Community
3ds Max Forum
Welcome to Autodesk’s 3ds Max Forums. Share your knowledge, ask questions, and explore popular 3ds Max topics.
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Lighting Analysis tool - Reliable?

22 REPLIES 22
Reply
Message 1 of 23
santiagogaray
474 Views, 22 Replies

Lighting Analysis tool - Reliable?

Hi,
This is a very important tool that we really wanted to use in in-house and one of the main reasons we bought Max Design 2009. Before using it on our projects we needed to prove the accuracy of the light values obtained so we performed some tests. We did some basic studies by measuring real light data and compared it with values obtained with the Lightning analysis tool. We first did tests with a number of surface fixtures in an enclosed room resulting in substantial discrepancies (2 to 3 times bigger than the real values).We then simplified the test to a one fixture, a Lithonia parabolic troffer with 3” louver and 3 fluorescent bulbs and measured the light at 6’ below the fixture. The discrepancies were still unacceptable, values differed from 80fc (obtained from Light Analysis tool) to 50fc (real data).
We did go through all the right settings as specified in this forum, still the differences are so big that I’m seriously doubting the reliability of this tool for any basic lighting study. I still want to assume I may be missing something, not a subtle thing though.
Please let me know if someone can help me with this, I’ll be happy to attach scenes, ies files, etc. and provide any data needed.
Thanks in advance.
22 REPLIES 22
Message 2 of 23
Anonymous
in reply to: santiagogaray

Hi,

can you post your specs ?
(a demo scene would help)

mr settings,
lights,
sky and
materials.

What did you use to get real light data for comparison ?

What kind of simulation were you involved in ? indoor only, daylight only, both. time frame etc.
Are you using a direct approach for comparisons or some kind of mean errors and such ?

Did you install last servicepack ?


I can say lighting analisys tools got better results than Radiance for typical daylight scene simulations.

max
Message 3 of 23

Hi,
i attached the scene that we used for the simplified study. Just one fixture(ies attached) in a conference room. We measured the actual light level in the room with a Mannix TLM-1337 light meter. As i said, the value we obtained right below at 5'6" from fixture was 49.8 fc. around 35 fcs less than the value registered by the virtual light meter.
The study is NOT including daylight as we covered the two openings (door and window) with balckout paper.
The comparison includes just this one direct value since I wanted to narrow down the variables to its simplest form.
Lastly, I haven't install the latest SP1, just the latest hot fix only. I need IT to find time to do so though I haven't read any relevant fix related to the light analysis or lights in general.
I hope the max file gives you the rest of the information needed.
Thanks for the help.

9312_jb0JIfPO6fe0ElkHs7BW.zip

Message 4 of 23
Anonymous
in reply to: santiagogaray

Thanks, I got the scene. I will test it asap.

Take care that with the last SP1 there's an important fix for LAT (lighting analisys tools). I'm not sure it's present also in the hotfixes.

max
Message 5 of 23
Snooze
in reply to: santiagogaray

In office we are using 4 methods of light simulations: MAX Design 2009, Max 9 Radiosity, Analog light labs, and AGI; all of which are coming out with relatively the same effects, but numerically way different. This concerns me. How accurate is the 3DM2009 light modeler?

The question that has risen is what are the light meters able to record? Is the light simulation able to measure diffused light from a cloudy sky? WSSP (Washington Sustainable Schools Protocol) requires that all rooms be measured with a diffused sky and I do not think MAX Design 2009 has a diffused sky. In the settings there are a few sky models, one of which is CIE Overcast Sky. You would think that this sky model would be diffused light, but seeing the renderings progress through the day.... I have to wonder. Also, I have heard that the Perez Sky Model is the correct one to use.

I have done extensive testing on 4 different room conditions with all the sky models in D2009 and the four other programs I mentioned above. All the data is graphically represented in excel to make easy comparisons. I am waiting for the correct days to go on site and duplicate the simulation to find actual numbers. I will post my results.

I noticed that the exported data from the D2009 light meters has direct light, indirect light, and the total lux numbers for each node. If I am concerned with indirect lighting, would I only record the indirect lighting numbers found by the meter? I would think that the total lux numbers would be the correct ones to use since direct lighting should be taken in account and recorded in total lux figure. Are the indirect light bounces from the direct lights glare in the room being measured and recorded in the indirect lighting column? This is significant when calculating the daylight factor. The requirements for LEED and WSSP are similar: 75% of occupied space to be lit by natural daylight. So, to keep contrast ratios down, direct light is shaded, therefore indirect/diffused light is what is most important. If the total lux numbers are used, then the results show that a large percent of the building is failing, but when only using the indirect light measurements, then the building starts to satisfy the requirements really well. In the long run the onsite measurements will tell all, but for now I need to have a conversation about these scenarios.
Message 6 of 23
Anonymous
in reply to: santiagogaray

In office we are using 4 methods of light simulations: MAX Design 2009, Max 9 Radiosity, Analog light labs, and AGI; all of which are coming out with relatively the same effects, but numerically way different. This concerns me.

Would be better to adopt a mean bias error to compare your various data. in daylighting, ratios are used more than absolute numeric values. this is also because daylight sims are based on time series of luminances and illuminances. ie. you measure a whole week in real-world, tracking down time series by 5mins each, then you simulate this in 3dsMaxDesign. You can take out a MBE by day and so make your comparisons.

Is the light simulation able to measure diffused light from a cloudy sky?

Daylighing is divided into direct and diffused components. the sun light and the 'blue' (scattered) light coming from the sky atmosphere. This sky luminuos distribution can be measured by a sky scanner or by a sky model. CIE and PerezAtAll do just that. LightMeter will record that. Then CIE is generally used for 'static' simulations (some selected days) while PerezAtAll is used for dynamic simulation (along a whole year). So what you'll use depend on what type os simulation you're going to run.

I am waiting for the correct days to go on site and duplicate the simulation to find actual numbers.

How do you know weather conditions for those days ? generally by looking for an EPW file (sky scanners are expensive) that reflect the same days you're inspecting. That EnergyPlusWeather file will be used by the PerezAtAll sky model to simulate the same weather conditions in the application.

If I am concerned with indirect lighting, would I only record the indirect lighting numbers found by the meter?

Dunno about the certification itself, but what you'll track down is what you have to compare with. If you have both direct, indirect measurements coming from real-world, you'll most likely go for both also when simulating daylighting on 3dsMaxDesign.

max
Message 7 of 23
Snooze
in reply to: santiagogaray

If the perez sky model is enabling me to measure both a cloudy sky and clear sky at the same time (indirect/direct measurements), then I would confidently be able to use the indirect light data exclusively and not worry about the direct light data to simulate an overcast sky. The same goes for CIE, I would think.

If you have both direct, indirect measurements coming from real-world, you'll most likely go for both also when simulating day lighting on 3dsMaxDesign.max


Are you saying that if I am only receiving indirect lighting throughout the entire room then I can use the indirect lighting data exclusively from the light meter? If that is correct, then what do I do if both indirect and direct light are present, do I use both the indirect and direct data independent of each other, or use the totals column? I would guess to use the two numbers independent from each other, but if I am doing this then my next question is are the indirect light measurements effected by the direct light that is entering the room?

So, if we figure out which data to use, I am ultimatly having a hard time with simulating an overcast sky. The CIE Sky Model shows this setting, but when doing a visual rendering and light measurement in D2009, a direct light pattern is still casted (showing crisp shadows). If the sky is truly overcast, then there should not be direct sunlight.. Right? This is leading me to speculate on using only the direct or indirect light measurements.
Message 8 of 23
Anonymous
in reply to: santiagogaray

.. the perez sky model is enabling me to measure both a cloudy sky and clear sky at the same time (indirect/direct measurements)
take care, everything coming from the sky is indirect light.

Are you saying that if I am only receiving indirect lighting throughout the entire room then I can use the indirect lighting data exclusively from the light meter?
nope. if in doubt, use 'total illuminance', really, it should be renamed as 'illuminance'.

are the indirect light measurements effected by the direct light that is entering the room?
well, that's for sure.

So, if we figure out which data to use..
you have to use the PerezAtAll sky model. With EPW files of the days you're gonna inspecting. PerezAtAll will take care also of your sun intensity, ie. if you have a cloudy day, you'll have no direct sunlight. Do not use CIE.


then take care, direct sunlight is generally not involved in daysim because you cannot state you have a certain daylight factor while you have direct sunlight hitting, for example, the user desks in your building, it's not 'allowed'.. you should use lightshelfs, special glazing systems, roller shades, venetian, my big fatty neighbor.. to 'mask' the direct sunlight as it's not that well good for office workers. I mean all depends by what kind of simulation you're going to run. daylighting is the proper utilization of daylight to reduce energy consumption, ok. but it's also taking care about direct light glaring and proper visual ergonomics. cool thing, is that you can simulate all of this in 3dsMaxDesign2009 and visually understand it better. if you see you'll have light peaks of 10000lux(direct sunlight) on desks, it will be better you'll adopt some shading things for your window; while if you have a coherent illuminance(around 500lux) for your interior over the whole working day, you're achiving a good light design. from there you can start to look at the requests of daylight certifications.

max
Message 9 of 23

Hi max,
I'm curious to see if you had any luck with the model I uploaded.
Thanks.
Message 10 of 23
ramy
in reply to: santiagogaray

As far as accuracy it can be debated. However, Max 2009 Design is recognized by LEED for getting lighting analysis credits, so that has to have some weight. Click here for an example of how we use the lighting analysis in our office. The colored light value image is just a screen capture from the max interface (no rendering). Hope this helps.
www.ramyhanna.com | www.tiltpixel.com
Message 11 of 23

We'll soon be posting the results of a validation exercise performed for us by the group that did the validation for Radiance a long time ago. We know already that our agreement with reality (per the carefully controlled experiment) is very close. Current results are suggesting it is as close as Radiance - if not better. However, we need to post the report before we go too much further with this.

We're very confident in the results from 3ds Max 2009 Exposure feature. You'll soon see why and can make your own decisions. We've asked for all the data and test setups to be published so that others can duplicate the results.
Message 12 of 23
fourthand11
in reply to: santiagogaray

Ken, have these results been posted?
Message 13 of 23

Ken, have these results been posted?


We had to run one more simultation to provide data to Harvard for a certain lighting case. Otherwise, we're 99% done and it is all good news. Stay tuned - keep asking until you see it from us! Maybe 2 more weeks?
Message 14 of 23

Ken.

I have just read the paper ‘Experimental Validation of ‘Autodesk® 3ds Max Design 2009’ and Daysim3.0® ‘
It is a very interesting read, but I would like to point out several errors with your Optical properties, Arch & Design Material Parameters.

Many of your materials have: #refl_func_low 1.0 ?

MoullionMetalSilver 7% specular reflectance = #refl_weight 0.004 ?

DoubleClearGalzing = #refl_func_low 0.0 ? #opts_skip_inside true ?

ExteriorWall = 0.1% specular reflectance = #refl_weight 0.006 ?

Plus other errors not listed here.

Also was the glass window modeled with the correct thickness to match the real window?

Thanks.

Lee.
Message 15 of 23

I've asked our expert to reply. Thanks for raising the concern. I'm sure there is an explanation.
Message 16 of 23
pfbreton
in reply to: santiagogaray

Hi!

Sorry it took long to reply because we did spent some time to revise all our material definitions prior to answer this thread. Her we go:


------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
3ds Max model: explanation of differences between the Radiance material and the 3ds Max A&D Material

Setting the correct diffuse color and specular color in 3ds Max Design:
The values used in the mental ray A&D Materials are different from the values used in the Radiance Materials because they don’t use the same algorithms.

The main difference is that the 3ds max A&D Material handle energy conservation internally, while the Radiance material don’t. Therefore, we need to compensate for this when we input measured data in the 3ds Max A&D Material.

For example:

Assuming a measured material of 77% diffuse reflectance and 4% specular reflectance we would use the following settings:

Radiance Settings:

void plastic InteriorBackWall
0 0
5 0.77 0.77 0.77 0.004 0

3ds Max A&D Material Settings:

#diff_color (color 197.13 197.13 197.13)  This represents 0.7730 on a scale of 0 to 1.0
#refl_color (color 255 255 255)  This represents 1.0 on a scale of 0 to 1.0
#refl_weight 0.004
#refl_func_low 1.0
#refl_func_high 1.0


First, we consciously chose to disable the Fresnel falloff (BRDF) functionality of the 3ds Max A&D Material by setting the BRDF Curve values to 1.0 at all angles. The reason for this was that we did not have full measurements of the materials and the specular effects are negligible in our case: we are interested into measuring mainly the diffuse illuminance at the sensor points (as the space is generally a diffusely lit space). Defining a Fresnel falloff on our materials would not introduce major differences in the measurements (except for glazing, which is discussed below).

We are left with 3 main parameters: the Specular Color, the Specular Weight and the Diffuse Color. Knowing that the Specular Color is multiplied by the Specular Weight, we leave it at 1.0 1.0 1.0 (White).

Then, we need to do a little bit of maths to figure out the proper balance between the Diffuse Color and the Specular Weight. Since the 3ds Max A&D material performs energy conservation internally, we need to apply the following formulas:

A&D Material Diffuse Color = - ( Measured Diffuse / (Measured Specular -1) )
A&D Material Diffuse Color = -(0.77 / (0.004 – 1))
A&D Material Diffuse Color = 0.773

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Glass vs Thickness:

Glass thickness was ignored. Light rays penetrating in the space are assumed to be attenuated with a Fresnel falloff depending on the incident angle. This attenuation occurs at every surface the light ray hits. The thickness of the glass is ignored and light rays will not be affected by how thick or thin the glass geometry has been defined in the model.

For more information on strategies around modeling glass correctly for lighting simulation, please consult the advanced Daylight Simulation White paper that can be found here:

http://usa.autodesk.com/adsk/servlet/index?siteID=123112&id=11360846

All the tricks are explained there.

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Errors in the paper:

We did notice some "errors" so we are rerunning the simulations right now to make sure we don't see major differences in the results. However, the errors we spotted are minuscule and are way below the overall errors described in the report. The "official" version of the paper will be published with the corrected data.

Hope this helps clarifying.
Message 17 of 23

Thanks very much for the reply.

It's a very interesting field and I have much reading to do.
I look forward to the revised final paper.

Regards.

Lee.
Message 18 of 23
Anonymous
in reply to: santiagogaray


So, if we figure out which data to use..
you have to use the PerezAtAll sky model. With EPW files of the days you're gonna inspecting. PerezAtAll will take care also of your sun intensity, ie. if you have a cloudy day, you'll have no direct sunlight. Do not use CIE.


I'm experimenting with the 3D max Design 2010 Daylight Simulation. For our own design process we will use real EPW files and test throughout the year in different conditions. For LEED 2009 Design and Construction the requirements I need to do a CIE clear sky at 9am and 3pm on march. or sept. 21st. My issue is that I don't know what to put in as values for the Diffuse Horizontal Illuminance or the Direct Normal Illuminance. Do you know of any charts where I can look those up? I can use a weather file to get the numbers in the ballpark, but I want to use the correct CIE Clear Sky. does anyone know those values for Pittsburgh,PA?

When I switch from Perez to CIE it places some values in there, are those the correct ones or some random defaults?

Message 19 of 23
gelbrich
in reply to: Anonymous

Has anyone answered this question yet? I'm facing the same problem.
Message 20 of 23
grue1970
in reply to: santiagogaray

Hello, I am still reading up on this for 3ds Max Design 2011, I found that you can go here: http://apps1.eere.energy.gov/buildings/energyplus/weatherdata_about.cfm then click on "Real Time Data" on the left, then enter your location (there's a search option for stations). You can then see what data is available and enter a time range.

I've seen some incomplete data, dates with no measurements.

Also from what I can tell the data for a specific date is actually a multi-year average for that particular date. So Sept 21st at 9:00am local time would be an average of that day/time over multiple years.
3ds Max all versions past and present | GPU Rendering on 2 machines | Standard, MentalRay, MentalRay IRAY, IRAY+, VRAY, Arnold

Can't find what you're looking for? Ask the community or share your knowledge.

Post to forums  

Autodesk Design & Make Report