[brodie_geers]

[edit: I just can't type as fast as 1/2 life. you get two answers for the price of one this time ]

1st regarding materials (if 1/2 life doesn't beat me to it ). Maxwell automagically converts 1.x materials to 2.x. I think what 1/2 life was referring to though is the fact that typically an automagically converted material isn't really optimized for 2.x. For that you really want to start from scratch using the wizard or something.

Editing IES data. You can open your .ies file in Notepad and you'll see a bunch of numbers. The trick is finding which one to edit. I usually set up a material before hand, do a preview, tweak a number and resave the file, and do another preview to see if the intensity changed. Typically the number you're looking for is towards the top and will be 1.0. In the example below I was tweaking a 75w emitter. We were using a very similar emitter but it was 100w so I upped it by 33%.



Regarding whether or not to model the emitter casing. If you'll see it, then of course you'll need to model it but you may be able to hide the inside from GI or something like that so that you're not creating a bunch of unnecessary caustics.

-Brodie

[fuso]

Hi again,

I've changed the EV of both camera's from 14 to 9 which is now the equivalent to f-stop 5.6, shutter speed 1/15 and according
to some tables a very realistic value for a work indoor space. After adjusting the lights it doesn't look much different I'm afraid.
I also had a look at your IES materials and noticed that the all have dispersion enabled, is that neccessary? Just wondering if
that could have an effect as I have dispersion disabled in my general render settings.

How small would you normally model your spheres? Mine are now 1cm in diameter and have a face count of 48. Btw, is it better
to enable smoothing? Gosh, Maxwell is starting to do my head in again, with all those workarounds and do's and no do's...

Anyway, thanks AGAIN for shedding some light into this.

[brodie_geers]

There may be something more than just the camera settings going on however an EV of 9 still seems a bit high. I always go by this guide which is pretty good http://en.wikipedia.org/wiki/Exposure_value

About half way down you'll see the guide with various scenarios. I would say a bathroom would probably be in the neighborhood of 7ish.

Your sphere should be fine. I would enable smoothing just in case, but I don't think it matters.

-Brodie

[Half Life]

Dispersion should have no effect one way or the other with IES lights as long as there is no BSDF with any transparency in the material as well.

For a scene like the bathroom I would start somewhere in the 5-8 EV range and keep all the emitters "real-world" values... it is more correct IMO to change the exposure value than to change the emitter values. I don't fuss with shutter speed because I don't have animation/motion blur with Sketchup and so I just ignore all that and lock exposure and use EV for all my exposure settings.

Your sphere should work fine.

Best,
Jason.

[Half Life]

Also, for the sake of speed I would hide most of the geometry in the other room when rendering either of these cameras -- Maxwell will render everything that is in the scene even if it is not visible and there is alot of stuff going on in the other room that will slow down your renders tremendously for no good reason.

Best,
Jason.

[fuso]

I had only unhidden everything as I exported the scene to mxs for you. When I render straight from 3DsMax I have only
switched on what's visible directly or for the sake of reflections, that has become my workflow since Maxwell V1. But it's
always good to remind people here on the forum about that...

@Brodie: Thanks for all your input as well so far. The table you've mentioned is exactly the one I used, that's why I changed it
to EV 9. Now I've changed it again to 8 and fiddling with all the lights again...

If I don't use IES like for all the other lights apart from the spotlights (the tubular neon lights behind the ceiling slots and above
the sinks), does one still have to multiply the initial wattage by the number of emitting objects sharing one emitter material?

[brodie_geers]

You no longer have to multiply wattages by the number of emitters you have. If you have a 100w emitter material mapped to 9 lights, each light will produce 100w of light.

That could be part of the issue with your IES lights, if you're mixing them with regular emitters, particularly if those regular emitters are unrealistically powered. Maxwell puts it's resources towards the high powered emitters first and saves weak emitters for later on down the SL road. So it could be that your IES lights are only dark at lower SL's. If you were to wait around long enough they may brighten up to realistic levels. The solution is to make sure all your emitters are close to the same strength (and make sure you don't have phys sky turned on because the sun is pretty bright).

-Brodie

[Half Life]

I went ahead and switched out all your emitter materials to the Maxwell preset "Compact Fluorescent lamp (cold) - 9W" (I changed the color temp to 6500k) and set the EV down to 5.51 and got essentially the same image in terms or lighting but with more real world settings for the emitters and exposure.

I also turned devignetting all the way up to 100% because of your low focal length.

Best,
Jason.

[Half Life]

I'm trying something I've never seen before which is to use a chrome material with a thick coating (with and Nd of glass) to simulate a mirror -- It will be interesting to see if it works

Best,
Jason.

[eric nixon]

if you want to make a realistic mirror you dont use coatings - coatings are for petrol puddles, soap bubbles and many other things where you get coloured light interference - however you can embedd a mirror plane inside a piece of glass.. try it looks nice.. to be safe you should add a tag to the glass so that it is hidden to gi.

for ies lights use lowish poly sphere, the smaller the sphere the sharper the more focussed the light pattern will be.. but a larger sphere will render quicker..

efficient interior light fittings which are seen directly by cam can be tricky to setup.. the first time.

[Half Life]

It's interesting your description of people who are only trying to help... and I'm not sure what my use of Sketchup has to do with anything related to Maxwell or more particularly materials (which are independent of geo anyway). Seems one or both of us have honked you off in some way but it is difficult to tell how since you have been less than verbose up to this point.

I don't disagree that there are better ways to make a mirror with proper geometry, I was just trying to see how far I could push the MXM since the geo I received did not allow for the technique you described... which BTW is just another compromise.

There's is also an issue with reflected light through real glass right now that might make that approach impractical... My thought process went something like this -- what we are looking for here is the reflectivity at the surface of glass (not metal) and a coating can give that. The idea that the coating will the exact same appearance never entered my mind. I just saw it as a better alternative to the raw metal.

Personally for me I'd rather just create some triangle groups of the back plane of the mirror "box" and set that for an appropriate material but there are some complexities there that I'm not 100% sure of when it comes to how the Nd and K values of the mirror "silver" would be effected by it's proximity to the glass surface.

The ideal approach would be to make it like it is in the real world which is to simply coat the back surface glass with "silver" -- but obviously that is not possible currently.

I've never heard that tip about smaller and larger IES spheres -- nor have I seen that in action but I'll be sure to test it.

Regardless bumping an old topic to take shots at Brodie and myself says to me that your motives have little to do with the topic at hand...

Best,
Jason.

[numerobis]

fuso, I think whats doing your head in, is the fact your taking advice from forum trolls who use sketchup - go figure...

i think, if this is what comes out when you open your mouth, you should better keep it closed again, like the last 5 years

[brodie_geers]

Thanks for the tip on ies geometry eric, hope that works.

-Brodie

[brodie_geers]

Here's a quick test on IES lights. http://dl.dropbox.com/u/384281/ies%20lights.jpg

Here's the setup. The wall is 10' high and there are 6 emitters.

1/8" radius - 14 sides
1/8" radius - 576 sides

1" radius - 14 sides
1" radius - 576 sides

12" radius - 14 sides
12" radius - 576 sides

The lowpoly sphere seems to show less noise and the smaller spheres do produce sharper patterns. However, to me it looks like the smaller spheres also produce less noise than the larger spheres (ie. small spheres render faster). I ran a secondary test just to be sure. I ran one render with just the 1" high poly emitter and a separate render with just the 12" high poly emitter. They rendered to SL 15 in essentially the same amount of time (2m26s vs. 2m23s with the small sphere actually being the quicker of the two).

If you just glazed over all that whole paragraph as I typically do, here's the short version. For ies emitter geometry, always use a very small low poly emitter. It will produce a clearer pattern and render faster.

-Brodie

[Half Life]

Thanks for the test Brodie, it saves me the trouble

BTW, the other problem with larger spheres is it cuts off the light (inside the sphere, as you can see in your test)... I've always avoided larger spheres for that simple reason.

Best,
Jason.