does Prolighting skies automatically give correct exposure - if i use white material 0.8 value(max diffuse),render output should look whitish(abt 0.9 value).also knowing time of day of hdr is useful.
Each "map" hrdi you get has a time of day, as far as sunlight you can choose the position of the sun via slider bar in the add-on. Some of the HRDI are very blue some are very yellow. There is an update with a HSV component so you can adjust to get the look you might be going for in your renders. Its not a one click stop and works every time but its better than anything else out there right now. Or at least anything that has come to my attention in the last week or so.ok.thanks.
Update:
After using Pro skies for a for a week now I have to say the GUI is its best feature. So easy to snap between hrdis and get the right "color" light i need. Time savings are starting to add up little by little.Thank you very much for the great tip! The only problem with this aproach is that it introduces fireflies. If you however bake those nodes into the image (make the same setup in compositor and save it as a corrected HDR image) and use the corrected HDRI as a color input for background node with a strenght set to 1 then the render looks good and the fireflies are gone.
Really? And everything else is the same, image res, brightness output of the sun etc? Iāve been fighting them with a bigger sun disk. Afaik any small and bright light source, especially if there are more of them, are very prone to produce fireflies.
Care to share one freely distributable in original and modified form that you know works?
Here is the file:
http://marekmoravec.cz/file/blend/Hdri-Sun-Compansation.zip
The HDRI is by Openfootage.net from this link:
http://www.openfootage.net/?p=1160
and it is licenced under a Creative Commons Attribution 4.0 International License
There is also a blend file with the compositor nodes to increase the dynamic range of the HDRI. It is a bit different to setup of Lucas but it is the same principle. With this HDRI I also noticed that the fireflies donāt go away completely (I tested that on a more cloudy sky before) but if you increase the MIS number to something like 4096 they go away quite quickly - with 300 samples you can get OK image.
Edit: I reuploaded the zip file because the corrected HDRI image had still too weak sun brightness. Now it should give realistic results.
Ok, thanks. Iāll test it out later.
maraCZ,
I tested OpenfootageNET_Gerlos_low.hdr using Lukas node setup and your baked OpenfootageNET_Gerlos_low_corrected.exr connected to a background node. Their is almost no difference. See below.
The OpenfootageNET_Gerlos_low.hdr with Lukas node compensation
The baked OpenfootageNET_Gerlos_low_corrected.exr
Here is my blender file for testing both hdriās
As you can see they look pretty much the same with a strength of 2.75 at 300 samples. The baking that your doing through compositor probably does not match exactly the Lukas node setup.
Attachments
Hdri-Lukas-test.blend (582 KB)
I tested that on a bit more complex scene - little bit more complicated geometry and more advanced shader - diffuse with glossy:
This is the corrected HDRI:
This is the Lucas method:
Rendered at 150samples with indirect clamp at 1.0, filter glossy 1.0, caustics anabled.
The brightness does not match perfectly but the renders are almost identical - the only difference is that the second version has more fireflies all over the sphere. The first render also has some fireflies - I dont know what is causing them. Need to test that further but I dont have much time right nowā¦
So Iāve had the pro-lighting skies ultimate for a few weeks now and I must say that I am a bit disappointed. Not with the add-on, that is really great, but with HDRiās themselves. By far the biggest problem for me, is the amount of exposure that was captured. Photoshopās own exposure filter says that they have approximately 5-7 EVs of exposure. Well, that is not enough as far as Iām concerned. A true HDRi has at least 12EVs of exposure. So in my eyes, these are not HDRiās but MDRiās (However cloudy HDRiās do not need the full 12EVs so at least some are decent in that regard). Yes it is true that they donāt take up insane amounts of space but consider that they are in fact 3000x1500 pixels big. And that for the price of the ultimate edition, you can buy yourself at least 16GB of ram (8GB for the lite version) and then use proper skydomes that give you accurate lighting without any node fiddling what so ever. No adding or multiplying to get good shadows, better light, etcā¦ A correctly made HDRi should work just by pluging it directly in the Backgound shader. If your image does not have enough information stored in it to start with, then no amount of processing will help you solve your problems or make the image better.
So what do you guys think? Are my concerns justified?
I am really considering getting my money back. I mean, if you are doing IBL, you might as well do it properly, damn it.
Hmmā¦ I agree. 5-7 EVs is not enough dynamic range to be considered HDR and 3000X1500 is too small, at least if you want decent reflections in highly glossy materials.
I canāt handle the memory requirements for a āproperā hdri. I have some on disk in the several hundred megabytes range (10k x 5x res, EV range unknown), and they wonāt even render (cuda crash). So proper for me means something my (modest?) hardware can handle. Even if I have to do tricks like either scaling the top up or adding lights.
I donāt use the addon or those hdriās myself, but I find it very weird that you think the addon is great but still want your money back?
The memory usage of a HDR in Cycles has nothing to do with its dynamic range. You have either 3 bytes/pixel for regular images (three channels with one byte each) or 12 bytes/pixel for HDRs (one 32bit float per channel).
The addon is good for quickly testing out the lighting, the browsing of HDRis is extremely fast. However I believe that the HDRis supplied with the addon are not of extremely good quality.
I remember trying out some of the HDRās Andrew had in the teaser packet for the architecture academy, and they werenāt all that good. Iām pretty sure that Greg Zaals HDRās are better.
So not that pro skies is perfect, there are bugs, maps not the best but they are work very well. Sure when you go all nerd critical on them they look not so great on paper. Keep in mind however that the reason its there is to help set up lighting fast and work well.
Andrew dose offer a money back deal but I think you have 30 days. Also after flinging emails back and forth with him theres more to come with this so its more of a Beta roll out I believe than anything right now. He is rather good and giving the early buyers freebies later
Yes, Gregās Bergen HDRi has nearly 15 EVs of exposure.
If and when Andrew improves the quality of HDRis I will gladly buy them again, but for now Iāve decided that I want my money back and Iāll buy some Peter Guthrieās skies. Maybe one every month or so or one per project. Yes it will cost me more in the long run but at least Iāll know Iām getting one of the best skydomes out there. And if Greg manages to launch his Haven project, I bet weāll see some really good quality HDRis that will also be cheap to buy.
If youĀ“ve ever photographed and put together HDRI you know why they cost as much as they do. The whole process is a pain in the ass.
Although it is kinda fun 
@BlackBart - Measuring the EV range like that in photoshop is not really accurate, thatās just the exposure offset from whatever the image was saved at. If you adjust the exposure and save the HDR, and then try that again, itāll obviously be different.
The only way to know is to have the original raw photos and calculate the EVs from the EXIF info. Or, if you can see that the sun is not completely solid white, then you know itās a good HDR with the full exposure range captured:
The sun should be slightly coloured (depending on time of day), and always brighter in the middle. If itās solid white/grey, the brightness is clipped and itās not an accurate HDRI. The left image is what you typically see with 12 EVs or less. The right image was captured with 22 EVs, by juggling filters and manually merging exposure brackets in the compositor (all software Iāve tried fail to correctly merge such high exposure values, and of course canāt tell what filter you used. Better to just do the math with nodes to merge them properly).
That particular HDRI (which I havenāt released anywhere yet, but will eventually be on hdrihaven.com) was quite taken late in the day in a dry dusty area, hence the very yellow sun. It gives this lighting, without any node hacks:
