think i've got the basics now.

left hand picture, Octane 2xGTX460's - 1h45mins.
right hand picture, Indigo, 1xGTX460 (no multi-gpu) + 8threads on i7 @ 3.6ghz with 12gb 1800mhz ram........10! hours and still wasn't done.

Octane has picked out the fine detail on the plywood much better. The cone isn't lost to blackness. The cork hasn't turned out quite as well but that's more a texturing issue. Octane seems to prefer very aggressive bump maps turned down VS delicate bump maps turned up. Hopefully i will learn how to make normal maps soon and see if this improves. Speed alone is good enough to make me buy another g-card!

my impressions.

i would like a white balance feature on the camera, had to tune all sorts of other variables to achieve something that only needed a white balance alteration.

i would like the preview balls to have some kind of proportional representation to the OBJ mess dimensions you've imported. If i'm using centimeters on the OBJ, i'd like to have an indication of a texture slider being in overall dimensions not just a scalar number. Instead I have to scale once, work on the material, then rescale in the final render which isn't very accurate.

[FooZe]

Hi,

Thanks for the feedback and your confidence in our product!
There are a couple of notes for getting new cards...
1) RAM won't stack (you will probably will have noticed that already since you have 2 cards). The scene limit is the lowest of the cards you are using to render (since the scene needs to be on all cards). So if you buy a card with a lot more ram than your current cards it won't all get used.
2) In Octane, the fermi core cards (5xx series) are still currently out performing the kepler cards (most of the 6xx) series. This is one of the top things on our list at the moment - addressing this and seeing what can be done.
3) Octane does not support ATI cards - it is NVidia CUDA only (you mentioned crossfire so just wanted to clear that up).

For the texture scaling - i have found it usefull to temporarily disable the material preview while i'm sorting out scale (and actually for quite a number of other reasons). This is the "Enable/Disable viewing of material nodes when selected" button on the Graph editor.
That way i can change material settings like texture scale and see the results on the scene, rather than jumping too and from the material preview.

Cheers
Chris.

[godsgifttoearth]

ok,

octane is definitely running quicker with Sli enabled VS not. some 0.2-0.3Ms/sec quicker with 2x GTX 460 768mb version.

Pci-e bandwidth problem?

[FooZe]

Hmmmm, strange.
Once you have disabled SLI, are you making sure both cards are selected under File -> preferences -> Cuda devices?

Cheers
Chris.

[godsgifttoearth]

Hmmmm, strange.
Once you have disabled SLI, are you making sure both cards are selected under File -> preferences -> Cuda devices?

Cheers
Chris.

yep. its also far less choppy with both cards running with SLI enabled vs not.

if this hasn't been reported elsewhere i'm just going to assume its a hardware problem my end. had a few crashes too, there doesn't seem to be a bug reporting system for me to submit reports?