Show Us Your Bryce Renders! Part 4
This discussion has been closed.
Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
Right... after close examination of Horo's detailed investigations into lighting - as covered by his last batch of video's I've come up with a new hybrid method that offers what I believe is more realistic light simulation (I've tried to check again Octane - which offers a good standard of light simulation) and not unreasonable render times. The process isn't perfect, there are noise issues associated with this approach, but the noise is "correct" in that the same noise was evident in Octane, but that Octane is capable of much higher levels of sampling the noise eventually goes away. Noise correction methods... I'm jotting that down on my "to-do". Here's the new method, video, tutorial... and renders for comparison. Bryce is on the left.
Bryce 7.1 Pro Experiments - Obscure Gel Lights - by David Brinnen
@ Horo, you are welcome.
@ Jamie, very good, I like all of them = nut in particular *_6.
@ BlumBlumShub, we nice render of the dragon and cross.
@ TLBKlaus, you have become a streaming light addict!
@ Dave, is good, but she does look a bit "hard", it gives her a painterly finish if you know what I mean?
@ IceScribe, good use of mounds, works very well, but I don't know that I'd want to drink that!
Hope you have a Happy Easter
Nothing ground breaking here... Just playing with OL checking out what settings can be altered and tweaked.
Here the saturation was turned down to get rid of a colour cast that was introduced by the HRDI.
@Horo: Thanks.
@David: Thank you. Videos are really good. Image turned out good as well
@bigH: Figure is really nice.
@Dave: Those turned out rather nice.
Nice lighting Bigh and the hair has come out very well.
Dave, isotropic effect working in your render, makes the material quite plausible, good lighting and material choices all around - as per.
OK let, doubled up the render size for Bryce and filtered the noise in PSP8 before scaling down again with more filtering. I'm thinking, render times aside, Bryce's TA is matching up well to Octanes pathtracing.
Fantastic videos. Horo! Not a waste of time in any way. Amazing work.
Indeed, much improved. I think the Bryce version does quite well especially with the post work corrections. It might not be a bad idea to keep track of common post work procedures. I'm curious myself the settings you used for the noise reduction. I use psp12 or something like that so I'm sure I can follow along.
Generally, the Bryce versions are missing light. Areas buried deep within geometric protrusions dont manage to gather enough light. The base of the cube where it meets the floor fails to gather sufficient light. This could be due to some sort of shading threshold under the hood with TA, or it could be a result of a lack of sufficient ray firings. It could also be caused by some degree of lingering bias in the scattering, or it could be a ray depth issue. Literally could be anything.
But most likely, it comes down to materials. Faking the SSS has to be considered and that mostly means finding ways of introducing some dare I say it.....ambient glow.... Not too much mind you, and it isn't by any means an accurate solution. But it might help in situations like this one. The way I see it, ambience is fine as the last resort, as the tweaker. In fact I think ambience is fantastic at tweaking once all the other lighting details have been worked out.
At this point I'd say some sort of global correction trick might be required. Not sure what that might be, but ideally it would help TA finish off the gathering process, perhaps exploring pathways TA currently omits.
Indeed, much improved. I think the Bryce version does quite well especially with the post work corrections. It might not be a bad idea to keep track of common post work procedures. I'm curious myself the settings you used for the noise reduction. I use psp12 or something like that so I'm sure I can follow along.
Generally, the Bryce versions are missing light. Areas buried deep within geometric protrusions dont manage to gather enough light. The base of the cube where it meets the floor fails to gather sufficient light. This could be due to some sort of shading threshold under the hood with TA, or it could be a result of a lack of sufficient ray firings. It could also be caused by some degree of lingering bias in the scattering, or it could be a ray depth issue. Literally could be anything.
But most likely, it comes down to materials. Faking the SSS has to be considered and that mostly means finding ways of introducing some dare I say it.....ambient glow.... Not too much mind you, and it isn't by any means an accurate solution. But it might help in situations like this one. The way I see it, ambience is fine as the last resort, as the tweaker. In fact I think ambience is fantastic at tweaking once all the other lighting details have been worked out.
At this point I'd say some sort of global correction trick might be required. Not sure what that might be, but ideally it would help TA finish off the gathering process, perhaps exploring pathways TA currently omits.
Agreed. Yes I think the first port of call will be to increase the ray depth. This clearly hampers the TA process in tight corners and that is exactly where you have identified an issue. TA is a brute force method and that has to be born in mind also, for all Octanes lofty claims to being unbiased, the render settings are adjustable in ways that suggest to me that it is "unbiased" within broadly defined parameters. And as such, a similar claim could be made for TA, except that I would have to add, within narrowly defined parameters. But unboosted and boosted, under the right lighting conditions can be matched quite closely with Octane (bearing in mind I my hamfistedness with this software) so I think the news is mostly good for Bryce's render engine - it needs tweak-ability adding not scrapping and starting again. If TA was given access to the kind of optimisation controls that Octane has - and casting my mind back, I recall this is exactly what Len suggested, it would fare better in terms of efficiency and final outcome. Still, plenty to test yet, this is my provisional view. It may change as I dig deeper.
I prefer the Bryce one, but it comes as no surprise, I do like to see darker shadows to hide in. :)
Off the top of my head, would introducing the slightest amount of haze help the light to bounce more and penetrate into the gaps better?
Introducing a volumetric material with TA is possible, but very render intensive and sometimes a bit crashy. Default Bryce haze could be used to fake ambient - somewhat towards what you are suggesting, but it would also interfere with the colour of the light interaction. It's an interesting idea, I'm not being dismissive, just if you are going to go down that route, there are more efficient alternative fixes - the aim at the moment is to try to create a system that works most of the time and isn't beyond most Brycers ability to replicate in their own renders.
Here a test using the Obscure Gel Lighting idea on that free model head. A weak Bryce sun has been used to introduce a touch of direct light - mostly for the benefit if skin specularity.
Another test of Obscure Gel Lighting...
This Obscure Gel lighting method is fantastic! The bad news is the noise (we had asked in the SC to increase rpp at least to 1024 but preferred 4096 from 256, but we didn't get it) and the longer render time.
If you export a dark set HDRI as LDRI for the gel, HDRI Effect has to be multiplied by the same amount Intensity was divided. The light source gets quite bright and it gives the main light. If the exported HDRI is relatively bright, i.e. as the backdrop Intensity is set, there is less contrast and more ambient light.
I had set Intensity to 30 and HDRI Effect at 60 for the test render. Then I set Intensity to 5 from 30 and exported the HDRI (not tone-mapped, of course) and had to set HDRI Effect afterwards to 360. This is the first render. HDRI used was Sunset. Note the light reflection on the ground from the sun reflected in the mirror ball. A fainter light on the ground is opposite. Here, the sun from the Utah Teapot is reflected to the mirror ball, which reflects it to the ground.
For the second render, I exported with Intensity 30 and could keep Effect at 60. There is more yellow, more ambient colour is retained. The reflection from the sun on the ground via the mirror ball is fainter but still visible.
@bigh - thank you. Why isn't she smiling? She's looking good enough to smile.
@Dave - really looking good.
@Rashad - thank you!
@David - the woven spheres - I'm with Dave, preferring the Bryce render. The Octane one may look more natural, but the Bryce one more artistically pleasing.
The head which pulls a face looks rather like having used some powder. The other one looks nice, the render, I mean.
Thanks Horo, good looking results!
The experiments continue...
Bryce 7.1 Pro Experiments - Hyper Surface Scattering - by David Brinnen
@David: I'll have to side with the Bryce rendered woven cube. As Horo said, it's more artistically pleasing. Skin texture on the first head, and the hair, really leave a lot to be desired. The second one really looks nice.
@Horo: Really nice looking pots. Is the color difference between both pots the result of lighting?
This one's been a WIP for some time, simply because I didn't know which direction to go with it. So I finally got off the pot and just tried something with it. Comments are gladly welcomed.
Indeed. I would go with a combination of things with perhaps the most prominent being an uncommon frame of reference. Luckily, in David's comparison there is a common reference point - namely Horo's HDRI.
In the comparison it is evident that in the Bryce shot the overall light level is lower, including the reflection of the garage in the sphere. Taking the dual image into PhotoImpact, I snipped out a slice of the Bryce reflected garage, placed it over the Octane reflected garage and adjusted brightness, contrast and colour balance to obtain a reasonable match. These settings were then applied to the entire Bryce half of the image.
The result is much brighter image, but clearly not representative of any Bryce settings and effects therefrom. Obviously the ideal place to match the garage-reflection brightness is in the Bryce setup prior to full render (should be possible with a bit of tooing and froing). With that done we can see what the Obscure TA Light does with the extra illumination.
Agreed. Yes I think the first port of call will be to increase the ray depth. This clearly hampers the TA process in tight corners and that is exactly where you have identified an issue. TA is a brute force method and that has to be born in mind also, for all Octanes lofty claims to being unbiased, the render settings are adjustable in ways that suggest to me that it is "unbiased" within broadly defined parameters. And as such, a similar claim could be made for TA, except that I would have to add, within narrowly defined parameters. But unboosted and boosted, under the right lighting conditions can be matched quite closely with Octane (bearing in mind I my hamfistedness with this software) so I think the news is mostly good for Bryce's render engine - it needs tweak-ability adding not scrapping and starting again. If TA was given access to the kind of optimisation controls that Octane has - and casting my mind back, I recall this is exactly what Len suggested, it would fare better in terms of efficiency and final outcome. Still, plenty to test yet, this is my provisional view. It may change as I dig deeper.
Octane may default to some larger number of light bounces like 30 or 100. I think it's fast even with that numbers because it doesn't split rays at each bounce into several ones. Bryce splits rays thus it becomes slower when more effects are added so lowering number of bounces is the greatest time saver.
Another thing that I'm going to make you bored to death is me repeating: "gamma". Check out last 3 images from this http://www.aversis.be/tutorials/vray/vray-20-gamma-linear-workflow_01.htm page - hdr map and simple shapes lit with it. That's what happens imho when you compare more contrasty Bryce output with heavier shadows to Octane rendered image in this case.
The experiments continue...
Bryce 7.1 Pro Experiments - Hyper Surface Scattering - by David Brinnen
That's an interesting one! I'm thinking about repeating it if I find some time.
Playing with clouds lit by radials. Orange on right and violet (subtractive yellow-green) on the left. Postpro in GIMP (sunrays, glows, colour correction, etc.)
I like to see the foreground more in this type of renders - you are at ground level .
if you were up higher then you would see more of the background .
just my point of view .
I've been meaning to ask for a while, why when setting up OL is it necessary to set the light gel to default grey and not pure white?
It's just that today I tried pure white and it gives better light returns.
For me, the last few days seem to have been an exercise in using white stuff and to get rid of colour casts introduced by various aspects of lighting. Having said that, my latest test render will not rid the noise even though rendered at 256RPP.
@bigH: Thanks for the suggestion. I have to be careful how high I place the camera since there's nothing on the other side of the tall mountain. I will give your suggestion a try.
@Dave: You test looks good to me. I especially like how the lady actually looks like she sitting on the ground due to the shadow cast by her hands and feet. I couldn't see the noise you spoke of, but that may be just me.
I am learning a few things, two being accidents can produce interesting images, and that I have yet to catch on how to produce real nice mirrored images. Here is the accident I spoke of.
Right, it doubles the light. But then, you can easily compensate with HDRI Effect.
I've been meaning to ask for a while, why when setting up OL is it necessary to set the light gel to default grey and not pure white?
It's just that today I tried pure white and it gives better light returns.
For me, the last few days seem to have been an exercise in using white stuff and to get rid of colour casts introduced by various aspects of lighting. Having said that, my latest test render will not rid the noise even though rendered at 256RPP.
I suggest resetting to default grey, simply to remove a step from the process, since if you need more light it's usually possible just to up the output from the HDRI. However... there's something else to consider which I've not really touched up because it needs testing and, even if tested, the answers are going to be hard to unpick. It is this, at what point does the HDRI effect of the light source become LDRI? If you think about how the render changes (or doesn't change) when you apply boost light that gives a clue as to what in particular is getting clamped. Somewhere at some point, a shift occurs. Now if the light is grey - but there is more light and this goes into the scene and then goes back out again and sees grey - does the grey get some light accumulation from this activity? So the light gets amplified within the light source? And if it does does having a white sphere with half the light, give the same results, or is more residual light amplified? Tricky questions.
Yes, the pot is default grey with 15% reflection.
Nice start with your red mountains. Difficult to say at the moment whether they are huge or tiny.
The blue render looks like an underwater scene.
@Peter - interesting test!
@dwsel_ - experimenting is always interesting. We either learn what can be done or what we should not do again.
PJF. In Octane Horo's HDRI looks a bit washed out - there are a couple of controls that I've just sort of noticed and slid around (only yesterday) that might correct this. As things stand in Bryce - the appearance of the backdrop in terms of contrast is hard to modify - in fact, I don't know how that works in HDRI's - in what way what we see is related to the light, there's some kind of connection obviously but... what is displayed on the monitor being only a narrow interpretation of the full range image, I'm not sure even if both the reflections looked the same if that would mean the outputs were the same. I usually leave all that horribly complicated stuff to Horo ;)
But it's a good idea... I'll investigate further.
Dwsel. Gamma - NOoooooooooooooooooooo! Only fooling, but it's just so difficult to get my head around. Every time I try to read up on this topic, I feel like I am reading most of it in some foreign language, I barely know enough of the basic terminology to grasp it properly. And also, another thing that deters me is that it seems like a terrible lot of faffing around - even more so than what we are used to.
Nice clouds! Now there's something I do understand. Clouds! Lovely fluffy clouds. More my speed!
The blueishness I'm guessing is haze. If you turn your atmosphere off it will remove this, then you'll need some light in there. To achieve that you could... just a suggestion, use a HDRI image, and exclude the mirror (if it is on a sphere exclude that object). or modify the mirror material so it doesn't cast shadows - that would work too. Remembering that Max Ray Depth here, will greatly influence the number of internal re-reflections.
@David - Octane: Image as environment, floatimage gives monochtome backdrop and light.
Float power changes light intensity and backdrop brightness, floattexture power as well.
Float gamma adjusts backdrop - unfortunately also the light , lower less contrast, higher more. There's no tone-mapping for the backdrop other than linear. Octane IBL is very simply implemented and doesn't give you a much of a choice.
Components made in Wings 3D exported to Bryce.
Scene assembled in Bryce exported to Wings.
Material groups assigned in Wings exported to Octane.
Victora exported to Octane from DAZstudio.
Victora combined with Bryce assembled scene and rendered in Octane.
I wonder if Horo can make a wide angle lens for Octane? Hmn... That would be nice.
Heh, well that was the other half of my 50 / 50 option, but I went with brightening the Bryce scene since its darkness was under discussion.
Interesting to see if lowering the brightness of the HDRI in Octane results in a darker scene.
Although the Octane version is more pleasing in an "ooh, ahh, fancy 3D" sort of way, I feel the Bryce rendition is more appropriate to that object existing in Horo's underground garage.
Off to work for a few hours now. Bleargh.
And Bryce has got something that Octane is sadly lacking... Horo's wide angle lenses!