Fun New Model - WIP

[kattkieru]

Heya! Good to see you on a new project!

The hardest part for me with the extrude method is the mouth and eye sockets. I find that they're damned near impossible to do with extruding, but they work really well by building them and then building out around them.

Secondly, the extrude method has one major flaw: the model you end up with does not have the facial topology required to be animated easily and realistically. There are naturally occurring concentric circles on faces, as well as lines that run along muscles in certain areas (for example, the one that goes from the sides of your nose past the sides of your mouth, which creases when you smile). The extrusion method only works in straight lines, which don't really occur in nature.

But keep at it! I want to see more. :)

[Hunkadoodledoo]

Update...For the detail that I wanted with the nose and the mouth, there were a lot of vertices that I had to consolidate to match it with the few vertices through the back of the head. That made the corner of the mouth and the nose a bit messy. I will continue to tweak, but this is what I have now:
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm1%28xanga%29.jpg

Long way to go yet.

[Hunkadoodledoo]

That work is as much as I can do in 2 days (I had a four day weekend), so maybe I can try each different technique for two days and compare the results of each...Hmmmm...Does anyone have any links to tutorials for modelling a head using reference photos with the different techniques (NURBS, spline, patch, box)? The more the merrier! Thanks for the help!

[Hunkadoodledoo]

I tweaked a little bit:
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm2.png
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm2%28side%29.png

Oh...I am still interested in trying different modelling techniques. If you have any good tutorials or books to suggest, I would greatly appreciate it! Thanks!

[BgDM]

Not a bad start. The head is way too narrow and it looks like you are not following good loop topology.

Take a look here -> http://www.elysiun.com/forum/viewtopic.php?t=21066&highlight=head+modelling

As well as here -> http://apieceofstring.com/kos/

BgDM

[Hunkadoodledoo]

Hip hop, hooray! I have brought this thread back from the grave! I started a couple of weeks ago with a different method and have had phenomenal results. I am not sure if it is just that I learned a whole lot from my first attempts or if this method is just that much better, but I am really excited about the direction of this model. I am doing the "learn as I go approach" because I don't really have money or time to go to school for CG right now, so my approach might not be the most effecient, but I am making progress. Here are a couple of the in progress shots:

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm4%28xanga%29.png

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm4%28wip%29%28xanga%29.jpg

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm5%28xanga%29.png

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm6%28xanga%29.png

Comments and critiques are welcome and encouraged! Also, big thanks to BgDM. I had seen that Elysiun thread before, but since I didn't necessarily have plans on animating this model, I didn't think I would need to adhere to a certain topology, but it turns out that that specific topology is just better all around. I think that this one is going to be a keeper!

[Joat]

There were a number of things that bugged me with the first head, but this looks like a keeper. Finish it with care, and you have a nice head to show to everyone.
As for texturing, try to find a wip thread "Mongrel girl/nudity" from the "big boys" wip/3d stills. There is a bump made for the head that pretty much nails it. And despite the looks it's not made with zbrush but by drawing it in photoshop. So I think it should be done in Blender as well.
Best of luck!

[Eckie]

Loved this one:

http://www.secondreality.ch/tutorials/modelling/head.html

...and I got this far with it:
(the first one having better topology, the second one having a better nose)

http://home.student.utwente.nl/r.eckhardt/Gezicht03.jpg http://home.student.utwente.nl/r.eckhardt/Gezicht04.jpg

I quit at this point because I kinda messed up the topology. (When I started the tut I wasn't aware the topology mattered, so the mesh is a little messed up and not worth completing.)


Anyways, the tutorial is great, and it's yet another method. (You didn't do box-modelling yet, right?) Have fun with it! :)

[TroutMaskReplica]

i don't recomment box modeling with blender. the current tools are more suitable to poly by poly method.

if you want to do the second reality box modeling tute by wiro i recommend downloading a copy of wings and using that instead. you can exchange files between blender and wings using .obj (or the native .wings format, but i've had some trouble with that so i just use .obj)

[Hunkadoodledoo]

Sorry for the lack of updates....I guess because I haven't made leaps and bounds, I didn't think that it would deserve an update. Some good things did happen, though. I got a bunch of better reference pictures! I'll post them along with the update. I think I am going to struggle so much when it comes to texturing. I get lost anytime I try to LCMS unwrap something. I bought a book called Building a Digital Human that might help me a bit. Anyway, I'll post whatever I can get rendered (I am on my lunch break).

Check it:
http://www.valdosta.edu/%7Ejcfaulkn/images/chetformodel8000.png

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm08%28side%29.png

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm08%28front%29.png

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm08%28wire%29.png

[Hunkadoodledoo]

Weird obstacle. I recently rendered an update with yafray with GI set to full and best yada yada yada, and I get (seemingly) random artifacts in my rendered image. Here's a screenshot:
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm09%28front%28big%29%29.png


There are three on the nose/cheek/upper lip, two on one eyelid, one on the other, one on the chin/lower lip and one on the neck. There doesn't appear to be anything on my mesh. They just popped up. Any ideas?

Edit:
I am getting tired of looking at this blank material, but I am not satisfied enough with the modelling to move on to full texturing yet. So, what I would be interested in is the method of frontal camera projection using the "sticky" button. I have had some advice on how to use it, but a minitutorial on it with pictures would be so greatly appreciated you don't even know.

[Hunkadoodledoo]

Well, I had remembered reading most of the discussion at this thread:

http://forums.cgsociety.org/showthread.php?t=256495

I guess I stopped reading before the tutorial link was posted, and after watching it, I succeeded at making a simple crappy camera mapped texture render of my model. I made a little video to make it more interesting:

www.valdosta.edu/~jcfaulkn/images/huge_chet_sticky5.avi

I really haven't done much as far as updating the model, but I have been retardedly (is that even a word?) busy at work these past few weeks. I could still use critiques if you have them......Please. I want affirmation! I want to feel that someone is out there that cares enough to tell me what is wrong. Right? Isn't that part of the nature of love? Giving someone what they need even if it hurts a little. Anyway, thanks in advance for your help. I'll be here if you need me.

[Inktvlek]

Wow! That's impressive! It looks scary, with the eyes inside the head :o)
Makes me wanna check out this technique aswell!

keep up the good work man! This looks really promising!

[Hunkadoodledoo]

Okay, if you thought that video was freaky, check out this halfway attempt at UV mapped textures. I only did the forehead and the eye sockets, so the rest of the texture looks REALLY WEIRD mapped to the mesh. I have to go to sleep soon, but here is a quick question. I understand that you can map at least one texture to the UVs using the UV/Image editor, but if you had different layers like bump, spec, displacement, where would you map them? I admit that I am too tired to look it up now, so if you have a quick answer or link I could follow I would be uber happy. Look at it:

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm10%28very_freaky%29%28xanga%29.png

Sleep cometh.

[TroutMaskReplica]

OMFG!!! that's CRAZY.

as for your question, create texture in the usual way, load it into memory, and in the map input panel choose uv. the map to panel is where you can assign your texture to one or more channels.

[Hunkadoodledoo]

Happy, happy day! Saturday is my friend! I spent some time today mapping the rest of the model (rest of the face, at least) to my reference image to see how difficult it was (it took an hour or two). I still have modelling to do - I plan to convert my SubSurf mesh to a real mesh before the final version. I guess I will try that for now. Until THAT is done, here is the most recent render:

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm11%28xanga%29.png

[Womball]

how did you make the face texture? is it an image of your face somehow flattened to fit on a uv map?

[Artaures]

womball it definitly looks like a photo of the dudes face that he posted recently. flattened, its a 2d image :/ ?

[Womball]

How do you create that though, I would like to map my face to a model.

[Hunkadoodledoo]

Well, I used the topology at this thread:

http://www.elysiun.com/forum/viewtopic.php?t=21066&highlight=head+modelling

And I used the LCMS UV Wrapping video tutorial found here:

http://www.blender3d.org/cms/Model_Material_Light.397.0.html

Then, I used the picture of Chet that I took in Oregon, and mapped the vertices like this:

http://www.valdosta.edu/%7Ejcfaulkn/images/chet_mapping.jpg

IMPORTANT NOTE: This is just a test, and is not the way it will be done for the final render. You should not do it like me.

Anyway, I have to wash clothes for tomorrow. Next weekend will be a three-day weekend for me, so I hope to make a serious attempt at finishing the model so I can focus full time on textures.

My Totino's pizza roll tastes like a taco!

[Hunkadoodledoo]

Okay...It has been a while. I have been busy doing other things, but I worked a little on the Chet head this weekend, and this is what I got:

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm12%28xanga%29.png


I know that I have a lot to do, so here is my to do list:
Finalizing the model
Make ears
Make eyes
Make teeth
Make hair (Come on 2.40 final!)
Better color map
Bump map
Spec map

Grrrr.....Long way to go, and no time to work. I will not abandon this project!

If anyone has a Blender specific face mapping tutorial, please post a link!

[Womball]

http://forums.cgsociety.org/showthread.php?t=293387

I'm working on a head too! I want to keep the overall topology really nice, and its also going to go photorealistic. Perhaps we can help each other out. I'm using blender 2.40 alpha 2 btw.

[Hunkadoodledoo]

Frustration....

I thought that I could get a better lighting result by rendering in Yafray, but when I rendered with my color map applied, I was getting a "Could not load image" error in the console. After much searching (literaly hours), I found that Yafray can only accept .jpg and .tga bitmap images for texture inputs. Yay. One battle won. So, I hope everyone remembers that and hopefully the information becomes more readily available in the future.

Battle #2:
So, after I got my color map working, I tried to render with my bump map (.tga as well), but now I am getting a "Could not load image" error again. Would anyone like to help me diagnose my problem? Thanks!

Edit:
Problem solved! Okay, this is how it went down. I split my color map into its component Red, Green, and Blue channels to use as my bump map. After saving the resultant image as a targa, I loaded into a texture channel to try to use it as a bump map. Error. So, after loading it back up in the GIMP, I saw that its color mode was Grayscale instead of RGB. So, here is the moral of the story:

When loading images into Yafray to use as a texture, it must be a JPEG (.jpg), Targa (.tga) and it must be saved in RGB mode in your image editing software.

I hope this helps you, and I will post an updated render tonight when it is done. Say, an hour or so from now. I need a new computer.

Edit:
Here is that render I was talking about:
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm16%28xanga%29.png

[Hunkadoodledoo]

So much good news! Where do I start? Okay, I got to hang out with Chet over my Christmas break and we went to the High Museum of Art in Atlanta! Yay! While I was there, it was overcast and pretty much the best possible lighting conditions for reference photos. Plus, all the photos I had didn't have the way his hair looks now. So, I got better reference photos with hair and even the back of his head! Okay, with the new version of Blender, I am super-duper excited about the hair/static particle function for his hair. I also have started modelling his ears! I would like crits on them so far...Ears are weird! I don't really need to know how accurate to his ears they are, but just how much they look like human ears. Let me know!

http://www.valdosta.edu/%7Ejcfaulkn/images/right_ear_wip.png

http://www.valdosta.edu/%7Ejcfaulkn/images/right_ear_wip2.png

http://www.valdosta.edu/%7Ejcfaulkn/images/right_ear_wip3.png

[RedSquirrel]

I like the render, only crit I can see is just above the ear lob, where the the ear cannel starts, the bit that looks like a stretched out "y" its looks a bit flat or sharp... when are you planning to join the ear to the head? Your WIP is coming along nicely :applause:

[Hunkadoodledoo]

Yay, criticism! I suppose I am more excited about my thread than most people, because this is my first attempt at a human head (or really any head for that matter). I haven't really had a whole lot of criticism so far (I am my biggest critic), so I am very excited to see that someone is willing to help me out!

RedSquirrel, I will be joining the ear to the head as soon as possible (maybe within a week or so). I am going to have fun fixing the topology for that (sarcasm). (I use parentheses a lot.) Also, could you GIMP or MS Paint the area you are describing as the stetched out "y"? I would like to make the change, but I don't really understand where you're talking about. Thanks again! Oh, I changed its orientation a little - here's an updated render:

http://www.valdosta.edu/%7Ejcfaulkn/images/right_ear_wip4.png

I feel like I take two steps forward and one step back, but at least I am making progress. I had a textured, unwrapped head without ears. Now, I have to redo the texture maps with my better images, but it will be an improvement AND he will have ears!

To help everyone (who reads this anyway) get more excited about where this is going, after I finish the head, I plan on rigging it and animating it for my very first mini-movie! As a teaser, I want you to think "Seis Million Dólares". I've probably already said too much. Anyway, for those of you interested, make sure to keep tuning in for updates. I will need a lot of help trying to figure out all the crazy improvements to the animation system when it comes time. The only thing I have ever animated is the blades spinning on a desk fan for about five seconds. This, I hope, will be at least four times cooler! Thanks again for the help, everyone!

[RedSquirrel]

After reading what I had posted I couldn't blame you if you didn't know what I was talking about, I do apologize for that, lol it looks like a "y" to me. :D

The blue bit in the attachment, shows what I was trying to indicate, and the last ear at the bottom has been smudged slightly to make it seem a bit more rounded.

[Hunkadoodledoo]

I feel held back by my hardware. I only have 256 megs of RAM, and when I am editing a 3000x3000 photo of Chet, it goes slowly. That discourages me and I end up not working as hard/as much as I really want to. Anyway...

Here is an updated color map from the new reference photos:

http://www.valdosta.edu/%7Ejcfaulkn/images/newer_texture_3b%28xanga%29.png

This is what I still need to do on it:
Take away the shadows and highlights
Make the neck blend better
Make the forhead blend better
Take off the nose
Remove the hair (to be made with Blender's static particles)

Any suggestions?

[Apollux]

Now I´ll have nightmares the entire week !! :argh:

That is some scary stuff !

[kattkieru]

That's a nice blend of those photos. The last render you posted was pretty nice too.

But how do all you guys keep doing the skin pores? That's the part that just escapes me.

Where is this going, Hunk? What's the end target?

[Hunkadoodledoo]

kattkieru,
To help everyone (who reads this anyway) get more excited about where this is going, after I finish the head, I plan on rigging it and animating it for my very first mini-movie! As a teaser, I want you to think "Seis Million Dólares". I've probably already said too much. Anyway, for those of you interested, make sure to keep tuning in for updates. I will need a lot of help trying to figure out all the crazy improvements to the animation system when it comes time. The only thing I have ever animated is the blades spinning on a desk fan for about five seconds. This, I hope, will be at least four times cooler! Thanks again for the help, everyone!

Okay, the render with the pores-ish look came from a bump map using a mix of an inverted, desaturated version and non inverted, desaturated version of the color map. That makes the beard stubble light and all the other pores dark. I did a real dirty job, but here is what the old one looked like:
http://www.valdosta.edu/%7Ejcfaulkn/images/chet_bump_map1.png

Wow, that looks bad. Anayway, I have to go get dinner before Subway closes. Thanks for the replies!

[RedSquirrel]

That's a nice blend of those photos I second that I really like how everything blends together in your map, hope to see more soon.

[kattkieru]

Okay, the render with the pores-ish look came from a bump map using a mix of an inverted, desaturated version and non inverted, desaturated version of the color map. That makes the beard stubble light and all the other pores dark. I did a real dirty job, but here is what the old one looked like.

Heh, I'm so A.D.D. ... I must have seen the above project description before.

Interesting way to do the pores. Looks like I need to buy a better camera to actually get that pore detail in there. ^-^

[Hunkadoodledoo]

Anyway, here is the jist:

Help me with my flow:
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm18%28ear_wire%29%28xanga%29.png

Here are the newest renders with ears:
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm18%28front%29%28xanga%29.png

(Note the repeating diamond pattern on the forehead. Weird.)

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm18%28side%29%28xanga%29.png

As always, comments and critiques are welcome and appreciated!

[BgDM]

I applaud you on your persistance with this one. Glad you are not getting discouraged and are trodding on with it.

Though I do have to say, it looks like you have way too many poly's on this mesh for the amount of detail that you currently have, which is minimal. You may want to look at the video tutorial done by mr_bomb last month in the Community Journal #2. It is a great video tute on head modelling and gets great results with very few poly's. You can see it at www.blenderartists.org in the articles section.

Your ear topolgy looks decent. Some more definition can be added by pulling some edges out. Looks too flat right now on the inner part.

BgDM

[Hunkadoodledoo]

BgDM, thanks for the reply! I am not extremely concerned with vertex density because 1) I want this to be very accurate and 2) I am a patient man and can wait for a render on my slow computer if I have to. Now, that I think of it, the only reason I have so many verts is because of the way Blender used to handle UV mapping when a model was subsurfed. It pulled the map in weird ways across verts and edges, so I just converted the subsurfed mesh to a real mesh. Now that it is fixed (or worked around, I'm not sure), I have come so far on the model in its high vert count form that I don't want to go back and get the old mesh. Plus, I suppose if you are mapping a homemade skin/bump/spec map, you can get away with few verts cause you can always clone/resize or redraw your map. Since I am working from an image, I thought it would be easier to move a bunch of verts instead of cloning/bluring/stretching my reference image in order to keep it at the best quality. I am kind of a perfectionist and would gladly sacrifice rendering time for accuracy. AND, this is my first attempt at a head and I want to do it right. Wow. I have excuses for everything...Oh, I checked out the link. Yay, tutorials! Right now, I have 2867 verts (about a third of those are in the ears and the inside of the mouth. That is just 3.25 times the 881 of the sample head from the tut. So, it isn't all that bad. Also, could you explain "pull out a few edges"?

[Hunkadoodledoo]

So, not much going on here...I tested the new live LSCM feature. It was pretty handy. I haven't really updated the model since the ear attachment (except deleting a rogue edge that had me going bananas), but I did test the new color map. I am sure that I am getting ahead of myself (get it!?), but I am VERY eager to see progress. So, here is an updated render with the new color map, UV subsurfing and an Auto-normalize in the GIMP:

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm19%28xanga%29.png

I really like the way the new map reacts to light. It was a great day to take pictures. The other map was tainted with incandescent light and too many shadows. This color feels more natural. Anyway, I just wanted to keep everyone updated.

[obelisk]

wow, this has come a long way. nice work. a couple of things though:

- if you're working this high poly, I see no reason not to model the eylashes. It's one of those touches that really make a huge difference in bringing a model to life.
- likewise, some kind of bump or normal treatment would help this heaps. It feels very porcelain at the moment.
- I think the ear leans back too much in the side view. I'd make it a little more upright, even if that differs from your reference.

keep it coming man

cheers

-O

[Hunkadoodledoo]

CGTalk maintenence lost my post. Frowny face.

Quick update.

It feels like I have done a whole lot since my last update, but not a lot seems to show. I added ellipsoids for eyes and mapped the color map a bit more, and voila! This is what I have:

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm20%28side%29%28xanga%29.png

So much to do, so little patience. I WANT A NEW COMPUTER! Grrrrr... Anyway, sleep cometh. Peace out.

[Hunkadoodledoo]

Good news! I may be building myself a new computer. This is what I have decided on so far:

MSI K8N Neo4-F Socket 939 NVIDIA nForce4 ATX AMD Motherboard
AMD Athlon 64 X2 3800+ Manchester 1GHz HT Socket 939 Dual Core Processor
CORSAIR XMS 2GB (2 x 1GB) 184-Pin DDR SDRAM DDR 400 (PC 3200) Memory
MSI NX6600 TD256ESP Geforce 6600 256MB GDDR2 PCI Express x16 Video Card
COOLER MASTER Real Power RS-450-ACLX ATX12V 450W Power Supply

Throw it all in a case I already have with some cheap SATA or EIDE drives, add some flavor of Linux and voila! New computer for about $800. Not that bad. Any suggestions?

Okay, small update. I added the lacrimal caruncles and mapped the skin a bit more. Here is a tip for anyone hand plotting vertices on a UV image map. If you want to keep up with the vertices you have plotted so you don't go back and plot them again, after you decide on their final positions, pin them with the P key as a flag that they are done and that you shouldn't mess with them again. Okay. Here's the picture.

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm21%28xanga%29.png

Kinda a bit off subject, but I also ordered the French artist Camille's album Les Sac Des Filles. I haven't received it yet, but I am greatly anticipating it. Okay. Comments and critiques welcome. I'm out.

[TroutMaskReplica]

how will you attach the ear's without destroying the UV map for the head?

looks good, but maybe just for the sake of the render and showing form, it might be a good idea to set up 3 point lighting.

[thondal]

just for the fun of it, i would throw in another 1 gig... i've been doing 1 gig for a time now, and i'm upgrading to 2, and i guess, i will move on to 3 or 4 in a year or so... :P


-thondal-

[FreakyDude]

how will you attach the ear's without destroying the UV map for the head?

Can't you unwrap (lscm or not) part of a newly attached mesh in blender without messing up the allready done part?!?!? If so that's really a let down. I really loved the way lscm and seams work in blender, but if you can't even attach new meshes.....? me going back to the 3Dmax referring thingy, but you can attach as many things there as you want without it messing up the existing uv for the original and the newly attached parts!!! That would really be a let down...........

[Hunkadoodledoo]

how will you attach the ear's without destroying the UV map for the head?

looks good, but maybe just for the sake of the render and showing form, it might be a good idea to set up 3 point lighting.

Well, the ears already are attached to the mesh. They have been seamed off of the head and I will be creating (cloning) some kind of texture for them later. This is what my UV mapping looks like for now:
http://www.valdosta.edu/%7Ejcfaulkn/images/texture_breakdown%28xanga%29.jpg

Yeah, still has a way to go, but I am going to erase all the hair and ears and shadows.

As far as 3pt lighting goes, I have heard people swear by it, but I really want to go for realistic and from what I have read, it isn't a very natural setup. I have heard artists use a 3pt setup to show off models and inspect them for areas that need work, but really haven't read too much on the subject to care. Wow. I don't think I will improve very fast at all with apathy like that. Tell you what. I'll look into it, and hopefully learn something I can share with everyone and we will all be the wiser.

just for the fun of it, i would throw in another 1 gig... i've been doing 1 gig for a time now, and i'm upgrading to 2, and i guess, i will move on to 3 or 4 in a year or so... :P

Even though 2 GBs of DDR400 is like 5 times cheaper than the eqivalent in RAMBUS (what I am using now), I am not made of money and I think that 2 GBs is a good step up from the 256 MB I am at now. I know that I will want more eventually, but we will have to wait for money to find its way into my bank account before I think about upgrading. 2 GBs and a dual core Athlon 64 should be quite a nice step up from what I have currently.

[thondal]

okey.. to upgrade from 256 - 2 gig, is a major step;) so, then it is another thing;) you'r gonna feel the difference at once.... never mind what i said then;)

-thondal-

[piro1]

wow very good job.
I'll keep learning

[Hunkadoodledoo]

Quick question for anyone in the know:

How exactly bump maps affect rendered nomals in Blender? I know the concept behind normal/bump maps, so I am specifically asking which model Blender espouses. Is it a black/gray/white = raised/unaffected/lowered, a black/white = unaffected/lowered or a black/white = raised/unaffected? Here is a recent render with the same image used mapped to Col and Nor:

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm21b%28xanga%29.png

It appears if the darker areas are raised, but are they really? Or are the light areas just recessed? Is it both? Arg. Anyone who knows for sure, please enlighten me.

Edit: Anyone? Bueller?

p.s. I am still studying up on 3-point lighting, and I will present my findings as soon as I find them.

Edit:
Three-Point Lighting: My Findings


History:

In my research, I found that 3pt lighting has been around a long time. Well, since time is relative, we'll say that it has been around twice as long as CG. At its simplest, 3pt lighting is "a very common lighting technique used in both still photography and in film...using three separate light sources [to] illuminate the shot's subject (such as a person) however desired, while also controlling (or eliminating entirely) the shading and shadows produced by direct lighting." It seems that it has fallen out of favor of mainstream photographers and movie makers, though.


Basic setup:

The basic setup is to use these three lights in your scene: a Key light, Fill light and Rim/Back(not Background)/Hair/Shoulder/Kicker light. They are positioned as follows:

Key light:
15 - 45 degrees to one side of your light
15 - 45 degrees higher than the angle of the camera
http://www.3drender.com/light/figures/3ptKey.jpg

Fill light:
+ or - 90 degrees opposite your Key light, on the other side of your camera, "but don't make all of your lighting 100% symmetrical!"
A little higher or lower of your Key light
http://www.3dtotal.com/team/Tutorials/Jenns3pt_tut/KeyFillScreenTop.jpg http://www.3dtotal.com/team/Tutorials/Jenns3pt_tut/1key1fillScreen.jpg

Rim light:
"[The Rim light] shines on the subject from behind, often (but not necessarily) to one side or the other. It gives the subject a rim of light, serving to separate the subject from the background and highlighting contours." I like that definition, and I like this explanation:
"The rim light is placed behind the objects being illuminated and is angled so that the light glances off the surface of the object at the narrowest angle."
http://www.andrew-whitehurst.net/tutorials/oranges_rim.jpg
http://www.3dtotal.com/team/Tutorials/Jenns3pt_tut/3PTscreenSide.jpg


Pros:

It seems that as far as composing a three-dimensional object for a two-dimensional image, 3pt lighting uses light to "'model' the objects so that on a flat surface we can tell what shape they are, and how one object relates to another." It seems that using 3pt lighting allows artist to compose a shot that has full contrast between light and dark, uses many shades of gray in between and separates the subject from the background to add dimension.


Cons:

Richard Harris made a good point in his study of Light:

"The biggest problem with this set-up is that it is artificial and doesn't reflect reality. The use of back lighting especially should only be considered if you are looking for a specific effect since it is so dramatic and recognisable. Back lighting can be very effective but it should be used with flair rather than blindly applied to every situation. The kind of light that 3 point lighting creates simply does not exist in nature and therefore it looks fake. The fact that it is taught in so many textbooks also lends it an air of cliché and it has therefore become tired and boring."


When to use it:

It seems that it would be appropriate to use 3pt lighting when you are trying to show the dimensions and realtions of objects to each other in a scene when it doesn't matter if it looks cliché, tired or boring. i.e. when showing off geometry on message boards to get critiques on the details that would normally be lost in shadow or black backgrounds.


When to not use it:

When trying to be original or realistic to the physical world, final renders, the third Sunday of every month with the letter "r" in its name.

Lessons learned:

I think I am actually going to give the setup a shot. We'll see how it looks soon enough.


References:
http://en.wikipedia.org/wiki/Three-point_lighting
http://www.itchy-animation.co.uk/light.htm
http://67.15.36.49/team/Tutorials/Jenns3pt_tut/3ptlighting.asp
http://www.3drender.com/light/3point.html
http://www.andrew-whitehurst.net/3point.html


Example 3pt lighting renders:
http://www.3dtotal.com/team/Tutorials/Jenns3pt_tut/final3pt.jpg
http://www.andrew-whitehurst.net/tutorials/oranges_final_comp.jpg
http://www.newtek.com/products/lightwave/tutorials/animation/3point-light/images/layered_composite.jpg
http://media.g4tv.com/images/ttv/graphics/thescreensavers/3382522.jpg

[Hunkadoodledoo]

You know, I am not sure what originally made me choose the lighting setup that I have been using for 90% of this project. It has just been two area lights with a size of 25 and shadow samples set to 12. That gave me two very large dim lights sucking up ram. At 100%, it would take 20 something minutes to render. With the 3pt lighting setup, I have more dynamic lighting range and it only takes 3 minutes to render. I suppose it has some pretty good advantages. Here is what I get:

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm21c%28xanga%29.png

I think it looks pretty good. The eyes still need major work, the texture maps are maybe at 20%, and I haven't even started thinking about the hair. So long to go. It looks like I will be postponing my new computer purchase a few months, by the way. With the new tech that will be coming out this summer, current prices will fall nicely. Time for dinner. Peace out!

Edit:
Oh yeah, if anyone knows the answer to my Blender bump map question, I would be very happy if you shared it with me. Thanks!

[kattkieru]

Heya,

With bump maps I've yet to figure out how to make them look the way I want every time. When I think I have it figured out the map I make seems to go the opposite way.

The best solution I've found so far is to position a test light somewhere in the scene so I can figure out how the bump is facing (in / out to black / white or vice versa, etc). Then I play with the Nor button and render it a few times. Generally Nor pressed in works, but at times I find I need to do negative Nor (press it so it's yellow).

I was told once that the system is black / 128 gray / white to recessed / unaffected / raised, but I really don't think that's the case. I wish I had a better answer for you...

[Womball]

Just a thought, if you use 3 pt lighting, make sure you turn the shados off on the other lights. You just need one shadow casting light (your main light). I'm find 4 points works better for me, and you should try area lights. They create softer shadows, but they are strong lights so you need to play with the settings a bit.

[Hunkadoodledoo]

Quickie.
Here is the my first render with an updated UV texture. I took off the hair and the ears - the texture on the ears is just preliminary. I monkeyed with the eye material. I think it looks too metalic now. If anyone has any eye material setting recommendations, please shoot 'em my way. Here we go:

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm22%28xanga%29.png

I think the next "big" step is to get rid of all of the stretching by finalizing the UV coordinates. Well, I gots to eat. I have tomorrow off and I am going to get my truck fixed. I might work on the Chet head a little bit. We'll find out tomorrow...dun, dun, DUN!

[TroutMaskReplica]

As far as 3pt lighting goes, I have heard people swear by it, but I really want to go for realistic and from what I have read, it isn't a very natural setup.

it's quick to set up and shows form and texture really well. i never meant you should use it in the final piece, just while getting crits on the modeling/texturing.

btw, looking really good.

also, 3pt lighting hasn't fallen out of favour. the basic concept of a key light and a fill light can be seen nearly all photography with a human subject.

[BrianH]

No offense, but I would really like to see a UV test image render on your face model. It seems you're just playing with the UV map until you get something that looks decent. And it looks like you only got one side of the face finished, because this image shows some really bad mapping on the right side of the image: http://www.valdosta.edu/%7Ejcfaulkn/images/fnm21c%28xanga%29.png . I would highly recomend using a UV map test image and aligning it to the face. It will make your mapping and unwrapping so much easier in the long run. Here's a tutorial that will help you greatly:

http://bgdm.katorlegaz.com/lscm_tute/lscm_tute.htm

And the UV test image for it is here:

http://bgdm.katorlegaz.com/lscm_tute/Tute_Images/UV.jpg

[Hunkadoodledoo]

BrianH,
Thanks for the comments!! I appreciate you taking the time to give me your opinion. Here are some renders with the internal test map:
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm22%28testmap%29.png
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm22%28testmap%292.png
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm22%28testmap%293.png

Good eye on the right side! My next big step of finalizing the UV coordinates will make sure to fix the crazy stretching. I would say that right now, I am at about 50%. As you can see, the bad areas are mostly the neck, the right side (his left side) of the face, the ears, the top and the back of the head. I think the hardest parts so far have been the eyes, nose and mouth. I suppose it has to do with the vertex density. Anyway, I might be able to get it done tomorrow or later this week, but until then I am only speculating. I'll let you know!

[Hunkadoodledoo]

Test renders:

Test static particle hair system:
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm23%28xanga%29.png

Test of the Indigo free, unbiased renderer (I don't think the bump values translate the same from Blender to Indigo):
http://www.valdosta.edu/%7Ejcfaulkn/images/fnm24%28xanga%29.png

Still working. Slowly, but stubbornly.

[timeless3d]

I just found your thread, awesome stuff. Can I ask what you are planning on doing with this? Do you have an application in mind?

Keep up the good work, I will be using this thread when I get to the point I am ready to model a human head.

[Milky]

It's looking really nice, great progress from the very beginning. Keep working. ;)

[Hunkadoodledoo]

I think I do too many test renders and not enough "work". Anyway, the biggest changes here (not even high quality changes, just a few things that were bugging me) are that I made the ears blend better and I made sure the baldness covers the whole scalp. So, this is just a test. Check it:

http://www.valdosta.edu/~jcfaulkn/dls/chetuv2.avi

[thondal]

it somehow feels as if the ears are too much backwards rotated... not sure how to explain it.. but might be just me... looking really good ;)

-thondal-

[Hunkadoodledoo]

I don't remember how long a thread has to be inactive before it is closed, but I wanted everyone to know that I am still working on this project. I have just been transferred to a new base in Warner Robins, GA, and my main computer isn't here yet. Hopefully, I'll pick it up as soon as I unpack and get a workin'. I hope 2.42 will bring a lot to the table for this project. I've got my fingers crossed.

[Hunkadoodledoo]

So....tired. I have been working on the Chet head a little bit recently, but nothing that deserves posting. I have been doing some eye tests that I think I am going to do a mini tutorial about, though. I will eventually adapt the method for Chet's eyes so he isn't a zombie anymore. Here are my results so far:

http://www.valdosta.edu/%7Ejcfaulkn/images/eye01%28xanga%29.png

http://www.valdosta.edu/%7Ejcfaulkn/images/eye02%28xanga%29.png

Keep in mind that these are still just tests. Any comments?

I will sleep now.

[TroutMaskReplica]

chet has pretty eyes. ;)

[Womball]

They look kind of like ping pong balls. Wheres the veins? Really nice iris texture. Also if your using blender 2.42, you should be able to use HDR images for reflections. This should be in the world tab. A transparent slightly larger sphere with ray traced reflection around the eye will provide a nice material for the HDR to interact with. It should create a wet filmy look to the eye. Also the topology in the pupil and a little bit of the iris should be extrude back into the eye. I find this to create nice depth to the iris and pupil.

[Ex-Nihilo]

I suggest making the spec harder. The eye looks a little dry to me.

[Hunkadoodledoo]

Update on the Chet head. I changed the material for both the skin and the eyes. What I had done was set the reflectivity of the skin to really low and jacked up the light intensity. That gave me okay skin results, but the eyes were way blown out. So...I upped the reflectivity on the skin and lowered the light intensity and bada bing! I must still say that the eye texture is in the testing phase. Chet's eyes aren't even that color. I still have to add veins and whatnot. I dropped the hair for now because it makes the rendertime about 25 minutes at full res (1024x768), and without it is about 2 minutes. That is a pretty big bottleneck to testing. Comments and critiques are welcome and appreciated. The only problem I can find right now (that I can't figure out), is that there are some artifacts on his left forehead that may be from a shadow buffer or something...Suggestions? Wait...I read something in Jeremy Birn's book Digital Lighting and Rendering about a shadow bias assisting in artifacts and self shadowing. I'll test out my hypothesis later. I need to sleep. Sleep well everyone.

http://www.valdosta.edu/%7Ejcfaulkn/images/fnm32%28xanga%29.png

[Hunkadoodledoo]

So, I guess the IT guys at my university got wise to the fact that I haven't attended class there in over a year and killed my webspace - broke all my image links. Boo. Photobucket it is. I updated the eyes a bit to include a better color representation of what Chet's eyes actually look like. I have been super busy going through school for my job, but this coming week is the last one. Then, maybe I'll post more regular updates. Enjoy!

http://i114.photobucket.com/albums/n270/hunkadoodledoo/CG/fnm34xanga.png

p.s. I keep getting these artifacts on his nose and forehead. On his nose, you can see a faint outline of the shadow on his left that is lighter than the surrounding skin and shadow. The one on his forehead is a kinda jaggy line right by the highlight above his left temple. Any suggestions?

[Hunkadoodledoo]

Then, maybe I'll post more regular updates.

I can resurrect a thread with the best of them! Yeah..."more regular updates." That's pretty funny. Anyway, I had a bit of inspiration to get some work done on the Chet head, so, I wanted to get some crits on what I have updated so far. So, here is the list of what still needs to be done:

1. Perfect UV mapping and finalze UV texture
2. Finish eye UV map
3. Normal map
4. Spec map
5. Weight paint for hair length and density
6. Bone structure
7. Driven facial expressions
8. Make a body
9. Teeth
10. Tongue

So far, I have done work on number 1 (I darkened his skin and removed some of the light and shadow information from the original texture), number 3 (copied the color map, desaturated, inverted (I think) and adjusted the brightness and contrast to get a better texture effect) and number 4 (again, copied, desaturated and inverted the color map - selectively blackened areas that don't need a high spec value). I have done some testing with weight painting, but I am still a ways away from having that down. I did see a great tutorial about doing a self portrait on Blendernation.com - http://terrier.infographie.free.fr/work/tuto_portrait/index.htm - it should help me regardless.

So, still much work to be done. I have yet to figure out the crazy artifact on his forehead where the rim and the fill lights intersect - or the funny highlight on his nose. Suggestions?

So, don't hate me for dredging up an old, almost dead thread. I am really just looking for any help I can get - the more people comment on my work, the more motivated I am toward finishing it. So, without further ado, here's the most recent iteration.

http://i114.photobucket.com/albums/n270/hunkadoodledoo/CG/fnm36.png

Wanna see a video of the Chet head in action? Check it out here:
http://www.box.net/public/c1u6b76qhk

It is an mp4, so you may need VLC or Media Player Classic to see it. As well, comments and critique welcome and encouraged. Have a wonderful evening!

[Hunkadoodledoo]

So, not much changed about the model. I changed my spotlights to arealights...It added some time to my overall render time, but it also changed the dynamics of the lighting. The highlights are more spead out. There isn't that weird highlight on his nose now...I also added an angular map to get a little reflection on the eyes and some color in the background. So, I am working on the weight painting for the hair, but I am waiting on a response from alt_ligury about how he did his hair in his self portrait. So...Enjoy!

http://i114.photobucket.com/albums/n270/hunkadoodledoo/CG/fnm37xanga.png

[Hunkadoodledoo]

Thanks to some inspiration from alt_ligury and some serious brainracking, plus help on irc (thanks, BeBraw), I figured out how to use one mesh with a static particle hair system and two textures - one for the mesh, and one for the particle strands. It wasn't clearly explained in alt_ligury's tutorial (I may post a quick write up later), but with it figured out, here is a preliminary render (I need to refine my VGroup weight painting for particle emission):

http://i114.photobucket.com/albums/n270/hunkadoodledoo/CG/fnm38xanga.png

Anyway, one of my biggest gripe about the hair system (I may just not understand it so well) is that the material settings for static particle hair strands are resolution independant. The start strand width and end strand width are set in pixels! That means that they will change the way the look when rendered at different resolutions. That seems kinda weird to me. If I am wrong, please correct me. It seems that it could be set in a percentage instead of pixels. Just my two cents. Anyway, thanks for reading!

[FreakyDude]

Thanks to some inspiration from alt_ligury and some serious brainracking, plus help on irc (thanks, BeBraw), I figured out how to use one mesh with a static particle hair system and two textures - one for the mesh, and one for the particle strands.

That's EXACTLY what I'm trying to figure out! How did you do that?

[FreakyDude]

Nevermind, figured it out. For those that don't know how:
under the tab particles, menu options for display (where mesh unborn died vect are) there is a slider that allows you to pick the active material.
Just create a new material index for your mesh, it doesn't need to have any vertex assigned to it, it just needs to exist, so you can create a material for it.

[Hunkadoodledoo]

Well, a couple of months have passed, and not a whole lot of work has been done. I guess I just got inspired this weekend, and although it may look like not much has changed, quite a bit of work went into perspective correction. I realised that all along I have been modelling in orthographic view, but as far as I know, there isn't a way to take orthographic reference photographs. Mine were taken with a 55mm camera lens. So, I modelled some cameras with 55mm lenses and corrected the model as best I could to match my references. SO, here is what I've got:

Here is the front reference photo:
http://i114.photobucket.com/albums/n270/hunkadoodledoo/CG/Chet_front_referencexanga.png

And here is a rendering from a camera with the same lens and approximate angle:
http://i114.photobucket.com/albums/n270/hunkadoodledoo/CG/fnm42.png

I know that the mouth isn't in the same position, and the lighting is really harsh on the right side of the model, but the perspective should be pretty darn close to reality.

After changing so much about the model, I may have to review the UV mapping (although it looks like most everything is still in place). I went ahead and rendered another turntable video which can be viewed here: http://www.box.net/shared/4isozovok8

On a side note, I have to say that Sculpt Mode has really stepped up to the plate and helped me out a lot. I really appreciate all the coding that went into getting it into 2.43. Thanks bunches!!

[Hunkadoodledoo]

Arg...I really want to make progress here, but I also want to figure out why light is shining through my model. I have turned off Double Sided rendering, the Alpha is at 1.0 and for the life of me, I can't figure out why the light I have set as the rim light on the Chet head shines through the back of his head and brightens his mouth and eye socket. I put it on a separate layer, and here are renders with that layer on and off:

Layer 5 Off:
http://i114.photobucket.com/albums/n270/hunkadoodledoo/CG/layer5off.png


Layer 5 on:
http://i114.photobucket.com/albums/n270/hunkadoodledoo/CG/layer5on.png


I have attached the .blend in case anyone wants to help me out. Super-duper thanks in advance for any help!

[fktt]

hmm.. that is WEIRD indeed.. :( turning the SBias all the way up to 0.25 helps a little but sadly it doesnt go any more up than that :(

perhaps specular-mapping could help? :)

[Hunkadoodledoo]

Thanks to SoylentGreen over at BlenderArtists.org, we found out that it was my clip Sta and End that should have been resized to be closer to my model. Then, I decided to use ray shadows and the problem went away altogether. But, here is an updated video with SSS. Enjoy!

http://www.box.net/shared/5ei6etp2vq (http://www.box.net/shared/5ei6etp2vq)

[fktt]

cool! :D why didnt i think about that, i mean i looked at the settings but i didnt think about that at all, like duh! :)

[Hunkadoodledoo]

Well, I am slightly retarded (sorry, if that offends anyone)...I have been using a Spec map to have customized specularity for probably the last 10 or 15 versions of my model, but I just realized today that the MapTo:Spec:Var setting has been at zero. So, it hasn't been affecting the Spec at all. Anyway, I changed that, I adjusted the hardness, I upped the SSS scale by 3.5x and made the scattering color a bit more pink. I think it does wonders. Here's a fun render:

http://i114.photobucket.com/albums/n270/hunkadoodledoo/CG/fnmSSS02xanga.png

I'd like to re-render the most recent video I linked to with all these changes, but it will take a a fair amount of hours for that. I'll post it when I get it done. As always, comments and critiques welcome and encouraged! Thanks!

[fktt]

lookin good mate! :D

[jinjenator]

That looks bloody good!!

[Hunkadoodledoo]

Thanks, guys! I'm rerendering the video as we speak, and I'll post it as an edit to this post when it gets done. I am at frame 133 of 300, and the last frame took 1 minute, 38 seconds to render at 640x480. I'll post it tomorrow! Have a great night, everybody!

Update:
Here's the video I promised. Nothing too spectacular. The more I look at it, the more I see I need to fix. Anyway, here goes!

http://www.box.net/shared/qm4tch5p0t

[fktt]

will be looking forward to it :)

[MSpace]

Love the still render above, but was unable to open your videos. Either the mp4 or the avi above. What codec did you use?

[Hunkadoodledoo]

Both were encoded directly through Blender, and I don't think I can remember what the settings were, but VLC should be able to open and display both videos. VLC can be found here: http://www.videolan.org/vlc/

Let me know if you still have trouble!

[MSpace]

Great thanks! Its seems that WMP and Winamp and DivX dont like Blender.... I guess I have a new one to use!

[Hunkadoodledoo]

Life happens. Months pass. Here's a quick update. I worked on it tonight, and the biggest change is that his mouth is no longer open in a goofy way. It is still slightly open, but it is much better...Enjoy!

http://i114.photobucket.com/albums/n270/hunkadoodledoo/CG/fnmSSS08.png

A little defocus in post....

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.