Tips on analysing the lighting quality of a space

I understand this is an extremely overly broad question, but I was hoping for some tips from the experts here :slight_smile:

As you may know I have modelled my real life living room as a learning exercise. I studied architecture, but I have done 3D modelling / rendering of the “artistic” sort in the past. I do not have any training in lighting design or light engineering. However, despite the lack of formal training, I’d like to analyse my designs more quantitatively.

Here is a luminance falsecolour (i.e. rendered regularly, and then run through falsecolor -i render.hdr | ximage.

I then produced an illuminance falsecolour (i.e. rendered with the -i option and then run through falsecolor -i render_i.hdr -l Lux | ximage. There is a slight mistake as the keyboard is abnormally dark because its material is excluded with -aE.

I am curious about what are perhaps the 10 things you might look for when analysing these types of falsecolours. Here’s my attempt after skimming through the IES lighting handbook with an untrained and inexperienced eye.

  • There is ~300-380 lux on the work table. I read that 250 lux is the minimum for a workspace and usually offices will aim for 320-500 lux. So, uh, it’s OK but not great I guess? Definitely can’t do arts and crafts on that table.
  • The luminance of the ambient area is much brighter than the work area. Not good. However, the luminance of the computer screen is generally much more than the environment.
  • I read that people enjoy a working plane luminance of 130cd/m2 ± 20cd/m2. This seems quite far below. Shame. I guess it’s a bit dark. Never really thought about my table that much.
  • There isn’t a large contrast of luminance within the working plane, which means it is nice and uniform.
  • Generally I am sitting facing a heavily illuminated corner of the room, which is nice.
  • The dominant illuminance is natural (my lamp is not even required), this is nice.
  • The non uniform illuminance provides interest and highlights the plants outside. I like plants.

In addition, what are the main ways you might analyse the space? How can I analyse the daylight factor using Radiance? Is there a way to calculate the daylight autonomy of the space? Do you check for glare in every space? How can you measure how “warm” the colour of the room is?

Hoping to learn more :slight_smile:

Hello Dion,

I have to say I have enjoyed your exercise, although I have not said
anything until now. I wish I could make renders as good as yours! I will
give you my impression… I have to say that I have not deeply investigated
people’s preferences and comfort, so all I know comes from the literature
and from talking to people as a teacher, developer, and practitioner.
Others in this list may give you a better idea, with more data and all.

Ok, so… analysing the lighting quality of a space is a very broad
statement. I will not focus on the fact that “comfort” is intrinsically
subjective (although sometimes you will find consensus, which allows
creating metrics), but on other issues.

  1. I understand that the minimum illuminance (lux) for workplanes have
    been mostly defined based on electric lighting experiments. That said,
    making simple (not rigorous) experiments with my students, I have found
    that 300 lux seems to be, indeed, a reasonable minimum illuminance for
    daylighting. I understand that C. Reinhart (I could not find the reference
    right now) made a much more rigorous work with his students, and also
    defined 300lux as a minimum illuminance… I believe this is the origin of
    the Daylight Autonomy with a minimum of 300lux.
  2. That said, people seem to tolerate much larger illuminance levels
    than 300lux. On an overcast day you can get 10 to 15 klux and you can walk
    outside without problem.
  3. If you will work with illuminance, it is faster to calculate a grid
    of sensors instead of an image… that is, you create the model (which can
    be somehow simplified, as no photorealism is required) and then, instead of
    calculating a million pixels, you calculate 50 points (or 500 or 5000 if
    you want). There are several tools out there that allow doing this!
  4. Calculating illuminance is a straightforward and somehow easy and
    pragmatic way of approaching to “comfort”, although it is wrong because our
    eyes do not see illuminance on a workplane. Instead, they perceive
    Luminance from the whole visual field. Accordingly, images and renders are
    mostly used for studying Glare and contrast. They intend to represent what
    our eyes see, so they are better for that. I usually try to avoid renders
    because they take a long(er) time, they are hard(er) to implement
    dynamically (i.e. calculate glare for every hour of the year) and they
    require me to choose, beforehand, where are people sitted, where are they
    looking at, how tall they are, etc… I do not feel comfortable making so
    many assumptions. Sometimes images are very useful, though! And Glare
    metrics are getting better! My respects to people working on that.
  5. I think we should always consider the temporal dynamics of daylight.
    You mentioned Daylight Autonomy, which is cool! There are other metrics out
    there that may help you design as well.
  6. Regarding metrics, my position is that they should try to represent
    what people want (or expect) and not the other way around… thus, choose a
    metric (and a workplane or grid) that make sense for what you are doing.
    Sometimes your workplane should be vertical, or in the floor, or not even
    plane. Also, you may want to use contrast instead of illuminance, or
    whatever… This is not to say that the current metrics are wrong, but keep
    in mind that you know the space you are working in. Do you like it? is it
    nice? When is it nice? If it is not nice on sunny winter mornings, then
    create a render for that moment and conditions and try to find the origin
    of the problem and fix it. Do not let your data you if you like your office
    (health is sometimes different, as we are not always capable of perceiveing
    what is making us sick). This point is purely my point of view.
  7. The way to calculate (illuminance-based) metrics is different to
    performing a render, and often use the so-called Matrix-based methods
    Radiance does not really offer you a very straightforward way of
    calculating the Daylight Autonomy (or other metrics), but allows you to
    create scripts that calculate them. Based on this capabilities, several
    Radiance-based tools have been developed that allow you to automate
    Climate-based Daylight modelling. For instance, Honeybee my own (much more modest) Groundhog
    http://www.groundhoglighting.comallow developing models in a 3D
    modelling tool, and run Radiance under the hood. Depending on what you are
    doing, these tools may or may not replace Radiance itself. However, it is
    always good to go through the Matrix-based methods tutorial to understand
    what they do.

My answer is perhaps even broader than your question, and is populated with
my own opinion. I hope some other people gets involved and provide their
positions as well.



1 Like

First, there has to be enough light (illuminance.) That is fairly well worked out–the IES Handbook has lists of levels for various tasks, or for just walking through a space.

Second, “glare” has to be at acceptable levels. “Glare,” however, is not just one thing. For instance, a sun patch on a table where you are reading a book may be too much glare – contrast. There are guidelines for contrast. But then we get into much more complex situations. The sun in your field of view is an obvious glare source. But a sunlit wall through a window can also be a glare source. Glare depends to some extent on the angular size of a source and how far above the direction of the gaze the glare source is. (Consider that the sun overhead is usually not a glare source.) The content of the glare source also makes a difference; a pleasant bright outdoor scene will usually be more acceptable than a large flat panel of uniform diffuse illuminance.

This is of course one reason we do renderings; there is still a use for just looking. If we could manage HDR VR displays they would be incredibly useful. (Has anyone connected one of the current VR headsets to rholo yet?)

There are also a couple of computational methods. Where daylight is concerned, the best current art is embodied in Jan Weinold’s “evalglare.” For electric light sources, there are various metrics (Visual Comfort Probability, CIE Glare Index, etc.) and these can be calculated by the “glare” script. These are covered in the IES Handbook. Do keep in mind that the bright points we see in some LED luminaires have yet to be studied as glare sources.

So that’s a start.

I agree with all of your commentary. I do wonder about the 130 cd/m2 target - was that specific to computer workstations or could have been based on more traditional workspace? Your lux values on the desk are pretty good so the luminance is mainly saying the color of the desk is dark in contrast to your screen or any papers.

If I recall correctly IES notes somewhere only 150 lux for workstations restricted to on-screen work, while 300 to 500 lux recommended for reading and writing depending on font size and writing implement (e.g. black ink is easier to read than red pencil).

To put it in a more concise assessment if you were designing a building and had to do this for many spaces:

You have enough lux on the desk for computer and basic reading tasks.
Your background environment is well protected from glare (no unobstructed view of the sun in front of you), and also not completely dark compared to your task.
You have a view outside for circadian benefit, and eye strain relief to be able to occasionally look out at distant things.