Best Practices: Finding the Right Perspective

Finding the right perspective is an art, not a science, and we can all get better at it by understanding a few basic principles. For some this may be a refresher. But, given Enscape’s great support for students by offering their software for free, it seems fitting to offer this “fundamentals” article as a way of helping aspiring designers get the best possible results when creating still images. And, since I don’t work for Enscape, I do not feel bad about mentioning not all software is free to students

This article will look at the various view types and composition considerations, as well as problems to watch out for. The following image represents a well-composed image where the edges are framed, the many vertical elements are not aligned or overlapping, a subtle ‘depth of field’ is applied, and the viewpoint is at eye-level.

Because finding the right perspective is more of an art, some of what I am about to profess may not be the right answer for everyone. Even if you don’t agree with some of the aspects covered, the hope is that everyone reading this will find some value and firm up their personal understanding of what constitutes a great rendered image.

There are three perspective types: 1-point, 2-point and 3-point as shown below. Enscape can create each of these view types. Additionally, your 3D modeling programs (Revit, SketchUp, Rhino and ArchiCAD) can display axonometric views.

In the next two images, you will notice a subtle difference between 2-point and 3-point perspective views. The point here is that, for 2-point perspective, the vertical lines are perfectly vertical. At the end of this article, you will see how the 2-point option is extremely helpful in the courthouse example.

Note that the horizon line is intentionally exposed in these images, and several like them in this article, to emphasis several fundamental concepts related to our topic at hand.

The vantage point from which a model is viewed is important to consider. Most people view architecture from the ground, standing on their feet, so that is the vantage point I prefer. In Enscape I will “fly” through the model looking for a good view, and then press the Spacebar to quickly set the vertical position at eye level; the exact height can be changed in Enscape Settings via the Spectator Height slider.

Aerial images also have their place, but it is important to keep in mind the extra work they often require. For example, compare the next two images, notice how much extra backdrop must be created, even with Enscape’s built-in environments. By contrast, there is a much smaller area to fill in for eye level views. From an eye level view, a large area of the backdrop is filled with sky, which Enscape handles well. Plus, just a few well-placed trees are sometimes all that is required to hide the horizon.

Another thing to note about the vantage point of a view, when set at eye-level, is that all the people in the view often have their head aligned with the horizon line. This is a very helpful detail when hand sketching but can also help to spot an Enscape-view positioned in a way someone would not normally look at a space. Knowing this helps with the realism of a final still image.

For eye-level renderings, meaning the angle of the view is as if a person were standing on the ground, the heads of most people will be at the horizon line no matter where they are in the scene, as depicted in the sketch below. As you can see, some people are very close while others appear far in the distance, but most of them have their heads aligned with the horizon. The exceptions are when a person is sitting, on a different level, bending over or just shorter than the person they are standing next to.

Here is the same concept visualized in Enscape. We have three people, all at a different distance from the viewer, but each of their heads are aligned with the horizon line.

Understanding real-world camera settings is very helpful in developing the right perspective in Enscape, as many of its settings are based on how a real camera works. For example, the default field of view (FOV) in Enscape is very wide and helpful for navigating a model on a computer screen. But for still images, a professional architectural photographer would not normally use that wide of a lens as they tend to distort the image, making a scene look less realistic. A common camera lens used for architectural photography is 24mm (Tilt-Shift) which is a 67 degree FOV in Enscape. The Enscape default is 14.5mm which is a 90 degree FOV. Check out this post for more on this topic: Lenses for Architectural Photography

Not only do people help to bring your sketches to life they also give the viewer a sense of scale. Continuing to look at the mechanics of a perspective, notice how we can use people in the scene as a sort of measuring stick, literally or subconsciously.

Here is the same concept visualized in Enscape. We can see how the woman’s body height can help us gage how tall the structure she is standing next to actually is. We can even project her height vertically and in perspective to, for example, place a 10’-0” (304cm) vertical line in the scene; each yellow line is the same height.

The composition of a view is a key ingredient in developing the right perspective. In the next image you can see some problems; our vantage point has us visually grazing the side of the main building and the outdoor fireplace covers a major edge of the main structural, leaving an odd portion of the roof exposed. Compare this with the next image, where the left side of the view is framed be the main structure, the fireplace does not cover any major elements and the top edge of the view is also framed by an umbrella. Even a small portion of a chair in the foreground helps to frame the view. I touched on this specific example in my ArchDaily article: 9 Ways to Make Your Renderings More Realistic

Now that we have talked about some fundamentals, let’s look at some applications. For most, we will have a poor, good, better and best example to compare.

Poor example:
If you only had one image to provide, this would be considered a “poor” example, as this is not how a person would normally view this project and the backdrop needs a lot more work.

Good example:
Now we are on the ground, making the vantage point better than in the previous image. However, there are still things we can do to make it better.

Better example:
In this case we have taken some artistic liberties and repositioned a tree, which does exist, so we have the sense of a branch framing the view in the upper left. This is a favorite technique in architectural visualization to ground the building and break up the vast amount of sky.

Best example:
This last example incorporates all the previous features as well as a custom Depth of Field to draw the viewers focus to the important part of the image – the building; not the person in the foreground or the city beyond. This is also great when your specific project does not have a custom skybox for the site. Blurring out one of the built-in Enscape options can help avoid questions and confusion by the client.

Now, let’s move inside this same project and look at an example in the open kitchen.

Poor example:
Here, the field of view is too wide and there are no people to help define the scale of the space. The scale may be difficult to understand as it is so open; there are no visible doors or objects in the foreground. Let’s look at how we can make this perspective a little better.

Good example:
First, we change the Field of View and vantage point. Now we have an object in the foreground, the dining room table, which helps frame the view and convey scale.

Better example:
Adding a few people helps bring the space to life and further implies scale. By the way, these high-quality examples in this article are from ArcvhVision’s RPC collection (AXYZ models).

Best example:
Adjusting the Depth of Field draws attention to the kitchen while still making the foreground elements visible but not the focus.

In this next example you will notice that two vertical edges are aligned. This makes it more difficult to quickly understand where the brick wall in the foreground stops and the pool house wall starts in the background. The view is also looking downward slightly, which makes the wall on the right looks like it may be sloping or not vertical. The second image corrects both subtle issues by moving the vantage point slightly and leveling the view.

There are always exceptions to the rule, as you will see with this last example. Sometimes we need to present a space with unusual proportions, like this courthouse foyer which is 22’-0” (6.7m) tall, but only has a floor area of 28’-0”x22’-0” (8.5mx6.7m). Not only that, but the floor and ceiling have important design features: a State seal and dome respectively. The first two images below use the techniques previously discussed but do not adequately represent the essence of the space. Let’s look at what we can do to properly capture the space, to make sure the clients and public understand the design intent.

To really capture this space, we need to use Enscape’s Architectural Two Point Perspective option. Additionally, we need to position ourselves about half way up in the space, not at eye-level as I often prefer. I backed up until I just passed through the wall, having gone too far, and then moved forward a little, which puts me as far back as possible. Notice each side is nicely framed by the ionic columns in the foreground. I was also able to use these side columns to level out the view, capturing a little of the seal on the floor and the dome above. The  field of view is set to 115 degrees. We must be careful, when adjusting the field of view, to not let the view or entourage get too distorted.

As you read this article, you may have thought to yourself that much of this is common sense. And you would be correct. But, as with most things it is good to have fundamental principles refreshed, so that they are better understood. In so doing, we will be able to more quickly develop our views and convey the design intent to those viewing our presentation graphics, especially seeing as it now takes more time to compose a view than to render it thanks to Enscape. In this regard, we should slow down and ‘smell the roses’ to create the best possible graphics!

So, the next time you are establishing a view in Enscape, remember these key points: view type, scale, field of view, depth of view, vantage point, alignment and overlap. And, once you find that perfect view, don’t forget to save it back to Revit or SketchUp so you can render it again in the future.