Thanks for the hint on the script mate, that will make life a lot easier!

Just thought of anther process too to simplify life! Brain in overload to make this process of trees easier!

Thought I'd have a play with creating some new low poly trees!

I ditched my old tree factory and developed a fairly complex to follow yet quick and rigorous massing of trunk and branches in polys / components then the smaller branches, twigs and leaves in geo and alpha maps components added to the simple order of branch components!

Though this could be Tree Factory 2 it's a task many would not wish to undertake! Though that said I've worked a few ways to really ease improve tree modelling!

The quick skeleton and leaf tests!

The skeleton isn't mapped yet - no mean feet in SU! But should be quite correctly mapped in the end! No bump yet on the leaves but a quick test with crap normal map proved to create a lot of additional variation in the leaves. The rendered ones need to be broken like the foreground imitation. This is a Gum Tree so the leaves will look pretty beaten and eaten!

I haven't arranged branches, twigs or leaves to the boughs though amasing thousands of leafs per tree and copied 20x I could still navigate in SU though graphics would resort to bounding box on nav - though I can put up with that to have some very smart trees!

Brings to mind - does anyone have ideas to what other tree this type of branch shape may suit? Oh and does anyone know of a taper follow me script?

Oh, And the geo is cleverly setup to cast near correct shadows in SU!

@unknownuser said:

Can you expand on your item #6. IE which program is linking to the MXM file. Is SU linking to and MXM file?

Yes phil, I just used the clip map B/W as the map in SU and then linked the SU material to the MXM. The MXM is pretty simple just colour, clip and displacement maps!

God I've got to fix the colour and brightness of my laptop - thats dark!

Hmmm?

I tried something different with trees to test what possibilities might exist to improve the rendered output from billboard trees!

1. Exported Onyx tree model
2. Imported to Maxwell, applied bark an leaf maps,
3. Rendered high res, low render time, sky dome,
4. Exported rendered image, alpha map, depth map (z buffer)
5. Created MXM (maxwell material) colour map, clip map, depth map as displacement,
6. Created SU billboard with tree mapped and linked to MXM
7. Exported render.

Here is the result, it actually created I think far better depth than just flats.

The nice thing is the shadow cast on areas of the tree itself and thus the correct response to sun direction as opposed to normal billboards. And best of all low cost in sketchup.

@remus said:

Im guessing this is what is being referred to: http://playuptools.com/

Yeah just worked it out! Dumb is me!

@nintari said:

Check out the latest alpha of PlayUp. You can export to SB2 directly from SU now, although collision isn't in yet.

Sorry mate what is "PlayUp" and "SB2"?

@thomthom said:

Not sure how things work in Maxwell, but if you use V-Ray for Sketchup, which runs from within SU and doesn't export anything this is a good aid to set the HDRI position as there are no visual aid provided. VfSU simply asks for offset values.

Mate maxwell is the same, studio (the standalone app) will allow you to position your background visually, though reflection, refraction and illuminaton channels are all by trial and error and lots of both!

For many the best way to avoid the excessive testing is to use a chrome sphere in the scene to replace the subject, run a test to see on the reflections where the location is and then adjust to correct, delete the sphere and unhide all other for the render, to get the reflections good enough to see this can take some lengthy render times!

I'm just reversing this process to speed it up and get accurate results first go!

@unknownuser said:

Richard,

I'm trying to wrap my head around the benefits of your method here. So the advantage is that now you will be able to place your HDRI visually in SU before exporting? Are there any other advantages? I tend to do a decent amount of work within maxwell studio (lighting, hdr, even placing trees and cars), so it sounds like this isn't really something that I would use, right? Or am I missing anything?

-Brodie

Mate I will often use studio if placing trees cars etc. but when I don't need to I find direct export saves me that step.

I'm pretty anal about shadow cast, I think the right angle for the light source can add a great deal of depth, mood and contrast to a facade render so I'm trying to develop a way to know exactly the result I will get or that I'm after.

90% of the time I use clipped trees in the foreground hidden from the camera so that I get there shadow cast, but no the tree - this gives the illusion of a well landscaped context but allows for minimal obscurity of the building! But I like to know I wont over drench the building in shadow! So light source location can be crucial - I think it is for a good shot anyway.

In studio although you can test quickly in the preview it doesn't support HDRI so any tests and it will often mean many (remember I'm anal), but this way I should get direct precise export first go and with very little additional steps!

Part of what I'm doing will be to generate a few HDRI maps of my own for different lighting hues and whilst generating make the light source in the same location as my default grid map on the sphere so that it cuts out many of the steps and the hue varient map selection through the plugin will be the only step besides positioning.

Just thinking about it this would also allow the positioning of the reflection channel also!

@ecuadorian said:

@richard said:

No renderer app has a simple way to do this now

I guess you are referring to integrated render plug-ins only, because in Kerkythea this is done interactively with a rotate tool, so it's dead easy to match the light source with the sun in the background. I used it to match the sun in this video:
http://www.youtube.com/watch?v=h3cwW2xkloA

Kerthy may allow you to rotate it but then still requires testing to ensure yu get the shadow cast your after! Testing with kerthy possibly isn't to expensive on time but Maxwell and other high end apps really are!

[quote="thomthom"]I shouldn't matter how you map the faces. The UV Mirror plugin should be able to transfer them anyway.[quote]

And perfectly it did mate, thanks a bundle! Great job!

@edson said:

where did you find alan's globe?

Mate in the UVTools thread!

Here is what I'm trying to do! Create a tool for positioning HDRI illumination images!

@thomthom said:

That's very interesting on what you're doing Richard. Would you mind making a post when you've perfected this?

For transfering UV mapped textured from front to back and the other way around you can use this plugin: http://www.sketchucation.com/forums/scf/viewtopic.php?f=180&t=14684

But I'm not sure if that's what you're asking for? I see that the top of your sphere's texture is distorted while the bottom isn't. Was this what you where referring to?

Mate I'm just not sure that would work with Whaat's spherical mapping?

I certainly shall keep posted! I worked out a simple way to get the HDRI sun in exactly the same direction as the SU sun so that the SU sun can be used as a guide for shadow cast from the final HDRI!

UPDATE

@ecuadorian said:

So the renderer you use requires you to create a mapped sphere in order to use an HDRI? That seems rather unpractical.

No mate - like most they require offset or rotational values! So all required some trial and error! I'm trying to take that out, if I open the chosen HDRI in PS I can drop my grid over the top and position the yellow dot where the main light source is on the HDRI then export just the grid/dot image as PNG, reload it into my model and then rotate the sphere to where I want the sun (dot) to be then read the offset or rotation values of the grid at the red indicator!

No renderer app has a simple way to do this now and this way it is rather easy! I can just save the grid/dot image for each HDRI I ever use so I don't need to do it again! Just drag in the globe component, reload the correct PNG file, rotate, read the values and delete all!

I also work out a moment ago how I can match the dot location to suit exactly the SU sun location so I can use that as a guide for shadow cast.

Hey has anyone managed to map a sphere the same on both the normal and reverse sides? Would love one mapped with a 2:1 image if they have mastered this!

Trying to create a sphere mapped with a PNG grid, so that I can develop a component that can be used as a tool for getting the correct coordinates for offset values when positioning HDRI maps in renderers! See the attached poor example.

The idea is that an Grid with numerical values (PNG image) can be placed over a HDRI in PS and the yellow dot placed over the main light source as a reference, the HDRI image hidden and the file exported and reloaded into the sphere mapping.

The Globe can then be rotated where you want your light source to be and the values read where they correspond to the red indicator and these values then used as offsets in the rendering app to correctly locate the HDRI illumination channel.

@ecuadorian said:

I read somewhere that UVTools only works in triangulated spheres. Try subdividing it.

Yes thanks mate I download Alan Fraser's globe and it's working a treat!!! Thanks again!

Hey all!

I've come up with a way I think to workout the correct offset values for maxwell and the like for getting correct placement of HDRI maps without all the trial and error and will post the SU component and instructions when done!

Though for now I need to get a hold of a correct spherically mapped sphere in SKP format, mapped with a 2:1 image (say 640 x 320 eg), I've tried Whaats UVTools plug though cant seem to get correct results, can anyone help out!

@unknownuser said:

There's a ruby that lets you, amongst other things, set the 'tilt' and the height of the camera. I use it all the time -works just fine with MW! Really handy. Forgot what it's called, though, and can't go check right now.

Anyone else know what it's called?

Hmmmmm? Well isn't that interesting!!! I'd like that one too if you can work it out later mate!

I agree with Stinkie - for the two point!

Mate do you run a full render or just use the preview output for the base? I find sometimes I like the preview render just as much due to it's crispness - if you up you screen res to max and undock the preview window you can drag it to be full screen and get fairly good output and infact much better than most other renderers anyway!

Actually for me if you could animate from studio I'd be just as happy with the output from the preview which would be fairly quick and for video the frames don't need to be as baked as a still!

Frenchy - YOU DA MAN!