Unity Review

Nice review at creative mac, apologies if this has already been posted.

http://www.creativemac.com/articles/viewarticle.jsp?id=82608

 

Although I just have to point out, only because it's so often and so consistent, that "it's" = "it is" or "it has", no exceptions. "Its" is like "his" or "hers"...note the lack of apostrophe. Kind of hard to read when I was cringing every few lines, sorry.

--Eric

 

Sorry Eric, those are my boo-boo's. I'm afraid I didn't do as good a job editing as I should have

Paul aka BigK

 

boo-boos

 

Ill just never learn. Oops, I mean, "I'll just never learn".

 

Interesting review. It's nice to find a solution that gets your projects moving!

I'm curious, are any of the other engines you tried well-known ones? Torque etc.?

 

Yes, I've used quite a few. Most notable include: Torque, Blender Game Engine, Shockwave (I've been using Director for over a decade), Quest3D, Deep Creator, Anark, and I've recently played with Beyond Virtual and Virtools.

 

Interesting review ... although it seems to end a bit prematurely. It kind of calls for a more in-depth followup.

Now to beat my dead horse

I've used a huge number of game engines and Unity seems to me to be the best overall tool by far.

However, I think Unity has one important (almost crippling) problem -- lack of support for common file types. (Note, I still think it's the best tool overall.)

It doesn't help that the Mac 3D market has been fairly slow to make the move to Intel (very odd given how apparently cross-platform all their code-bases are; why on Earth isn't Lightwave or Maya available native for Intel yet?). As a prospective Unity developer I have a choice of Cinema 4D R10 -- which doesn't currently export to Unity properly, Maya -- which won't run on my main dev machine (Intel) or my old dev machine (G4), Lightwave -- which will run under emulation but isn't fully supported, or a PC-based solution which will also not be fully-supported.

Assuming I wanted to tool up a small dev studio right now, I can't buy a Mac that runs the current version of Maya, which is the "golden path" for Unity content development. This rather dire situation isn't OTEE's fault, but it's unfortunate nonetheless.

While Unity may be good enough to justify using a Mac for development, this doesn't mean that your artists will necessarily use Mac software.

I hope that in the long term collada support and/or some other open file format will eliminate this issue. I assume that Unity went the FBX route because it represented a quick win, but ultimately FBX does not seem like the best way to do unless Autodesk open sources the tools or at least documents it.

It seems very similar to the route taken by Macromedia with Director's 3D engine, which required every 3D tool vendor to implement export to Shockwave. Not one of the exporters ever worked very well and as a consequence Director 3D has remained pretty marginal despite the still large Director user base (I myself have been using Director since it was called Macromind Videoworks).

There are a couple of quite impressive, easy-to-use, and (relatively) cheap PC game dev tools (3d game studio, blitz3d, and some more expensive tools whose names I forget) out there but they only allow targeting PC/MS platforms. Every one I've seen has far better file support than Unity.

Possible quick wins include:
Support for textured .objs
Support for .X files (OK that's asking a lot)
Support for .md2 and .md3 files
Support for VRML files (ignore primitives; just meshes)
Providing a Unity file format that is well documented and has a text version (e.g. the way .3DMF could have a binary or text representation) which it would be fairly easy to target using a scripting language.

If you compare Unity to Blitz3d (for example), the latter does all of the above except for VRML support and having a non-binary version of its documented file format.

 

Just for the record I ran both Maya and Lightwave on a 1ghz G4 powerbook for years. I had no problem using Maya (up to v7) for all my viz work on that hardware. And Lightwave has always run well on low end powermacs and laptops.

cheers.

 

We run Maya on many Intel macs here at SG - on the PC side, Max exports FBX just fine ;-)

 

>>>
It seems very similar to the route taken by Macromedia with Director's 3D engine, which required every 3D tool vendor to implement export to Shockwave. Not one of the exporters ever worked very well and as a consequence Director 3D has remained pretty marginal despite the still large Director user base (I myself have been using Director since it was called Macromind Videoworks).
<<<

I just want to add that the 3dsmax export plugin is okay for director as it doesn't include all the junk which you get out from other exporters like maya for instance. Secondly if you've bought plasma you were also able to export the physics side defined in the modeller which worked pretty well for my projects.

Obviously they could have done better but on the other side these routes work.

 

>>>
Try with Collada (.dae). That's one open, XML-based, and widely supported format.
<<<

And also so a bloated format.

 

lightwave's ub is in beta right now. fwiw, they have stated on their forums that once the ub is released, some sort of support for unity (and others) is on their radar. native support certainly would be a welcome convenience but in the meantime fbx works fine. don't really know what they have planned though.

[edit: just looked back at exactly what newtek wrote about it and it wasn't as strong as i remembered it. so i edited the above to be a little less certain of what they have in mind. i know their Mac guy knows we'd like unity support. here's to hoping...]

 

Thank you for the feedback. A followup is planned, but it probably won't go too much more in depth than what you see in my other articles. The reasons are: a) online magazines (my target publishers) prefer short, to the point articles, and b) "in depth" reviews tend to lean towards one genre of game development and I wanted this article to be more generalized. Of course if my publishers ask for an "in depth followup" I'll be more than happy to oblige.

Cheap? Yes. Impressive? Maybe. Easy to use? Hmm... I suppose that depends on your game development background. It's my opinion (based on my own experience and from what I've read on the game dev forums) that a major obstacle for many beginning game developers is the inability to write code, and in particular, C++. Of the two you've listed I've only tried 3D Game Studio, and 3DGS uses their own "flavor" of C, which IMHO makes it not particularly easy to use.

Of the engines that I used that didn't require programming in C (a partial list is in my earlier post), none came close to Unity's overall ease of use, speed of development and stability (with extra emphasis on stability!).

I personally think that OTEE's decision to support FBX and Collada will, in the long run, be the right one. But unfortunately at present, there are still a few "gaps" in media support. Most noteably is the lack of a good, inexpensive way to get animated 3D characters into Unity. Blender would be the best solution for this, IMHO, and I commend OTEE for their efforts to "motivate" the Collada Blender developer to finish his plugin. Hopefully that short coming will be short lived.

 

Well, you can't say it this easily but the most used modellers/3d packages for games (very generalised) are still 3dsMax and Softimage. Softimage XSI Foundation is just a steal for it's price.

Professionals might draw their displacement maps in zBrush or other tools but it mostly lands back in one of these two apps. Both are sadly not available on OSX but that's what's beeing used by a larger number of developers as far as i know.

 

Just for the record: ZBrush is available on OSX we use it every day at SG to do characters...

 

I never said that zBrush isn't available on OSX. If you reread my posting then you will notice that i said "both" which is regarding to 3dsmax and softimage.

However a downside of zBrush for some people might be that you don't get a crosslicence like you do for modo. But that's a total different story...

 

Ahh - Sorry about that.

 

No problem! You're welcome... :O)

 

Before going on a random walk through the 3d tools market let me say that Unity 1.6 looks incredible. Great job. I especially appreciate Cinema4D R10 support, which gives me a viable production pipeline at last. Woohoo...looks like I need to learn another 3D app.

"And Lightwave has always run well on low end powermacs and laptops."

I meant that it isn't fully-supported by Unity.

"lightwave's ub is in beta right now..."

Right so all they need to do is (a) release a UB and (more importantly) (b) produce files Unity can successfully import.

"Well, you can't say it this easily but the most used modellers/3d packages for games (very generalised) are still 3dsMax and Softimage."

3dsMax has 200k licenses, no question it's #1. Softimage? Even by dropping their price to a quarter of Maya's they still can't quote marketshare. Maya is almost certainly second, and then there's all the others (Lightwave, XSI...).

From a post on Google Answers:

I visited a bunch of game dev houses in Japan in the late 90s and was amazed that most were using Lightwave and XSI. I'm guessing that the driving factor is what tools are best supported by the Console tools market.

If you want proof that XSI's market share sucks, check out this "FAQ" from their own website:

First of all, it's clear that Softimage gets a lot of questions about its lousy marketshare.

Second, it's clear that they have no good answers.

Yes, market share is a "great opportunity" for XSI. I bought a $495 XSI license a couple of years ago; in a market where usability is supreme (every high end 3d modeling tool has, essentially, the same feature set these days) it is the least usable 3d program I've ever seen. There are two approaches to gaining marketshare, ideally if you're trailing in the market you do both: improving the product and reducing the price. XSI has been trying a LOT of the latter.

"Of the two you've listed I've only tried 3D Game Studio, and 3DGS uses their own "flavor" of C, which IMHO makes it not particularly easy to use."

It's about as much like C as C# or JavaScript or PHP or any other of a huge number of C-like scripting languages.

Again, I said that Unity is (in my opinion) the best tool overall, but its greatest weakness is support for third-party file formats. Arguing that it's easier to use than Product X is kind of moot; I agree with you. Ease of use is Unity's strong suit.

I think that instead of supporting, say, a third scripting language, spending some effort on supporting a few well-known file formats might be a little more helpful.

If you compare this situation to Blitz3D (for instance) -- it has a simple, well-documented native 3D file format (.B3D) which has allowed numerous third-party developers to produce first rate importers and exporters for it.

"I personally think that OTEE's decision to support FBX and Collada will, in the long run, be the right one."

Check out this statement from a thread on the FBX File Format specification (or lack thereof) by a game developer trying to work with it:

FBX strongly resembles Macromedia's W3D format in that rather than provide a nice spec for the format you get a slug of binary code that talks to the format. I'd say the internals of FBX are so horrible that they're embarrassed to publish them. Shockwave 3D has been out for, what, five years, Director has over 100,000 users, and it's still not well-supported.

Collada looks good in that it's essentially a fairly straightforward XML-based 3D format. It surprises me how slowly it's gaining traction.

"Hopefully that short coming will be short lived."

Hope in vain -- it's already been quite long-lived, especially considering how long it takes to write a .MD2 parser (or find open source code lying around).

"Blender would be the best solution for this"

No -- support file formats not applications.

 

Geez Podperson, get up on the wrong side of the bed or what? :wink:

 

I don't know if there are any licensing issues, but MD5 support (the DOOM III file format) could be a huge quick win.

http://www.modwiki.net/wiki/MD5_(file_format)

It gets you Blender, Max, Maya, XSI, and Lightwave support, and it's text based (unlike earler MDn formats).

 

I agree that probably would be a nice generic format to support in Unity... but how many of those MD5 plugins are Mac compatible? And I can't see a Maya exporter, only an importer (not that it matters, but it's the principle)...

Unity is a Mac tool and first and foremost it should support tools that are available on the Mac, then worry about the other platforms and their file formats. And at this stage the best 3d tools on the Mac are C4d and Maya, with Lightwave a reasonable second. And once Blender's collada format has animation that should be the tool of choice for those on a budget or no budget at all.

I'd like to see Modo format supported natively in the future as it is a very powerful app and is cross-platform with the same serial number! Which is fantastic. Also last year they previewed some of Modo 3's animation tools. And it looks like they will have procedural animation similar to apps like Endorphin (http://www.naturalmotion.com/)... which will make it a great game asset tool. Obviously one shouldn't buy an app because of future features.... but if Unity supports 'lxo' natively sometime in the future I'll be very happy!

I'm sorry for those PC users coming over to the Mac to use Unity and are having problems with either the Mac as hardware or software support, but at the end of the day it is nothing that we Mac users haven't experienced in bucket loads going the other way... trying to have either good W3D support on the Mac (what ever happened to Mac Maya's w3d support?), what no Mac Plasma? to bad Mac Director 3d developers... Sorry we don't support a Mac format of that export plugin for our 3d game engine, etc, etc. So PC users having problems you will have to forgive some of us who gloat a bit, or feel smug now that we have our own Mac game engine that is fantastic and is only getting better!

But at the end of the day I feel Unity needs a format that is cross-platform and has an open SDK. I agree collada is it and I believe they support it, now we just have to wait for the app devs to catch up. But please don't suggest formats that don't have Mac support equal to their PC counterparts.

Yes i may have got up on the wrong side of the bed today... :wink:

Cheers.

 

I agree. Just throwing in compatibilities regardless of what the pipeline for those formats looks like is not a Mac-like solution. There are some discrete and soon to be complete paths for getting content into Unity, ranging from free (Blender through Collada), to indie (Cheetah 3D and its upcoming character animation abilities), to pro (C4D and then Maya). Having a few "content rivers" makes the playing field Mac-centric, closed, and reliable, just like iTunes->iPod and similar models.

 

I tossed MD5 out as a single, fairly simple file format that would instantly give you character animation pipelines for Blender, Max, XSI, and older versions of Lightwave, Maya, etc. It would also let you use tons of existing MD5 format content out there as placeholder artwork during development. I'd say, put in MD5 support and ignore Collada for 18 months and we'll all be ahead.

"how many of those MD5 plugins are Mac compatible?"

The Blender one is a python script. The Maya one is a mel script. In any event, MD5 is a relatively simple, text-based format so, unlike FBX or Collada, writing your own is viable.

"Just throwing in compatibilities regardless of what the pipeline for those formats looks like is not a Mac-like solution"

Actually transparent support for multiple file-formats both Mac and non-Mac is almost the definition of Mac-like.

"There are some discreet (huh???) and soon to be complete paths for getting content into Unity, ranging from free (Blender through Collada), to indie (Cheetah 3D and its upcoming character animation abilities), to pro (C4D and then Maya)."

I.e. if you want to develop games using stuff you can buy retail today you need to use Cinema4D. Otherwise, you need to use second-hand hardware (e.g. to run Maya) or future products with no stated release dates such as Maya for Intel Macs (2007?), Cheetah 3D with Character Animation (2009?), or a functional Blender/Collada pipeline (2019?).

There's a ton of really incredible functionality in Unity. It can't have been trivial to implement. Why hasn't more effort gone into supporting a few more widely used 3d file formats?

The W3D Analogy

The Director W3D issue is exactly the problem I have with Unity3D. FBX is just as lame a format as W3D. Autodesk could drop support for it on the Mac tomorrow and we'd be screwed.

3d formats == PC formats

The reason that it seems like folks arguing for more file format support is a "PC user argument" is that 3D has been dead for so long on the Mac. The only good 3D packages on the Mac right now are ports from UNIX (Maya) or Windows/Amiga (Lightwave, Cinema 4D). The top Mac 3D tools (ElectricImage, Strata) aren't credible for game development because they never targeted that market (and both companies have died and been resuscitated, and haven't delivered significant new functionality since).

One Last Thing

I'm not a PC user (well, I do use PCs... but only when I have to). I learned to program on an Apple ][+ and wrote three games for it. I've been a Mac user since 1985, and a registered Apple developer since 1990. I was at one of Apple's announcements of the AIM alliance. I was an invited guest at their System 8 release in Australia. I was in the theater to see the 6100, 7100, and 8100 launched. I have the Rhapsody DR1 CDs in a box in my closet along with System X Public Beta. I may be pathetic, but I am a pathetic Mac user

 

"But at the end of the day I feel Unity needs a format that is cross-platform and has an open SDK. I agree collada is it and I believe they support it, now we just have to wait for the app devs to catch up. But please don't suggest formats that don't have Mac support equal to their PC counterparts."

Why not, if the formats still have better Mac support than the file formats we currently have? The fact is there isn't any 3d package on the Mac (worth mentioning) that isn't available for Windows, so no file format satisifes this criterion.

I can't tell if Unity actually supports Collada because I can't export Collada files using any software I have access to.

Here's a priceless quote (about FBX from the Collada folk) from http://www.3d-test.com/interviews/collada_1.htm:

So it's proprietary or vapor. Can we have a third choice please?!

 

Well by that argument then FBX is the format to use... it is supported by Maya, Max, XSI, C4D, Lightwave and several other pro level 3d apps. Maya, Lightwave and C4D are available on the mac so it looks like OTEE made the right decision. Collada is the up and coming format challenger and it may take market share away from FBX in the future... but here and now FBX is the solution. So Maya works on Intel for most people, and no doubt will turn up UB and maybe even have 64bit support after Leopard is released in 2007. C4D is UB and so is Modo. And they both support FBX or have native Unity connections.

But do some research and suggest to us a format that is robust, has animation support, multiple texture and uv support AND has good to excellent support in Mac based 3d apps (which doesn't require me begging for or coding my own plugin). If such a beast exists then bring it to OTEE's attention and it will probably turn up in 1.7! :wink:

I hope collada takes off, as it seems to be a good format for long term backups... But time will tell no?

cheers.

 

I think you make some really good points, podperson. I'm not really convinced about MD5 because it won't "just work" without hunting down the correct version of an export plugin if one exists for your software package, but perhaps Unity could install these plugins itself like it does with C4D (and Maya?). It's still only compatible with predominately PC softwares (XSI, Max, Blender are the only three "usable right now"), though.

I think what's wrong is just that the situation isn't so good on the Mac right now for a complete 3D pipeline.

Sorry, I meant "discrete."

 

I do have a question regarding all these formats. Maybe someone knows this?!

Theoretically i would say a X -> 1 -> X relationship is always the best solution in terms of time all parties need to invest to get communicating to each other properly. The better if it's an open not bloated format with a future where someone is behind bugfixing/maintaining/enhancing the format also in the future.

Isn't there such a format or why is there a native support of a special format at all? I really would love to have max3d/lxo support due to that most of the timne it's just easier to export to each proprietary format but this also does take development time away on otee's side for implementing and keeping up, doesn't it?

Reasons i can think of why there's native support:

a) There is no (working) support for the X -> 1 conversion from a company for a certain product.
b) It's easier to use the home format as you just click your shortcut keys - no property dialogue menus to go through.
c) It's more reliable to use the companies proprietary format as they care about it.
d) A companies format includes features which otherwise would get lost
e) Not beeing fond of format 1, not easy to deal with.

What's the point with each tool (cheetah, maya, c4d) which got native support in unity so far and why is it there?

I always feel a little bit dissapointed when i read that one of the big features is for instance support of the next c4d version native as it's simply useless to me and others who do not use c4d at all. Now i don't know how much work has to be put into keeping up with native formats but i guess someone else could feel the same about when we get lxo support on day. Great for the modo users but doesn't make a c4d user really smile i suspect.

Obviously it would be the best if everything and all would be supported and maybe it's just a matter of time to go there for all the bigger apps but otee isn't about 100 people who don't know what to develop next, so this is something which results to missing something else in unity which could have been there.

 

I just assumed it was a matter of how cooperative the larger companies were, or how complete their plugin APIs were (the C4D import is actually just something that converts to FBX in the background, avoiding extra steps for the user, so it's not really native).

Cheetah 3D's developer is interested enough in Unity to have worked with OTEE on direct support. Interested enough to browse these forums not infrequently.

Newtek, however, I can see not considering Unity worth their time. Maybe that's simply the case with Luxology as well.

 

Yes but why does the cheetah developer and all the others just don't care about supporting one standard format and make sure it's implemented as a seamless process in their application instead of trying to get their own propritary format through?!

This seems to me as the most effective and best solution for all.

Can you imagine a working internet without standard defined protocols?

 

Some may be interested. Problem is that you need to be able to get agreement. Look how DirectX passed OpenGL. Some reasons standards tend to come around are:
- Market dominance: a defacto standard such as X86 or the QWERTY keyboard
- Early in a market where companies perceive that the increased interoperability will allow them to sell more or increased interoperability greatly increases the overall market - ex. email
- Later in a market where a few dominant companies are left and interoperability helps all. Ex. the railroad guages in the US
- Government imposed standards - IRS accounting standards in the US
- Standards to prevent Government imposed standards when the "market" requires standards. Ex. FASB accounting standards

The problem right now is there is no dominant player in the 3D space of enough power to create a defacto standard and the current top tools do not see enough of an advantage to spend the time and resources on creating an open standard. In fact, they may worry about loosing marketshare.

 

In the long term perspective it's always a benefit to be available to as many users/tools as you can. Companies like Discreet for sure a big enough to speak a word on this issue and get things done which they partly did like they're part of the Khronos Group together with others and worked on Collada but there is no real pressure behind it to get it done right!

I would suggest each producer. Do your own format if you wish and can't life without it but do also implement one standard format pretty well. This way your flexible to do your own stuff but also provide a tool which is useful in a wide field. Those without the ressources should just stay with the open format.

And just because you named it qwerty/z sucks.
This unlogical order due to the history of incapable mechanical typewriters.
I hope that we soon will have a change and get used to reconfigurable keyboards like the russian lcd keayboard.

And we're only talking about some technical stuff here... :O/

When unlogical and/or unpleasant things happen i often have to think of one of manu chao's songs:

Hey Bobby Marley
Sing something good to me
This world go crazy
It's an emergency

 

Is that the one that's going for a mere US$1200? (But will drop to under $1000 by September so that's OK then....)

--Eric

 

That LCD keyboard is about the coolest thing ever... but I fear it's not LCD, but OLED, in which case it will fade and discolor quickly. (I'm led to believe that OLED is still only suitable for things like phones and music players, that are lit only a few minutes a day.)

http://www.artlebedev.com/everything/optimus/

 

I've no idea but i guess time after time there will be cheaper versions and with alternative display technologies if that's what makes it expensive. Actually it's the philosophy behind.

I meant this one: http://www.artlebedev.com/everything/optimus/

 

Yep, that's the one. It is actually now LCD, not OLED (for that reason, I guess). However, it's also no longer color, just greyscale, and yeah the price is $1200. Not so cool anymore unfortunately.

--Eric