Friday, August 3, 2012

Sky - Looking Up and Down


Venus Drawing Room Ceiling, Versailes, by R.A. Houasse
Look up at the ceiling of any Baroque Church or Palace and you will see chubby cherubs floating around in a framework of clouds.  The other allegorical figures vary, but the cherubs and clouds are a constant.  These ceilings must have driven the artists crazy because, although the viewer was looking up, you didn’t want to see only feet and tush, or conversely, topsy-turvy people.  This is not a post on perspective foreshortening so suffice it to say that figures were oriented to the walls of the space, while cupids and clouds were allowed more freedom to float.
This (the clouds, not the cherubs) applies to architectural rendering in that clouds have different shapes when seen from the side and from below.  The photo above was taken at an angle of 30 degrees off the horizon.  The clouds at the bottom are near the horizon and the clouds at the top are nearly overhead.  The horizon clouds are, well, horizontal.  The clouds overhead are freeform shaped.
If you go out to sketch clouds you will get the same effect as the camera – well, you should if you don’t let the paper and pencil get out of hand.  And to tell the truth, letting things get out of hand in a sketch is not such a bad idea.  After all you are trying to capture a feeling, not record in high fidelity.
Even the most curious cloud formations are believable when they conform with perspective reality.  The above clouds are a curious chevron zigzag, but they are perfectly normal to any viewer.  You could almost locate a couple of vanishing points using the parallel “zigs” and “zags”.
As with any simple perspective grid a “sky grid” can be constructed with a central vanishing point and an off screen diagonal vanishing point.  One warning I should point out here is the potential for distortion at the top edge of the grid.  This can be easily corrected by shifting the diagonal vanishing point further out to the side, flattening the squares a bit more.  When dealing with clouds by themselves this distortion can be ignored – after all, a freeform cotton ball looks fairly normal whether distorted or not.
In the above photos the building is viewed at such an angle that the clouds could be any sort of free form without looking out of place (image at left).  Only a set of receding parallel lines similar to horizon clouds would look unusual, as you can see in the same building image to the right.  
When you are viewing a non-rectilinear object such as a tree, the orientation of the underlying perspective is less sure and the correctness of any specific cloud formation is harder to determine.  The photos above are looking up toward the top of a ridge.  The left image suggests a low ridge because of the flattened clouds which are normally found near the horizon.  The center image suggests a steeper ridge because the cumulus cloud peaking out between the trees on the right would normally mean you are looking up at a steeper angle.  The right image is rather confusing in that the clouds suggest that you are looking up at a very high angle.
Following are some renderings and sketches showing various approaches to high angle skies.
Century Hotel NYC (detail)
Guangzhou Tower Sketch
645 Fifth Avenue, NYC (detail)
Petronas Tower Sketch, Kuala Lumpur, Malaysia, (detail)
Republic National Bank, NYC (study)
Times Square North, NYC (detail)
Union Square North, NYC (detail)

The reverse of looking up is…
…looking down.
The aerial, or bird’s eye view, is essentially a flipped sky view.  And the same perspective grid can be used as a framework for any aerial layout.
This rendering of Canary Wharf in London is just one of many aerial views I have completed over the years.  It is unique in that it includes a sliver of sky at the top.  The clouds are as flattened as the distant countryside, which is reasonable since the viewpoint is at an elevation of half a mile or so.
I will have to write a post on the various problems in creating aerial perspectives; but that will have to wait for later.

1 comment:

  1. This comment has been removed by the author.

    ReplyDelete