Thanks Hasdrubal. I know a little about photography, and a stress the word “little”, but even with that knowledge it does not seem to help much with these Cheetah settings. I saw the camera does indeed have a lot of those settings but they only seem to apply to motion blur. They made no difference in lighting exposure.
There is indeed a small area light with that florescent fixture, but you are right, I must have missed turning it on in that first image. There are also four 6” LED dome lights (more area lights) on the ceiling which are not in view.
I am not even worried about composition or aesthetics yet, just trying to get something that looks somewhat natural. Once I get a baseline down I will definitely have to do more research into this aspect of photography though. This is one area I have a lot to learn. I should also note that this particular scene is just a combination of models I threw together for testing. The majority of the bathroom items are from my home remodel project, the room itself is just a quick sketch and the character was from a game I have been working on. This all started when I was planning on baking lighting into textures for an interior model and quickly learned the lighting was not going to be as simple as adding a HDRI image.
It looks like this is going to just take a lot of experimentation. And to be honest, I rather like experimentation for the most part. The problem I am having here is just the time it takes to do the experiments. These render times are brutal.
Speaking of which, I have been at this all day again today. I still have no idea what I am doing, but I have learned a few things. I am going to document this stuff as I go just in case someone else finds this useful.
First and foremost: Render threads.
This one really through me for a loop. I have been using all, (see 0), threads since we have had the option and never thought there would be any reason to change it. Here I have been abusing my poor iMac for no reason at all. Hours and hours of 100% CPU usage and insane fan noises.
I started doing testing with less threads and found that when I used 4 threads instead of all 6 my fans obviously do not get nearly as loud. “Well no crap Sherlock”, but here is the kicker. The same render only took 10% longer to render rather than 33% like it should have been. The speed differences are more noticeable on really short <5 minute renders, but the long >1-2 hour renders I have been doing really starts to close that gap.
So this thing must have been thermal throttling itself the entire time. I know Apple does not prioritize cooling on all of these fancy overpriced products in creates, but it should at least be able to run under full load without thermal throttling.
Really long story a little less long, the extra load and heat may not be worth the little time it saves. An extra 12 minutes on a 2 hour render is worth it to me when my iMac does not sound like a jet fighter in full afterburner. This has to be a lot easier on the expensive hardware as well right? Another added bonus is being able to do other light work as it renders without slowdowns.
Second: Render samples and brightness.
This is another one that really caught me off guard. I was doing all of my light testing at 100 samples thinking I was getting a good idea how things would look before a “final” render. This was not the case. The same scene at 250 samples is almost twice as bright as it was at 100. Although the lighting difference really starts to diminish after that. The difference between 500 and 1000 samples is hardly noticeable. I should note this is more prevalent with low light and high exposure settings.
Third: Light value vs exposure. There is one key difference. When the actual light source is in view the higher exposure setting seems to blow out the light in question. If you are going for a light that “glows” in the shot go for the higher exposure. If you are looking for a more subtle light, like a florescent light in a well lit room, go for the higher light value and lower exposure.
The exposure really works well with the sky light. Even with insane light values on the sky light at the default 1 exposure it just does not have the correct lighting effect. With a higher exposure it looks a lot more like actual sunlight but it also blows out the other lights on screen. As you can see with the florescent light that also has a light value of 1. My next test is to leave the Sky light at 1, changed the interior lights to a fraction 0.1-2 or so, and leave the exposure at 50. I am getting closer to what I am trying to achieve, but this is slow going.