Posts by TeddyBear

I like it, nice job.

There are two things that looks out of place to me - the low-res road texture and the rock (maybe it's the normal map?). Other than that it's smashing darling, smashing!

A half days work so far, no normal map for this one but I think it can stand on its own two legs.

[Blocked Image: http://img381.imageshack.us/img381/8003/head043vq9.jpg]

I really dig it!

Thanks

Ah... yes, the results are quite alike. However, the methods differ alot. For instance, Debevec makes use of equipment worth great sums of money and big studios, my method makes use of a cheap digital camera and nothing more. Also, I'm quite excited about the new addition. It is a program that reads any displacement/height/bump/whatever-you-want-to-call-it map and calculates a well structured model of it.

This can of course be applied to more than photographically generated displacement maps. With this tool you can for example simply paint structures that are complex to model and then convert the texture information to a model with a highly tailor-made structure.

It needs shadows, looks like it's floating. The lightmap scale on that house wall needs to be smaller as well. Otherwise great as always

killer & d3 - I really like it! There's just something about HL mapping.

Thank you all very much! (I'm finally able to answer, mapcore's been down for me all day)

Quote from PS_Mouse

While that is goddamn awesome, I understand the technique you used to generate that normal map uses a video camera. Where would that leave you when you're generating the normals for an object or creature that isn't real? Would you use clay models, or an entirely separate technique?

Well this is where things become interesting; this is where the creative side comes through. As replicating reality tends to be mechanical, creating concept models allows for interesting ideas and associations. Then I usually think about objects that I can combine and use for the construction of my unreal object. When I've gathered all normal information it's all about photoshop manipulation which is a challenging and fun task. Clay modelling is an alternative but then you will get unrealistic texture structure.

Quote from Zacker

If I understand correctly Teddybear, then that head model and texture maps are all done by a script?! If so, GIEF!

Well, besides from the photoshoot which is far from scripted (), yes, it is.

In fact, I thought maybe you guys would be interested in this:

Me and a friend of mine have been developing an algoritm (since the lunch break yesterday) which loads a heightmap and analyzes its volume and structure. From that analysis it is able to construct a perfectly symmetrical edge looped model with a polygonal budget specified by the user.

We were thinking about selling the idea to whomever was interested. Combined with the height map scan method you can easily instansiate real world objects as computer graphical models. Combined with the normal mapping technique you have a precise replica of the real world object.

Thanks alot guys

Now with diffuse! And animated!

http://media.putfile.com/Fabian

Original res 2048^2. The diffuse is pale because it's in the process of being zombiefied

Application slots closed.

Normal + parallax (latter just for testing, not very accurate)

[Blocked Image: http://img381.imageshack.us/img381/4935/renderme8.jpg]

As I've written a script for this it was done in 15 minutes total (photo session + selecting + photoshopping + processing).

Well at least I really think the HDR gives the wine cellar the right touch (but it still has overexposed spots that need fixing).

So warby - your map is one of the best ones I have ever seen in amateur history but there are some issues that I think need to be addressed and these are:

- HDR - either subtle or nothing at all.

- 3D skybox houses - do not look realistic model or texture wise. This is a pretty bad thing as they are meant to work as a nice framing touch for the map. Although I think it is a very easy problem to fix as you have already placed them in a nice manner. Just resculpt them and you're all set.

- Volumetric light in the Colossus vs Tripod scene - remove it. It ruins the stylish and artistic touch of the scene.

If you fix these things I'm very sure I will never see anything as good ever again.

Warby, listen: your map is great but the HDR totally ruins everything.

Yeah. Maybe a bit more like this (judging by the choice of shadowing)?

[Blocked Image: http://img140.imageshack.us/img140/8705/genvivi2co5.jpg]

or this?

http://img241.imageshack.us/img241/9202/genviviob3.jpg

But then again, I like the daylight setting of your scene. Just reduce the yellow a bit (sunlight is really more white than plain yellow) and the contrast of the shadows and you have a kickass render. The models and textures are A+

Perfect! I think we have enough applicants for the project now, I will wait one day more to see what may come up. Then we are pretty much ready to start

Yeah, that one is really, really good. Philip is one of the hottest cg game artists out there imo.

This is _the_ map I'm really looking forward to play. I genuinly like the atmosphere you've created. If you need help with textures just ask!

Black Widow Games

- Teddy "TeddyBear" Bergsman - Lead 2D artist

Cool guys! Please check this out: http://www.mapcore.net/forums/viewtopic ... 169#128169

Photographically generated displacement information - test group

Please feel welcome to sign up! Application slots closed

For those who have no clue what this is about, please read the following:

During the past two weeks I have been developing a method which basically scans real-world objects by computing and merging separate frames of a recorded sequence of digital footage. Yet I have only conducted one laboration in the purpose of proving my hypothesis - a scan of my own hand. During the past week I have improved the frame calculation process and I feel ready to begin my second testing session.

This method is currently adjusted for the way I work and function but as I will release this to the public, public input is of great importance So, the purpose of this test group is to see what suggestions you have for the method appliance. It could be technical remarks, scanning improvements & tips, suggestions on what to scan, areas of use, etc.

For me, it would be interesting to see how the method applies to better - or worse - equipment than my own. However, the most interesting part is to see which different object scans you come up with.

Everyone who joins this group must understand that the method has to be kept secret. I don't want to spoil anything with an inferior appliance.

[Blocked Image: http://img172.imageshack.us/img172/5563/handfixeskh2.th.jpg]

I will have private, separate discussions with everyone involved. Group consultations may not occur. If you want to join, please contact me and ask me any questions you might have.

Regards,

Teddy

Test group applicants:

SnipaMasta - Panasonic digital video camera *.avi (clean 720x576, 25 FPS)

st0lve - Fuji Finepix S5000 *.avi (nosiy 320x240, 30 FPS)

aevirex - Canon Powershot a510 *.avi (clean 640x480, 10 FPS)

Meotwister - Canon PowerShot A620 *.avi (clean 640x480, 30-60 FPS)

Sascha Herfort - Canon Powershot A95 *.avi (clean 640x480, 15 FPS)

LuftWaffle - Kodak EasyShare DX7630 *.mov (clean 320x240, 20 FPS)

Nice but I'm not so sure about the rounding. Also, the lighting feels too yellow and smooth. With that intensity I'd expect strong sunlight but the shadows aren't crisp enough.