Rendering, Post Production, The Next Steps

[benjaminbourgoin]

Hello,

 

My name is Ben, I'm an apprentice architect in Seattle. I did my undergraduate in architecture at the University of Washington and now i'm getting ready to go to graduate school. As I've progressed from simple line drawings to fully detailed renderings over the years i've hit some what of a plateau. I'm currently working on a 120k sf building and we just got through our design review process, when I sat down and really looked at the images I had produced for it, I realized that I don't like any of them. So I got to thinking, why don't I like them, what is it about a rendering or any visualization that draws you into them. What makes you feel an image. So far I've come up with several things... Images must have a focal point, specifically one focal point. When we look at things in real life, our brain allows us to focus on a single entity in our field of view, anything else is perifery (spelling?) therefore for an image to read as realistic, there must be a single focusing action. Agree/Disagree? Examples? I'm ranting now. Back to the topic.

 

The Plateau. I use sketchup, and Vray, and Rhino, and Grasshopper. These are adequate tools for fast work, the kind done in school. Most of the weight of the images are achieved in post production. I'd like to move away from that, especially as I've become more and more interested in animation, 3ds max seems like the next logical step for my learning curve. So i'm looking for advice/ comments on the image I made accompanying this (copyright Gordon Fleener Architects) and suggestions for moving forward in my visualization learning.

 

Also, one specific question: I see this effect all the time in animations, notably in Alex Roman's the Third and the Seventh, when a dark scene is progressively illuminated by lights flickering on as though someone were switching on each floors breaker one at a time. HOW DO YOU DO THIS?! I assume it has something to do with key frames where a light is activated but I don't understand how the flicker effect is achieved. Anyways, these types of effects are impossible to deal with in Sketchup with Vray and it's begining to cramp my creative flow.

 

I'm also sick of the way Sketchup handles large models or complex scenes. It's unfortunate 32 bit architecture not only limits itself but also the plugins that operate in it (and yes, I checked, Vray for sketchup runs its threads inside of the program so it can't operate in a 64 bit environment outside while rendering).

 

Can you tell I'm frustrated?

Any empathy?

 

Best to all,

Ben.

[carlotristan3d]

The things that gets lost on visualizers is that we tend to focus more on the technicalities, the message should never be lost on the medium. You should focus more on the design. The image IMHO is trying too hard to be 'artsy'. There's just too much gimmick in it, DOF, monochrome, athmosphere, etc. Take the best POV that best showcases the design, for starters. We're all guilty of these, trying too hard to make a beautiful rendering that we forget our end goal is to sell the design. take off your visualizer hat for a moment and put on your architect's hat- you'll see your image in a different way. Good luck.

[heni30]

I think there are some depressing aspects about the rendering. The trees look dead, all the same and monotonous. The building is too mono- value.

The open space that has the most interest potential is covered by trees. I just did a couple of highlights in PS in the glass and changed the tone to a warm sepia and already that makes a difference - 10 minutes.

 

In other words your next step should be more about expanding your aesthetic sense and attitude rather than focusing on tools as Scott Schroeder has mentioned in another thread.

Edited December 8, 2012 by heni30

[benjaminbourgoin]

[ATTACH=CONFIG]48537[/ATTACH]I think there are some depressing aspects about the rendering. The trees look dead, all the same and monotonous. The building is too mono- value.

The open space that has the most interest potential is covered by trees. I just did a couple of highlights in PS in the glass and changed the tone to a warm sepia and already that makes a difference - 10 minutes.

 

George, thanks for your post, an interesting interpretation of my image, although to me, the improvement doesn't come from the sepia, or the sky adjustments, it comes from what you commented on originally, the monotone. You desaturated the image to black and white before adding the sepia filter am i correct? When an image is black and white it often helps a lot because you're no longer so focused on materiality as you are on lightness and contrast. I agree with you, this helps, and if I had it my way most of my renders would be in black and white with very little color.

[heni30]

You're right about the desaturation of the tones - that's why the trees blend in better. Actually, hitting the colorize button does that automatically - everything becomes gradations of whatever color you have as your foreground color in the tool box.

 

But also I think the glass highlights add a lot. That window on the left comes alive with the PS dodge. Also the line of balconies has a dodge 2/3 of the way down which breaks the regularity and suggests light and atmosphere coming from outside the image. I'm also a big fan of this style but I haven't had time to explore it. I guess Hugh Ferris was the master of black and white atmosphere.

Edited December 8, 2012 by heni30

[Valtiel]

Agh, accidentally backed out of this page while typing this reply the first time and lost the whole thing. Hopefully I restored it all but I took several days to come back and retype it >_

 

This discussion on desaturation does remind me of some advice I often hear from concept artists and matte painters whom work primarily in 2D. It's something you can use to develop your viz products independent of software limitations. I'm not attempting critique your posted piece with the below points though I do like the directions you're exploring with it.

 

1: Consider the silhouette of your model in black & white when choosing a camera angle and position. This is primarily done by character concept artists when designing a figure. Before going into the more rendered features, they work in B&W to find a silhouette that is interesting and expresses a distinct character from a basic humanoid figure. So in Archviz the application of moving a camera around a set structural form and evaluating the silhouette of your subject with an alpha channel saved to a B&W jpg is different than the quick, gestural brushstrokes they use in 2D to create novel geometries but I think the principle applies. Said principle is simply that the elements of colour, lighting, and details all overlay the structural geometry and inherit, for better or worse, its quality; A good silhouette is a fundamental part of a good composition and the best lighting, colour, details, plants, materials etc. will still feel awkward on a poorly conceived sillhouette.

 

2: Desaturate everything temporarily - Work with values. This is more in line with the thread discussion and is a technique you would use in post-production to evaluate the compositional and lighting qualities of your piece. Even when your final product is in full-colour, it helps to simplify your evaluation by removing colour and looking at a desaturated copy and seeing how the luminance values perform by themselves. It may help you find issues in contrast or subject framing that may have been obscured by the play of colour, which introduces it's own dynamics of contrast, focus, and detail.

 

3: Flip your image horizontally - Does it still perform aesthetically? This can be combined with working copies of your composition in post as well as the sillhouette's aesthetic performance in the first point. The idea is that after working on an image for so long we develop a familiarity/comfort/bias with what we're working with and may become less able to identify issues with the composition. Temporarily flipping the whole image horizontally allows us to look at our piece with a fresher set of eyes, more subjectively removed from it, and better able to find things we've become complacent with. I've heard several 2D artists say that if your composition doesn't look right while flipped left-right, then there's something wrong with it's original orientation.

 

Hope that inspires something, I've always found it fascinating advice.

 

Cheers!

Riley

Edited December 11, 2012 by Valtiel

[Nic H]

I think if you are serious about it you need to stop using sketchup and rhino - and get into 3dsmax...

 

if your aim is alex roman quality (a lofty goal!) you need to develop your eye in tandem with improving your tools - there is no real shortcut except hard work!

 

i have spent the last 2 days learning mudbox and its pretty sombering once all the tutorials and exercises are done - the realisation of how much more time and work its going to take me to become professionally competent in it.

[Chris MacDonald]

I don't agree with that at all. You can get out of Sketchup what you put into it. Look at Peter Guthries works for example.

[Francisco Penaloza]

Hi, I think is not a point to argue here if Sketch up can or cannot do good renderings, for me it is always about the artist/designer to see how far a tool or a set of tools can go.

Now regarding your questions, Arch Viz has advanced so much these past 10 years, that stepped from being a different way to show a concept or idea, to an almost cinematic way to show and sell a project.

Now these 2 basic styles still work together, I produce tons of renderings weekly and they go from simply quick to full video presentations. Using Sketchup, Max, Aftereffects and what not; again the point is preparing the image for what it is need it. Sketchup is a great tool and if you add VRay to it, or Maxell or any other 3rd party rendering you can get great result, but yes it is limited to 32 bits memory, it cannot handle animations, it cannot handle proxy or very large environment, but again Sketchup was not designed for that. 3dsMax, Cinema 4D, Modo, Blender, Maya and more are very robust 3D applications that you can go way further creating 3D, but the main thing is what you need your image for.

If you feel that your image does not look impressive, that is good, that men that you want to go further, that you recognize your limits and have hunger for more, Focal point is only one small part of what can create a good image, you can create a focal point not only with Shapes as mentioned by Riley, but also with Lighting, or colors, or effects such, DOF, Fog etc.

For you I would recommend before to try to imitate a style, trend or renderings, research Architectural Photography, Basic concept of Still Life painting, Color theory, Focal points and projections, some of these concept will help you to develop your own technique, and express really what you want to show about your building. IMHO this can be more helpful than try to copy someone else rendering style, because yes you can figure out how he/she did desaturated the image, but you won’t know the why, plus that technique won’t apply always, what if a client ask for a sunset shot?? or a nice sunny day?

For example, in your first image you got a blended monotone image great, but the trees in front of your building, call the viewer to look at them before your building, also all your composition is leaning to the right, Your building need camera correction (Sketchup is missing this) to make it look straight and more accentuated, also you have a raining day with the rain also moving to the right, it would help if the rain is moving in the opposite direction of your building, the summer dress people under the rain is seems strange, also cloudy rainy days does not produce hard shadows, your glass does not reflect environment, plus the lack of interior lighting does not create an inviting feeling to the observer, to feel like he want to run inside of the building, it is raining outside I better go it, it seems warm, live, fun, I don’t know.

Not that I want to say you did everything wrong, just proving my point of technique is not everything, it has to go together with theory and application and improvisation too, who is the owner of this building, what is the use of this building, is really a raining day the best way to display this building?? Well if you are up north in Europe or Scotland I don’t blame you 

I don’t want to take more of your time sorry for the long answer, but you get the idea, keep posting here though, there is tons of great artist that will help you to elevate your rendering skills.

 

Fco.

[benjaminbourgoin]

Alright, here's another stab at it. The last image was produced for the master use planning process, and the city of Seattle is super anal about showing planting, now that I'm just doing renders for our client I can provide a little more emphasis on the buildings themselves. Any more feedback people have for me I'm so happy to receive. I can't believe how helpful all of you are!

 

Thanks,

Best,

 

Ben.

[BVI]

The light glows from the car headlights, breaks, rear lights etc never look good when you comp in a photo and then bring down the transparency - Better to do these as PS effect combined with a photo. Shane Gee did a nice tutorial here: http://i167.photobucket.com/albums/u143/FreeAgent84/Headlightstreakstuto.jpg

 

Also, pull up your contrast and up the down the brightness to get more definition on the image.

 

What I found with most projects is that the scene should be about 50% of the image, by that I mean that you should add lots of detail to the environment as your would the building. You can add loads more people walking around - more planting on the sidewalk, a few cars driving around etc. Your curb detail needs work - it's right in the foreground. Also, the sky image is very low res and the background plate is at an odd angle.

 

Every commercial image that you do should be a place you would like to hang out and visit.

[Scott Schroeder]

The light streaks are also at ankle level according the the lonely ghost man in the corner. Speaking of which, either remove him or give him some friends. He just looks so lonely, like he died before he could own his first Audi and now his ghost just stands on the corner.

 

I'm not a fan of the overall look. Nothing about it excites me. It just makes me sad. As Nicolas points out, this is a public place. You need to convey to the viewer that you actually want to go there and not just zip by the building. There is a time a place for mood and artsy-fartsy stuff, but when you overkill it your image just looks campy and poorly thought out. Sort of like a guy with too much cologne on.

 

Your sky cloud streaks also form almost a perfect 90 degree angle with your roof, that's drawing my up there there where the modeling detail is far too low.