- This topic has 6 replies, 3 voices, and was last updated 2 years, 6 months ago by
.
-
Posts
-
Per title: Light differences (volume, intensity) between DS versions, with older scenes created in older DS versions being overlit when rendered/previewed in newer versions of DS.
Issue seems between DS Ver 4.15.X.XX (roughly) and 4.20.X.XX+, with a massive light increase between scenes when created in the former and rendered in the latter.
(If have not seen or experienced, try an older scene created in 4.15.X.XX and render or preview with Iray in 4.20.X or 4.21.X+ to replicate.)
Anyone know what I'm talking about? Giant light flooding (and also translucency differences, but not the main issue) between versions for some reason?
Cannot be fixed with simple Exposure Value adjustments from tone mapping for scene-wide compensation; has to be fixed with actual light/lumen decreases, light by light.
When did this start? What is the fix? Is there a calculation or recipie for lumen dropping per light to match lighting of older scene from within older DS version?Too lazy to continue to try and parse the convoluted and denialist DAZ forums for the history of this bug. (They don't even admit to bugs being bugs, of course. And anyone calling out a "problem" is a heretic and will be lucky to have their post kept up much less get a straight answer.)
I know DAZ has tried to "fix" DAZ Iray between versions. And has had a fun, rollercoastery history for users and PAs in the process (with lots of posts about assets not coming out at their intended baselines when crossing over into new DS and Iray versions). But in regards to actual light volume complaints and fixes it seems different, or buried; with the only peeps coming up as a lack of real-world light metrics being any part of DS lighting or settings (another real problem).(If this has been posted in Zone forums somewhere, or if "known issue", please point thanks.)
Could this be caused by HOW you light your scenes? Maybe the type of lights used? The few times I've opened and rendered an older scene in 4.2-something I've not noticed a big difference. I tend to use spotlights (diffuse and off camera) for all my lighting, a hangover from my days as a studio photographer.
Scene is the same used across both scenarios. No changes. Only difference is DS version and Iray version (which is the higher suspect of the two for the difference in output).
Example: Spot 1 (of 3) is isolated (all other lighting disabled). Comparison renders between DS 4.15.X.XX and 4.2X.X.XX is made at identical resolutions and sample sizes. Result: Massive overlighting of 4.2X output (whether Iray preview or render), requiring individual light (Spot 1) to be reduced considerably in lumens to match 4.15.X output.
Consistent overlighting across all lights and conditions between DS (and Iray) versions.
(Or "underlighting" within 4.15.XX, if the 4.2X results are to be considered a "fix" or "upgrade" of Iray. The problem with that definition, of course, is that assets made or intended for DS 4.15.X or earlier will have a different look than later DS versions, and vice versa - as some have complained of in the D-forums.)Didn't the change log mention that they updated the iray version they were using? Not 100% sure, and too lazy to look... but that may explain it.
I stopped using spotlighting long ago, and mostly use either HDRI lighting or simple emissive plane primitives (or both) to get what I want. Used to be that I would tag bulbs and lamp surfaces with emissive properties, but I eventually figured out that doing so meant longer render times with not much different results.
It does seem to be a spot-specific problem. (Obviously more consistency is going to be gained from HDRI-only lighting regardless.)
It may or may not also be associated with Spectral Rendering. At least, that's what a couple of problematic scenes seem to share in common. (However, turning Spectral Rendering off, or even saving out older 4.15.X scenes without it seems to make no difference.)Noticing as well that converting scenes to Blender (via Diffeomorphic) is still handled fine from 4.15.X, but try the same past a certain point with other DS versions (and subsequently Iray versions) and there is a significant drop in light volumes within Blender when converting from 4.2X.X. Seemingly the same amount required to tone down older scenes in newer DS versions. I've experimented with the exact same scene between DS versions, and the only difference in conversions is lighting volume compensation (manually) in 4.2X+ releases. Same, converse, issue in Blender as in later DS versions: Pumping up or down the light volume, per spot, to match what is not necessary from 4.15.X.
...So, if Blender likes Diffeomorphic conversions from 4.15.X, but not later versions, then what changed?Really seeming more and more like an Iray version or implementation problem within newer DS releases.
(I, myself, am only recently discovering any discrepancies because I had purposely relegated myself to 4.15.X versions to avoid bugs and asset incompatibilities in newer releases. Now that I'm dabbling with more recent versions, the complaints make more sense.)(Maybe just converting and rendering using 4.15.X is still the better pathway.)You could try the conversion route, but if that doesn't work, it may mean having to re-think the lighting entirely.
You do have me thinking that maybe I should shake-out some really (really, really, really!) old 3Delight stuff, and re-do the lighting just to see what I get.
I'm just interested in seeing if this is a universal problem. Something having to do with spotlights and DS and Iray versions.
I find it very interesting, and possibly telling, that older versions of DS/Iray do well with light (spots) conversions to Blender, but newer ones do not. And that seemingly the exact amount of decrease required for old scenes with spots to match up in new versions of DS is what is then needed to be increased in newer DS scenes converted to Blender.
But I'm only one end user, so the test group is too small to determine.
- You must be logged in to reply to this topic.
