Apoco 77 Making of

Posted by on Dec 13, 2016 in Portfolio | 6 comments

azulosa

The models and assets for this scene are available at my turbosquid : http://www.turbosquid.com/Search/Artists/zuliban

Before i start i would like to address some of the most asked Questions over Facebook and the Forums.

Q:What computer do you have? did you used a render farm for this work?

A: I Used a quad core i7 cpu with 32gb ram and a GTX 970 yep… i took along time to render all these many images.
I rendered these images overnight while on the day i worked in something else so each image is the same render time, around 8 hours render time.
That means 2 months of render time overnight, i did not wanted to upgrade to another computer in the middle of the project
and i was thinking of rendering it a render farm but i was never in a hurry since this was a personal work.

Q: why did  you use Corona render?

A: Im always looking to try new things and learn new things either to have a better workflow or just for curiosity and understanding,
i found Corona when it was beta stage i had my doubts about stability and bugs but Corona was really good even in beta so i went for it and
rendered the final scene in Corona also…. i liked the name Corona since in my city they build a beer called Corona too ;).

What i liked of Corona when i made this scene was:

– super fast.
– no need to tweak settings this gives me more time to focus on the artistic side and not on the technical side.
– beautiful glass shader.
– superior Highlights reflections to other renders.
– ”unbiased” Quality that show all the small shadows from indirect lighting and light bounces and so on.
– for me this is a big plus more than the other points, Energetic  Developers great forum community.

Q: why Nuke and not Photoshop?

A: I honesty would use Photoshop over Nuke i like more Photoshop and its interface the problem is Photoshop right
now is limited in the 32 bit workflow there are a lot of non usable tools.

Q:what is the advantage of 32bit vs a traditional method?

A:in 32bit you have the maximum manipulation of your images that can be like tone mapping, color correction , burned areas,
effects such glares and flares without any of the downsides of artifacts noise or ”bands” around colors.

the downside is you need to render in lineal space to not ”damage” the HDR data that means that your images need to be
as neutral as possible this is bad too since you don’t have a general ”feeling” of how your image will look until you do
post production.

Q: How long did this took?

A: When i work in personal projects im never constant i worked this one for example 2 weeks non stop and  stopped working on it for like 3 months and so on until finished i think this one would have taken  me 1 month to do everything and like i said like 60 days of overnight renders for each image.

———————————————————

EARLY STAGE

When i start any scene i want to have a general idea of what i want  for example in this scene
i wanted an exterior  with a pool, i wanted the lighting to be a bit warm with a touch of blue tones to make it look more fresh.
pure black colors and vivid colors.

I want to say i never pay any attention to realism and it never cross in my mind when i do any of my works i put in the highest priority the
idea and the overall mood, colors, etc , realism is something extra that comes by itself sometimes i want more stylized renders than realistic ones.

I started this scene on the kitchen area and moved from there.
i always do basic models sometimes place holders, most of the time i never put attention at this early stage to any kind of texture
i focus mainly in the shape of the objects and the lighting overall feeling but i never stay there too much just do one or 2 quick
renders and focus on the shapes if i feel something does not look right i discard the model and model something else until im happy.

Then i move to another area and do the same for me this is like sketching  and this gives me general idea of what the scene will look like
i never do any serious modeling or texturing until i know for sure what i will do will fit in the scene even with this kind of mindset i know
at the end some models will not look good in the scene even if they are polished they will look out of place their color or shape will not match
and so on doing these ”sketches” avoid most of this trouble.

Staying so much in the same area will slow you down until you feel you are not moving or making any progress.

This is the most important part for me as soon i see the work start to look like i want now i can really focus on modeling.

Here are some images of the Early parts of the scene  and how it evolved:

rad

xxlight-flos-mini-glo-ball-t-table-lamp-xl4

Here are some more parts of the scene from early stages i always do this workflow in every kind of 3d work i do, it does not matter if is
an interior, character or environment spending more time in the early stages sometimes feel like you are not progressing
but it is the pillar that support the whole finished work.

xcdasdxxxvphoto_1412300243087dggasdef
When i feel i have this early stage finished now i can finally focus on modeling , unwrapping and texturing.

I never touch rendering again until the end i just do quick renders for each model to see if they will look good on the scene.

Now i will show the most asked models and parts of the scene that people asked me for.

This is the wireframe and low poly of the Standard chair and it’s highpoly version.
1It is very important to have all the edges smooth for closeups  and have a correct mesh and wire flow.
2After i have the modeling part done the next step is to have a non distorted UV, this is a time consuming task and sometimes is boring but is as important as any other step.
5This is a screenshot of the low polygon version and how it looks with the UVW’s on the side.
3this is the final version of this part with turbo-smooth added and how it looks.

4The next Step is this metal structure.

6These are the seams of the UV on this metal object , the main problem is is to have  continuity from the textures that will be here if done wrong when the texturing is applied it will show these
seams.

7i created these red lines to know if my seams have a continuity a good flow for the textures, it is important to spend sometime here and know that when you apply any texture they will look
seamless.

8As you can see the red lines don’t show any kind of distortion  when they go through the green UV seam lines.

9When this step is done now the main focus are the little weld parts of the chair there are done in Zbrush and exported back to max as bump or displacement maps.

10It is very important to store  a ” Morph Target ” since Zbrush will alter the mesh when you model  in this way you have protected the mesh from max and is needed to extract the displacement map correctly.

12The modeling of the weld is nothing hard or fancy is done with a standard brush and flatten brush .

13This is how it looks at the end, the white parts are the weld areas.

silla1This is a preview rendering  i do these only when i consider the models finished i just use a HDRi to see if it will look good in the scene most of the times i do test several HDRis to see it in a different light condition.
silla2silla3silla4silla5Here are another two examples of using this same workflow.
Basic modeling in Max—->check UV’s—–> detail in Zbrush. In this next one i used polypaint for the mask of the material then do some quick HDRi test.
v1v2v3v4dddHere is another one
v5v6v7v863The next request i got asked alot was the outside wall.

Here is a screenshot of the Photoshop layers.
wall-psAs you can see i use a lot of layers for the best control of my textures,i usually do a base layer and in top of that i start adding detail.
Is good to have the maximum control over your textures in case you want to change something like colors or make the texture cleaner and not too dirty etc etc .

Here is a step by step on how i did this texture:
pic1pic2pic3pic4pic7pic8pic9
The Sofa was another requested model , i start with  a low poly model with rigid edges at this point it does not matter if it looks too rigid since it will be re shaped in Zbrush.
v1Here is how it look after some time in Zbrush as you can see  there are not straight lines anymore and shapes are more chaotic.
v2After the main sofa model is polished i used Marvelous Designer for the main shapes of the pillows these pillow will be further detailed in Zbrush and corrected.
v3This is the marvelous designer topology in Zbrush as you can see the mesh is very messy and the topology incorrect it does not have a flow even when MD now works with Quads  if you want to detail this pillow more it will create alot of pinched polygons.
v4After playing with Zbrush remesher and decimator i created a better looking mesh keeping the same shape  now this model can be detailed even more.v5v6After detailing the models adding textures and so on these are the test renders i made and considered this model and area finished.
omnevvvxxdLast part is the lighting , rendering and post production.
For the lighting there is no magic trick i just try to keep it simple add the light sources never use any kind of ”trick” like light panels in front of windows or filler lights.
These are all the lights on the scene  just the sun and the artificial lights the indirect lighting from the  sky is a gradient map  i did not use a HDRi since most of the commercial HDRi are badly done and create a lot of noise.
pic4For the rendering  Corona is too simple i did not move any kind of settings i used mostly bucket rendering for previews all my renders where rendered and saved as EXR for postwork i did not add any kind of lighting mood in this stage since im working in 32bit and it is better to have a very neutral image.

Right now Corona 1.5 is out and you have LUTs, glares and flares and a lot more of post production was added that you don’t even need an exterior post production program .

I used nuke instead of Photoshop because Photoshop is still not fully 32 bit ready .

Nuke have way too many settings and to add this in this tutorial i will need a lot of time to do a step by step or do a video , i can easily say i use around 120 or more moves to create a single image for this i will add a small evolution of my image from RAW to finished .

Heres one example.
9101112
Here is another example.

13141516
So this is the final part of this ” making of ”thank you for reading it and i hope you find this useful and of your liking.

 

6 Comments

  1. Excelente trabajo Zuliban, eres un gran 3dArtist, muchas gracias por compartir esta valiosa información.
    Saludos.

  2. hI zULIBAN…I CAN ONLSY SAY…NO WORDS..AS AAALLL YOUR WORKS!!!…i’d like to ask you if possible to publish a screenshot of sun settings and gradient map used in bk. other question..the artificial light have a .ies profile or just default (sphere, plane or so) and standard color and intensity?..
    thanks for all your work..

    lillo

  3. Very learning stuff,
    Thanks for sharing…!

  4. Nice work! I do not use these tools but it looks very interesting! 🙂

  5. Zuliban you are the Best! thank you for inspiration

  6. Impresionante…
    Me intriga el que uses un gradient como luz indirecta y me surge una duda ¿preparas un gradient a 32b en photoshop o usas el default de Max?

Leave a Reply

Your email address will not be published. Required fields are marked *


7 − 5 =