Daz Studio HSS, US and US2 Tutorial planned but want feedback first

123468

Comments

  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    Szark said:
    Mustakettu85 I will try those lights....there didn't work for a while so I forgot about them.

    Missing shadows, you mean? There was an easy fix actually, but I guess most people overlooked the thread where it was posted. They have been okay for the last handful of builds, though.

  • MEC4DMEC4D Posts: 5,249
    edited December 1969

    Hi Szark , no problem :) always a pleasure :)
    Using Gamma correction On and Gamma set to 1 make not really sense , the best results was around 1.50-2 to max 2.20
    however to compress the specular too much so it affected the main specular color , if you see to much changes in the main textures try to increase the Gamma up or down to find the best result with your lights as nothing is here a standard and each light setting need own gamma value
    DS point light are one of my favorite , as it bring out the surface so nicely especially if used SSS
    I always love your scene setups , can't wait to see what you cooking this time :)

  • SzarkSzark Posts: 10,634
    edited December 1969

    Rogerbee said:

    It's nice to see that, now and again, the folk that make the tutorials need them as much as we do. You never stop learning.

    CHEERS!

    exactly way many planned tutorials haven't been finished yet because I was fooling myself in thinking I had a full understanding and the more I got in to it the more I realised I didn't know as much as I thought. But I like that as it has taught me a lesson. I love learning like this now as for the first 4 years of using DS I was self teaching, hardly asking questions now I have nothing but questions. :) Also other plus side to this is meeting great helpful level headed mature people.
  • SzarkSzark Posts: 10,634
    edited December 1969

    Szark said:
    Mustakettu85 I will try those lights....there didn't work for a while so I forgot about them.

    Missing shadows, you mean? There was an easy fix actually, but I guess most people overlooked the thread where it was posted. They have been okay for the last handful of builds, though.Overlooked, no just missed it as I got used to using Uber lights and those other lights got put to one side in my thoughts. Thanks for bringing those back out. LOL

  • MEC4DMEC4D Posts: 5,249
    edited December 1969

    Hi, yes when you use a spotlight is different story , it don't even need gamma correction
    it work most best with UE2 or when too much lights are in the scene
    I never use it on a point or spot light

    below is a test I just made with UE2-Direct light and 1 spot light and shadow
    the B without Gamma is over exposed ( burned image )
    the A is compressed with the right white balance and wide gamma of the gray scale and that is what gamma correction should do , the results can be adjusted to the desire point .. so just example , also do you see the lines in the B image in the darker areas ? very annoying when you get this in a render and hard to edit ;)

    Mec4D said:
    Most of the monitors are NOT 2.20 but simple 1.00


    Hi Cath,

    You mean the dynamic range of the monitor, right? Because whatever devices are using sRGB (i.e. almost all of them) are not linear. That's for historical reasons.
    http://www.arcsynthesis.org/gltut/Illumination/Tut12 Monitors and Gamma.html

    thanks to Gamma correction ON and Gamma 2.0 the render will be compressed and overexpose reduced ending with a brilliant color and balance

    DS does it the other way actually. Take a look at the file attached.
    Here's a wiki quote that explains why there may be confusion:
    "Without context, a value labeled gamma might be either the encoding or the decoding value. Caution must be taken to correctly interpret the value as that to be applied-to-compensate or to be compensated-by-applying its inverse. In common parlance, in many occasions the decoding value (as 2.2) is employed as if it were the encoding value, instead of its inverse (1/2.2 in this case), which is the real value that must be applied to encode gamma."

    gamma-correction.jpg
    877 x 574 - 27K
  • SzarkSzark Posts: 10,634
    edited December 1969

    Mec4D said:
    Hi Szark , no problem :) always a pleasure :)
    Using Gamma correction On and Gamma set to 1 make not really sense , the best results was around 1.50-2 to max 2.20
    Yeah I am seeing that clearly now.

    Mec4D said:
    however to compress the specular too much so it affected the main specular color ,
    Sorry can you rephrase please I am not understanding what you mean. As far as I am aware all Bump, Displacement, Opacity and Specular maps are all set to 1.00 in DS automatically. Diffuse maps are set to 0.00 at default leaving the Gamma Correction ON to guess the Gamma of the Diffuse Texture Maps. I have noticed the need to adjust specular settings in testing that RoboSpider Image with Uber Surfaces. It sort of put me back in few steps with getting to grips with new Uber Surfaces settings and light balance with using Gamma correction ON at 2.2.

    if you see to much changes in the main textures try to increase the Gamma up or down to find the best result with your lights as nothing is here a standard and each light setting need own gamma value
    DS point light are one of my favorite , as it bring out the surface so nicely especially if used SSS
    I always love your scene setups , can't wait to see what you cooking this time :) Me cooking, nothing just testing out all this on my last image. It gives me instant feedback on the differences between Gamma on at 1.00, ON at 2.20 and one with Gamma correction OFF

    Mind you I have a Bird image cooking for Ken's contest this year. ;)

  • MEC4DMEC4D Posts: 5,249
    edited December 1969

    With the specular it depends of the setting , I use mostly Ubersurface shader as base and not standard default so it looks totally different when the gamma correction is on or off .. I would use higher values when I use gamma correction as it getting compress .. specular from light is a fake product so it does not get so well together for that reason I prefer reflection maps or true reflections as specular if possible .
    There is no standard , each scene with different light will give different results and all you have to do is trust your own eyes while creating
    See the gamma correction as Image Curves in Photoshop nothing more nothing less .

  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    Mec4D said:

    it work most best with UE2 or when too much lights are in the scene
    I never use it on a point or spot light

    below is a test I just made with UE2-Direct light and 1 spot light and shadow
    the B without Gamma is over exposed ( burned image )
    the A is compressed with the right white balance and wide gamma of the gray scale and that is what gamma correction should do

    Cath dear, for the life of me, I can't seem to reproduce the effect you're getting. Here's a test with UE2 (AO, soft shadows, no map, 5% intensity), a spotlight (no decay, 100% intensity), specular surfaces... I keep getting expansion with GC on at 2.2. I tried using UberSurface for the materials, and got the same.

    What are you viewing/editing your renders in? Maybe there's some extra colour management going on somewhere?

    GCtest_SPIDER_.png
    1000 x 500 - 414K
  • MEC4DMEC4D Posts: 5,249
    edited December 1969

    I use 100% intensity UE2 and 100% 1 spot light with raytraced shadow
    UE2 without map just plain base white , Ambient occlusion with direct shadow

    UE2 on 5% is nothing use just the base and do not change any settings for the best result let the gamma correction do that for you
    I did not edited my renders they are straight from DS , and I get the same results in all of my renders

  • SzarkSzark Posts: 10,634
    edited March 2014

    Thanks to you both and Millighost, I think the haze of confusion is lifting.

    Shame no one thought to point me to that thread when I asked about Gamma Correction a while back, it could have saved me months. And shame on me for missing it. At least I am getting there now thanks to this recent help.

    Sorry for my absence but this is what I am like, a dog with a bone and had to get it all straight in my head.

    This is where I am at now, see image below (untouched straight out of DS). I thought let’s try a simple scene to get my head around this.

    UE 2 GI (Global Illumination) with 30% IDL Strength (thought it could do a little higher IMHO to get more bounce light), 75% AO
    2 X dzPointLights
    Scene inside a Cube, PGE tool to make the floor a separate mat zone, the rest of the cube has a light material hence the reflections being strong.
    Small Cube Sub-D X 2 res with Smoothing and Collision Applied.
    Uber Surface applied to all mats except for the floor which is a new Shader Mixer shader from the Store and all Specular turned off on all Uber Surface Materials. I have only used Diffuse, Diffuse Roughness, Bump, Displacement, Refraction, Reflection and Fresnel.

    I rendered this with Gamma Correction ON at 2.20, Pixel Filter Catmull_Rom as IMHO the colours and reflections seem to be better than Sinc, Max Ray Trace Depth 3, Shading Rate 0.20.

    Just over 12 hours to render at that res but I did have UE2 very high, Quality wise given it is GI. I still think it needs higher quality settings. Because some cannot use GI I am going to test with UE2 Base IDL (not GI) and see what happens.


    Now to keep chipping away at Uber Surfaces and see what other metals I can do. To me this looks like an Aluminum or Alloy type of metal. But I think if I change the environment to something more colourful with more contrast it will start to get a better look. I need to play with Fresnel a bit more but overall I am happy in the direction.

    I mentioned earlier and know one contradicted me but for clarity am I correct in assuming that Gamma Correction is a too fold solution in getting the best results from the surfaces with physically based light and to get a print that looks like the rendered image when viewed on a monitor?

    RoboNewTest1.png
    1400 x 787 - 2M
    Post edited by Szark on
  • MEC4DMEC4D Posts: 5,249
    edited December 1969

    Nice done Szark ,
    the Gamma correction will keep the render white balance the way it should and bring a lot more details since the range get higher .. recommended for printing .

  • SzarkSzark Posts: 10,634
    edited December 1969

    Thank you Cath. I think I am getting better results using the Base UE2 and IDL than GI. Hopefully have another test render done tonight.

  • MEC4DMEC4D Posts: 5,249
    edited December 1969

    Mostly I use UE2 and IDL , GI can create nice effect but not for all scenes and surfaces

  • SzarkSzark Posts: 10,634
    edited December 1969

    I have never used GI before yesterday but I am liking IDL more with this test.

  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    Szark said:

    UE 2 GI (Global Illumination) with 30% IDL Strength (thought it could do a little higher IMHO to get more bounce light), 75% AO

    I'm not sure if IDL strength setting would affect GI mode. It may, but I haven't tested. I'm fairly sure, though, that AO strength should not affect GI. AO is a completely unrelated shader operation.
    Overall "light intensity" might influence GI (try upping intensity scale). But I forget if I've ever tested this to make sure.

    What will certainly influence GI brightness is max raytrace bounce. But it will also increase render times with every extra bounce. I don't know if there are optimizations in the shader code for subsequent bounces - you could probably ask Omnifreaker about it. With optimisations, it's possible to only have the time increase by about 10% with every extra bounce. Without them, it's a whole new pass: 3 bounces, thrice the time.


    I use 100% intensity UE2 and 100% 1 spot light with raytraced shadow
    UE2 without map just plain base white , Ambient occlusion with direct shadow

    UE2 on 5% is nothing use just the base and do not change any settings for the best result let the gamma correction do that for you
    I did not edited my renders they are straight from DS , and I get the same results in all of my renders

    I actually did start with 100% UE2 at unmapped white (although I see no point in enabling what should be more costly shader code, i.e. directional shadows, when there is no map that the shader would sample for brightest regions to cast those shadows...). It didn't really change anything apart from making the transition between spotlight and ambient way less noticeable.

    Are you by any chance using reflections on those surfaces in your example? That's the only thing I can think of that would explain the effect you're getting...

    PS Linear workflow would be the correct approach regardless of light intensities actually.
    http://www.pixsim.co.uk/downloads/The_Beginners_Explanation_of_Gamma_Correction_and_Linear_Workflow.pdf

  • SzarkSzark Posts: 10,634
    edited December 1969

    Yes IDL does work in GI. Yes got Max Ray Trace at 3 and yes I know what it does and how it relates to GI etc etc and yes render times, oh my.

    No I am liking UE2 base and IDL better, soft shadows. I never use Directional Shadows.

    Uber Surface doesn't have a Max Ray Trace setting but Uber Surface 2 does.

  • SzarkSzark Posts: 10,634
    edited December 1969

    UE2 IDL with Soft Shadows using the Soft Box preset.

    3 dzpointlights

    I am liking this one more than the first.

    This is a low quality render. I will let it render overnight and see how it looks in the morn.

    robofinaltest.jpg
    1500 x 843 - 973K
  • SzarkSzark Posts: 10,634
    edited December 1969

    here we go full high quality settings. Took about 14 hours to render but I only had Max Ray Trace Depth at 2.

    What you think....good, room for improvement, crap? Please let me have it, it is the only way to learn. :)

    RoboFinalA.png
    1500 x 843 - 2M
  • SimonJMSimonJM Posts: 5,942
    edited December 1969

    Szark said:
    here we go full high quality settings. Took about 14 hours to render but I only had Max Ray Trace Depth at 2.

    What you think....good, room for improvement, crap? Please let me have it, it is the only way to learn. :)


    Looks pretty good to me. Apart form the total pain in creation/posing and the likelihood it'd look too 'fussy' a little chain cobweb at one corner of the block might be fun. The middle section of the 3rd leg from front, right side of the front spider looks like it has caught a shadow, a dark reflection or the surface texture may be a little off.
  • AndySAndyS Posts: 1,434
    edited March 2014

    Hello Pete,

    what were your parameters for the "low res" picture? That even looks good - without these noisy shadow zones you normally get for low quality renders.
    And your robo-spiders are great !

    Btw:

    3 dzpointlights

    I have a problem even adding a second one. Everytime I try, it replaces the first one.
    How to?

    Andy

    Post edited by AndyS on
  • SzarkSzark Posts: 10,634
    edited December 1969

    Thank you.

    The low res...hum can't remember I will have to take a look and get back to you.

    To load more than one light is to hold down Ctrl while loading and a new dialogue box will appear, you need to choose "ADD" instead of "Replace".

  • Mustakettu85Mustakettu85 Posts: 2,933
    edited March 2014

    Szark said:

    What you think....

    The spiders are super pretty =) The black box looks as if it were made of rubber to me, is it what you intended?

    PS You wrote in an earlier post "...To me this looks like an Aluminum or Alloy type of metal. But I think if I change the environment to something more colourful with more contrast it will start to get a better look. I need to play with Fresnel a bit more but overall I am happy in the direction."

    Have I said that according to physics, metals should not have Fresnel?

    ...and one more message for Cath - another thing apart from reflections that would make sense to me re:your effect, that would be specular colour maps. If you're using any of these, then I understand how it works =) If not, then it's a mystery to me =)

    Post edited by Mustakettu85 on
  • SzarkSzark Posts: 10,634
    edited December 1969

    Rubber/foam I didn't really have a direction, just what looked good. :)

    There seems to be two trains of thought on cgSociety, one group says yes for Fresnel and some say no. I understand dielectrics and non dielectric surfaces and like you thought this too. But after reading many pages on the subject I am not so sure now.

  • SzarkSzark Posts: 10,634
    edited December 1969

    Szark said:
    Mustakettu85 I will try those lights....there didn't work for a while so I forgot about them.

    Missing shadows, you mean? There was an easy fix actually, but I guess most people overlooked the thread where it was posted. They have been okay for the last handful of builds, though.can you point me to that thread please Mustakettu85 as soft shadows don't seem to be working http://www.daz3d.com/forums/discussion/38362/

  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    Szark said:
    Rubber/foam I didn't really have a direction, just what looked good. :)

    There seems to be two trains of thought on cgSociety, one group says yes for Fresnel and some say no. I understand dielectrics and non dielectric surfaces and like you thought this too. But after reading many pages on the subject I am not so sure now.

    It does look good. Would you share the settings, please?

    The question would be, what background does that "pro-Fresnel" group have? If they're trained artists, that's one thing. It means that we, in turn, as artists, can disregard their opinion because it is what it is, "whatever looks prettiest".
    Now if those guys are somehow physicists/engineers, then it's more interesting. I, for one, have never heard about Fresnel laws in context of metals and light. But I did not major in optics. Maybe there is a handful of specific cases or whatever.

    As for the lights, my fix was for the time when the shadows did not work at all, and it's here: http://www.daz3d.com/forums/discussion/12191/#182158

    The guys in the thread you linked basically told it like it is, though: a point light won't cast shadows that are much soft, by definition. There's that intrinsic conceptualisation limit. Now if you imagine a naked lightbulb and a photostudio softbox, which would cast softer shadows? Softness of shadows depends quite a lot on the emitter surface size. Think of a softbox as consisting of an infinite number of infinitely small pointlights spread over a surface; each casts sharp shadows, but as there's a lot of them and some are far apart, their shadows overlap and hence the edges are blurred... something like that.

    Here's a link that specifically contrasts "traditional" point light sharp shadows with softer ones and gives some interestic ideas about setting up surfaces to better match either: http://www.peachpit.com/articles/article.aspx?p=516590&seqNum=6

    There are "point" (i.e. omnidirectional) lights in UberSoft Lighting Kit, but given that they are morphable into directional lights, and given that the shadows from all of them are calculated by the Master one, I won't even try to guess what those "soft" shaders could be under the hood.

    Here are two renders of a 3 meter tall cylinder atop a 5 meter side plane. The one on the left is illuminated with a single point light; the one on the right - by four identical ones (intensity 25% each). I've moved them apart a little, so if you look hard, the shadow gets marginally softer. More pointlights spread farther apart - more softness kicking in earlier. Hope that helps!

    dzpoints_.png
    1000 x 500 - 202K
  • evilded777evilded777 Posts: 2,437
    edited December 1969

    Hey szark, have you tried this combo: UE2 set to occlusion with a shader mixer IDL camera and using the 3Delight point cloud script? To me it seems to be about the fastest IDL plus occlusion setup.

  • SzarkSzark Posts: 10,634
    edited December 1969

    LOL that went straight through one ear and out the other evilded. Got any links to show how to do this? I am presuming it is with the external 3delight render engine (free)? I know of threads discussing this but I get lost at page 2 in the main one. There are no Step 1 > step 2 > etc.

  • mjc1016mjc1016 Posts: 15,001
    edited December 1969

    Szark said:
    LOL that went straight through one ear and out the other evilded. Got any links to show how to do this? I am presuming it is with the external 3delight render engine (free)? I know of threads discussing this but I get lost at page 2 in the main one. There are no Step 1 > step 2 > etc.

    In the render settings, the selection box for Render Engine, there is a selection for Scripted 3Delight. If you choose that, there is another group of options listed below...the scripts you can use. One of them is the Pointcloud render.

    Just for the first time...take a scene and flip that on and render...don't change anything, just render...then come back and we'll move you on to the next steps.

  • Takeo.KenseiTakeo.Kensei Posts: 1,303
    edited December 1969

    The question would be, what background does that "pro-Fresnel" group have? If they're trained artists, that's one thing. It means that we, in turn, as artists, can disregard their opinion because it is what it is, "whatever looks prettiest".
    Now if those guys are somehow physicists/engineers, then it's more interesting. I, for one, have never heard about Fresnel laws in context of metals and light. But I did not major in optics. Maybe there is a handful of specific cases or whatever.

    For what I know you have Fresnel everywhere and as you have reflections in metals you have a Fresnel term http://sirkan.iit.bme.hu/~szirmay/fresnel.pdf

    The question is then : do you work with Physically Plausible BRDF Shaders or not ?

    We don't have shaders in DS in which you calculate the Fresnel and other terms from the IOR of the material you want to imitate. So then the question is How do you do with non plausible shaders and a Fresnel term that is not IOR dependant?

  • SzarkSzark Posts: 10,634
    edited December 1969

    Mjc got that. I remember seeing it but never done it. Will this render be inside DS? I will give this a whirl tomorrow evening. It is getting late here and I am heading to bed soon.

    Takeo this is what I got from reading many things about cg in general and from what I could make out it does depend on the shader and software been used. Thanks for the link I will give that a read tomorrow too.

    Thanks to everyone that is helping me with this I am having so much fun and learning loads

Sign In or Register to comment.