Discussion About Arch_Vis - PhotoRealism - HyperRealism

[alfa smyrna]

I am working on a image for the last 2 weeks, only for practice and self-training purposes and have decided to send this for the CGArchitect Awards. (Hope to publish soon )

 

When I made this decision, my confusion started.

Because I am trying to make my image as realistic as possible; to challenge myself, to experiment and learn new techniques and for self-training purposes.

 

To make it realistic, I paint textures with dirt, deform corners of geometry etc etc. Lİke the real world. And it really looks more realistic!

 

But at that point, I started to question myself, what are the lines that define architectural visualisation.

Because a client ask us to present their buildings in the best possible way. Like in tv commercials.. Trying to make the product as appealing, nice, sweet, attractive as possible.

And dirty old look is not part of this.

Honestly I couldn't find the answer.

 

I am aware that my scene is for practice purposes and there is no client and I can do whatever I like

I like to go for hyper-realism and it is fun to study and experiment on this.

 

But I wanted to know what your opinions are about this matter.

 

Is hyper-realism against nature of ArchVis?

Or are we yet at the begining of the path and because of this are we exaggerating the imperfections yet but sooner, with practice will we find our ways to make buildings realistic as a photo but also nice and appealing?

 

( I dont know if it is necessary to add but to avoid misunderstading : I am not against photo-realism and I dont criticise any photo-real works. Instead I love photo-real works.

Just opening up a discussion to hear different ideas on this paradox)

Edited April 22, 2009 by pixela

[Dave Buckley]

The nature of Arch-Viz is whatever the client specifies.

 

There is no right or wrong as such, but generally your clients are going to want their buildings/developments to look as nice as possible, which generally means clean, and can therefore sometimes lead to images looking not as realistic as they would if you took it to the Nth level of detail with regards to dirt/cracks/deformities etc.

 

But it really does depend on the client and also the task in hand. It also depends on what the images are being used for i guess, marketing or planning etc etc

 

Don't limit yourself to one style/technique. I made the mistake of thinking there was a set way to do things, and i was looking for a none-existent answer, i spent far too long trying to answer the question you have just asked.

 

However, people may disagree with me, this is just my opinion

[STRAT]

exactly what Dave say.

 

in the industry you draw to satisfy your client. generally, clients dont want to see dirt and grime, which might be more real in some instances (not all of course, just think of a nice brand new building. no dirt and grime there). so you do what your client wants. no right or wrong.

 

and for personal work, again, you do what makes you happy.

[alfa smyrna]

Thanks for replies

I understand you and get your points.

 

And about brand new buildings : Actually there are imperfections on them too. For example: their surface is never pure vertical as in cg but has thicker or thinner parts on stucco (not the bump ) , or craftsmanship is never pure perfect and paint is always wavy with a sharp careful eye. These example can be infinite.

 

But I dont think there are details a contractor or anyone who is selling the building would like to see and have there.

Yet they will make the new building look more realistic, like a real photo has been made after building is finished.

[dean@pikcells]

Personally I tend to find a compromise between "CG realism" and "Actual realism".

 

For example, when creating roads, I always add tyre marks, but I wouldn't add pot holes

 

Arch vis industry generally uses our skills to create images to sell buildings. Other areas of CG are different, and so different approaches are needed.

 

You might get a more varied response from other forums such as cgtalk, where there are multiple industries who have different views on realism. A games company won't push for photo real graphics, but they have other ways of immersing the user into their games, so you believe what you are playing. This could be seen as realism.

[Dave Buckley]

Yet they will make the new building look more realistic, like a real photo has been made after building is finished.

 

'Will' is the key word in the above statement. I disagree, it purely depends on the clients style and again back to what the image is being used for. Some clients may like photoreal to give a true representation, whereas others may prefer a more artistic style. So they won't necessarily want to make the image more realistic. As already said, it purely depends on the task in hand. i believe the term 'Arch-Viz' is a very broad subject that you can't narrow down to one style - hence why if you look on most viz company websites, they split there portfolio's down into categories displaying a number of different styles

[alfa smyrna]

'Will' is the key word in the above statement. I disagree, it purely depends on the clients style and again back to what the image is being used for. Some clients may like photoreal to give a true representation, whereas others may prefer a more artistic style. So they won't necessarily want to make the image more realistic. As already said, it purely depends on the task in hand. i believe the term 'Arch-Viz' is a very broad subject that you can't narrow down to one style - hence why if you look on most viz company websites, they split there portfolio's down into categories displaying a number of different styles

 

Yes I totally agree with you and you are right.

Actually my target was mainly to discuss and hear opininons about Realistic Style itself. "PhotoRealism in Arch-Vis".

In my works, I also used a hybrid style like dean@pikcells.

Added some small details to make it more realistic, but so small details that hardly noticable, hardly seen but perceived.

I tried to go after this in my works.

But I was wondering what other friends think who are working in this style.

[BVI]

It's a balancing act - Check out Alex Romans awesome illustrations - they are just amazing and the pinnacle of real - but we would never be able to do that for a client - the heavy DOF and focus on foreground elements would have them pulling their hair out. We recently did some heavy 'Romanesque' post on a project - only to be asked why the images look blurry... Alex even states that most of his photo real stuff is personal work.

 

Having said that, its always nice to add some realistic dirt mapping etc...

[rmccoy]

I personally prefer to try for photorealism whenever possible. Depending on what phase the project is in can dictate what style should be used. if it's at the beginning of a project then artistic styles lend better to getting people interested. If you are submitting your building for review to a city/county/state then if you have it photorealistic it can actually end up being a bad thing. They will expect to see what is in the rendering for the final project even though things may/will change between this stage and how the building will actually be made. Keeping things a little vague can actually help people get a better understanding of the overall concept of the thing. Towards the end of a project getting to the hyperreal stage can be better so you can show the details that have been devloped since the project first came along

 

It all varies depending on if you are working freelance, for an arch vis company or as an inhouse 3d person for an architect or engineer. In the end though it's who is paying your bills that gets to dictate what they want.

[morri]

I personally prefer to try for photorealism whenever possible. Depending on what phase the project is in can dictate what style should be used. if it's at the beginning of a project then artistic styles lend better to getting people interested. If you are submitting your building for review to a city/county/state then if you have it photorealistic it can actually end up being a bad thing. They will expect to see what is in the rendering for the final project even though things may/will change between this stage and how the building will actually be made. Keeping things a little vague can actually help people get a better understanding of the overall concept of the thing. Towards the end of a project getting to the hyperreal stage can be better so you can show the details that have been devloped since the project first came along

 

It all varies depending on if you are working freelance, for an arch vis company or as an inhouse 3d person for an architect or engineer. In the end though it's who is paying your bills that gets to dictate what they want.

 

Have you noticed how "drawing" has made a comeback? I find a little CGI fatigue out there. Everyone can do them so nobody is wowed by it any more. I would look for style in a rendering over hyper- realism. A less defined image allows people to see what they want to see by completing the details themselves.

[William Alexander]

exactly what Dave say.

 

in the industry you draw to satisfy your client. generally, clients dont want to see dirt and grime, which might be more real in some instances (not all of course, just think of a nice brand new building. no dirt and grime there). so you do what your client wants. no right or wrong.

 

and for personal work, again, you do what makes you happy.

 

No dirt and grime is not true. There is dust, there is rain, there are on going construction elements that create dirt and grime before-from the time construction is put into place. That is a CGVFX theory-application for realism. Even the most pristine of objects has imperfections and physical effects from it's enviroment.

 

I personally do not condone heavy handed use of CG dirt and grime, very light use of those techniques can alter an image from obviously CG perfect to realistic. Barely percievable use. Just enough to effect uniform pixels and vary them up. Somewhat akin to underpainting. Not CSI Miami underpainting but just a touch to break up uniform areas of pixels.

 

It's all about subtle layering. Subtle being key operative. My current job is pretty much all about P-lam cabinets... Overseeing production, shop drawings and oddly enough frequently 3D visualaztion for a number of uses. So my stuff is plain jane boxes, in relative real world lighting, work very hard to get accuarate reflection, specularaity, frensel and colorations. Even with all of that it's still to CG, by it's very nature.

 

AO pass is manadatory. I can't stand linear work flows, waste of my time doing iteration upon iteration of renders and tweaking. It's a personal problem The Ao pass in multiplied or overlayed in post. I'll tweak the levels on the AO pass if it seems off, to heavy or light. But by applying just enough visiblity it brings the image to life. As does using varied noise and 'grunge' maps as part of the material-shaders, subtle use of.

 

This light handed use of CG dirt and grime can help yield the 'Illustrated' mind filling in the blanks function-technique for a CG render. Well in my most humble experience

[alfa smyrna]

Thanks for replies

It is very nice to hear your ideas and approaches. As there is no right or worng answer for this question, and it is the reason I wanted to name it as discussion. It is very interesting for me to see different perspectives of the matter.

[Fran]

Trying for realism to visualize a design for a client is a different proposition than doing so to try to win a competition or impress your peers in the cg industry. A very good way to improve your skills and in turn benefit your client work, is to try to impress your cg peers. If your cg peers are nice people, they will offer helpful criticism. Not all of your cg peers are going to be nice people.

 

I agree that most clients will not want to see dirt and grime, but I think you are on the right track regarding subtle irregularities introduced either by geometry or materials. I only use ambient occlusion for massing studies and shaded views to allow the client to approve geometry and camera views. I personally don't like the effect it has on a color render.