Graphics Team

Ok guys let me just clear something up. (I think i know what nick's trying to get at here)

-Perlin Fractal Noise Map- (according to wikipedia)
"Define an n-dimensional grid. Each grid coordinate stores a gradient of unit length in n dimensions. To sample, determine which grid cell you're in, and then compute the n-dimensional vectors from the sample location to each grid coordinate of the cell. For each grid coordinate, calculate the dot products of the corresponding distance and gradient vectors. Finally, interpolate these dot products using a function that has zero first derivative (and possibly also second derivative) at both endpoints."

Too much math, doesnt need to be on the art forums. I'm sure Nick has realized that by now. We all learned something today. (btw these perlin fractal noise maps are very very commonly used in video games these days).

By the way guys, let's all try to focus on the artwork here on this forum. We can ask questions by pm'ing the mods or posting it on the Misc threads. I know this post also counts as one, but I can't stand to see this nonsensical talk continue. its been like 2 pages and theres less than a couple models that actually fit the described outline set by the team leads. If you think of an idea post your thoughts and how you will go about making it and then see what feedback you get. once you all seem to have a *concrete* idea of what to make (and the approval of the team leads), make it in the proper program and post it according to the set format. Then if the model needs adjusting, you'll be told. It's as simple as that. You guys went on a huge tangent with that perlin fractal noise, and I see really random questions and lots of models being posted without approval of the leads (remember thrivingcheese said these >>>are too be discussed<<<.)
that's all
savvy?

I've set to work on modeling some vertebrae.

EDIT: Here is a simple vertebra, a few modified versions, and a rib.

Vertebra:

Here's the file download.

EVanimations these're awesome models, with skills like that the game can become far better looking than expected. I am curious to see what the microbe stage could look like with your models

Bawhaw, shucks. Make a man blush.

Microbe stage models should be relatively easy, seeing as they're organic blobs and organelles. I intend to do as much for this game as I can

Great work EVanimations, but remember to offer to do something first and confirm that its a necessary model, because so far everything on the list is "to be discussed".

Threw together a chloroplast Quickly, could probably use some more work but here we go;

Low Poly Model:

High Poly Model:

Placed inside 3 hexagons as per example from the concept animation, each hexagon has a radius of 5 meters and a height of 2 meters. Texturing is not done of course. Tell me what I need to work on.

A chloroplast has already been done by Thriving Cheese here (https://thrivegame.canadaboard.net/t1131p180-art-team-we-need-you). I commend you for doing this work, but just make sure to discuss next time before getting to work, to avoid modelling something that's already done. You wouldn't want to waste any time.

No, thank you for pointing it out, I haven't seen any lists for completed works, so thank you for point this out to me. Would it be possible to update one of the title posts with a list of completed works, just so that this problem doesn't happen again.

Yup. (I'll fix that soon when I get home).

It occurred to me that we should probably list off the polygon counts for 3D models we submit. Detailed models are good and all, but we do need to keep check on our polygon counts. Especially since the engine will likely be fairly unoptimized.

I think the OP states a recommendation of ~100 polygons per model. However, if the guidelines can be improved in any way, please say so.

This is a conversation carried over from the Art Team: We Need You thread.

Tritium wrote:

Version 1:

Spoiler:

Version 2:

Spoiler:

I don't know guys i really don't like this orange-red color...

Its not just personal preference, endoplasmic reticulum always have a colour somewhere between pink and orange. There's just something about the colour and texture of the first ER that didnt translate well into the second one, even after the colour change, but unless someone wants to take a shot at making another version we can settle with this for now. However, I like the nucleus, it was done well.

I guess that's what you expected? Less transparency and a texture- the only things i didn't change in the second model.

Spoiler:

https://www.dropbox.com/s/99y0wm0fmg5xbxn/Kernel4.blend

NickTheNick wrote:

I think the OP states a recommendation of ~100 polygons per model. However, if the guidelines can be improved in any way, please say so.

I think that listing the specifications of the models created is just useful to have just for the sake of statistics. Kind of a courtesy when showing off your model.

Ohh, I read your post wrong, you mean that an artist should post their poly count when they submit a model. Yes, I agree.

Ok i need some info on this 100 polys refers to number of triangles or quad faces? Look at this ball

Spoiler:

It has 56 faces or 112 triangles. If this is the limit the models will be very very simple.

I believe as a general rule, most people list Tris.

That is true, that is a very low count, however that is what cheese said and I'm pretty sure polys are counted as triangles. Does this mean very simple models for the opening builds?

We should be using simple models for opening builds because we want to showcase the engine more than the assets at first. If we wish we can release a functional build that would allow for most computer to run it and release a separate build that would be used to showcase some of the other assets depending on what the programmers want.

The opening build should be first and foremost about the functionality of the game. Think of the current assets as place holder I suppose.

Thinking of trying out the sand particles. How should they be done? Any reference images or specifics that you guys want?

Just thought about something.

How are we going to create the bones for the limbs of the creatures? I would go about it by modeling several different bone ends/caps/joints or whatever and generating some polys between the caps.

For now we need to leave that alone because the OE game design document needs to be finished.

Tritium wrote:

Ok i need some info on this 100 polys refers to number of triangles or quad faces? Look at this ball

NickTheNick wrote:

That is true, that is a very low count, however that is what cheese said and I'm pretty sure polys are counted as triangles.

Ops, I've must have missed a '0' at the end there, a more resonable Tris amount for the organells of the microbe stage would propably be ~1000 depending on how big quantities of organells on the screen at the same time that we'll have. This estimation is mostly based on tf2, where the maximum amount of polygons for most hats are ~1000, and for most weapons the maximum's ~6500 polygons, and from how I've understood it we'll have some couple of organells in en each cell/microbe and maybe something around 10 - 25 microbes at screen at the same time, but not all the organells won't be equally  big so the optimal polygon count will be different depending of the average size of the organell in-game. Altough I don't know if our engine's faster or slower than tf2's version of the source engine, wich ain't the most modern of engines from what I know.
     Polygons includes bothe Quads and Tris/Triangles, but most game engines only uses triangles/tris in-game (most of them automativcly converts the Quads to Tris/triangles) and thus what really matters for game engines are the tris amount.

NickTheNick wrote:

Does this mean very simple models for the opening builds?

So maybe we'll have to start with simpler models in the begging to test how much our engine can take at a time.

Don't worry too much about polygon count at this point. The graphics engine won't be the fastest, but from my limited understanding of 3D modelling, it's easier to remove detail than to add it after the fact. So, if we find at some point that the 20k polygon nucleus model is a little too intricated and makes two nvidia Titans in SLI give off SOS in the form of smoke signals, we can still reduce its complexity. (As a side note, please don't use 20k polygons for an organelle).

Nimbal wrote:

Don't worry too much about polygon count at this point. The graphics engine won't be the fastest, but from my limited understanding of 3D modelling, it's easier to remove detail than to add it after the fact. So, if we find at some point that the 20k polygon nucleus model is a little too intricated and makes two nvidia Titans in SLI give off SOS in the form of smoke signals, we can still reduce its complexity. (As a side note, please don't use 20k polygons for an organelle).

It's sometimes easy for us modelers, well at least me, to get lost in making the details just perfect :P . Keeping poly count in mind helps me avoid that.

I've just got a few little tidbits on Polycounts for the use of whoever sees fit to use them:

http://www.rsart.co.uk/2007/08/27/yes-but-how-many-polygons/

http://forum.beyond3d.com/showthread.php?t=43975

http://www.polycount.com/forum/showthread.php?t=50588

http://www.ericchadwick.com/examples/provost/byf1.html

http://www.ericchadwick.com/examples/provost/byf2.html

Bon apetit.