General Retopology question
I'm terribly sorry for posting an ignorant question such as this, but I really wanted the opinion of experienced modelers.
I'm essentially creating a model of a human figure that is to be used purely for illustrative purposes. Meaning, I'm not using this model for games, or even for animation.
I've sculpted this figure in Zbrush and it's 4 million + Polys, I want to take one of the lower subdivisions, export it, UV map it, and bake the high rez detail on it as a normal map, then proceed to texture and beautify it. I do plan on somehow rigging this (I actually don't know how to rig or pose a model) so I can pose it and put it on hand painted backgrounds, that way I can make posters and canvas prints out of my 3D character I've modeled...
Okay so... knowing what I am doing and how I am going about this, do I even need to Retopologize this model? When not working with a limited poly video game model, is there any reason to retop a Zbrush sculpture?
Thank you so much for your help. Polycount has been so nice to me and has been a life saver thus far.
I'm essentially creating a model of a human figure that is to be used purely for illustrative purposes. Meaning, I'm not using this model for games, or even for animation.
I've sculpted this figure in Zbrush and it's 4 million + Polys, I want to take one of the lower subdivisions, export it, UV map it, and bake the high rez detail on it as a normal map, then proceed to texture and beautify it. I do plan on somehow rigging this (I actually don't know how to rig or pose a model) so I can pose it and put it on hand painted backgrounds, that way I can make posters and canvas prints out of my 3D character I've modeled...
Okay so... knowing what I am doing and how I am going about this, do I even need to Retopologize this model? When not working with a limited poly video game model, is there any reason to retop a Zbrush sculpture?
Thank you so much for your help. Polycount has been so nice to me and has been a life saver thus far.
Replies
I personally find it easier to work with the features of a human face if it has a nice edgeflow to it. The bigger reason to consider retopology for illustration would be to maintain a consistent density. It can help with sculpting, but it'll also help make sure you're not wasting exponential amounts of polygons in areas that aren't benefiting from it. That may not even be a problem if you're happy with the model at 4 million though (zbrush can more than handle that amount even on old machines, and if you do wish to take it to an external program for baking, the decimation master plugin can hack it down to a lighter size).
Rigging would be something beneficial to dabble with if you ever do plan on making characters for games or animation, but if you want to stick purely with the illustration side you might want to look into using Transpose in zbrush, or the zspheres rig. Then you can quickly pose a model, do a quick pass to sculpt in any changes you might want to make for that pose (folds, muscles and skin stretching or bunching, etc), and render away.
But I must ask, if I go this route, will it still be possible for me to beautify my model with various maps? Like texture maps for skin detail and such?
Or rather... Considering my lack of knowledge, maybe the better question would be, if I render in Zbrush, what would be the best way to go about further beautifying my model with skin detail and color and such...
thanks!
Generally for texture maps the only ones zbrush makes use of are the diffuse and displacement maps. And even then you're probably just better off sticking with raw polygons and polypaint if you're just aiming for a pretty render. Zbrush relies more on having the detail in the geometry itself along with the different materials/shaders blending together to create the kind of surface information that normally gets faked and baked.
One handy thing with BPR is that is also breaks down the render into passes, which can give you seperate images for things like AO, depth, SSS, and shadows incase you wanted to composite the result in another program.
This sounds like a MUCH more sound alternative for my purposes. And it is also much more artist friendly, which gets a high score in my book.
Thanks once again! You have been a great help.
Pixologic certainly keeps expanding on it, but ultimately you still have big shortcomings that fall on your shoulders, as the artist, to fix. It's certainly possible to still get decent results, but chances are you'll have to paint and composite your way out of lighting, shading, aliasing, and/or color space issues.
Either way I have both Maya 2012 and Zbrush 4R2 now. So I will try the new Zbrush I just upgraded to and see if it suits me at all.
But thank you Crispy4004 for your input! I will keep it in mind as I continue to explore and research my own personal workflow.
But take for instance Marco Menco's Beta Examples. Now look at his Vray results. You'd have to be a glutton for punishment to attempt matching those results with Zbrush BPR. I see nothing wrong with using it at say the pre-vis stage, think it's well suited there, but any further and you'll likely be wasting time fighting the pitfalls. After all, many of the most convincing examples still look distinctively like a product of Zbrush's rendering. People are naturally going to push it, but not because the results are worth it. It'll be either because they like the convinence of staying in Zbrush, view it as a challenge, or more realistically, want to get around learning the nuts and bolts of a more capable render engine.
Marco modeled it, someone else textured and rendered it. It's entirely possible the latter was more familiar with rendering in Vray than Marco was at using features that were currently in the beta cycle of their life span? When you go through some of the other galleries (beta or existing), the results that can be achieved with a few passes are mind blowing and hard to tell how it was rendered unless they say so.
Perhaps the real problem is a matter of preference. If you want or need physically accurate lighting in a 3d scene, then yeah, render it somewhere else. If you just want a cool looking image of your model to composite, zbrush can more than deliver.
Anyways, the aliasing, materials, and lighting inconsistencies are usually dead giveaways that something was rendered in Zbrush, that's absolutely still true of the beta images. If they get close it usually says a lot more about the artist's post skills than what BPR is capable of as a render solution. Passes in of themselves are not a surefire recipe for success. When your materials are not energy conserving, your reflections and glossiness are matcaps, your GI faked, your lights lack proper attenuation, and worst of all, your color space disregarded, they can only take you so far. It becomes an unnecessary balancing act to try to fix. Great exercise for your artistic eye, but not exactly practical.
I don't think it's a matter of preference so much as what makes more logical sense. Need to present a concept quickly? Use BPR, it's quite powerful for that. Want to get a high quality awesome render in less time and effort? Go with something else. BPR is just too limited and cuts too many corners to be considered a fully capable alternative.
I don't think anyone is arguing that render engines like VRay or Mental ray aren't more advanced or more physically accurate, the issue is are such things needed to achieve the desired visual result. Because if he's piping his work over to another app in order to set up the camera, lights, materials, passes, and whatever optimization he needs to make to the 4mil mesh to get it over, and the result for all his time is just a few hairs difference, it might not make the most logical sense depending on what he is aiming to achieve.
You're not quite understanding my argument here. Matcaps, lightcaps, and everything else BPR cuts corners with, are not adequate alternatives to the real deal. BPR is absolutely phenomenal if you need quick mock-ups, but we are not talking a few hairs difference here. There is a massive gap in nearly every aspect of the rendering that stacks the odds against you with BPR. It's a completely detached method, in which many standard lighting, shading, and linear workflow relationships are broken. That's a very challenging environment to work with. Trying to fake your way to the same quality results is going to cost you more time in the long run, not save on it. It's called "Best Preview Render" for a reason. That's what it's good for.
I get your argument, I just don't see why you insist on repeating so much as I've yet to suggest BPR is more superior to MentalRay/VRay/etc. I've even been agreeing with you for the most part. But what hasn't seemed to click yet is that the end goal is to achieve the final result that the artist wants for the image, which is something entirely up to the artist. Some artists might want the results that could only come from knowing how to use a fancy software renderer, some get the results they're looking for with photoshop and zbrush, others even like to use game engines/real-time shaders to get the look they want instead. It's up to the artist to decide what works for them. Yes, matcaps and the like cut corners over more physically accurate rendering solutions. You've said it, I've said it, you're still repeating it for some reason and probably will again at this rate. It doesn't change anything. BPR is still capable of phenomenal results, and if it happens to accomplish the desired goals of the artist in a time-efficient manner, then it's simply the logical choice for that situation.
The guy intends to handpaint a background too, should we warn him that photoshop's paintbrush doesn't have real GI either?
Just show your stuff here, you'll get more accurate advice.
Look cryrid, you haven't suggested BPR is better just like I haven't suggested it cannot still be used to produce stunning art. I 100% agree good work ultimately comes down to the artist. After all, there are impressive examples out there with Max Scanline and Maya Software, despite some similar setbacks. Our disagreement here though is purely based on how practical BPR is as a rendering tool, not the merits of the artwork being created with it. With regard to game engines, someone could just as easily make a similar case for using Marmoset, Crytek, UDK, etc. over BPR to present game models, and not just because they're in game.
Last, comping a render with some hand painted backgrounds does not mean nothing is accurate so anything goes. You would still ideally want the best results to work off of. The only exception is if you are taking a diverging NPR approach like your Olivier Thill example.
The guy wants to create an illustration using a model, there are a near infinite number of approaches to be had and it all comes back to artistic merits. The practical choice is always going to be the one that creates the result that the artist is looking for for that particular piece, you still don't seem to be getting that seeing how you think it is now marmoset vs bpr.
Again, the artist is free to use whatever tool he wants, but the situations where BPR is the best and even easiest choice for that final shot are quite limited. After all, one of 4leaf's concerns was Zbrush's lackluster support for maps. BPR just sort of sits in no-man's-land. It's almost a little surprising that Pixologic hasn't switched to rendering using real-time methods, given BPR is hitting closer to those results in more time.
I would like to start off by stating the fact that my 3D skillset leaves much to be Desired. I know how to get around Maya on a sub-Professional level, and I know and have worked through the general pipelines for making video game models buffed with maps to make them look better.
My love is for traditional art. I've spent many days gesture drawing figures at a local art cafe, and working with traditional clay. When I'm working with Zbrush, It feels most like what I know and what I'm good at.
The second that I leave Zbrush with whatever I create in it, I have to spend hours re-topping, UV mapping, and doing things I absolutely despise in programs I hate with every fiber of my being. it's frustrating, it's tedious, and those aspects to modeling are aspects I hate. (As nessesary as they are for many fields, they simply aren't my cup of tea)
Zbrush's system of rendering is a cleaver way to fake the more powerful and realistic scenes that you can get from other programs... But frankly I don't want to have to deal with the aspects involved to get my 7 Million poly model over the those programs and out of Zbrush. (Btw... The only other program I own is Maya 2012... and god does 2012 suck ass...)
So yeah... call it stubbornness or whatever, I simply am willing to go many many many extra miles if it means staying in zbrush. It's just more convenient and preferable for me personally.
Also, my clientele aren't CEO level employees of large animation studios... They are niche groups of average people who wouldn't likely notice the differences in rendering.
Anyways, thanks for all of the advice, as it has helped me so much. If I ever do feel the need for more powerful and accurate results, I will be sure to drop my stigmas to go for the more powerful alternatives.
Many of your concerns here are actually not as big of a deal as you might think. Getting a better render inside of Maya isn't terribly difficult or time consuming, especially with Decimation Master, UV Master, and/or GtoZ making the transition easy from Zbrush, but the issue is of course you have to know what you're doing in Maya. Mental Ray can definitely be a bit overwhelming will take plenty of research, but the payoff is worth it.
To put it in a traditional art perspective, if plein air painting bores you to tears, you are likely are not going to have the drive to learn the intricacies of Mental Ray. Of course though, good rendering, like good plein air painting, is an absolutely killer skill to have. What I would do in your situation is go ahead and see what you can do with Zbrush, but dedicate some time to on the side to learning Mental Ray. I guarantee in time Zbrush BPR will actually become the more time consuming and frustrating process. Especially as you start exploring lighting and shading as a means of creative expression.
And I have never actually used Decimation master. I have a little experience with Mental Ray rendering... I used it when I was using the "Making of Varga" tutorial to help me along... but honestly, I just followed the tutorial mindlessly, Mental ray was a scary scary endeavor.
In time I'll definitely look up more on Mental ray and such...
And I actually do enjoy Plein air Painting. Except when It rains on me...
My workflow these days is to unwrap it with this, run it through UV Master in Zbrush using existing seams to further optomize, then do some final touchup back in Maya. Before I was using Headus UV layout, which is also a great application, but I found it a bit unstable on my machine and UV master seems to do a pretty good job optimizing.
Decimation Master is a great tool. It reduces the polycount while doing an excellent job of retaining the details. Perfect for bringing it into Maya for a render when the topology is not important.