[michaelplogue]

I noticed in the buildings here in pittsburgh there are some fairly typical ways brick buiildings stain (water stains off of sills - base of the building is dirtier then then most areas - stuff like that).

For some of these specific stain areas, you can try the trick I did for the messy green paint stains around the window. Rather that making a huge textue and mask to use in conjunction with the other wall textures, I simply placed a small plane around the window, and used a simple single layer with transmittance map. Saved me tons of work trying to match a single large map with my geometry.

[Edit] You would want to use a dual BSDF / Ghost layer material for this, so you don't completely obscure your main brick material.

here's a question though: what is a typical multiplier for an .hdr map? it seems every hdr I load is extremely dark (even when I convert to .mxi) and I have to punch up the numbers.

Every one I've tried, I have had to used different values. However, typically I set the intensities as follows:

Background: between 3 and 5
Reflection: 2 - 4
Refraction: 2 - 4
Illumination: 8 - 12

For HDR's you really have to experiment to get something that looks right.

.

[lebbeus]

The building looks a bit like the ones you'd find in Oakland, near Pitt or CMU or the Carnegie…so

are the curbs granite or cast concrete?
Are the stones sawn, thermal or split finished? (there's probably others but the sample in front of me only has these three)
what are the mortar joints? raked, concave, v-shaped, weathered, flush, etc?

oh yeah…where are the pigeons?

here's some more masonry/concrete specific things:

efflorescence:http://en.wikipedia.org/wiki/Efflorescence
spalling:http://en.wikipedia.org/wiki/Spalling

[misterasset]

It's funny you mention the scope of the project, because that was my first thought. I thought: this would take me months to do right. Just the randomizing of faces/bricks/tiles/gutters, etc. geez... Tora's work, too, especially comes to mind in terms of time spent on details. Love his stuff.

Yeah, he does great work, but months for one rendering to look "just right" isn't very cost effective or timely as far as most of our clients are concerned. They always want it in like 3 weeks or so. How much of this effort would you put into a project that scale. Keep in mind we're also an architecture firm so I get to spend at most half my day doing 3D work. Thanks.

Chris

[mverta]

I hear ya... I work under no shortage of deadlines, too. If I sleep 5 nights any week, it's a freakin' miracle, and I haven't seen a holiday in 7 years. That being the case, what I've done is build a library of materials as large as I can, so I'm building the fewest number/doing the least painting and can spend most of the time tweaking geometry. But honestly, there's no shortcut to the best work, time-wise. You can't throw money at it. You can sometimes throw people at it, but it depends on the scene. So it's very hard to say. In my job, if it's not "photoreal", it's useless, so usually there's a schedule for it, even if it's on the outside of ridiculous.

_Mike

[wagurto]

After reading this post I just understood the meaning of photorealism. I think that's not what I want or this is not what my clients will be willing to pay. Would you call the Gianni Melis work photo real? if is not how do you classify his wondrefull work? that's all that I want for my work, I would love to get that level of quality.
Thanks for this post

[mverta]

No I would not call GM's work photoreal. They're highly stylized. They're beautiful, but not photoreal. Photoreal is where no matter how hard you try, no matter how much you scrutinize the image, you simply can't tell it's CG. Let me rephrase that: no matter how much the most experienced, critical eye examines the image, it can't be distinguished from a photo. (There are a few people who call just about everything photoreal.) It has to pass muster with the most experienced, critical viewer. When absolutely nobody can tell, then it's photoreal.

There are VERY few images -EVER- that truly can be called photoreal, and a ton that are VERY close. Personally, I don't think I've ever actually gotten photoreal, I've just gotten very close a few times. Close enough for my clients, anyway. Not necessarily the highest bar, sometimes.

_Mike

[lsega77]

Just added a new wip to the top of the thread. did some brick 'painting' just trying to see if I'm on track with the brick.

Luis

[iandavis]

Yeah, I concur with mike. photo real needs to be indistinguishable from a photograph. I have only seen a couple images that fit that bill. (movies don't count). It's relatively easy to fool someone if they only have 3 seconds to analyze a shot. A still image is a whole different and much higher bar.

I have seen many images that were shockingly real, but after a few seconds I usually find 10 little signposts to it's synthetic origin.

Mike, your R2 is pretty drn close. If I for a second thought R2 actually existed, well, then it would be a whole different ballgame.

Most people shoot for "realistic-ish". Realistic enough to fool MOST people. Most of my clients can be fooled quite easily, and unless your a CG artist you just don't see most of the issues. Lucky for us really... if they all were as critical as we are, I'd likely be out of business.

this is my idea of photorealistic,
http://www.iandavis.ws/maxwell/mini_rear.jpg
Only one of the many clients I have shown this image thought it was CG, the others assumed it was one of my photographs. This image has zero effort in the grunge department, all textures are straight maxwell with no texture maps save the background and the floor/ground.
In other words, one doesn't need to spend weeks and weeks getting everything just right. Getting the lighting and most of the critical bits right and 90% of clients will just freak. The trick is learning which bits are the biggest giveaways, after that it's much easier.

cheers

------------------

The newest version is much better. The lighting on my screen still looks a bit low contrast (brighter than natural shadows) however it could be my mac display. The addition of clouds and window reflections are excellent. I'd say you have moved this forward quite a bit. The curb looks about the right scale too. I understand it's a huge amount of work... keep it up, it's looking great.

[Peder]

Luis! I hope you don't mind if I took your image and played a little with it as an exercise for myself?

Anyway I thought I'd post it back and see what you think. I added some more colour and warmed up the brickwork -it seemed the image looked very desaturated. Also I randomly burned your bricks to give variation. Further I straightened out the perspective since I thought the building looked a bit squat. Unfortunately that was impossible to do without loosing the corners in this image, but that could easily be fixed if you rendered the view with a bit more background. Also I cropped the image to focus more on the building and loose some of that large ground in the foreground.



Regards,
Peder

[Tea_Bag]

Is it better to create textures in photoshop by creating the multiple layers as a single texture then use it in maxwell? Just wondering cause I thought creating multiple layers in maxwell would be more realistic!?

[mverta]

Yes. No. Maybe. Depends. It's best to build textures in layers in Photoshop so you can split portions out to build better layers in Maxwell, if that's what the material calls for.

_Mike

[sidenimjay]

seems every response is so fixated on textures. while they are at the top
of the list for achieving photoreal images, they are only a third of the equation

the other 2 are lighting and modelling.

if you look at your model, there are way too many straight lines, IMHO

it looks as though one could cut themselves on the sharp edges of the
buildings. therefore it seems as if everything in this image was made
by a machine instead of human hands, or even human hands using machines.

the sidewalk is perfect, the edges of the bricks are perfect , the tree is so
straight, no knots or twists etc.... gives it a feeling as though noone lives
there to muck things up a bit
(just noticed the tree trunk is not in frame, its the traffic light)

i have seen plenty of lambert shaded renders that look damn good , purely because the model is not so perfect..... case in point is the clay renderings
of the house set from maxwell home page. not sure if it is still there, but the
modelling along with the lighting makes that image look at least as though it were a set from nightmare before christmas that has yet to be painted.

so on top of the texturing , loosen up the edges a little to make it a little
less perfect, think of it as dirt maps for modelling, i think it will help alot.

hope this gives some added direction for you

[lsega77]

@sidenmjay

thanks for the comments on the model. I've been choppin it up for a bit now.

Luis

[Tea_Bag]

Cheers for that Mike!

[iandavis]

Sidenimjay,

nice observation. I suppose texture is harped on because it can provide the easy quick fix, where modelling and lighting is much harder and has often a more subtle effect on the outcome.

For sure the big three are modelling, lighting and texturing. Personally being a photographer I believe that lighting and texture are equal most of the time. In fact, bad texture and low poly models can be very convincing with good lighting. So, yeah, lighting is important.

Problem is I suppose it's a case of 'one thing at a time' if we talked about everything all at once, there would be brain aneurisms all around.