Only MayaLT and FBX ? I want to export a standard maya file because currently that is the only way I can get vertex alpha into Unity, fbx does not support vertex alpha to Unity, although vertex colours come across.
Or how about working with the Unity Dev. team and asking them to support the MayaLT format ?
I have to say though, if your taking suggestion to the team, the current method for applying vertex colours / vertex alpha in maya is woeful, the tools need a complete overhaul, I've personally developed my own colouring suite with script to get round the current workflow. I need to see if that script trick posted earlier in this thread works with my personal scripts.
I'm using the mobile transparency vertex colored shader out of the box, colours appear but no alpha ? I load the fbx files back into Maya and sure enough the vertex alpha has been removed ?
EDIT : Just to confirm, I can save a standard maya file to the desktop, import it into Unity, use the exact same shader and my vertex colours AND alpha are present no problems. Go figure ?
Question hoping that some AD guru will help me with.
On regular Maya you can create a script called userSetup.mel in the script (My Documents...) folder that loads some custom preferences.
In this case I want the new modeling toolkit to load and stay ALWAYS on the left side (closer to all the other buttons). Normally once you drag it manually it resets position every time on open/close app or sometimes mid session. Annoying as *^&$&
I got this script to work with regular Maya but LT ignores it. Did some of the commands/names change in this regard?
file -f -new;
loadPlugin "C:/Program Files/Autodesk/Maya2014/bin/plug-ins/modelingToolkit.mll";
dockControl -e -a left NEXDockControl;
***
I found a workaround but... it's autodeleted by LT.
I copied the script to pluginPrefs.mel (deleted evalDeferred("autoLoadPlugin(\"\", \"modelingToolkit\", \"modelingToolkit\")");) and when I start it all works as intended but on every exit LT reverts the pluginPrefs to default.
Any idea what to do with this?
***
Just to add, modifying modelingToolkitProperties.mel doesn't do a thing. I can delete the whole thing (lol) and it still works opening on the right.
I've used Maya for years...even taught it at college.
I'm working with the trial. Being Freelance, even $50 is a good chunk, but it's doable. The polygon limit, obj export (I use ZBrush in my pipeline) and no python/mel support makes it...difficult and frustrating....and most hesitant to make the payment/purchase.
One of the most significant things that Maya LT along with the ridiculous $200/month rental fee for the full version has done is make me seriously consider an alternative like Modo, or Blender even. It's made me thankful that ZBrush isn't owned by Autodesk as well...
I don't mean to sound overly harsh - Autodesk has done some great things in the past by offering a free educational license, and I think Maya LT and the rental options may have started with good intent, but...overall, it's left a bad taste in my mouth.
I guess we judge Autodesk so harshly because they have a monopoly - with 3ds Max, Softimage, and Maya.
Hi! I'm a new member here. I have been in Map making/modeling and all other things associated with game development as a hobbie since 1997. Being just a hobbiest for years, software prices are an extreme concern. I think Autodesk is missing the mark on Maya LT, and it's current set of features.
I had my credit card out twice ready to buy, and stopped at the last second, and decided to do more research on the product. I am so glad I did.
People want software that makes life/work easier, not harder for no reason. Right now Maya LT seems like it actually makes things harder like with a 25k poly limit, scripting etc.
For starters the product needs:
Minimum of 50k polys, but more realistically 75k or up on FBX export
Mel and Other language Scripting
OBJ Support
Lastly with this being a new product with changes hopefully coming quick, the first few product upgrades should be free, until it's more into a refined state. Paying 800 dollars now, and then you upping the polygon limit, etc and having to pay more money for that is terrible. I won't be purchasing LT until these things are in the product for that reason.
/wallet closed
I am greatly interested, and would buy it, but right now is not the right time.
Hi! I'm a new member here. I have been in Map making/modeling and all other things associated with game development as a hobbie since 1997. Being just a hobbiest for years, software prices are an extreme concern. I think Autodesk is missing the mark on Maya LT, and it's current set of features.
I had my credit card out twice ready to buy, and stopped at the last second, and decided to do more research on the product. I am so glad I did.
People want software that makes life/work easier, not harder for no reason. Right now Maya LT seems like it actually makes things harder like with a 25k poly limit, scripting etc.
For starters the product needs:
Minimum of 50k polys, but more realistically 75k or up on FBX export
Mel and Other language Scripting
OBJ Support
Lastly with this being a new product with changes hopefully coming quick, the first few product upgrades should be free, until it's more into a refined state. Paying 800 dollars now, and then you upping the polygon limit, etc and having to pay more money for that is terrible. I won't be purchasing LT until these things are in the product for that reason.
/wallet closed
I am greatly interested, and would buy it, but right now is not the right time.
Agreed with your points. I finished my trial and came to the same conclusion:
For me, it's:
1) Unlimited Polys (do we really need a limit?)
2) Mel/Python Scripting
3) obj support
The strategy should be to get legitimate licenses out there for individuals/freelancers/indies - market penetration, and a smooth upgrade path to Maya Complete
Agreed with your points. I finished my trial and came to the same conclusion:
For me, it's:
1) Unlimited Polys (do we really need a limit?)
2) Mel/Python Scripting
3) obj support
The strategy should be to get legitimate licenses out there for individuals/freelancers/indies - market penetration, and a smooth upgrade path to Maya Complete
I don't really think you can upgrade smoothly as LT files do not open on Maya (or other way around not sure).
I really like the idea of LT but found it way too restricted for now.
I would expect the maya lte to have the following features included:
- no polylimit or at least higher the bar up to 100,000
- mel support at least in such a way that you can use standard melscripts found in the communitys and that you can have a tech-artist / scripter on site using a full version of maya to develop tools / scripts which then can be used by LTE users.
- rendering should be possible, maybe with a resolution limit and singleframe only
Unity and Unreal both have a 65K vertices per object limit.
That's different. Those are lowpoly, realtime, game models. In production you make highpoly models and then bake those into normal maps to be used on the ingame lowpolies.
It's kind of like saying that an allen key and screwdriver are all that's needed to build IKEA furniture. A lot of equipment and factorywork comes beforehand.
That's not to say a package is useless if it doesn't give more than 65k tris, for example all I do is ~1000 poly models with handpainted textures and that's plenty, but there are indie game workflows that have highpoly thingies. Stuff like [ame="http://www.youtube.com/watch?v=A-wW5KP0-LA"]Strike Vector[/ame] for instance.
I wouldn't be surprised if Maya LT 2015 has a bumped up poly count, .obj and at least some scripting support. I can't imagine they're making too many sales with it as it is.
I wouldn't be surprised if Maya LT 2015 has a bumped up poly count, .obj and at least some scripting support. I can't imagine they're making too many sales with it as it is.
I doubt it, or rather it wouldnt fit in with their latest anti-consumer move, which not to have it as a "purchase" but rather a rental if this is to be believed:
"Autodesk today said it would discontinue the ability to purchase software on Feb. 1st 2015. The move is part of its long-term strategy to move revenue from purchases to subscriptions and rentals"
I'd say that the current Rental scheme isn't really targeted for hobbyists really. It's all about having some flexibility to manage costs without an up-front license investment or long-term commitment.
Some people might only need additional licenses for a specific period, but don't want (or need) to pay the cost for the full perpetual license.
Bellsey: It's a dick-move for anyone who does want to use the software for a longer amount of time, because any lower-level company or freelancer will end up paying much more than they would if they'd just buy it. These people don't upgrade to the latest superficially different version, because that's just not cost-effective, and the versions since 2009 have been getting steadily worse.
Bellsey: It's a dick-move for anyone who does want to use the software for a longer amount of time, because any lower-level company or freelancer will end up paying much more than they would if they'd just buy it. These people don't upgrade to the latest superficially different version, because that's just not cost-effective, and the versions since 2009 have been getting steadily worse.
Ya agreed I intend to stick with my current version of Maya for long while.
Bellsey: It's a dick-move for anyone who does want to use the software for a longer amount of time, because any lower-level company or freelancer will end up paying much more than they would if they'd just buy it. These people don't upgrade to the latest superficially different version, because that's just not cost-effective, and the versions since 2009 have been getting steadily worse.
The key thing with the rental schemes, is to do your sums around what your need and requirement is. I think that if you are looking to use a product long term, then you're probably right, rental isn't an option. And we're not saying it is, that's why customers are still able to buy a perpetual license in the usual way
"Autodesk today said it would discontinue the ability to purchase software on Feb. 1st 2015. The move is part of its long-term strategy to move revenue from purchases to subscriptions and rentals"
I think people should maybe read the article again....
Autodesk today said it would discontinue the ability to purchase software upgrades on February 1, 2015. The move is part of its long-term strategy to move revenue from purchases to subscriptions and rentals.
Ahh, I see. Apparently copy and pasting text accurately is beyond some people. Serves me right for not looking for myself. Sorry about that.
So basically pay full pop and keep a version forever or subscribe. OK...
There was an issue with the copy and paste (just wasnt working at the time), so it had to be typed manually on my end, indeed I missed a word though it was unintentional. This info was taken mostly from the modo forums initially.
That said, what is an upgrade in the eye's of AD? The more important part of that post though is this: "The move is part of its long-term strategy to move revenue from purchases to subscriptions and rentals."
Whether "upgrade" is there in the opening sentence or not, if their goal is to get everyone on subscriptions and rentals, then expect to see an Adobe CC business model. No more upgrades can imply that anyone with a current version of Maya can not upgrade to future versions because simply future versions will exist only, again, as part of a rental or subscription service. This doesnt mean they will continue to offer perpetual license purchases for future versions of the software, rather it implies quite the opposite.
I think there's a lot of crossed wires here and people are seemingly desperate for Autodesk to literally follow Adobe's model, even though Autodesk haven't even said that.
Regarding upgrades and what the term actually means, this is the ability to upgrade an Autodesk product from a previous version to the current version. So for example, someone has purchased a product and they may have stopped their subscription (if they bought it) for a period of time, and they then wish to upgrade to the most current version of their software.
Autodesk currently allow customer to upgrade their software to the current version, for a fee. Until this year, there were different upgrade pricing depending on how old the software version was, someone wanted to upgrade from. Also, (if I recall) there was no limit to how old a version of software was, that someone wanted to upgrade.
As of this year, the upgrade policy was changed and basically simplified. Only the previous 6 versions will remain upgradeable. Owners of older software versions who wanted the current version would need to purchase entirely new licenses.
If you did have a version eligible for upgrading, a single pricing structure was put in place. User upgrading to the current version, would have to pay 70% of the new license price for an upgrade.
Essentially, the idea of staying on an old version of software and then just paying to upgrade to the current version when you thought it was necessary, becomes detrimental to actually just keeping on subscription. To keep up to date and have previous version usage, it actually makes more sense to remain on subscription.
I think there's a lot of crossed wires here and people are seemingly desperate for Autodesk to literally follow Adobe's model, even though Autodesk haven't even said that.
On the contrary, thats the last I think most users want (including myself). Again there's a trust issue at play, and the track record hasnt been in AD's favor. It was already stated that there will not be an "upgrade" for the LT line. Its also been stated that their goal is to move away from upgrades and go towards rentals and subscriptions (obviously because more revenue can be had this way).
This leaves a pretty obvious picture in the direction of future AD products including Maya LT. If you delayed on upgrading for a few years, regardless you would have to pay the full amount to get the latest version of the software. This isnt something I desperately want to happen, the opposite in fact. Given whats been said over the past two weeks though... it doesnt paint the image you are suggesting.
If they want more people on rentals and subscriptions, how do you think that will play out in the long run? Corporate will do what benefits them the most.
The information I've posted is the current upgrade policy. It would seem to me, that based on the information that was announced from the investors day, that Autodesk will be ending their upgrade program. As of Feb 1st 2015 (still over a year away), users will be unable to upgrade old versions in the way I described.
new polylimit 65k
and no polylimit with the send to Unity feature...
Hi all,
I am super excited about the reception of Maya LT thus far, its off to a great start and indie developers are putting it through its paces.
For those of you that are not familiar with Maya LT, its a version of Maya aimed at indie game development at a price that is affordable for all. For $50 a month, you get full maya modeling, animation tools, shader FX, texture baking and much more. Its just the beginning and we have a lot in the works that will continue to add value games workflows.
Great updates are coming for Maya LT in the next couple of weeks that I wanted to share. First, we increased the FBX poly limit to a per object of 65k polygons on export, allowing you to create advanced characters or assembling large worlds. Additionally, we added a special workflow with Unity that allows you to "Send to" Unity directly. This allows you to access your project files from Maya LT and export UNLIMITED polycounts to Unity projects. Thats right, there are no limits what you can do with LT and Unity.
This allows you to access your project files from Maya LT and export UNLIMITED polycounts to Unity projects.
I do not know the capability of unity, but that still screws anyone trying to highpoly bake right? That was the concern/crippling point that everyone was huffing & puffing about with the poly limits.
I mean what if you use a similar tool to turbosmooth? Would it refuse to smooth the model? How would it act when you reach the 65k limit?
think that more about bringing in whole levels or multipart objects and not one single object with 1mio poly...
a car for example hast hundreds of parts and all togehter about 250k poly... you could now send a car over in one go...
I guess they really, genuinely believe that the whole polygon limit thing was a good idea to begin with ...
So basically ... they still don't understand that even indie devs might need to bake models in Xnormal. Oh well.
PR talk is such a strange thing. "Yay, we heard your feedback ! Except, we really didn't !"
Inside the Maya LT scene there is no polycount limit at all.
So if you want to have a 3 million polycount model inside LT, that is not a problem. You can "Turbosmooth" it as much as you want.
It is only when you export to FBX that there is a 65k limit.
So if you want to take your base mesh to Mudbox, as long as it is below 65k then you can and do your detail painting and baking there.
If you want to bake in Maya LT, then you can use as large a high poly mesh as you want. No restrictions.
The only thing you cannot do is export the high poly from LT. But you should not need to do that, because the game won't use the high poly and in something like Mudbox you use the lower poly.
As for obj support. I don't know if that is coming, but I do know that FBX is far more developed so it would probably be beneficial for Zbrush to support FBX as an import format since many game engines tend to support FBX these days.
If you want to bake in Maya LT, then you can use as large a high poly mesh as you want. No restrictions.
In theory, that sounds great (and I suppose that this is something that product managers at Autodesk genuinely believe). However, in practice Xnormal exists for a reason, and it is so widespread in both the indie and non-indie spaces that I think it is a bit silly of Autodesk to ignore its existence and penetration (and therefore, the need to be able to export to OBJ without limits)
Now to be fair, Maya can actually be pretty speedy when it comes to baking a simple normal map. But for anything else, Xnormal runs in circles around it performance wise. Adding to that that converting a SmoothPreview model to raw geo for baking can cause huge issues and slowdown, it creates a scenario where it is a huge relief for the artist to be able to be outside of Maya for anything baking related.
I also understand that AD wants to push for FBX instead of OBJ. But again, I would bet that a survey within game professionals would give OBJ as a clear winner when it comes to the question of what is the most used format for raw geo export.
I am not saying that FBX is bad ; just that its history has been pretty bumpy, and it affects its perception as a reliable format. I have memories of older FBX plugins just failing to import or export for no apparent reason, whereas OBJ has always been rock solid.
The only thing you cannot do is export the high poly from LT. But you should not need to do that, because the game won't use the high poly and in something like Mudbox you use the lower poly.
What they are saying, and I think they're saying it very clearly so I'm not sure the confusion here, is that if they create a subdivision mesh, say a hard surface mesh for a gun, they will not be able to bake in a program outside of Maya.
After they model the weapon, they will have to apply several smooths to it, to get the gun smooth enough for a bake. But at a very low, low number of polygons for this high-poly mesh, they are no longer allowed to export - so they will not be able to bake it in another program, say Xnormals.
This is a huge restriction for a lot of people, because exporting the smoothed high version to bake is a very necessary step in almost every pipeline. Without being able to get the high-poly mesh out of Maya, they can't do anything with it.
I don't even do that much hard surface in Maya and this would even hinder me when I go to make basic stuff for characters.
Number wise - if I want to create a high poly object for baking in xnormal (a very huge part of my workflow), I am limited to 937 quad polygons in Maya. With three subdivions, I hit the 60K limit. So if you're creating a 10K quad mesh for a gun, you're limited to a single subdiv/smooth, before you run up against the limit, which is not nearly, nearly, nearly enough for baking out good low res maps.
I don't see how this functionality isn't seen as super simple stuff.
I guess they really, genuinely believe that the whole polygon limit thing was a good idea to begin with ...
I think it is more one of the strategies to not cannibalize regular Maya.
We want to offer Indie devs some of the great tools inside Maya at an affordable price, but we don't want to make it unfair for the regular Maya user.
It makes no sense for LT to take away Maya customers.
What we want is for LT to find a home in Indie devs that would not otherwise use regular Maya.
There is no evil plot behind it. And as you can see we are already tweaking what LT offers to try and align it better.
Oh I totally understand that there is no bad intention behind these limitations. But as I mentioned it earlier in the thread, I just don't see any reason why LT should have any less features than regular Maya to begin with.
The distinction could very well be made at the license level - a studio could qualifiy as "LT eligible" if its workforce is under a certain number of seats, for instance. Then this studio would simply get "LT pricing", for a full, unrestricted Maya experience.
But back on the subject of which features to restrict : I personally think that preventing full resolution export is a very unwise idea. Even individual freelancers working in, say, movie previz often need to quickly export dense subdivided meshes for a quick Keyshot render, for instance.
Number wise - if I want to create a high poly object for baking in xnormal (a very huge part of my workflow), I am limited to 937 quad polygons in Maya. With three subdivions, I hit the 60K limit. So if you're creating a 10K quad mesh for a gun, you're limited to a single subdiv/smooth, before you run up against the limit, which is not nearly, nearly, nearly enough for baking out good low res maps.
Good point yes.
I think at the moment the xnormal workflow you describe is not supported for LT. You'd have to use the baking tools inside Maya.
I think it would be worth describing why the baking tools in LT are not up to the job. Is it just performance in xNormal? Or are there other baking problems in Maya's baking tools too?
The distinction could very well be made at the license level
I was of this opinion too, but I don't think Autodesk could really make that work. You cannot go police 1000's of small users and see if they may have broken the license agreement. Some may not even fully understand or read the license agreement.
I think it would be worth describing why the baking tools in LT are not up to the job. Is it just performance in xNormal? Or are there other baking problems in Maya's baking tools too?
Caveat: I'm still on Maya 2012, so I'm just assuming that the baking process is pretty much unchanged in later versions.
Baking normal maps in Maya is pretty decent. However, baking AO is a massive pain. It uses a different renderer (Mental Ray), takes forever, often crashes and has some weird unintuitive problems where you need to hide the low poly model for it not to occlude itself.
It's also missing the vast array of other baking options that Xnormal has, including the ability to bake vertex colour to texture straight from a polypainted Zbrush model and baking curvature maps (both very useful).
Unifying the baking process to use tha same renderer for everything would be a good start. Get a large sample of workflows from artists in the industry and see what they think is missing or could be improved.
If you want to bake stuff like normal maps, vertex maps, etc, etc, then you're better off using Turtle. It's far better (imo) and easier then using Mental Ray.
Turtle is included in MayaLT and now also Maya, not just Suites only like it was.
Xnormal is good (no denying that), but in my experience there are alot of Maya based studios also using Turtle, and have done for years.
As for formats, again OBJ is good and tried and tested. But, despite this, it's old and still kinda limiting. FBX is a way better and more flexible option (and there's an SDK for it as well).
I think for some people, FBX might still has this stigma attached to it. They tried it in previous years, encountered problems and then never went back to it. I can sympathise and relate to this, based on my own production experience. However the FBX today is a different beast to what it was. I've seen many studios now go back to it, and adopt it as their exporting format and for Unreal/Unity, it's become the best way to get data into their tech.
I think one of the main issues with using mayas normal map baking is that maya doesnt seem to handle high poly meshes very well (eg 4 million poly zbrush model). Maya seems to really struggle or crash with these meshes, where xnormal just does the job, am I wrong?
Hi,
I am just a hobbiest, and have no plans for commercial use atm. I am very interested in Maya LT. But the product right now is not quite where it needs to be. Now upping fbx polygon export to a 65k limit is a welcome start.
My concern is, I do want to buy the product as a single perpetual license, and no subscriptions. If I buy it now, am I locked out of the coming changes unless I pay more money? I'm sorry, not being rude, but that's the way I understand the pricing model at AD. Buy your product, and if you don't subscribe you don't get any updates.
I'm not prepared as a hobbiest to pay money for a product that doesn't seem to have hit it's full stride yet, to just have to pay more money to get it up to speed.
I think one of the main issues with using mayas normal map baking is that maya doesnt seem to handle high poly meshes very well (eg 4 million poly zbrush model). Maya seems to really struggle or crash with these meshes, where xnormal just does the job, am I wrong?
Many years back, I might have agreed with you, but not now. In last several versions of Maya, I've had no problems in importing high poly meshes. I was testing one only last week with Maya and MayaLT. Character assets, multiple objects -12m polys no problem. And I believe that asset originally came from ZBrush.
As for baking though, as I said, you can use Maya's/Mental Ray, but I'd use Turtle. I find it cleaner and more efficent.
The idea of 'not cannibalizing regular Maya sales' is a silly one'. Set qualifications a studio has to meet (income/size/idontknow), and limit the sale of MayaLT to them.
'Oh but what if a larger studio cheats and gives us the wrong information?!!11' Uh... then they don't have a legitimate license. You know, like those people I hear about who use pirated software?
Offer studios the option of buying your software at rates that are affordable for them, or live with the fact that they'll consider other software.
Speaking of the no-script thing (which uh.. wtf? if any software relies on scripts it's Maya), together with the export limits, I suppose that completely kills the GoZ workflow?
Many years back, I might have agreed with you, but not now. In last several versions of Maya, I've had no problems in importing high poly meshes. I was testing one only last week with Maya and MayaLT. Character assets, multiple objects -12m polys no problem. And I believe that asset originally came from ZBrush.
As for baking though, as I said, you can use Maya's/Mental Ray, but I'd use Turtle. I find it cleaner and more efficent.
I think it is great that you took the time to make such a test, but unfortunately it doesn't really paint the whole picture of the highpoly to lowpoly asset creation process, which is what is being discussed here. Being able to import a dense mesh from Zbrush is great (and actually, quite necessary) but then there are many other steps involved in the process.
For instance, let's say that this dense mesh imported just fine. Which retopology tools were then used in order to create the final ingame mesh ? Were the Maya Live tools speedy enough to even allow the user to create the lowpoly mesh while conforming to the high ? If not, then how would such a highres mesh be then exported to Topogun for accurate, high performance retopology ? In order to export it under the 65k limit, the user would certainly have to rely on some polygon reduction tools. Did Maya behave well when processing that high resolution mesh for reduction ?
OBJ has the main limiting factor of only supporting a single UV set. Blech.
This is not relevant to the problem at hand. In the context of this discussion, OBJ is being used as a rock solid way to export high density polygonal data, eventually along with simple material properties (for per-chunk region mask baking) and vertex color information (to bake diffuse to texture). The lack of secondary UV channel support is not a problem in that pipeline.
Now of course FBX has these features and more. But I have to admit that I do have a bit of an icky feeling when I have to use that format, based on bad experiences in the past. Sure enough, these problems might be solved by now but once you lose the trust of the user it is hard to gain it back - whereas OBJ never ever failed me.
Please note that I am not trying to be excessively defensive here, and my intention is not to ignore or refute arguments just for the sake of doing so. I just want to help, bringing up the issues that I see with the current direction of Maya LT, from the perspective of 8 years working at game studios, a good chunk of which spent with Maya as a primary 3D tool.
I think it is great that you took the time to make such a test, but unfortunately it doesn't really paint the whole picture of the highpoly to lowpoly asset creation process, which is what is being discussed here. Being able to import a dense mesh from Zbrush is great (and actually, quite necessary) but then there are many other steps involved in the process.
For instance, let's say that this dense mesh imported just fine. Which retopology tools were then used in order to create the final ingame mesh ? Were the Maya Live tools speedy enough to even allow the user to create the lowpoly mesh while conforming to the high ? If not, then how would such a highres mesh be then exported to Topogun for accurate, high performance retopology ? In order to export it under the 65k limit, the user would certainly have to rely on some polygon reduction tools. Did Maya behave well when processing that high resolution mesh for reduction ?
I didn't have to do a specific test because I knew it worked. But I have also been doing stuff with Maya and Maya LT recently.
The modelling tools in MayaLT have full parity with those in the full Maya, and that includes all the improvements made with new modelling toolkit (built upon NEX). Personally I've never had any problems with Maya's modelling (it always has been very good), but the new improvements have been well received and there's hopefully more to come. There are plenty of videos showing the new tools, including the retopo tools.
Also, included in these improvements, is poly reduction and again, this is also in LT. For some time Maya's poly reduction wasn't that great, so rather then reinventing the wheel and in the spirit of sharing best practices, we've taken reduction algorithm (and related featureset) from Softimage and added it to Maya. This has improved things greatly, because (imo) the poly reduction feature in Softimage was awesome and the best of all the 3.
The performance is better, and I can dynamically reduce a mesh even if it's skinned to a rig and being animated.
as a bump to this tread, the Service Pack and Extension Release for MayaLT 2014 is now the live and can be downloaded.
Please note that it's the Extension release that includes the new 65K export limit and the built in Send to Unity option.
Only customers with active Subscription will be entitled to the Extension for Maya LT 2014. Rental plan customers of Maya LT will have access to extension releases for the duration of their rental plan.
Only customers with active Subscription will be entitled to the Extension for Maya LT 2014. Rental plan customers of Maya LT will have access to extension releases for the duration of their rental plan.
While the changes are welcome, only allowing customers with subs to take advantage of the changes, on a piece of software that is so new, and still finding its footing is terrible, just terrible. Having to pay more for patches for basic functionality is terrible.
If I misunderstand AD's policy feel free to enlighten me. I'm not touching the software until AD gets it into a stable state...... if then.
Replies
Or how about working with the Unity Dev. team and asking them to support the MayaLT format ?
I have to say though, if your taking suggestion to the team, the current method for applying vertex colours / vertex alpha in maya is woeful, the tools need a complete overhaul, I've personally developed my own colouring suite with script to get round the current workflow. I need to see if that script trick posted earlier in this thread works with my personal scripts.
EDIT : Just to confirm, I can save a standard maya file to the desktop, import it into Unity, use the exact same shader and my vertex colours AND alpha are present no problems. Go figure ?
On regular Maya you can create a script called userSetup.mel in the script (My Documents...) folder that loads some custom preferences.
In this case I want the new modeling toolkit to load and stay ALWAYS on the left side (closer to all the other buttons). Normally once you drag it manually it resets position every time on open/close app or sometimes mid session. Annoying as *^&$&
I got this script to work with regular Maya but LT ignores it. Did some of the commands/names change in this regard?
***
I found a workaround but... it's autodeleted by LT.
I copied the script to pluginPrefs.mel (deleted evalDeferred("autoLoadPlugin(\"\", \"modelingToolkit\", \"modelingToolkit\")");) and when I start it all works as intended but on every exit LT reverts the pluginPrefs to default.
Any idea what to do with this?
***
Just to add, modifying modelingToolkitProperties.mel doesn't do a thing. I can delete the whole thing (lol) and it still works opening on the right.
I'm working with the trial. Being Freelance, even $50 is a good chunk, but it's doable. The polygon limit, obj export (I use ZBrush in my pipeline) and no python/mel support makes it...difficult and frustrating....and most hesitant to make the payment/purchase.
One of the most significant things that Maya LT along with the ridiculous $200/month rental fee for the full version has done is make me seriously consider an alternative like Modo, or Blender even. It's made me thankful that ZBrush isn't owned by Autodesk as well...
I don't mean to sound overly harsh - Autodesk has done some great things in the past by offering a free educational license, and I think Maya LT and the rental options may have started with good intent, but...overall, it's left a bad taste in my mouth.
I guess we judge Autodesk so harshly because they have a monopoly - with 3ds Max, Softimage, and Maya.
I had my credit card out twice ready to buy, and stopped at the last second, and decided to do more research on the product. I am so glad I did.
People want software that makes life/work easier, not harder for no reason. Right now Maya LT seems like it actually makes things harder like with a 25k poly limit, scripting etc.
For starters the product needs:
Minimum of 50k polys, but more realistically 75k or up on FBX export
Mel and Other language Scripting
OBJ Support
Lastly with this being a new product with changes hopefully coming quick, the first few product upgrades should be free, until it's more into a refined state. Paying 800 dollars now, and then you upping the polygon limit, etc and having to pay more money for that is terrible. I won't be purchasing LT until these things are in the product for that reason.
/wallet closed
I am greatly interested, and would buy it, but right now is not the right time.
Agreed with your points. I finished my trial and came to the same conclusion:
For me, it's:
1) Unlimited Polys (do we really need a limit?)
2) Mel/Python Scripting
3) obj support
The strategy should be to get legitimate licenses out there for individuals/freelancers/indies - market penetration, and a smooth upgrade path to Maya Complete
I don't really think you can upgrade smoothly as LT files do not open on Maya (or other way around not sure).
I really like the idea of LT but found it way too restricted for now.
- no polylimit or at least higher the bar up to 100,000
- mel support at least in such a way that you can use standard melscripts found in the communitys and that you can have a tech-artist / scripter on site using a full version of maya to develop tools / scripts which then can be used by LTE users.
- rendering should be possible, maybe with a resolution limit and singleframe only
That's different. Those are lowpoly, realtime, game models. In production you make highpoly models and then bake those into normal maps to be used on the ingame lowpolies.
It's kind of like saying that an allen key and screwdriver are all that's needed to build IKEA furniture. A lot of equipment and factorywork comes beforehand.
That's not to say a package is useless if it doesn't give more than 65k tris, for example all I do is ~1000 poly models with handpainted textures and that's plenty, but there are indie game workflows that have highpoly thingies. Stuff like [ame="
I doubt it, or rather it wouldnt fit in with their latest anti-consumer move, which not to have it as a "purchase" but rather a rental if this is to be believed:
"Autodesk today said it would discontinue the ability to purchase software on Feb. 1st 2015. The move is part of its long-term strategy to move revenue from purchases to subscriptions and rentals"
http://gfxspeak.com/2013/10/02/autodesk-sales-strategy-includes-discontinuing-upgrade-purchases/#sthash.czT5o9Qq.dpuf
Some people might only need additional licenses for a specific period, but don't want (or need) to pay the cost for the full perpetual license.
Ya agreed I intend to stick with my current version of Maya for long while.
The key thing with the rental schemes, is to do your sums around what your need and requirement is. I think that if you are looking to use a product long term, then you're probably right, rental isn't an option. And we're not saying it is, that's why customers are still able to buy a perpetual license in the usual way
"Autodesk today said it would discontinue the ability to purchase software on Feb. 1st 2015. The move is part of its long-term strategy to move revenue from purchases to subscriptions and rentals"
Autodesk today said it would discontinue the ability to purchase software upgrades on February 1, 2015. The move is part of its long-term strategy to move revenue from purchases to subscriptions and rentals.
So basically pay full pop and keep a version forever or subscribe. OK...
There was an issue with the copy and paste (just wasnt working at the time), so it had to be typed manually on my end, indeed I missed a word though it was unintentional. This info was taken mostly from the modo forums initially.
That said, what is an upgrade in the eye's of AD? The more important part of that post though is this: "The move is part of its long-term strategy to move revenue from purchases to subscriptions and rentals."
Whether "upgrade" is there in the opening sentence or not, if their goal is to get everyone on subscriptions and rentals, then expect to see an Adobe CC business model. No more upgrades can imply that anyone with a current version of Maya can not upgrade to future versions because simply future versions will exist only, again, as part of a rental or subscription service. This doesnt mean they will continue to offer perpetual license purchases for future versions of the software, rather it implies quite the opposite.
Regarding upgrades and what the term actually means, this is the ability to upgrade an Autodesk product from a previous version to the current version. So for example, someone has purchased a product and they may have stopped their subscription (if they bought it) for a period of time, and they then wish to upgrade to the most current version of their software.
Autodesk currently allow customer to upgrade their software to the current version, for a fee. Until this year, there were different upgrade pricing depending on how old the software version was, someone wanted to upgrade from. Also, (if I recall) there was no limit to how old a version of software was, that someone wanted to upgrade.
As of this year, the upgrade policy was changed and basically simplified. Only the previous 6 versions will remain upgradeable. Owners of older software versions who wanted the current version would need to purchase entirely new licenses.
If you did have a version eligible for upgrading, a single pricing structure was put in place. User upgrading to the current version, would have to pay 70% of the new license price for an upgrade.
Essentially, the idea of staying on an old version of software and then just paying to upgrade to the current version when you thought it was necessary, becomes detrimental to actually just keeping on subscription. To keep up to date and have previous version usage, it actually makes more sense to remain on subscription.
On the contrary, thats the last I think most users want (including myself). Again there's a trust issue at play, and the track record hasnt been in AD's favor. It was already stated that there will not be an "upgrade" for the LT line. Its also been stated that their goal is to move away from upgrades and go towards rentals and subscriptions (obviously because more revenue can be had this way).
This leaves a pretty obvious picture in the direction of future AD products including Maya LT. If you delayed on upgrading for a few years, regardless you would have to pay the full amount to get the latest version of the software. This isnt something I desperately want to happen, the opposite in fact. Given whats been said over the past two weeks though... it doesnt paint the image you are suggesting.
If they want more people on rentals and subscriptions, how do you think that will play out in the long run? Corporate will do what benefits them the most.
new polylimit 65k
and no polylimit with the send to Unity feature...
I do not know the capability of unity, but that still screws anyone trying to highpoly bake right? That was the concern/crippling point that everyone was huffing & puffing about with the poly limits.
I mean what if you use a similar tool to turbosmooth? Would it refuse to smooth the model? How would it act when you reach the 65k limit?
I can't see it acting nice to say the least.
a car for example hast hundreds of parts and all togehter about 250k poly... you could now send a car over in one go...
So basically ... they still don't understand that even indie devs might need to bake models in Xnormal. Oh well.
PR talk is such a strange thing. "Yay, we heard your feedback ! Except, we really didn't !"
I don't understand your point.
Inside the Maya LT scene there is no polycount limit at all.
So if you want to have a 3 million polycount model inside LT, that is not a problem. You can "Turbosmooth" it as much as you want.
It is only when you export to FBX that there is a 65k limit.
So if you want to take your base mesh to Mudbox, as long as it is below 65k then you can and do your detail painting and baking there.
If you want to bake in Maya LT, then you can use as large a high poly mesh as you want. No restrictions.
The only thing you cannot do is export the high poly from LT. But you should not need to do that, because the game won't use the high poly and in something like Mudbox you use the lower poly.
As for obj support. I don't know if that is coming, but I do know that FBX is far more developed so it would probably be beneficial for Zbrush to support FBX as an import format since many game engines tend to support FBX these days.
In theory, that sounds great (and I suppose that this is something that product managers at Autodesk genuinely believe). However, in practice Xnormal exists for a reason, and it is so widespread in both the indie and non-indie spaces that I think it is a bit silly of Autodesk to ignore its existence and penetration (and therefore, the need to be able to export to OBJ without limits)
Now to be fair, Maya can actually be pretty speedy when it comes to baking a simple normal map. But for anything else, Xnormal runs in circles around it performance wise. Adding to that that converting a SmoothPreview model to raw geo for baking can cause huge issues and slowdown, it creates a scenario where it is a huge relief for the artist to be able to be outside of Maya for anything baking related.
I also understand that AD wants to push for FBX instead of OBJ. But again, I would bet that a survey within game professionals would give OBJ as a clear winner when it comes to the question of what is the most used format for raw geo export.
I am not saying that FBX is bad ; just that its history has been pretty bumpy, and it affects its perception as a reliable format. I have memories of older FBX plugins just failing to import or export for no apparent reason, whereas OBJ has always been rock solid.
What they are saying, and I think they're saying it very clearly so I'm not sure the confusion here, is that if they create a subdivision mesh, say a hard surface mesh for a gun, they will not be able to bake in a program outside of Maya.
After they model the weapon, they will have to apply several smooths to it, to get the gun smooth enough for a bake. But at a very low, low number of polygons for this high-poly mesh, they are no longer allowed to export - so they will not be able to bake it in another program, say Xnormals.
This is a huge restriction for a lot of people, because exporting the smoothed high version to bake is a very necessary step in almost every pipeline. Without being able to get the high-poly mesh out of Maya, they can't do anything with it.
I don't even do that much hard surface in Maya and this would even hinder me when I go to make basic stuff for characters.
Number wise - if I want to create a high poly object for baking in xnormal (a very huge part of my workflow), I am limited to 937 quad polygons in Maya. With three subdivions, I hit the 60K limit. So if you're creating a 10K quad mesh for a gun, you're limited to a single subdiv/smooth, before you run up against the limit, which is not nearly, nearly, nearly enough for baking out good low res maps.
I don't see how this functionality isn't seen as super simple stuff.
I think it is more one of the strategies to not cannibalize regular Maya.
We want to offer Indie devs some of the great tools inside Maya at an affordable price, but we don't want to make it unfair for the regular Maya user.
It makes no sense for LT to take away Maya customers.
What we want is for LT to find a home in Indie devs that would not otherwise use regular Maya.
There is no evil plot behind it. And as you can see we are already tweaking what LT offers to try and align it better.
The distinction could very well be made at the license level - a studio could qualifiy as "LT eligible" if its workforce is under a certain number of seats, for instance. Then this studio would simply get "LT pricing", for a full, unrestricted Maya experience.
But back on the subject of which features to restrict : I personally think that preventing full resolution export is a very unwise idea. Even individual freelancers working in, say, movie previz often need to quickly export dense subdivided meshes for a quick Keyshot render, for instance.
Good point yes.
I think at the moment the xnormal workflow you describe is not supported for LT. You'd have to use the baking tools inside Maya.
I think it would be worth describing why the baking tools in LT are not up to the job. Is it just performance in xNormal? Or are there other baking problems in Maya's baking tools too?
I was of this opinion too, but I don't think Autodesk could really make that work. You cannot go police 1000's of small users and see if they may have broken the license agreement. Some may not even fully understand or read the license agreement.
That was my understanding of the situation.
Caveat: I'm still on Maya 2012, so I'm just assuming that the baking process is pretty much unchanged in later versions.
Baking normal maps in Maya is pretty decent. However, baking AO is a massive pain. It uses a different renderer (Mental Ray), takes forever, often crashes and has some weird unintuitive problems where you need to hide the low poly model for it not to occlude itself.
It's also missing the vast array of other baking options that Xnormal has, including the ability to bake vertex colour to texture straight from a polypainted Zbrush model and baking curvature maps (both very useful).
Unifying the baking process to use tha same renderer for everything would be a good start. Get a large sample of workflows from artists in the industry and see what they think is missing or could be improved.
Turtle is included in MayaLT and now also Maya, not just Suites only like it was.
Xnormal is good (no denying that), but in my experience there are alot of Maya based studios also using Turtle, and have done for years.
As for formats, again OBJ is good and tried and tested. But, despite this, it's old and still kinda limiting. FBX is a way better and more flexible option (and there's an SDK for it as well).
I think for some people, FBX might still has this stigma attached to it. They tried it in previous years, encountered problems and then never went back to it. I can sympathise and relate to this, based on my own production experience. However the FBX today is a different beast to what it was. I've seen many studios now go back to it, and adopt it as their exporting format and for Unreal/Unity, it's become the best way to get data into their tech.
I am just a hobbiest, and have no plans for commercial use atm. I am very interested in Maya LT. But the product right now is not quite where it needs to be. Now upping fbx polygon export to a 65k limit is a welcome start.
My concern is, I do want to buy the product as a single perpetual license, and no subscriptions. If I buy it now, am I locked out of the coming changes unless I pay more money? I'm sorry, not being rude, but that's the way I understand the pricing model at AD. Buy your product, and if you don't subscribe you don't get any updates.
I'm not prepared as a hobbiest to pay money for a product that doesn't seem to have hit it's full stride yet, to just have to pay more money to get it up to speed.
Many years back, I might have agreed with you, but not now. In last several versions of Maya, I've had no problems in importing high poly meshes. I was testing one only last week with Maya and MayaLT. Character assets, multiple objects -12m polys no problem. And I believe that asset originally came from ZBrush.
As for baking though, as I said, you can use Maya's/Mental Ray, but I'd use Turtle. I find it cleaner and more efficent.
'Oh but what if a larger studio cheats and gives us the wrong information?!!11' Uh... then they don't have a legitimate license. You know, like those people I hear about who use pirated software?
Offer studios the option of buying your software at rates that are affordable for them, or live with the fact that they'll consider other software.
Speaking of the no-script thing (which uh.. wtf? if any software relies on scripts it's Maya), together with the export limits, I suppose that completely kills the GoZ workflow?
I think it is great that you took the time to make such a test, but unfortunately it doesn't really paint the whole picture of the highpoly to lowpoly asset creation process, which is what is being discussed here. Being able to import a dense mesh from Zbrush is great (and actually, quite necessary) but then there are many other steps involved in the process.
For instance, let's say that this dense mesh imported just fine. Which retopology tools were then used in order to create the final ingame mesh ? Were the Maya Live tools speedy enough to even allow the user to create the lowpoly mesh while conforming to the high ? If not, then how would such a highres mesh be then exported to Topogun for accurate, high performance retopology ? In order to export it under the 65k limit, the user would certainly have to rely on some polygon reduction tools. Did Maya behave well when processing that high resolution mesh for reduction ?
This is not relevant to the problem at hand. In the context of this discussion, OBJ is being used as a rock solid way to export high density polygonal data, eventually along with simple material properties (for per-chunk region mask baking) and vertex color information (to bake diffuse to texture). The lack of secondary UV channel support is not a problem in that pipeline.
Now of course FBX has these features and more. But I have to admit that I do have a bit of an icky feeling when I have to use that format, based on bad experiences in the past. Sure enough, these problems might be solved by now but once you lose the trust of the user it is hard to gain it back - whereas OBJ never ever failed me.
Please note that I am not trying to be excessively defensive here, and my intention is not to ignore or refute arguments just for the sake of doing so. I just want to help, bringing up the issues that I see with the current direction of Maya LT, from the perspective of 8 years working at game studios, a good chunk of which spent with Maya as a primary 3D tool.
I didn't have to do a specific test because I knew it worked. But I have also been doing stuff with Maya and Maya LT recently.
The modelling tools in MayaLT have full parity with those in the full Maya, and that includes all the improvements made with new modelling toolkit (built upon NEX). Personally I've never had any problems with Maya's modelling (it always has been very good), but the new improvements have been well received and there's hopefully more to come. There are plenty of videos showing the new tools, including the retopo tools.
Also, included in these improvements, is poly reduction and again, this is also in LT. For some time Maya's poly reduction wasn't that great, so rather then reinventing the wheel and in the spirit of sharing best practices, we've taken reduction algorithm (and related featureset) from Softimage and added it to Maya. This has improved things greatly, because (imo) the poly reduction feature in Softimage was awesome and the best of all the 3.
The performance is better, and I can dynamically reduce a mesh even if it's skinned to a rig and being animated.
Please note that it's the Extension release that includes the new 65K export limit and the built in Send to Unity option.
Only customers with active Subscription will be entitled to the Extension for Maya LT 2014. Rental plan customers of Maya LT will have access to extension releases for the duration of their rental plan.
While the changes are welcome, only allowing customers with subs to take advantage of the changes, on a piece of software that is so new, and still finding its footing is terrible, just terrible. Having to pay more for patches for basic functionality is terrible.
If I misunderstand AD's policy feel free to enlighten me. I'm not touching the software until AD gets it into a stable state...... if then.