Jump to content

Film-Quality Renderings?


Wayne Ken
 Share

Recommended Posts

Dear All,

 

I was browsing on XSI website yesterday, when I come across with a term there, something like "film-quality rendering".

 

I've been in architectural visualization industry for more than 6 years by now, and seen so many nice photoreal renderings out there. But the thing is, we're rarely see a "film-quality" architectural renderings so far.

 

What I mean with "film-quality" here is like this: When you see a nice photoreal rendering, imagine that you're watching a widescreen movie in cinema (with probably Tom Cruise walking around in your scene :p ). Then suddenly you will feel that those nice photoreal renderings are not real enough.

 

So the issue is what's the key items that differentiate our architectural renderings from those 3D scene produced for films? Is it the software, the techniques, the media or the users?

 

Any opinions or tips'n tricks are welcomed. :)

 

Regards,

Wayne Ken

Link to comment
Share on other sites

Time-Money-Talent...

 

Give me a a year to complete, a couple million dollars, a team of specialists, a 600 computer renderfarm...

and I'll give you that film quality.

 

Sorry, don't mean to sound like a smartass, truth is, we get a week to do it, not a huge amount of money, and sometimes a one man team...

Link to comment
Share on other sites

i think we try to recreate the architectural space in its reality, and obviuosly the artists style will be left on the image to some degree, but in the films the architectural spaces are a tool that can be used to enhance the story. they can be altered and manipualted to end up at something that would not occur as strongly or obvioulsy in reality. i think films bend the limits of reality to enhance story telling - although often too much of this can ruin a film (mission impossible / new Bond etc) when the story line becomes a means of added massive effects rarther than the effects adding to the story line.

Link to comment
Share on other sites

There is something you can add to our work to get that cinematic feel and it doesn't cost any money or need any extra technology or processing power and it's quite easy to do.

 

Just add dirt! It has no place in our work apart from a subtle application for realism but look at any movie and you'll see it everywhere.

It's in the textures, the weathered models and even the atmosphere.

 

Post your best Arch viz work on CGTalk or similar and you'll get the same reply: "add dirt".

 

The scale of the scene is another thing.

 

Check out Yiran's work posted here.

Link to comment
Share on other sites

yYrin's web site .

Just thought i would let you know.

Everytime i go to the website for yarin's company to look at their work, our office security system blocks a trojan virus coming in.

We have a very good system.

They should ge it sorted.

 

So be warned or tell them, as I have blocked it out completely.

 

phil

Link to comment
Share on other sites

Dirt - that's immedieately what came to mind when I read the question. The goal in ArchVis is to showcase and sell a realistic, yet aesthetic or flattering view of a room, building, etc. The goal in film is simply selling the shot to make it look as real as possible, ray-traced marble floors be damned! Also I think they tend to utilize a lot more compositing and post effects in film than you might see in an average arch render. You can add a lot of atmospheric elements such as dust that would add to the realism. Just the fact that a scene is moving and can use motion blur, would probably seem more 'real' than a still interior. Just my two cents. Christopher Nichols would probably know the most about this subject.

Link to comment
Share on other sites

OK... wrote some notes:

 

Modeling:

- bevel all the corners of buildings to catch speculars on the edge

- even a perfect building is never perfect

- only model what you see

- model to a pixel accuracy not a 1/8th of an inch

 

Texturing:

- UV unwrap everything

- Provide high res textures for everything

- Provide at a MINIMUM color, bump, and specular for every surface

- dirt is pretty, dirt is detail, dirt is scale

 

Rendering:

- Use high quality AA and texture filtering

- use motion blur on every shot no matter what

- Render at LEAST 10bit log cineon, if not full HDR or EXR format

- Render LOTS of passes for compositing

 

Compositing:

- Match the focus with a circle convolve (not just a blur)

- Carefully match the film grain in all channels (blue is the hardest)

- color shift all the blacks to match the color timed plate. Remember there is no such thing as 100% black.

- Edgeblur all the CG elements to make the less CG in your scene. If you used proper fresnel reflection with your plates, you can reduce the amount that is needed.

Link to comment
Share on other sites

Time-Money-Talent...

 

Give me a a year to complete, a couple million dollars, a team of specialists, a 600 computer renderfarm...

and I'll give you that film quality.

 

Sorry, don't mean to sound like a smartass, truth is, we get a week to do it, not a huge amount of money, and sometimes a one man team...

 

I have seen amazing stuff being done with very little money and a team of 10 people. Battlestar Galactica comes to mind. I think it goes beyond money, man power, and render farms. A lot is related to knowledge, training, and experience. Don't want to tut my own harn but the work I did for the freeway scene for my DVD was done by myself, in a week. Many would consider that "film quality" including many of my co-workers.

Link to comment
Share on other sites

That was pretty awsome Chris.

 

Could you elaborate a bit more on pixel accuracy? I assume you mean something along the lines of Make it Look Good as opposed to accurate but I don't know.

 

Also I know this goes around a lot here but how do I use multi passes? What "Math" do I need to use for the specular, diffuse, ambient occulsion etc..

 

Thanks

Link to comment
Share on other sites

Could you elaborate a bit more on pixel accuracy?

 

I believe Chris' point is that your accuracy should not exceed anything that will be represented as less than a pixel in your final comp. For example, forground detail is modelled to 1/4", scene detail to 1/2", background detail to 2" accuracy (I made up the numbers to illustrate). As items get further back in the scene, they are represented by fewer pixels in the image/comp and can therefore have less detail.

 

The difficulty in comparing Arch ill to film/tv work, is that often arch ill views will be changed at the last second, thereby screwing all strategy to conserve detail/accuracy. The people I know in the film biz, generally have the camera path/ angles worked out from the director before they begin.

Link to comment
Share on other sites

Ok that was what I was wondering as almost every client I have ever talked with completely disregards all of my advice about setting a view first. I try but I always get to change things at the last minute. I am so friggin lucky go ahead rub my belly for luck maybe you will be as lucky.

Link to comment
Share on other sites

The difficulty in comparing Arch ill to film/tv work, is that often arch ill views will be changed at the last second, thereby screwing all strategy to conserve detail/accuracy. The people I know in the film biz, generally have the camera path/ angles worked out from the director before they begin.

 

That is very true... but there are changes that take place in the film industry as well. You can always do the same thing that the filmbiz people do and solve those issues with previz. Build a super low res version of your scene and solve all your views etc there. Then put it in your contract that anything outside of the previz session is an added service.

 

It could be like this:

Start with a previz and charge them for that. Go through the previz session and come up with a cost based on what is decided there. Then if changes are made after that, it will cost them extra. That is the way it is done on films. Often the previz people can end up being a different company than the primary.

Link to comment
Share on other sites

 

Also I know this goes around a lot here but how do I use multi passes? What "Math" do I need to use for the specular, diffuse, ambient occulsion etc..

 

Thanks

 

That really gets into compositing. I can only recommend a training DVD that woudl be a good start, if you are really interested in the subject.

 

http://www.thegnomonworkshop.com/dvds/rne01.html

 

I should note that Rob is a friend of mine and I have worked on several features with him.

Link to comment
Share on other sites

OK... wrote some notes:

 

Modeling:

- bevel all the corners of buildings to catch speculars on the edge

- even a perfect building is never perfect

- only model what you see

- model to a pixel accuracy not a 1/8th of an inch

 

Texturing:

- UV unwrap everything

- Provide high res textures for everything

- Provide at a MINIMUM color, bump, and specular for every surface

- dirt is pretty, dirt is detail, dirt is scale

 

Rendering:

- Use high quality AA and texture filtering

- use motion blur on every shot no matter what

- Render at LEAST 10bit log cineon, if not full HDR or EXR format

- Render LOTS of passes for compositing

 

Compositing:

- Match the focus with a circle convolve (not just a blur)

- Carefully match the film grain in all channels (blue is the hardest)

- color shift all the blacks to match the color timed plate. Remember there is no such thing as 100% black.

- Edgeblur all the CG elements to make the less CG in your scene. If you used proper fresnel reflection with your plates, you can reduce the amount that is needed.

 

 

Wow! This is great! Thanks for sharing, Chris!!

Link to comment
Share on other sites

I have seen amazing stuff being done with very little money and a team of 10 people. Battlestar Galactica comes to mind. I think it goes beyond money, man power, and render farms. A lot is related to knowledge, training, and experience. Don't want to tut my own harn but the work I did for the freeway scene for my DVD was done by myself, in a week. Many would consider that "film quality" including many of my co-workers.

 

There's nothing wrong with tuting your own horn, but why do you think there is such a lack of film-quality in arch/viz field ?

 

Also when you talk about very little money, how much are you talking about ?

Link to comment
Share on other sites

Also when you talk about very little money, how much are you talking about ?

 

Well it is hard to give specific $$. But you hear stories all the time. Volker Engel, who I met several times, started his shop with exactly that in mind, where he made all the VFX for the Coronado some insanely small budget.

 

http://www.uncharted-territory.com/

 

I spoke to someone yesterday that mentioned that the nightcrawler effect of X2 was written (software from scratch) and put into shots my mainly one guy in around 2 months.

 

Then if you want to super low cost, there is always this little piece of history:

 

http://www.405themovie.com/Home.asp

 

On the other hand, trust me, I have seen my share of ubber expensive million $ shots as well. I worked on Day After Tomorrow.

 

There's nothing wrong with tuting your own horn, but why do you think there is such a lack of film-quality in arch/viz field ?

 

Well besides the reasons I gave, I think there are other issues. Time mangement is one of them. Architects spend 95% of their time modeling. I think architecture models are generally over built, and details are put in the wrong place (this is a very general statement when comparing the VFX models). I think that many people live in the fear that the model will change and the view will change and they over build and overplan. I also remember when I did archviz, I spent so much time dealing with poorly made models. I used to blame it on the fact that they made it in FormZ, but I think it is generally just poor modeling as well. Coplanner faces, etc. If they were more efficient at modeling, and instead of doing the work in 1 week, do it in 2 weeks, I think the level of quality could go up a lot more.

 

But many new skill would have to be adapted. Proper texturing, more time on shaders and lighting, better compositing skills (not just A over B)... learn how the properly move cameras for animations.

 

The other advantage that VFX has is that they are working for the film industry, so we already have access to high quality cinematography. Camera moves are very... well... cinematic. The DP (director of photography, aka the man behind the camera), has planned out the lighting perfectly, all we have to do is match it. The added bonus, is that we have learned from all these things.

 

Sure you can't do it all in archviz, but I tell you it gets better every day. The CGarchitect Siggraph party shows me that the field is getting better every year. Archviz has matured and grown a lot in the last 3 years. You can thank places like cgarchitect for making that happen.

Link to comment
Share on other sites

Would that DVD be applicable to non-Nuke apps? Especially combustion or heaven forbid photoshop?

 

Thanks

 

Found this very helpful.....

 

Math anyway....Digital Compositing for Film And Video-Steve Wright-Focal Press. It works in Phot0Chop, After Effects, Combustion...really I mean really helps to understand shake. It's the the math/process that is compositing. Way cool, mind melting at first, but I'd say it's a must read if you want to get on top the compositing curve. and read again, and again and agian LOL :p

 

WDA

Link to comment
Share on other sites

Found this very helpful.....

 

Math anyway....Digital Compositing for Film And Video-Steve Wright-Focal Press. It works in Phot0Chop, After Effects, Combustion...really I mean really helps to understand shake. It's the the math/process that is compositing. Way cool, mind melting at first, but I'd say it's a must read if you want to get on top the compositing curve. and read again, and again and agian LOL :p

 

WDA

 

That is a good one. For books on Compositing, besides the one you mentioned, there is also this one "The Art and Science of Digital Compositing" which is somewhat the bible of compositing. Starts with a lot of the history of in camera stuff... really cool.

Link to comment
Share on other sites

That is a good one. For books on Compositing, besides the one you mentioned, there is also this one "The Art and Science of Digital Compositing" which is somewhat the bible of compositing. Starts with a lot of the history of in camera stuff... really cool.

 

Have to take a look at that one....more technique/methodology? The history of rotoscoping in there too?

 

WDA

Link to comment
Share on other sites

It's a pretty big book, has a lot of everything... roto work is hard. There was a guy at DD that had done Roto from the old days when they painted every frame on glass.

 

LOL..did some roto with Curious GFX (color person/BW BG from one plate) and was thinking if this is supposed to be one of the the easier method/apps....my god did they all drink heavily to numb the pain in the days of painting the rotos. Guess it's no different than drafting large commercial or areospace, pencil-pen-bottle of Jack Daniels :p

 

Thanks for the info!!!

WDA

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...