@EQ,
Thanks for the clarification
so one smoothing group should be ok for that mesh?
the picture above is from the viewport, though the render displays similar results.
Should the normal map compensate for these errors 100% in the viewport and the renderer?
I am going to make some generalizations here, so bear with. I'm sure there are counter examples, but I'm looking more at trends than specifics.
I think that the point that is often missed is that philosophically, from a design standpoint, Blender is very different from Maya and Max (I assume, as I use Maya and have only a passing familiarity with Max). Because of this, evaluating the packages on a tool-by-tool level is largely counter-productive. Part of Blender's hard to learn, easy to use paradigm comes from it's design as a series of tightly integrated systems. This is very different from Maya at least which seems to have many very loosely connected systems from the way in which things have been added over the years (at least on the UI level, I understand the flexibility of the nodes). Because of this, it is very easy to learn any one of Maya's parts, UVs, Modeling, Animation, etc, but each part must be learned individually because not a lot is shared between the components. On the other hand, while it is difficult to get into Blender initially, a large amount of the tools, interface, and conventions, etc. transfer seamlessly between different modes. If you know how to model in Blender, you probably know 90% of the entire program, because it's very easy to then move to UVs, or animation from there. Systems like vertex groups (which control anything from skin weights, to hair stiffness and length, to modifier values, physics properties, particle properties, etc.) are especially relevant to this example.
Because of this, what it really boils down to is where you want or need to invest your time. Blender has a large up-front cost that pays out over the long term, whereas some of the more traditional packages have a very low initial cost, but you need to continue to learn other systems within the package over the long term. Deciding whether to learn Blender should really be based on considering which of these philosophies best suits your needs (along with the other more concrete necessities, $, tools, etc.)
As was mentioned earlier in this thread, the developers of Blender have realized that ease of use and easy to learn are not mutually exclusive, and all of the work over the past 3 years of blender development has been aimed at fixing that. All of this said, Blender has quite a bit of work to do with many of it's tools and lower level stuff, but philosophically, it is on solid ground.
On a more personal level (read: personal preference), I would argue that Blender is a more artist friendly tool than Maya, and, from what I've seen, Max. I mean this in the way that, when modeling in Blender, I have a much more tactile and direct connection to my model than I do when working with Maya. Within Blender, I model almost exclusively in a single, full-screened perspective view (which is even easier now that I can tear stuff off and throw it onto my second monitor), and, using almost exclusively short-cut keys, I can pretty much pretend there is no interface. It's fast, but more importantly, it feels good in the way that sculpting feels good. Just unobstructed and direct. When I think about how much time I spend in these programs, and how much more time I will spend, this almost becomes a quality of life issue. Where am I most comfortable.
Don't get me wrong, sometimes within Maya I can work in a similar way, but too often it feels clumsy and frustrating.
@EQ,
Thanks for the clarification
so one smoothing group should be ok for that mesh?
the picture above is from the viewport, though the render displays similar results.
Should the normal map compensate for these errors 100% in the viewport and the renderer?
I will post the .blend soon
Check the maya examples in that thread, and try the same mesh with 3point Shader + Quality normals in max.
You'll see the results you get when everything is synced up correctly. In maya there was still some artifacts on the bottom, it truly is a "worst case" mesh, but you can get much better results when its being done correctly.
Ah yes, in which case my advice is also not entirely correct. The need for multiple smoothing groups for clean bakes has --through experience-- become so lodged in my brain, I was highly skeptical of anything else working.
That being said, unless you're lucky enough to work with an engine that uses 3ps' Quality Normals, you're still stuck with 'the old way'.
Any engines that already do, by the way? I imagine you're working with your clients to implement this, and I'm looking forward to seeing it in something I get to work with.
Anyway, just breaking the normals along all uv-splits goes a long way, and doesn't actually impact your workflow at all, unlike the other way around - breaking apart your uv's at all normal breaks, which makes texturing a bitch. That is, assuming you're breaking normals at every radical change in angle, which does have the added benefit of giving you a bake that's slightly easier to work with because you don't have the gradients on planar faces.
edit: I'm talking about non-organic assets, ofcourse. Hence the 'radical changes in angle', which you tend to get on mechanical stuff.
Ah yes, in which case my advice is also not entirely correct. The need for multiple smoothing groups for clean bakes has --through experience-- become so lodged in my brain, I was highly skeptical of anything else working.
That being said, unless you're lucky enough to work with an engine that uses 3ps' Quality Normals, you're still stuck with 'the old way'.
Any engines that already do, by the way? I imagine you're working with your clients to implement this, and I'm looking forward to seeing it in something I get to work with.
Anyway, just breaking the normals along all uv-splits goes a long way, and doesn't actually impact your workflow at all, unlike the other way around - breaking apart your uv's at all normal breaks, which makes texturing a bitch. That is, assuming you're breaking normals at every radical change in angle, which does have the added benefit of giving you a bake that's slightly easier to work with because you don't have the gradients on planar faces.
edit: I'm talking about non-organic assets, ofcourse. Hence the 'radical changes in angle', which you tend to get on mechanical stuff.
Any advice that is given freely on this forum by people with much greater experience than myself is most humbly received and appreciated.
Its all good to know and it really helps clear things up and stops be second guessing myself.
Sometimes its hard to filter the good info from the bad as it is so intermingled but i can always count on PC to come up with the goods eventually
I'm Blenders 'release manager'. I'd like to address a couple of general comments that have been made in the thread.
First off, the artists will be pleased to hear the positive comments on the trailer. There is a lot of work yet to be done on every shot, and improvements that will be made in simulation (the team feels a lot of improvement needs to be done on the hair solver yet), tweaks, for textures, lighting, animation, etc. yet to go.
Regarding Blender and its capabilities.
Like all major 3D packages we have our strengths and weaknesses.
We have long been aware of professional users of other packages frustration with our user interface, and non standard navigation and short cuts. Unfortunately to make the changes for custom shortcuts and easier UI customization required a massive effort - about two full developer years on the front end for our two most experienced developers. Thus it just wasn't possible to make those changes until relatively recently.
I do hope once our next release or so comes out you will give the latest Blender version a try. There will be a huge amount of written and video documentation that will come out within a few months after the release. Also there will be custom keybindings presets for Maya and other 3D applications.
I would encourage all of you to stop by our booth at Siggraph this year. We will be doing demos the entire time, and I'm fairly confident you will be blown away.
metalliandy.
as to the 'need to focus on the core tools' - be aware that a lot of things like smoke, fluids, and sculpt are done by folks who have zero interest in working on anything but those tools. So advancements in those areas generally have no impact on the rate of development in other areas.
Of course others would also argue with you on what 'core tools' to focus on. Personally my own interest is getting our sculpt, retopology, and paint tools competitive with the market leaders. Although I do think getting our mesh tool upgrade completed (probably will happen after Sintel is completed) and fixing weaknesses in our UV tools is also highly important.
I'll pass your list of suggestions onto the dev team regarding UV and mesh tools (mesh tools issues are waiting on our major overhaul being ready 'bmesh').
pior,
regarding setting up Blender as a dedicate modeler/sculpting tool/video editor/video compositior/uv unwrapping/texturing/rigging/animating etc tool - the interface is now about 99% defined in python. Some time after the next release or two I'll be working on dedicated simplified interfaces.
Regarding not enough high profile quality output from Blender - definitely agreed. The quality work is being done, it just doesn't get posted anywhere except for one of the hundered of Blender user forums around the net. This is a problem we are aware of - and will get it taken care of probably after this summer.
Mightypea,
ior: There's a thread on the blenderartists forum where someone is setting up trades between artists and coders, where an artist can do art for a coders project and the coder will code in return.
The idea is mine and is called 'art for code' - the website for it is being set up now by BTolputt. Within about a month it should be ready to go.
Excellent news, I'll be looking forward to that website.
Personally, while I'm mostly interested in seeing the modeling tools improve, I think a very good implementation of retopology could lure professionals to Blender very easily. We already use topogun/3dcoat/etc. for just this purpose, and if Blender could do it equally well that could make it part of every modeller's toolset.
Glad to see you on this forum, by the way. For us to register at BlenderArtists (where there can at times be an aggresive attitude to any mention of the commercial software we're used to) is a bigger step than just talking about it here, with the people we know.
That, I think the problem has been its zbrush'y way of doing things, meaning quite different, so users of other apps or such will never really catch on to that whole thing (although its changing.
However the modelling tools and uv-editor has kicked ass for quite some time while I'm still waiting for the big apps to do something about their ca. 1994 uv-editors.
I've been doing published game art for quite some years in blender now, so everytime I see someone say "It's not quite there yet" I just facepalm and wonder if I've been living in a dream world where my art never actually happened.
It's all about the artist, and the blender community might have been lacking a bit in the grand master artist per user ratio, (except for belias, that guy works magic with the canvas)
I'm really determined to get myself used to it but as you said, the interface is so.... abstract...
But, it's scriptable and open and all that, and getting more powerful. I hope one day i can make full current gen pieces in blender without feeling awkward
Thanks for coming here to polycount tom, its great hearing feedback and opinions from people that are actually involved with the softwares direction!
One thing that i had a little trouble with the other day was that i made a mesh to export to another program, but i had a problem somewhere on the mesh where i had some random face connected to the wrong edge making 3 faces coming off one edge.
Does anyone have a good way of locating such problems quickly or easily? I resorted taking it to max just to fix the small problem .
Problem with the blender artist community and lack of good artists is mainly blenderartists.org, that place is a shut in community where few really develop, and people sprinkle around false reinforcing comments and feedback in threads.
It even got to the point where a special forum part called "focused critique" had to be created back in the days because people got butthurt over any critique.
My experience is that people who have come from there and gets to polycount will start learning way faster and grow in ways they never did before, (I know I did).
That place needs to be way harsher, since it's almost the official place for blenderartists to go.
Example (and nothing personal against the artist):
Thanks for taking the time to post LetterRip
I really appreciate that the devs of the smoke and physics are separate from the other tools and think they all do a great job at what they do...it just seems there is much more interest in developing those sort of tools than everything else as last time that a major modelling tools overhaul happened was so long ago (2.35?) but im hoping that BMesh will fix that issue.
I really think that Blender is a fantastic tool, but it just needs a little refining.
Going forward it will be easier for users of other app to use Blender but until camp Blender is torn down, most people probably won't cross pollinate. It's like wearing an ascot, walking into a biker bar and ordering a martini with a twist of zucchini. It's probably easier to just to go a martini bar.
Some of the best forums (and apps) I've seen have a wide diversity of users and suggestions come from all angles. Forums where users are free to say "here's how I would do it in XX app..." and it isn't met with scorn but inquisitive appreciation. It often sparks new workflows and tools to be written. You end up with the best possible way to do something being propagated to all the apps instead of "the only way to do it in XX app" which might be a crappy work around at best but because no one knows of a better way its the only way that's ever suggested. Silly and unhelpful but allows for a constant positive feedback loop.
I'll generalize and I'm sure I'll get some flak for it especially now that the blender crowd is gathering pitch forks and torches in hand.
- Artists post because they want to improve.
- Blender artists post because they want someone to say something nice about them. Read that as: They personalize their work so much that a knock against their work is a personal attack, not helpful advice.
The post that eld linked to happens many times over. A lot of people falling down, tripping over themselves to give praise to what desperately needs a critical eye. They shower praise because that's what they want in return.
How can I trust what they say when all I ever see them do, is praise junk? Why would I turn to them for advice or even think to post things there? More importantly how can anyone ever really get better if everyone is too busy blowing smoke up each others asses? There is a nice way to give critiques, but even that seems to be bad form. Those that attempt it are drown out or shushed. No thanks...
I don't think you'll see any of that change Vig. Pretty sure the blender community is larger than any other user base, and well, there is a certain stigma to 'free' that will never go away. I really don't get it though. For all the anti-blender talk I see regarding the app and its community, there's just as much ignorance and attitude on the commercial side of the fence. I've always been on that side, but not with the attitude. I've found many useful tools and workflows in blender and very much look forward to seeing it grow. The new UI changes in 2.5 are far more forward thinking than anything I've seen in maya, max or xsi in recent years. Just speaking objectively, as someone who chooses to look past opinions, affiliations or financial investment in a product and try things for myself before making my own judgements.
As far as their community goes, I don't see it as any different than commercial software forums. People posting work less for crits than just positive feedback and a general banding together and forming of their community around a common interest. This happens in all the software forums I've visited over the years - just not on nearly as large a scale, since the blender community is massive. It's also why I don't really take part in the communal back-patting in places like that. I'm not there to make friends and feel good about my software choices or have someone tell me my shitty art is good. You still have to come to places like polycount for any real critique.
@Vig
I use Blender as my main application so i could be counted in that demographic but i certainly don't consider myself to be a "Blender Artist" in the way you describe it...i post because i want to get better at what i do.
I think perhaps its more likely that many Blender artists are just starting up and therefore what looks good to them isn't what looks good to more experienced artists.
[sweeping generalisation]It is true however that the Blender community (as any community that is dedicated to a single application) can be very incestuous and perhaps a little bigoted with a massive, almost fanatical bias toward Blender and everything open source and are sometimes seemingly unable to receive any criticism for the way Blender does a certain thing.[/sweeping generalisation]
If i remember correctly LetterRip did a nice presentation on being an effective Blender advocate at the last Blender conference
I have gotten flamed upon many occasion by suggesting that perhaps something that Max (or any of the big apps) does, is a better way or should be integrated in some fashion into Blender.
One example i can think of currently is i posted on BA suggesting that there should be an option in the colour picker for RGB values of 0-255, rather than using 0-1 as i was tired of copy and pasting Hex codes to get an exact colour.
Needless to say it was frowned upon.
Personally i feel that if another application does something better than im all for it and if i can afford the program i have no problems making the purchase.
Day to day i use Blender, Silo, Zbrush, Unwrap 3d Pro, MeshLab, Polygon Cruncher, FAOGen, xNormal, Milkshape (exporting only :P) and CS3 (among others im sure that im forgetting) and though i would love to use Max, the cost is unfortunately just not justifiable or affordable (£3k ($4.3k)).
I hope that the Blender community does change for the better but it will take a good amount of time i think.
Vig: That's close to what I brought up to the BlenderArtists webmaster last Blender Conference. I asked him what the purpose was of a software specific forum, since if I create a model, I'll want feedback from Max and Maya users as well, especially since many Blender people are new, inexperienced, and hobbyists. The answer that I got was that the good people will migrate from the BA forums, where they can get better criticism, while new people would find the forums handy because they can get help with the problems they have with the application.
Whether that's a lame excuse or downright admittance of the forums' poor quality, I'll leave up to you. I posted artwork there once, got the comment that I shouldn't be using triangles for game meshes, and stayed with Polycount from then on.
(And so it might also well be possible that the better artists don't like visiting the BA forums either, skewing the perspective even further...)
Eld: Oh, my comment was mostly spurred on by Tom saying that was an area he wanted to focus on. I only wanted to say that, while to me personally modeling tools are most important, any work on retopologising could make Blender a part of our workflow. But no, I hadn't yet tried it, although I must say that --if I'd use seperate retopology software at all--, it'd be Silo, which I already own. As soon as I can configure Blender to be more like Silo I'll be taking more regular looks at it, though. Silo is really the best example of how to 'get it right', workflow-wise, as far as I'm concerned. It's only too bad they're stopping development on it, as it has so much room to grow, still.
Speaking of configuring it that way, is selection highlighting an option in Blender?
Anyway, it's amusing you'd post that gun, as I've been making fun of the hype around that thing ('I thought it was real!' 'amazing!') for a while now. It's a prime example of what's wrong with that community.
Silo is really the best example of how to 'get it right', workflow-wise, as far as I'm concerned. It's only too bad they're stopping development on it, as it has so much room to grow
what maddens me was they had really major names in modeling behind them. then they just kind of lost heart. I've been watching Blender really closely for a while now waiting for the feature freeze, and B mesh to see if writing up some tools like silos would be practical.
Most of the stuff in silo was about the tool logic. The cut tool for instance is awesome, it cuts connects and divides in 1 context sensitive keypress. To maintain my sanity moving from silo to max i wrote my own silo cut, break and merge in a max script. the difference it makes is nuts.
...It's a prime example of what's wrong with that community.
Because Blender is free the community is probably 90% kids, noobs and complete amateurs so is it really that surprising that the signal to noise ratio is so out of whack. You also have to figure in that people don't want to know "why" any more, they just want to know "now", synonymous with the "make art" button, something that always prevalent where there's no initial investment involved.
metalliandy: No, the only official word they've given is that they are focusing on other projects in addition to silo as well as using it themselves for asset production in those projects. Along this line they are still committed to it. For some reason people equate this to no development at all. It's not the same level of commitment and frequent updates we saw during the 1.x cycle, so I'm sure it's disappointed a few who either missed that boat and bought in after it had sailed or who simply got used to the way it was. They've released 2 updates in the last 6 months or so however, both of which improved the software in my daily use.
My take on it all is that Silo still rocks, even with slowed development. Considering how far other apps have come in the last few years with much bigger teams and budgets to put into development I still prefer to use Silo, which makes me a bit more disappointed in the expensive modeling tools I've paid for and hardly use. I'm actually more disappointed with the developers of those tools and their inability to surpass or at least allow me to reproduce the workflow I have in Silo.
One thing that excites me about blender is that with the new customizable UI and bmesh integration, blender will probably allow me to come the closest to reproducing my silo experience than any other app. Something to look forward to at least.
Ahh cool...i was a little scared for a second :P
Its a shame that they feel Silo has matured...i remember the 1.xx cycle with updates every other week.
Its really is a shame
Yeha I wish they were still as into it as they used to be. It's really exciting to be a part of software development, watching and helping it grow into something you enjoy using. I still think they could turn it into a more full featured modeler that could compete with studio modeling tools if they wanted to. But I kinda see where they ended up and why. As a solid core toolset, they really produced a great app, and that's probably all they wanted. It caters more to a free flowing approach to modeling. This puts it in somewhat of a niche though, when other people who need more precision tools and modeling history, modifiers, etc can't really justify adding it to their toolsets. Would have been cool to see them try to tackle that side of things.
Anyway, it's amusing you'd post that gun, as I've been making fun of the hype around that thing ('I thought it was real!' 'amazing!') for a while now. It's a prime example of what's wrong with that community.
Oh god, I just noticed, It's in the Forum gallery even.
"A collection of the best artwork posted on these forums (selected by our moderators)"
Well no matter how broken that gun looks I think its cool that he managed to get convincing imagery out of it!! I mean its a terrible photoslap job with no hope of working in a game ... but it put some light on the kind of quick and dirty FX used for movies - whatever works works! I think its kinda cool ...
Yeha I wish they were still as into it as they used to be. It's really exciting to be a part of software development, watching and helping it grow into something you enjoy using. I still think they could turn it into a more full featured modeler that could compete with studio modeling tools if they wanted to.
Totally. the surface tools were a step in the right direction. I still find it better than 3D Coat sometimes for retopo work. I cant believe theres no smoothed preview in 3dc.
I don't see any mention on this thread of possible IP issues around a studio using Blender. Is there any concern that Blender may not have been overly careful to avoid patent infringement in the code that goes into it?
Of course, if you're a one-man shop, then this isn't something you're going to worry much about. If your philosophy is that good ideas shouldn't be patented, then you also probably don't care. I'm not sure any of this would stop a studio from using Blender, but it might give them pause.
I'm not saying Blender has or hasn't violated any specific patents, but I've heard people mention that when it comes to rendering, it is hard to avoid stepping on patents unless you're very careful. Perhaps Blender has been very careful.
I'm mostly wondering if anyone has ever given any thought to this issue when looking at Blender or if it really is a non-issue.
There should be no IP issues when using Blender, unless the source code is modified and publicly distributed (in which case it must be redistributed with all builds of the program or be publicly available). Internal builds of Blender do not need the sources redistributed. More information about licensing with Blender can be found here: http://www.blender.org/education-help/faq/
From what I understand, all developers are aware that when they contribute source code, it will be released under the GPL. I doubt very highly that there are any patent issues, as most open source communities are very highly sensitive to licensing and IP rights (since those really are the bedrock of the development model) and nearly all features that are committed to Blender are publicly reviewed. In addition, from what I've seen, the developers' commits are subjected to a high level of scrutiny, as the entire repository and commit log are open to the public.
I think many of the improvements and new features in Blender come from publicly available sources such as white papers so generally Patents are not really an issue.
I don't see any mention on this thread of possible IP issues around a studio using Blender. Is there any concern that Blender may not have been overly careful to avoid patent infringement in the code that goes into it?
We don't implement known patented technology. I have looked up the patents listed on the Mudbox startup and they are all about copy protection for instance so not an issue.
Generally companys patent stuff that is obvious enough and broadly used enough that every software company on the planet that tries to implement a user interface or other commonly used system is likely to implement the same obvious ideas. So any software you use is likely to have the same patent infringements.
Also the majority of the features you see in most commercial software (as well as Blender) are based on papers from academia which are often stuff that they won't be able to patent. Ie ZBrush UV master, and the transpose tool both had academic implementations first.
I think that part of the issue with Blender not being used in commercial studios is that no one gets trained to use Blender ahead of time. All the schools are teaching the popular commercial software (Max, Maya, SoftImage) so fresh recruits for commercial studios are all familiar with those programs. Any studio that wanted to base their pipeline around Blender would have to re-train both their exisiting personel, and any new hires. With the current state of the industry, everyone is wanting to spend less on training, not more. And of course, we know how resistent 3D artists can be to switching programs...
So established studios will not adopt Blender due to the training burden. The only real possibility for the commercial use of Blender is in new studios. Fresh startups could adopt Blender in order to reduce their overhead. (an even more weighty consideration for startups) Since they would have smaller teams, their training wouldn't cost nearly as much, and they would mainly just have to train new employees.
I think that part of the issue with Blender not being used in commercial studios is that no one gets trained to use Blender ahead of time. All the schools are teaching the popular commercial software (Max, Maya, SoftImage) so fresh recruits for commercial studios are all familiar with those programs. Any studio that wanted to base their pipeline around Blender would have to re-train both their exisiting personel, and any new hires. With the current state of the industry, everyone is wanting to spend less on training, not more. And of course, we know how resistent 3D artists can be to switching programs...
So established studios will not adopt Blender due to the training burden. The only real possibility for the commercial use of Blender is in new studios. Fresh startups could adopt Blender in order to reduce their overhead. (an even more weighty consideration for startups) Since they would have smaller teams, their training wouldn't cost nearly as much, and they would mainly just have to train new employees.
I dont know, more and more studios are using tools like Modo, but no schools are teaching it. While im sure this is relevant, i doubt it really has the much affect. Simple fact is that if a studio thinks a tool will improve their pipeline, they will consider using it.
Blender simpy hasn't found its niche. It tries to be like 3dsmax, Maya, etc where its a tool that does EVERYTHING, however i dont know that it does any one thing a reasonable amount better than any of the other apps. And honestly that is the way to break the cycle of the big three(Max, Maya, XSI), it has to really stand out and do something better. So where is the motivation to switch, you can say price, but that is about it, and i mean to replace a full-feature app, as that is how most people view blender, not as an addon for specific things. Again modo, topogun, uvlayout, xnormal, crazybump, all these programs have sort of a Niche and that is how they make it into studios.
I dont know, more and more studios are using tools like Modo, but no schools are teaching it. While im sure this is relevant, i doubt it really has the much affect. Simple fact is that if a studio thinks a tool will improve their pipeline, they will consider using it.
I have never used Modo, so im not sure how easy it is to pick up but Blender is pretty notorious for its steep learning curve.
How did you find learning Modo EQ?
It would be interesting to find out how long it took people in this thread too become fluent in Blender, so we could get a good average on a time period.
I think that part of the issue with Blender not being used in commercial studios is that no one gets trained to use Blender ahead of time. All the schools are teaching the popular commercial software (Max, Maya, SoftImage) so fresh recruits for commercial studios are all familiar with those programs. Any studio that wanted to base their pipeline around Blender would have to re-train both their exisiting personel, and any new hires. With the current state of the industry, everyone is wanting to spend less on training, not more. And of course, we know how resistent 3D artists can be to switching programs...
So established studios will not adopt Blender due to the training burden. The only real possibility for the commercial use of Blender is in new studios. Fresh startups could adopt Blender in order to reduce their overhead. (an even more weighty consideration for startups) Since they would have smaller teams, their training wouldn't cost nearly as much, and they would mainly just have to train new employees.
Also I know my school was looking to implement blender, and I have heard of some other schools that already have implemented it, not as a replacement for one of the big three programs, more of an extension of the big 3, because not all students can afford one of those 3 programs (even with student discounts) I feel like this is a good idea on paper, but in actuality it may not be the best course of action. If you get familiar with a program like blender then try to move to Maya/max/xsi I feel like you'd experience just as much confusion/frustration as you would moving in the reverse as I have tried to do. Again not many major studios are using it, so learning it as your primary program might not be the best idea. Its much of a catch 22, you don't want to train people in it if the major studios are not using it, but you're right the studios definitely don't want to use it if no one knows how..
I have never used Modo, so im not sure how easy it is to pick up but Blender is pretty notorious for its steep learning curve.
How did you find learning Modo EQ?
It would be interesting to find out how long it took people in this thread too become fluent in Blender, so we could get a good average on a time period.
I learned Modo in about 30 minutes, no exaggeration. However i came from using Lightwave for about 8 years prior, and Modo is built on a lot of the LW stuff, its basically everything good about LW but without all the bad(i haven't used LW since about version 6 or 7 so it may be quite a bit better now).
I think most people can pick it up in a day or two however, its a simple, straight-forward app. A lot of people coming from max miss the complex stuff you can do with the modifier stack and such, modo lacks a lot of flexibility in that regard but makes up for it in tools that are sensible and just sort of "work".
It would be interesting to find out how long it took people in this thread too become fluent in Blender, so we could get a good average on a time period.
I sat down for an afternoon to get a general overview of blender from modeling, to texturing, to rigging and animation and even some rendering using one of the tutorials from the wiki. It pretty much walks you through the entire process of building a simple character and animating it including lip synch and animation blending. After that I felt pretty comfortable with the UI, and workflows that it offered. This was with 2.49.
@EQ
wow that is crazy fast!
I have thought about learning Modo but keep getting put off by the fact that there is no modifiers
I only hear good things about it though, so i prob. will end up taking a closer look eventually
@James
nice!
The Wiki is a great place to start
The n00b to pro docs are pretty good too (or used to be at least...i have not checked them out in ages)
I first picked up Blender seriously around the early 2.3x release (2.34 i think) and it took me about a year before i could model with any sort of speed and was confident with the program, though i was coming from a few months of Milkshape and was basically learning to model from scratch.
At the time this crazy thing called Subd was so alien to me so i cant really give an accurate time-scale on the learning curve for experienced artists.
ZBrush took me a few days to get up to speed, so something like that would prob. be reasonable if i picked it up today.
Replies
Thanks for the clarification
so one smoothing group should be ok for that mesh?
the picture above is from the viewport, though the render displays similar results.
Should the normal map compensate for these errors 100% in the viewport and the renderer?
I will post the .blend soon
I think that the point that is often missed is that philosophically, from a design standpoint, Blender is very different from Maya and Max (I assume, as I use Maya and have only a passing familiarity with Max). Because of this, evaluating the packages on a tool-by-tool level is largely counter-productive. Part of Blender's hard to learn, easy to use paradigm comes from it's design as a series of tightly integrated systems. This is very different from Maya at least which seems to have many very loosely connected systems from the way in which things have been added over the years (at least on the UI level, I understand the flexibility of the nodes). Because of this, it is very easy to learn any one of Maya's parts, UVs, Modeling, Animation, etc, but each part must be learned individually because not a lot is shared between the components. On the other hand, while it is difficult to get into Blender initially, a large amount of the tools, interface, and conventions, etc. transfer seamlessly between different modes. If you know how to model in Blender, you probably know 90% of the entire program, because it's very easy to then move to UVs, or animation from there. Systems like vertex groups (which control anything from skin weights, to hair stiffness and length, to modifier values, physics properties, particle properties, etc.) are especially relevant to this example.
Because of this, what it really boils down to is where you want or need to invest your time. Blender has a large up-front cost that pays out over the long term, whereas some of the more traditional packages have a very low initial cost, but you need to continue to learn other systems within the package over the long term. Deciding whether to learn Blender should really be based on considering which of these philosophies best suits your needs (along with the other more concrete necessities, $, tools, etc.)
As was mentioned earlier in this thread, the developers of Blender have realized that ease of use and easy to learn are not mutually exclusive, and all of the work over the past 3 years of blender development has been aimed at fixing that. All of this said, Blender has quite a bit of work to do with many of it's tools and lower level stuff, but philosophically, it is on solid ground.
On a more personal level (read: personal preference), I would argue that Blender is a more artist friendly tool than Maya, and, from what I've seen, Max. I mean this in the way that, when modeling in Blender, I have a much more tactile and direct connection to my model than I do when working with Maya. Within Blender, I model almost exclusively in a single, full-screened perspective view (which is even easier now that I can tear stuff off and throw it onto my second monitor), and, using almost exclusively short-cut keys, I can pretty much pretend there is no interface. It's fast, but more importantly, it feels good in the way that sculpting feels good. Just unobstructed and direct. When I think about how much time I spend in these programs, and how much more time I will spend, this almost becomes a quality of life issue. Where am I most comfortable.
Don't get me wrong, sometimes within Maya I can work in a similar way, but too often it feels clumsy and frustrating.
My $0.02.
Check the maya examples in that thread, and try the same mesh with 3point Shader + Quality normals in max.
http://boards.polycount.net/showthread.php?t=72861
You'll see the results you get when everything is synced up correctly. In maya there was still some artifacts on the bottom, it truly is a "worst case" mesh, but you can get much better results when its being done correctly.
That being said, unless you're lucky enough to work with an engine that uses 3ps' Quality Normals, you're still stuck with 'the old way'.
Any engines that already do, by the way? I imagine you're working with your clients to implement this, and I'm looking forward to seeing it in something I get to work with.
Anyway, just breaking the normals along all uv-splits goes a long way, and doesn't actually impact your workflow at all, unlike the other way around - breaking apart your uv's at all normal breaks, which makes texturing a bitch. That is, assuming you're breaking normals at every radical change in angle, which does have the added benefit of giving you a bake that's slightly easier to work with because you don't have the gradients on planar faces.
edit: I'm talking about non-organic assets, ofcourse. Hence the 'radical changes in angle', which you tend to get on mechanical stuff.
Any advice that is given freely on this forum by people with much greater experience than myself is most humbly received and appreciated.
Its all good to know and it really helps clear things up and stops be second guessing myself.
Sometimes its hard to filter the good info from the bad as it is so intermingled but i can always count on PC to come up with the goods eventually
Thanks for the help man!
I'm Blenders 'release manager'. I'd like to address a couple of general comments that have been made in the thread.
First off, the artists will be pleased to hear the positive comments on the trailer. There is a lot of work yet to be done on every shot, and improvements that will be made in simulation (the team feels a lot of improvement needs to be done on the hair solver yet), tweaks, for textures, lighting, animation, etc. yet to go.
Regarding Blender and its capabilities.
Like all major 3D packages we have our strengths and weaknesses.
We have long been aware of professional users of other packages frustration with our user interface, and non standard navigation and short cuts. Unfortunately to make the changes for custom shortcuts and easier UI customization required a massive effort - about two full developer years on the front end for our two most experienced developers. Thus it just wasn't possible to make those changes until relatively recently.
I do hope once our next release or so comes out you will give the latest Blender version a try. There will be a huge amount of written and video documentation that will come out within a few months after the release. Also there will be custom keybindings presets for Maya and other 3D applications.
I would encourage all of you to stop by our booth at Siggraph this year. We will be doing demos the entire time, and I'm fairly confident you will be blown away.
metalliandy.
as to the 'need to focus on the core tools' - be aware that a lot of things like smoke, fluids, and sculpt are done by folks who have zero interest in working on anything but those tools. So advancements in those areas generally have no impact on the rate of development in other areas.
Of course others would also argue with you on what 'core tools' to focus on. Personally my own interest is getting our sculpt, retopology, and paint tools competitive with the market leaders. Although I do think getting our mesh tool upgrade completed (probably will happen after Sintel is completed) and fixing weaknesses in our UV tools is also highly important.
I'll pass your list of suggestions onto the dev team regarding UV and mesh tools (mesh tools issues are waiting on our major overhaul being ready 'bmesh').
pior,
regarding setting up Blender as a dedicate modeler/sculpting tool/video editor/video compositior/uv unwrapping/texturing/rigging/animating etc tool - the interface is now about 99% defined in python. Some time after the next release or two I'll be working on dedicated simplified interfaces.
Regarding not enough high profile quality output from Blender - definitely agreed. The quality work is being done, it just doesn't get posted anywhere except for one of the hundered of Blender user forums around the net. This is a problem we are aware of - and will get it taken care of probably after this summer.
Mightypea,
The idea is mine and is called 'art for code' - the website for it is being set up now by BTolputt. Within about a month it should be ready to go.
again thanks to everyone for their comments,
Tom Musgrove
LetterRip
Excellent news, I'll be looking forward to that website.
Personally, while I'm mostly interested in seeing the modeling tools improve, I think a very good implementation of retopology could lure professionals to Blender very easily. We already use topogun/3dcoat/etc. for just this purpose, and if Blender could do it equally well that could make it part of every modeller's toolset.
Glad to see you on this forum, by the way. For us to register at BlenderArtists (where there can at times be an aggresive attitude to any mention of the commercial software we're used to) is a bigger step than just talking about it here, with the people we know.
I'm really determined to get myself used to it but as you said, the interface is so.... abstract...
But, it's scriptable and open and all that, and getting more powerful. I hope one day i can make full current gen pieces in blender without feeling awkward
One thing that i had a little trouble with the other day was that i made a mesh to export to another program, but i had a problem somewhere on the mesh where i had some random face connected to the wrong edge making 3 faces coming off one edge.
Does anyone have a good way of locating such problems quickly or easily? I resorted taking it to max just to fix the small problem
In 2.49 in 3d window, go to spacebar menu -> select -> non-manifold or with shift+ctrl+alt+M
In 2.50go to the 3d window's header, select -> non-manifold while the shortcut is the same.
It will select parts of the mesh where an edge is shared by more than 2 faces.
Problem with the blender artist community and lack of good artists is mainly blenderartists.org, that place is a shut in community where few really develop, and people sprinkle around false reinforcing comments and feedback in threads.
It even got to the point where a special forum part called "focused critique" had to be created back in the days because people got butthurt over any critique.
My experience is that people who have come from there and gets to polycount will start learning way faster and grow in ways they never did before, (I know I did).
That place needs to be way harsher, since it's almost the official place for blenderartists to go.
Example (and nothing personal against the artist):
http://blenderartists.org/forum/showthread.php?t=164370
There's barely any comments helping that guy in the right direction.
I really appreciate that the devs of the smoke and physics are separate from the other tools and think they all do a great job at what they do...it just seems there is much more interest in developing those sort of tools than everything else as last time that a major modelling tools overhaul happened was so long ago (2.35?) but im hoping that BMesh will fix that issue.
I really think that Blender is a fantastic tool, but it just needs a little refining.
Some of the best forums (and apps) I've seen have a wide diversity of users and suggestions come from all angles. Forums where users are free to say "here's how I would do it in XX app..." and it isn't met with scorn but inquisitive appreciation. It often sparks new workflows and tools to be written. You end up with the best possible way to do something being propagated to all the apps instead of "the only way to do it in XX app" which might be a crappy work around at best but because no one knows of a better way its the only way that's ever suggested. Silly and unhelpful but allows for a constant positive feedback loop.
I'll generalize and I'm sure I'll get some flak for it especially now that the blender crowd is gathering pitch forks and torches in hand.
- Artists post because they want to improve.
- Blender artists post because they want someone to say something nice about them. Read that as: They personalize their work so much that a knock against their work is a personal attack, not helpful advice.
The post that eld linked to happens many times over. A lot of people falling down, tripping over themselves to give praise to what desperately needs a critical eye. They shower praise because that's what they want in return.
How can I trust what they say when all I ever see them do, is praise junk? Why would I turn to them for advice or even think to post things there? More importantly how can anyone ever really get better if everyone is too busy blowing smoke up each others asses? There is a nice way to give critiques, but even that seems to be bad form. Those that attempt it are drown out or shushed. No thanks...
As far as their community goes, I don't see it as any different than commercial software forums. People posting work less for crits than just positive feedback and a general banding together and forming of their community around a common interest. This happens in all the software forums I've visited over the years - just not on nearly as large a scale, since the blender community is massive. It's also why I don't really take part in the communal back-patting in places like that. I'm not there to make friends and feel good about my software choices or have someone tell me my shitty art is good. You still have to come to places like polycount for any real critique.
I use Blender as my main application so i could be counted in that demographic but i certainly don't consider myself to be a "Blender Artist" in the way you describe it...i post because i want to get better at what i do.
I think perhaps its more likely that many Blender artists are just starting up and therefore what looks good to them isn't what looks good to more experienced artists.
[sweeping generalisation]It is true however that the Blender community (as any community that is dedicated to a single application) can be very incestuous and perhaps a little bigoted with a massive, almost fanatical bias toward Blender and everything open source and are sometimes seemingly unable to receive any criticism for the way Blender does a certain thing.[/sweeping generalisation]
If i remember correctly LetterRip did a nice presentation on being an effective Blender advocate at the last Blender conference
I have gotten flamed upon many occasion by suggesting that perhaps something that Max (or any of the big apps) does, is a better way or should be integrated in some fashion into Blender.
One example i can think of currently is i posted on BA suggesting that there should be an option in the colour picker for RGB values of 0-255, rather than using 0-1 as i was tired of copy and pasting Hex codes to get an exact colour.
Needless to say it was frowned upon.
Personally i feel that if another application does something better than im all for it and if i can afford the program i have no problems making the purchase.
Day to day i use Blender, Silo, Zbrush, Unwrap 3d Pro, MeshLab, Polygon Cruncher, FAOGen, xNormal, Milkshape (exporting only :P) and CS3 (among others im sure that im forgetting) and though i would love to use Max, the cost is unfortunately just not justifiable or affordable (£3k ($4.3k)).
I hope that the Blender community does change for the better but it will take a good amount of time i think.
Whether that's a lame excuse or downright admittance of the forums' poor quality, I'll leave up to you. I posted artwork there once, got the comment that I shouldn't be using triangles for game meshes, and stayed with Polycount from then on.
(And so it might also well be possible that the better artists don't like visiting the BA forums either, skewing the perspective even further...)
Lol! that's a classic.
Speaking of configuring it that way, is selection highlighting an option in Blender?
Anyway, it's amusing you'd post that gun, as I've been making fun of the hype around that thing ('I thought it was real!' 'amazing!') for a while now. It's a prime example of what's wrong with that community.
what maddens me was they had really major names in modeling behind them. then they just kind of lost heart. I've been watching Blender really closely for a while now waiting for the feature freeze, and B mesh to see if writing up some tools like silos would be practical.
Most of the stuff in silo was about the tool logic. The cut tool for instance is awesome, it cuts connects and divides in 1 context sensitive keypress. To maintain my sanity moving from silo to max i wrote my own silo cut, break and merge in a max script. the difference it makes is nuts.
Silo is a sweet app.
My take on it all is that Silo still rocks, even with slowed development. Considering how far other apps have come in the last few years with much bigger teams and budgets to put into development I still prefer to use Silo, which makes me a bit more disappointed in the expensive modeling tools I've paid for and hardly use. I'm actually more disappointed with the developers of those tools and their inability to surpass or at least allow me to reproduce the workflow I have in Silo.
Its a shame that they feel Silo has matured...i remember the 1.xx cycle with updates every other week.
Its really is a shame
Oh god, I just noticed, It's in the Forum gallery even.
"A collection of the best artwork posted on these forums (selected by our moderators)"
Totally. the surface tools were a step in the right direction. I still find it better than 3D Coat sometimes for retopo work. I cant believe theres no smoothed preview in 3dc.
http://www.pepeland.com/
Of course, if you're a one-man shop, then this isn't something you're going to worry much about. If your philosophy is that good ideas shouldn't be patented, then you also probably don't care. I'm not sure any of this would stop a studio from using Blender, but it might give them pause.
I'm not saying Blender has or hasn't violated any specific patents, but I've heard people mention that when it comes to rendering, it is hard to avoid stepping on patents unless you're very careful. Perhaps Blender has been very careful.
I'm mostly wondering if anyone has ever given any thought to this issue when looking at Blender or if it really is a non-issue.
From what I understand, all developers are aware that when they contribute source code, it will be released under the GPL. I doubt very highly that there are any patent issues, as most open source communities are very highly sensitive to licensing and IP rights (since those really are the bedrock of the development model) and nearly all features that are committed to Blender are publicly reviewed. In addition, from what I've seen, the developers' commits are subjected to a high level of scrutiny, as the entire repository and commit log are open to the public.
For example, the Approximate Ambient Occlusion came from Nvidias GPU Gems 2
We don't implement known patented technology. I have looked up the patents listed on the Mudbox startup and they are all about copy protection for instance so not an issue.
Generally companys patent stuff that is obvious enough and broadly used enough that every software company on the planet that tries to implement a user interface or other commonly used system is likely to implement the same obvious ideas. So any software you use is likely to have the same patent infringements.
Also the majority of the features you see in most commercial software (as well as Blender) are based on papers from academia which are often stuff that they won't be able to patent. Ie ZBrush UV master, and the transpose tool both had academic implementations first.
So established studios will not adopt Blender due to the training burden. The only real possibility for the commercial use of Blender is in new studios. Fresh startups could adopt Blender in order to reduce their overhead. (an even more weighty consideration for startups) Since they would have smaller teams, their training wouldn't cost nearly as much, and they would mainly just have to train new employees.
I dont know, more and more studios are using tools like Modo, but no schools are teaching it. While im sure this is relevant, i doubt it really has the much affect. Simple fact is that if a studio thinks a tool will improve their pipeline, they will consider using it.
Blender simpy hasn't found its niche. It tries to be like 3dsmax, Maya, etc where its a tool that does EVERYTHING, however i dont know that it does any one thing a reasonable amount better than any of the other apps. And honestly that is the way to break the cycle of the big three(Max, Maya, XSI), it has to really stand out and do something better. So where is the motivation to switch, you can say price, but that is about it, and i mean to replace a full-feature app, as that is how most people view blender, not as an addon for specific things. Again modo, topogun, uvlayout, xnormal, crazybump, all these programs have sort of a Niche and that is how they make it into studios.
I have never used Modo, so im not sure how easy it is to pick up but Blender is pretty notorious for its steep learning curve.
How did you find learning Modo EQ?
It would be interesting to find out how long it took people in this thread too become fluent in Blender, so we could get a good average on a time period.
Also I know my school was looking to implement blender, and I have heard of some other schools that already have implemented it, not as a replacement for one of the big three programs, more of an extension of the big 3, because not all students can afford one of those 3 programs (even with student discounts) I feel like this is a good idea on paper, but in actuality it may not be the best course of action. If you get familiar with a program like blender then try to move to Maya/max/xsi I feel like you'd experience just as much confusion/frustration as you would moving in the reverse as I have tried to do. Again not many major studios are using it, so learning it as your primary program might not be the best idea. Its much of a catch 22, you don't want to train people in it if the major studios are not using it, but you're right the studios definitely don't want to use it if no one knows how..
can plebs like me get that script? I used to use silo but max is preferable as it is more industry standard now that silo is rather quiet.
I learned Modo in about 30 minutes, no exaggeration. However i came from using Lightwave for about 8 years prior, and Modo is built on a lot of the LW stuff, its basically everything good about LW but without all the bad(i haven't used LW since about version 6 or 7 so it may be quite a bit better now).
I think most people can pick it up in a day or two however, its a simple, straight-forward app. A lot of people coming from max miss the complex stuff you can do with the modifier stack and such, modo lacks a lot of flexibility in that regard but makes up for it in tools that are sensible and just sort of "work".
I sat down for an afternoon to get a general overview of blender from modeling, to texturing, to rigging and animation and even some rendering using one of the tutorials from the wiki. It pretty much walks you through the entire process of building a simple character and animating it including lip synch and animation blending. After that I felt pretty comfortable with the UI, and workflows that it offered. This was with 2.49.
wow that is crazy fast!
I have thought about learning Modo but keep getting put off by the fact that there is no modifiers
I only hear good things about it though, so i prob. will end up taking a closer look eventually
@James
nice!
The Wiki is a great place to start
The n00b to pro docs are pretty good too (or used to be at least...i have not checked them out in ages)
I first picked up Blender seriously around the early 2.3x release (2.34 i think) and it took me about a year before i could model with any sort of speed and was confident with the program, though i was coming from a few months of Milkshape and was basically learning to model from scratch.
At the time this crazy thing called Subd was so alien to me so i cant really give an accurate time-scale on the learning curve for experienced artists.
ZBrush took me a few days to get up to speed, so something like that would prob. be reasonable if i picked it up today.
Interesting stuff!
That is version 1.68...well before it was open source.