Author Topic: Improved behavior of IES lights  (Read 1150 times)

2020-01-10, 10:08:47

ankerstjerne

  • Users
  • *
  • Posts: 4
    • View Profile
Hi!

According to my searches, a few years have passed since IES behaviour was brought up on the forum, and it seems nothing has happened in terms of improving the weird behaviours noted by users. I am slightly dissapointed that it isn't more of  a priority.
I realise that this topic isn't crucial for people in Archviz, but for many people working with corona in lighting design, product design, interior design and architecture, I believe it is important.

Firstly I would like to mention that I work as a lighting designer, and regularly use photometric files (incluing IES) in other calculation software, so I feel like I have quite a good understanding of what an IES file is, what information it contains, and how it can behave in software.

I just spent some time in a test scene checking the behaviour of corona lights and 3ds max photometrical lights with and without IES files.
Below is a list of things that I find strange..

- IES files contain information about the intensity of a fixture. This information is not used when loading an IES file with a corona light, but must be input manually. It is however used when loading an IES file with a photometric light, which seems more logical to me.

- If I try to set up a corona light, and a photometric light, loading the same IES file into both, and adjusting the lumen values of both to be identical, the resulting intensity in my rendered image is not the same. This is problematic because you can't count on intensities when using corona lights and photometric lights in the same scene. I often need to do this in scenes.. For me photometric lights work better with IES files, and Corona lights work well for generic spots, where you can quickly adjust directionality. I use corona lights as well for simulating linear fixtures such as LED strips.

- When using a corona light with an IES file, the light can be set to disc and sphere with the same result as long as intensity is set using the 'default' setting. When using the 'lumen' setting however, the disc and sphere lights have different intensity. This seems very illogical?

My feature request to improve IES file behavior in Corona would be the following:

- Corona lights automatically read the intensity (lumen) values in an IES file.
- When an IES file is loaded into a corona light, it should disable the shape picker drop down. The IES file should lock the shape of the light to a sphere with a predetermined size. IES distributions are by default spherical.
- A corona light and a photometric light loaded with same IES file, with the same lumen intensity, should render the same.
- When using corona lights without IES files, it would be great to be able to adjust light intensity as lumen/area. This way it would be possible to create rectangular lights of different lengths in a scene and give them the same intensity (think LED strips that have a certain lumen output pr. meter). I know this is possible to do with a corona light material, but comes at a cost of lower render performance.
- Many european lighting manufacturers don't provide IES files, but instead provide Eulumndat files (.ldt). IES is an american standard and LDT is a european one. It is possible to convert ldt to ies and vice versa, but it would be more elegant if corona lights could load both file types.

- Finally on a more general note: Could the Interactive LightMix interface have a GUI option to look like a fader panel (as in Maxwell render multilight GUI). Optimally with the option to control it via physical faders using a MIDI interface. This would be a game changer in presenting architectural lighting design to clients.

All feedback, corrections and comments are welcome.
Best regards - Christian
« Last Edit: 2020-01-10, 10:17:05 by ankerstjerne »

2020-01-14, 21:56:20
Reply #1

zaar

  • Active Users
  • **
  • Posts: 20
    • View Profile
It would be nice if the intensity was manually loaded, with the option to override of course. But are you also saying, or suspecting, that even when inputting the correct intensity into a Corona Light, the output intensity is not correct? How do we know that the photometric light is not wrong?

2020-01-15, 10:18:35
Reply #2

GeorgeK

  • Corona Team
  • Active Users
  • ****
  • Posts: 286
  • George
    • View Profile
Can you please provide some comparison screenshots?
“Every artist was first an amateur”

2020-01-16, 09:56:49
Reply #3

OJ

  • Active Users
  • **
  • Posts: 10
    • View Profile
I did some testing because i was also interested in this matter. When using photometric light it uses the correct intensity.
And when setting the same intensity in Candela for Corolight everything seems very wrong.
If I conver the photometric light's intensity to Lumen and use that value in Coronalight as Lumen the difference is not that big but they are not the same.



Here is image of the ies photometric chart from IES Viewer which esures that the intensity should be around 54000 candela.


I not well experienced with Candela Lumen conversions so i can't check if the conversion that happens when changing photometric light  intensity type is correct.
PS. Also not sure if i used the wrong intensity from the start but the one I used was automatically picked when i chose my IES file for photometric light.
« Last Edit: 2020-01-16, 10:07:31 by OJ »

2020-01-16, 14:50:30
Reply #4

GeorgeK

  • Corona Team
  • Active Users
  • ****
  • Posts: 286
  • George
    • View Profile
Thanks for the shots OJ, so apparently this is a known and reported issue, also found in https://corona-renderer.com/forum/index.php?topic=13397.0

(internal id=236026421)
“Every artist was first an amateur”

2020-01-20, 17:28:18
Reply #5

ankerstjerne

  • Users
  • *
  • Posts: 4
    • View Profile
It would be nice if the intensity was manually loaded, with the option to override of course. But are you also saying, or suspecting, that even when inputting the correct intensity into a Corona Light, the output intensity is not correct? How do we know that the photometric light is not wrong?

I have no way of really testing this. But based on my experience of corona lights generally being buggy with IES files, I am inclined to believe the photometric light is behaving 'correctly'.

Can you please provide some comparison screenshots?

Here are some shots from my test file. All lights in these images are loaded with the same IES file from the iguzzini website (renowned manufacturer).
For the sake of clarity, I want to point out that lumen is the unit used to measure luminous flux which is the total amount of light emitted from a light source regardless of direction.
The IES file is a capture of how this light is distributed in different directions. Therefore I consider lumen the logical unit to use when inputting intensity for an IES file: The lumen value defines how much light is emitted. The IES file decides where the light goes.

Left: Corona light set to 25mm disc, intensity: 50(default). Right: Corona light set to 25mm sphere, intensity: 50(default). Result is identical.


Left: Corona light set to 25mm disc, intensity: 800lm. Right: Corona light set to 25mm sphere, intensity: 800lm. Result is not identical.


Left: Photometric light, intensity: 570lm (loaded from IES file). Right: Corona light set to 25mm sphere, intensity: 570lm. This is the issue OJ also showed in his images. NB. exposure was raised in this image to make up for weak intensity of photometric light.


To summarize my problem: It doesn't really matter to me whether Corona or 3ds max lights are 'correct'. I have other software for lighting calculations that need to give me precise values hitting a surface. In a render, the exposure can be adjusted anyway, as with a camera. BUT as long as corona lights are buggy and the user interface for IES files is opaque (as illustrated above), I am still more confident using the photometric lights for IES files. And this is where I have a problem: not being able to use corona lights and photometric lights in the same scene with any confidence that the lumen values are reliable.

Thanks for taking part in the discussion! Again, all suggestions, comments and corrections are welcome.
Best - Christian.
« Last Edit: 2020-01-20, 17:32:47 by ankerstjerne »