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
I love hearing your professional advice and wisdom, Phil! ;)
That was a beautiful three posts, and the finale of that third post is simply and breathtakingly gorgeous!
See?
Now I'm glad that you've responded after my reluctance to remain on this path. Remembering back, right after this thread began, I was using GC = 2.2 for nearly anything I did, and got quite used to it. I have been recommending your studio product a lot, as I remember how much fun I was having with GC = 2.2 using much from that kit as a good starting point. Perhaps it is a matter of old, bad habits that keeps pushing me back off? Maybe. After reading your above quote, two things must occur: I need to get that book in my hands and I really should get back into my investigation towards Gamma Correction.
Now that I'm thinking of it, I seem to recall that many of my findings on the matter were of faster renders, too. Being almost only interested in animation processing, this is a huge factor for me. Like SciFi Funk, I have a vision that I'd like to see through to fruition within my relatively short lifespan! :)
It's that render speed that often drags me away from other rendering changes. But this one actually landed me faster render times - which simply blows me away! More real, and faster? Eeee Gads! That's cool!
For the most part, I believe it was when I wanted those stark, crisp space images that ended up causing me to convert back. But maybe black doesn't need to be that black. And, again, perhaps my lighting setup was off. I'll definitely reopen this can o' worms though! It was fun then, I'm sure I'll love it now!
I need to get that book as well. It seems other things keep coming up to sap away at my finances whenever it seems as if I'm getting ahead.
Regarding the 2.2 conversion, I'm still not convinced. I still think that it will depend mostly on the scene. So Dart, if the scene calls for deep, dark shadows, then it calls for deep dark shadows.
The thing about quotes is that they can be taken out of context. For instance, what bad habits did animators get into with CG lighting in the early nineties when CGI was still in its infancy (Jurassic Park notwithstanding)? It isn't covered because it's a purposely selected quote that fits Phil's personal perspective on the subject and reinforces his argument in favor of the GC 2.2 setting. There isn't anything nefarious or dubious about selecting the quotes or posting them, it merely fits his opinion and interpretation of the subject. My problem is that they just don't relay the whole context of the subject, which could span a paragraph, a few pages or an entire chapter. Salient information may be missing that also needs to be understood when discussing this issue.
That is a good point EP. Honestly, I haven't tried out Phil's GC trick yet - so I can't pick a side.
I do have the 2nd edition of Birn's book though.
Well even though it has a huge impact on the appearance of the output, after all, that's what it is, it seems that (and this seems to be explained better than I could earlier in this, excellent thread) GC = 2.2 also fires in some sort of texture map efficiency mode that just makes that render engine come alive with speed! Perhaps I'll test it when I have the time to mess around. But I recall 'feeling' a noticeable difference in speed when setting up and using the simple setup Phil was calling for in his experiment he wanted us all to try.
Whether I end up using it for my main production remains to be seen - but I sure am intrigued by what it does to Carrara.
I also like to compare the renders I'm getting with frames from, say, Clone Wars, which I really love the look of. I love how they've crafted their look and feel and the professionalism of the movements, camera movements, expressions, lip synching, individualism, etc., just blows me away. But so now I'm going to go and learn whether their black is really black, or is it a trick, making me think it's black. And then I want to learn whether I can achieve black using GC = 2.2, if so.
Colour me intrigued...must try playing with this over the weekend.
Well, I suppose I did select the quotes, but there was no "context" that was qualifying what he was saying. He devotes over 10 pages in the new edition of the book to Linear Workflow, so I obviously couldn't just quote the whole thing as that would contravene his copyright.
If a scene calls for deep, dark shadows, there is nothing in Linear Workflow that prevents you doing that. The perception that Linear Workflow = pale or washed out images couldn't be further from the truth, as I have been at pains to point out from the start of this thread.
The bad habits that he refers to include things like using lights with unrealistic falloffs (because they look better than inverse square falloff without using gamma correction) and adding extra fill lights to compensate - now we've all done that!
I understand that I am just one voice, but when Linear Workflow is at the heart of all physically based renderers and has become the default in programs such as Cinema 4D, as well as being an option in all major 3D programs (not to mention being unequivocally endorsed by Mr Birn, which in turn means the likes of Pixar) , I think you have to sit up and take notice.
Colour me intrigued...must try playing with this over the weekend.Earlier in this thread, Phil gets pretty informative on the subject, and offers a link to a video where a fellow takes his time to do his best to explain Gamma Correction, which is really a fun watch. But in what Phil says and what I've observed, it seems true to me, that there's more to it that what it might seem on the outside. Like controlling a GC slider of a game or something. For a good demonstration of this stuff, I'd suggest starting this thread at the beginning. Once you feel you've read enough to get a good grasp of what's going on, and have picked up on Phil's request for us to try, then go ahead and skim forward, stopping at some of the posts that interest you, images that make you go... wait a minute... lemme see that one again, and get to the link to the video, set yourself aside some twenty minutes, I think, and watch the whole thing, paying attention - even if after hearing a few initial paragraphs you feel you've got it. Just watch the thing - all of it. It's worth it. Then continue through the thread, or get to the experiment that Phil has set up for us in the first post or three.
One area that GC has really shone beneficial is on interior CGI. Ambient light has always been improperly demonstrated - no matter the software - without GC. And some apps might require a different setting than 2.2, which is a specific trigger, it seems, which triggers a true "Linear Workflow" in Carrara's render engine. GC is nothing new, and I'll agree with anyone on that. But I still feel that anyone opening any part of this thread should at least try Phil's simple experiment request. It's fun and it may just change how you look at lighting your scenes, by adding simplicity galore.
Dart wrote "And then I want to learn whether I can achieve black using GC = 2.2, if so." Yes, you can have black, and other dark shades, and dark shadows, and high contrast, and night scenes and whatever you want.
I doff my cap to you Sir Phil!
(as usual!) :)
I am definitely scrambling to right my financial ship and start buying more 3D stuff again. If I get Birn's 3rd edition, maybe I can send the 2nd edition copy up to EP.
That is a good point EP. Honestly, I haven't tried out Phil's GC trick yet - so I can't pick a side.
I do have the 2nd edition of Birn's book though.
It's not a side really. It's just my opinion. I have tried it, and I didn't care for it. It may be my monitor calibration, or the way Macs handle gamma, but if I were to use it, I found a lower value suited my taste better. Then if you add the issues mentioned with hair and other things, things which make me want to use Carrara, and that it works best with GI, which is slow for my aging machine, it's just not worth the extra work. The results don't equal the trade-offs.
I tend to not go for hyper-realism either. When I design a dark scene, many times I don't have a background or backdrop. I want the image to disappear into the inky blackness. The gamma correction kind of hoses that effect as well.
I wasn't trying to imply that you were being disingenuous or suggesting you violate copyrights by posting the entire text (although, that would save me a few dollars). ;-) It's just that I don't have the book, so I don't know the whole of the discussion.
As to using lights with unrealistic characteristics because they look better, is not a bad thing if they serve the image. If you were after a hyper-realistic render to composite with a live action plate, then it would be a bad habit. Up or Toy Story, not so much.
I would point out that film uses all sorts of unrealistic practical lighting. A big one in the '50s and '60s was to shine a spotlight with an opaque gel and a little slit into the eyes of the actor or actress in close up shots to highlight the eyes. The good films, you don't hardly notice it, but with low budget films and TV shows of the era where quick light set-ups were required, the effect can be somewhat jarring. The original Star Trek is a good example of the not-so-subtle method.
Another prime example is night lighting. Masters of cinema such as Spielberg were guilty of this. Look at many films and TV shows shot in the '80s and '90s for this one: It's night in the forest. You see the actor stumble over the rise. How? Because of the bank of klieg lights hidden behind the rise and backlighting the actor. Many times there's no explanation or source of the light mentioned in the movie. Sometimes there's a reason, such as a UFO, truck or whatever, but I would bet money, in most cases it was just to service the look of the picture. Many lower budget creature features are good examples, as well as the X-Files TV show.
Now, I'm not saying that lazy lighting practices are a good idea, but some of the methods you mention I'm hesitant to call bad practices.
Again, it's just a difference of opinion. I respect the hell out of what you do, so please don't think otherwise.
Just a quickly thrown together image to demonstrate that you can do dark images and shadows with Linear Workflow. EP - I believe that Mac monitors are set to 1.8 gamma, so that value should work better for you.
Hey kids !!
Sorry to butt in…just stopping by to check whether there are any 8.5 updates or, more importantly, any betas for C9 which is expected apparently in a few months? (1Q 2014 ?).
Wow, the nonlinear discussion is now at 35 pages ??
I skimmed thru the last few pages and it seems like there are still some basic misunderstandings about what it’s all about, so I’ll put in my 2 cents to see if I can help to clarify it a bit. But then I promise I’ll butt out…
As I mentioned, I believe, long ago in this thread, linear workflow and gamma correction are what I call CALIBRATION SETTINGS. Either you need it or you don’t, based on whether or not each component of your scene and hardware and software are linear or nonlinear.
What does that mean? Well, let me give a simple example…
Let’s take JPEG images for example.
Now, if you start with an RGB, uncompressed image in which each pixel is described by an RGB value, and you want to compress it into a JPEG to make it a smaller file, how do you do that? Well, one of the ways is you take advantage of the way our eyes see things. And one of the interesting aspects of how our eyes work is that they are far more sensitive to low lighting levels than to high brightness levels. For example, when you’re in a very dark room, as you gradually increase the lighting level your eyes will notice the change. But if you’re in a very, very bright room, and change the lighting level by the same amount, your eyes won’t notice, and all they see is BRIGHT !!!
What does that mean? Well, it means that the RGB image you produced will look the same to you if you remove some data from the image which describes high brightness levels. And less data means a more compressed file. In effect, it’s almost like any parts of the image over a certain brightness are just described with one bit of data that says BRIGHT !!!
What you have just done is converted your linear RGB image into a nonlinear one. The data describing the high brightness levels is small, and the data describing the low brightness levels is great.
And that’s one reason why JPEG images are, by definition, nonlinear. There’s not a 1 to 1 relationship in the description of the brightness levels from black to white.
Now that’s not the best example, and of course there are a lot more things that JPEG compression does to an image, but I hope it makes the point.
And other components of your scene and hardware and software can be nonlinear (computer monitors, texture images, etc.). In your final render you should make sure that all of those components with any nonlinearity have been taken into consideration, and re-converted back to linear for the final image if they need to be AND IF YOU WANT THEM TO BE.
The point is this: you need to UNDERSTAND each component of your system, scene, hardware, and software, and how it handles nonlinearities so you KNOW whether you need to take any action to convert nonlinear components to linear ones (eg, implement a gamma correction). IT’S NOT ALWAYS RIGHT, AND IT’S NOT ALWAYS WRONG !!! But as Jeremy Birn says, you should implement a linear workflow, which means you should UNDERSTAND and CONSIDER nonlinearities and decide how and whether to deal with them at each step along the way. But that DOES NOT mean you always apply GC=2.2 in your Carrara renders. Not even close.
Now, until someone really understands how Carrara handles all the components of the scene and the renders with regards to nonlinearities/gamma, as well as the other components of your particular system, you’re just guessing. And to my knowledge there is no detailed description of how Carrara deals with those nonlinearities. On the other hand, with more professional apps, many of them go into great detail with explanations and options for dealing with nonlinearities throughout the workflow. Not so with Carrara. There’s only one Gamma Correction setting. And I have no clue what that does. Does anyone here know exactly what it does?
Now, I ASSUME that Carrara automatically applies gamma correction to JPEGS internally, in which case you wouldn’t need to do anything for gamma correction. But who knows?
Again, the point is that you need to UNDERSTAND each component of your system/workflow, and whether you need to apply any correction to that component. And until people truly understand nonlinearities and their particular workflow, this thread will probably continue for another 50 pages…
Okay, I’ll butt out now….
Huh. That's what Carrara is set to by default on my Mac. Macs really started the desktop printing revolution, and as such the Mac screen resolution and fonts were designed with print house requirements in mind. That's where Mac screens are 72 dpi as opposed to Windows which are 96 dpi or something. I know I had read up on it years ago.
Anyway, my point being, Macs were also very heavily into digital photo manipulation before MS based systems. Photoshop, AE, Flash, etc. were all originally Mac only or Mac first applications. Perhaps the Mac 1.8 Gamma is a holdover like the 72 dpi thing?
I wasn't trying to imply that you were being disingenuous or suggesting you violate copyrights by posting the entire text (although, that would save me a few dollars). ;-) It's just that I don't have the book, so I don't know the whole of the discussion.
As to using lights with unrealistic characteristics because they look better, is not a bad thing if they serve the image. If you were after a hyper-realistic render to composite with a live action plate, then it would be a bad habit. Up or Toy Story, not so much.
I would point out that film uses all sorts of unrealistic practical lighting. A big one in the '50s and '60s was to shine a spotlight with an opaque gel and a little slit into the eyes of the actor or actress in close up shots to highlight the eyes. The good films, you don't hardly notice it, but with low budget films and TV shows of the era where quick light set-ups were required, the effect can be somewhat jarring. The original Star Trek is a good example of the not-so-subtle method.
Another prime example is night lighting. Masters of cinema such as Spielberg were guilty of this. Look at many films and TV shows shot in the '80s and '90s for this one: It's night in the forest. You see the actor stumble over the rise. How? Because of the bank of klieg lights hidden behind the rise and backlighting the actor. Many times there's no explanation or source of the light mentioned in the movie. Sometimes there's a reason, such as a UFO, truck or whatever, but I would bet money, in most cases it was just to service the look of the picture. Many lower budget creature features are good examples, as well as the X-Files TV show.
Now, I'm not saying that lazy lighting practices are a good idea, but some of the methods you mention I'm hesitant to call bad practices.
Again, it's just a difference of opinion. I respect the hell out of what you do, so please don't think otherwise.
You are right about the "cheats" in live action lighting - heck, I know that a lot of so called night scenes were shot in daylight and then simply toned down and maybe given a blue color cast to give a night-time feel. And I agree that if the image looks good, that is the main thing. I have just personally noticed a huge jump in the quality and ease of putting together lighting since using gamma in my own images. If others choose not to use it, at least that is an informed choice rather than not knowing that the option exists. And I have a huge respect for you too!
Hi Joe, I hope you are well.
I totally support that you should ideally understand what you are doing with Linear Workflow. What I would add is that most image formats designed for display on a screen, and that includes images from a digital camera, most internet images etc, are designed to be non-linear, and that includes JPG, BMP, PNG etc. The main exception that I am aware of are HDR image formats that are by definition linear. When I say "designed to be non-linear" I mean that they have a gamma of 2.2 built into them, according to a standard called sRGB which was agreed across the makers of monitors, cameras, TVs and software so that images can be captured, shared and displayed consistently.
When using 3D software, the calculations all happen in linear space - twice the value equates to twice the brightness. So what you need to do is de-gamma any images (and even simple colours if they were selected from a screen display!) before the calculations take place, and apply a gamma after the calculations in order to correctly display the results. I am as sure as I can be that this is what Carrara is doing when you use a gamma value.
It should not be applied to images that are used for bump maps, displacement, normal maps and also if the image is an inherently linear format such as HDRIs. And we know that Carrara does not apply correction to backgrounds, backdrops and hair color so this needs to be done manually - maybe Carrara 9 will include gamma correction as a factor that can be interactively applied to all images.
But I firmly believe that if you are after a realistic result, you should always use the gamma correction in Carrara. If nothing else, it makes your lighting behave more like the real world. Below is an example - no GI, no images maps or anything clever in this, it is just two spotlights overlapping on a grey plane. What should happen is that where they overlap, it should be twice as bright as the areas where only one spot is shining. I think it is clear that without gamma, the overlap area is much more than twice as bright (and theory suggests it is over four times as bright), while in the gamma corrected image, the lighting adds in a much more natural looking way.
It is this kind of effect that makes lighting without gamma so difficult, especially in interiors.
thanks for the explanation joe, I did a brief search yesterday 0n linear and nonlinear workflow (as I have recently bought a hdr prog oloneo photoengine - it was nothing to do with this gamma thread) and the people explaining it didn't seem to have a clue .... your explanation was very good so thanks for that. It will be interesting to see how that relates to getting our renders out in print.
philw, great and simple visual explanation. When I first used the gamma thing that was what I noticed - that my highlights were less blown out while I was getting more details in the shadows .
now I understand why.
cheers from here :)
Yes, if you think of classic portrait lighting, you have a key light that provides the main illumination, if you then add a fill light that partially overlaps where the key light is shining (so a similar situation to the pair of spotlights above), without gamma it will light up the already lit area far more than the area in shadow, which is what you wanted that light to do. No wonder lighting has been so difficult in the past! This is resolved with gamma turned on so that the lighting is far more natural and realistic (and predictable, which is also an important factor).
last year I received a few responses from publishers saying my work looked very cg
(they didn't say that was bad but that is what I interpreted it as)
so hopefully this will help.
thanks again for bringing this up Philw
headwax,
That's an odd comment from publishers, as I would say your work looks less CG than a lot of what I see. Its very whimsical, and I do not see many artists doing 3d work in that vein.
Your work has always looked very painterly to me, so perhaps it is in finding ways to refine your technique that will help your work stand out a bit more, as if it does not already.
headwax,
That's an odd comment from publishers, as I would say your work looks less CG than a lot of what I see. Its very whimsical, and I do not see many artists doing 3d work in that vein.
Your work has always looked very painterly to me, so perhaps it is in finding ways to refine your technique that will help your work stand out a bit more, as if it does not already.
thank you cdordoni, I really appreciate that kindness! cheers :)
with the 2.2 gamma correction I've been finding my carrara hair blown out - ie very highlighted - reducing it to 1.8 correction almost fixes this, any other ideas? I've dropped highlight to nil in the shader as well . This whole linear thing is very interesting. I have oloneo hdr software will also compress tonal range and bring back detail to highlights. Highly recommend it.
thank you cdordoni, I really appreciate that kindness! cheers :)
with the 2.2 gamma correction I've been finding my carrara hair blown out - ie very highlighted - reducing it to 1.8 correction almost fixes this, any other ideas? I've dropped highlight to nil in the shader as well . This whole linear thing is very interesting. I have oloneo hdr software will also compress tonal range and bring back detail to highlights. Highly recommend it.
Yes, the hair shader seems to be a noted exception in that it does not automatically adjust for the gamma correction in the way that normal shaders do. I have found that applying a multiplier of around 40% to the color works OK to bring the tone back down.
thanks PhilW, appreciate your advice -will try that :) cheers!
Just a bit of experimentation, and I'm still working out her shaders for this, but here I am lighting Rosie 4.2 using nothing more than a bi-gradient in the background, GI and IL on. The first one uses a bi-gradient backdrop to block the view of the bi-gradient in the background. The second is the same scene, no backdrop. The third is the same as the second, but with a darker bi-gradient BG.
velly interesting, thank for sharing that
one thing that draws my eye is the shadow at the décolletage.
how it is as not as dark as that at the recesses of the neck (first two images) - almost as if light is going through the fabric at the chest
so do I understand correctly that bi-gradient in the background is contributing to the lighting because you have GI and IL on?
Never play with it (render times?) but it makes a good effect
If you have Sky Light turned on, whatever you have in the Background will be used as an environmental lighting source, which could be a flat colour, a bi-gradient, an image or an HDRI. Realistic Sky will also contribute to the lighting. The lighting from some of these can end up looking a little flat, as effectively the same light is coming from all directions, but it is heaps better than using Ambient Light, as it produces correct shadowing in dips and creases of your model. I think the best use is to use this in conjunction with some direct lighting so that the form is brought out more.
I absolutely agree, Phil. I wouldn't want to rely entire on a bi-gradient. Too huge of scale - the whole world boiling down to four colors, after all. But, like HW, I have never really tried this type of lighting solution. Not really solution, enhancement. So I wanted to try it out. I was working on this character, and she's trying to go linear. Her hair continues to improve, I think, from what I've had before, and I'm really liking what global illumination can do for shaders. Now I want to try this out with more of a complete scene and see if it taxes the render time to a point that I won't like it (?) because it doesn't seem slow at all with the figure alone, compared to when I had a special light rig on her.
Of course, I won't be over doing it with a bi-gradient, since there will be scenery in the way. Outdoor scenes can use the atmosphere and sky.
But I did forget to add in the original post above, that I do have a single spot light in that scene a foot and a half away from her face at a brightness of 11 set for forty feet with a range falloff of 90. Without that, the whites of the eyes weren't getting light. Now, being a full, nice sleep later, I know why. I would need to turn on Light through Transparency in the GI settings. How embarrassing! I spent a long time working on the shader for the sclera to try and get then from being black. Adding glow does something that I just don't like - which removes any of the depth you might get from real lighting, highlights become wrecked...
I want to see how far away from cheating (with lighting) I can get, while still holding on to my Carrara-Fast rendering. Going linear, this would be the perfect time for me to implement such a thing. So I can use this method across the whole production. Something really cool too, if we want to, we could actually use our 3d paint talents to paint a lighting solution onto a spherically UV mapped sphere and use the resulting image to light the scene with in the background channel. You may not see it very well, but in images 1 and 2 above, the ceiling color of the gradient is red, the floor's center is light bluish green - both ends fading to that near white gray. Additionally, Holly Wetcircuit - (3D.Wetcircuit.com) has an excellent tutorial for IBL (image base lighting).
So I want to put some of these ideas to the test and see what kind of animation performance I can get from Carrara compared to the ways I've grown to use previously, which is this: I would light the scenery items how I want the scenery to look. If elements of the scenery appear as if they should cause illumination, I would estimate how much and what color, and add the lights in conjunction with the glowing shaders. If my characters needed a boost, I'd boost them. It became apparent to me that I always wanted to highlight my main characters, not so much the background clutter people. So I'd attach a highlighting rig directly to each character. I am certainly not afraid to use this method still - as I really like how easy it is for me to see how I'm lighting things in Carrara. It's my favorite software ever for that. I must admit that I have yet to get 'expected' results from shape and tube lights. But I think that is just me - and what I 'expect'. So I just use spot lights and adjust their angular and distance falloffs to blend them into shapes or lines (tube) of light as I need them - as the spot lights can be very predictable.
Design Acrobat was just mentioning the importance of IES lighting (and included links in his post), which is a whole range of lighting solutions I've never even tried - not once!
Using GI with IL, however, I now want to see what happens with glowing portions of objects as far as lighting. Aura can be a fun effect to use. I know that it has a bad rep with some of you, but when used subtly, it can really add to your glowing thingies. However, if the glow is just a portion of the mesh, the aura will actually get placed around the mesh - not just the glowing domain. For lighting though, I doubt this would be a huge issue because the light, itself, is not visible, just its affect on nearby objects.
We'll see.
Good luck with your experiments! I'm with you on shape lights, they do not behave as I would expect them too, so you are not alone in that. As you know, I have taken to using glowing objects as very controllable soft lights with GI turned on. I am amazed at how quickly some of my renders with full GI are these days, I was playing with lighting a figure earlier and it was rendering in around 30 secs with full GI.
Just curious about the cpu you are using or you access a renderfarm?