Yes mate nice job, makes me realise I really do need to invest some time in Podium!

Cheers, Richard

Sone updated images of Low Poly SU cars I’ve been working on for direct export to maxwell with only one material.

Still got a fair way to go but think I’m much closer to getting the maxwell materials right. Needed to pull some neat tricks with the materials and mapping to get it this far.

Mate I think I use the scale tool more than any other tool in SU! Seriously!

I've pleaded with the @last guys throught various beta tests to improve the functionality of the scale tool. my biggest wish is for non scaling textures and non scaling components ie: a window where the frame can be made of non scaling (user defines no scaling axis), then when the window is scaled the components of the frame stay the same in the thicknesses bu their length changes. I can see SO MANY uses for this like a garden bed of non scaling plants that could be scaled to fit its place but the plants themselves don't scale!

@gaieus said:

So you mean no matter what "kind" of cares they are, they will share the same (less important and apparent) components like chairs and such? Good idea! This way even though they count for about 500-600 kb each, when using more of them in a model, the bytage won't add up in reality.

(Or have I misunderstood something?)

Exactly i did this with my low poly cars, in those all the wheels, interior etc were all the same component just scaled to suit and the model size using multiple new cars onlt went up marginally.

Hmm actually thinking about it dam it won't work withou introducing new materials - what I wanted was only one material needed and one object per car, bugger! hmmmm? a rethink!

I have achieved a way with weight mapping in maxwell with hi poly that all cars in the scene can be done on one map and all cars together in the scene are one onject in maxwell.

@gaieus said:

Looks really nice, Richard! What would be the poly count (and the file size) of this beauty? (Just to be able to give a "real" feedback). Also I'd wonder what would be the file size without the texture (just as a comparison).

Mate the car weighs in at about 4000 entities and nude 500k, textured 600k! So mate I think fairly light!

Like i did witth my previous low poly sketch cars I'll make common components of the interior, break disks etc which aren't shown in the model above!

I've always had a bit of an ongoing personal project of getting really high quality cars at the lowest processing cost (low poly count, high render quality)! And I think I've finally cracked it!

Although as you can see I've still got a fair way to go this is just the progress so far! Please only consider the side of the car the rest is not mapped with high res images as yet. Rendered in maxwell with only one material.

This project has tested to the maximum all my skills in mapping in SU, photoshop skills and maxwell studio and materials..

I'm after feedback and sorry don't want to give too much away on how I have come up with the hellish process for these as after years and years of coming up with new techniques in SU I think I've finally found something I might be able to sell on as a bit of a hobby!

And today I really sat down and worked it all out!!!! I'll post some later work soon!

@unknownuser said:

Totally irrelevant: NIN, eh? I went to school with Trent Reznor's niece. She told me Reznor's "the sweetest guy".

Man that is because it was his niece you were talking to, my niece thinks I'm sweet - oh well then again so does everyone else who wears a dress! If only they knew the real me! - Muhahahahaha!

Mate I'm interested in your methods. Did you develop your own course structure or use one developed by google? I have negotiated a position teaching SU at our local community adultt larning college and did approach google for something of a course structure based upon their own model but with no response!

@kwistenbiebel said:

Snob ! (referring to Eeepc)

Nice to see the Boulder guys amazed by Whaats spherical mapping script.

[EDIT]: When I think about it, the Boulders guys should have given the correct 'expected' answer: "very nice scripting, but we already implemented it ourselves in SU7 together with other new mapping methods, modeling tools and faster viewport reaction on high poly models...you know, we are on top of things.".

Or WOW, lets give Whaat a really high paid job!! Or wow we really need to get on top of things obviously!!!!

i had a play with the plguin and found if i reduced the texture scale just a bit the mapping at the poles tended to clear.

Also exported to Maxwell with no issues!!!

Whaat again you do some absolutely amazing stuff mate!

One more here!

From the Sunny Central Coast (well not today)!

@pav_3j said:

even my parents are excited about this, and they can't even spell sketchup.

Pav, I just about wet myself mate!

Whaat mate you have done it again!

I was only whinging the other day somewhere here about SU needing spherical mapping. I'm currently wishing to use a sky dome with a spherically mapped sky that would be linked to an MXI file for rendering from maxwell though with the tools at my disposal now I have to do this through Maxwell Studio.

Cheers, Richard

I think the two mapping techniques being discussed here are different or should be! the rewrapping versus pinned deformation!

This function of wrapping and rewrapping is quite common to other modelling and texturing apps. Maybe someone with one of these could show example!

Beibel

Mate yourn the money again! I thought about this the other day, even for simple models like context buildings or for examples like your idea for baked textures from podium. A simple form could be built and textured in SU - unwrapped - image edited to add details such as windows, doors etc and then rewapped to include the added details. Or when modelling with photomatch the unwrapped image could be edited easily to take out detail such as plants etc in the photo as needed to clean up the model.

Seems to me like a must for direct integration of SU for google with regards to google earth purposes!

I must say this really drives home the general need for better mapping functions in SU like spherical mapping!

@jasonh1234 said:

Those trees are awesome. Gonna have to look up this Onyx prog.
I wish SU allowed for LOD placeholders.

Sorry to sound dumb! LOD?

I would have to recommend ONYX. I'm amazed at the variety of species you can model, the ease of use and then check out the billboard abilities where you can look at the tree from any angle or elevation to suit your scene angle, change the sun direction to suit and get the ground shadow layer too. Great for compiling scenes later in PS! The other beauty is once you have set the parameters for one tree you can just randomise new versions of the same species.

I don't have any poly count problems if i import OBJ format to maxwell studio from Onyx where I can map leaf images without issue but all that is lost upon import of 3ds or dxf to SU!!! Bugger!

I've asked if any of the ruby guys could invest some time in creating a script for SU that works to make a component of one leaf then search for any similar geometry and replaces that with this new component, this would be an absolute gem for many other components bought into SU. no takers yet though unfortunately!

@adamb said:

@richard said:

@jim said:

The biggest challenge is identifying the geometry for a leaf. My first thought was that if the leaves all have a surface area that is similar, it could make finding them a lot easier - you could find them by area.

Can you post a sample skp of an imported tree?

Maybe we can get you half-way there, if not all the way.

Thanks for responding jim!

I can't load up an onyx tree due to the user agreement, though the same effect would be buy grabbing any SU tree with component leaves and exploding down to base geometry.

I would imagine the script would either search for alike areas or simply alike geometry. The script would only need to locate one similar face and then select all connected and replace!

I would imagine the script wouldn't be so hard to do but could be a great time saver where ever a model contains numerous instances of the same geometry.

My 0.02 of your earth pence worth...

Area might be a good key, you might also want to group connected geometry, find its OBB (oriented bounding box) and then match on that. And of course then you'd have the necessary local transform "for free" too!

The biggest problem is that processing geometry in Ruby is spectacularly slow and if you're going to be using dense geometry you might be waiting a long time.. Oh but you do raytracing don't you, so you're used to that.

Adam

thanks adam!

Yes mate using the likes of maxwell I'm used to waiting LOL!!!

I think though that the benefits of this idea given SU's lack of handling large hi poly models without taking advantage components and groups is worth the consideration of this time loss early rather than later!

When I bring an onyx tree into SU the leaves are generally in 2 or 3 groups and need repainting. just exploding these three groups for regrouping into one can take my machine an hour now! And then I have no ability at any stage to map images to the leaves without doing them one at a time, 10000 leaves NOT A CHANCE!

@plot-paris said:

I am a great fan of drag-and-drop actions (even if I am no mac fan ).

a cool option would be just to drag a material and drop it onto another, which brings up a prompt if you want to replace the material with the dropped one...

cool, isn't it?

Very!

I ahve to agree with you too pete!

Though I think as one starts to venture into the area of photo realism the limit of your output (aside from skill and motivation) is the tools at ones disposal. Free by no means equals crap!

I take a look at Indigo and the development there by one guy!! And the development of the plugin by Whaat. You make a suggestion, he confirms it and then exceeds your expectation in it's implementation! And it's free!

I like to get specific (just a habit). I think if someone does address the tutorial for indigo - explaining mostly the terminology that differs alot from the norm and a preview engine is built into indigo as currently unlike it's main competition maxwell you need to wait quite some time before the render cleans enough to determine your success, then it will make it a rather hard one to beat!

@jim said:

The biggest challenge is identifying the geometry for a leaf. My first thought was that if the leaves all have a surface area that is similar, it could make finding them a lot easier - you could find them by area.

Can you post a sample skp of an imported tree?

Maybe we can get you half-way there, if not all the way.

Thanks for responding jim!

I can't load up an onyx tree due to the user agreement, though the same effect would be buy grabbing any SU tree with component leaves and exploding down to base geometry.

I would imagine the script would either search for alike areas or simply alike geometry. The script would only need to locate one similar face and then select all connected and replace!

I would imagine the script wouldn't be so hard to do but could be a great time saver where ever a model contains numerous instances of the same geometry.

Got to bring this back up for some attention PLEASE!!!