Understanding My Place with TG2

[rcallicotte]

To Planetside -

This is a sincere question and hopefully doesn't generate a poopy-thread.   

My expertise in TG2 is somewhere between a brand spanking newby and someone like JimB.  I would prefer enough insight and grasp of this and other software products to someday reach the caliber of a JimB or an njen.  To head in that direction, I need to ask something that I hope you (Planetside) will answer clearly.  I don't mind others adding their input, but I would really like to hear this from Planetside and I'm sure there are others who could benefit from a clear answer to this.

I see a dichotomy in TG2's utilization in the IT industry.  Of course, this might just be my lack of understanding (see above statement about my expertise).  Here are the two points of apparent bifurcation -

1.  TG2 is a landscape rendering system (and a very nice renderer).
2.  TG2 can export pieces of its mesh to be used in other high-end software packages.

Here is my confusion - both of these things are true, but in both cases someone like me (who isn't a professional...yet) struggles to see how I can make either of these ways work to produce packages for $$.  Of course, there are professionals who have accomplished this through heavy modification.  But, is there a goal Planetside has that will meld these two worlds or will we always need to find a way to fit #1 into something like a Photoshop scenario or #2 into a somewhat strenuous pipeline of expensive software? 

In other words, what should a buyer of TG2 think is the easiest thing, the logical thing, the point of Planetside's venture into the entertainment and advertising world of 3D?

Oshyan recently stated, "The intended workflow with complex objects is to work with external applications and composite, or use rendered out environment maps from TG2."  Are these two basic options the main goal of TG2?  If so, how can we develop some tutorials that are straightforward enough to help people like me?  If not, please explain Planetside's purpose for the media software industry.

[gregsandor]

Terragen is and has been a landscape renderer.  It now includes features that permit those landscapes to be more realistically detailed and animated.

Nobody has ever asked me "What can you render for me in Terragen?"  Its usually more like, "Here's what I want.  I don't care how you do it."

A potential client has a need for art and describes it to you.  He says:  "I want to look at Jupiter's moon Europa from outer space, and fly across its icy surface.  I need thirty seconds of this at x resolution, 3 months from now. I have the following resources to accomplish this on my side... and what we would expect from you is the following... ."

You scratch your head and think about each of those elements and what it will take to accomplish them, and to put them together to form the whole.  You research what kind of data is available.  Also at this stage you think about how you will accomplish the shots and you make some tests, including software. 

Note that I did not say "at this stage you decide how Terragen, how Paintshop, how any program will make the shot.  Define the work, then choose the right tools for the job.  If you want to render stuff in Terragen, then do that.  If you want to produce art, do that with the best tool for the job.  That is an important difference in your mindset. 

Here is my confusion - both of these things are true, but in both cases someone like me (who isn't a professional...yet) struggles to see how I can make either of these ways work to produce packages for $$.  Of course, there are professionals who have accomplished this through heavy modification.  But, is there a goal Planetside has that will meld these two worlds or will we always need to find a way to fit #1 into something like a Photoshop scenario or #2 into a somewhat strenuous pipeline of expensive software?

Software is a tool.  That "heavy modification" that you mention with a hint of distaste is exactly what we do.  It happens that Terragen makes certain kinds of heavy modifications look good.

The difference is like that between two guys who each have a hammer.  One is a carpenter, who uses a tape to measure, a saw to cut, a vise to grip, and the hammer to drive nails. The other guy is what, a hammerer? 

[rcallicotte]

gregsandor, you are saying things I already know.  I work with software, but I don't work in the media industry.  If you do, speak on a level of someone with a technical background.

As for my real answer, it needs to come from Planetside or it's going to get confusing.  I'm not asking for pep talks or dissertations on what someone assumes I don't know.  Just an answer to the question.  As of right now, that hasn't been answered.  I'd rather get an answer to what I'm asking, which is fairly straightforward and I expect a no bullshit answer.

In essence, I'm asking what Planetside's goal is for this tool.  You know?

[gregsandor]

Your questions were "how I can make either of these ways work to produce packages for $$." and "what should a buyer of TG2 think is the easiest thing, the logical thing, the point of Planetside's venture into the entertainment and advertising world of 3D? "  I am qualified to answer both as I am a "buyer of TG2" who has used Terragen in feature film production, games, and scientific visualization for more than a decade.  See my first reply for the answer.  It was intended to be helpful toward your goal.

[rcallicotte]

"But, is there a goal Planetside has that will meld these two worlds or will we always need to find a way to fit #1 into something like a Photoshop scenario or #2 into a somewhat strenuous pipeline of expensive software?"

This is the only question that I would like answered from Planetside, since TG2 is a completely new package that hasn't even been released yet.  I'm attempting to nail down exactly what TG2's plans are for people like me who are attempting to grasp where to put their time and money.

@gregsandor - I appreciate that you've been in this business for awhile using Terragen.  Some people in your industry could create something out of a TV tube and some sprockets from a Lego set, but I'm not one of those people.  I need some basic instruction / tutorials on things.  So...if Planetside says that #1 is the option, I'll need to learn how to do that.  If Planetside says #2 is the only option, then I'll need to consider that.  If it's both, then I'm hoping someone somewhere will give some useful tutorials on a TG2 pipeline that can do the likes of something short of working on a movie like Star Trek.  This would give me, and possibly others here like me, an opportunity to learn and then with some diligence we would be able to fit into your business relationship description above.

[Matt]

Hi Calico,

I'll be answering your question shortly. Please consider this a placeholder

Matt

[Cyber-Angel]

I understand the frustration of someone trying to ascertain the target market(s) for TG2 and where planetside wish to position it within its market segment, what ever that is intended to be; I say this because I have asked time and again, at this time I am no closer to any understanding of these matters.

Clearly, there must be short, medium and long term goals for TG2 as a commercial product again these aims are not clear from what the developers have said on these forums in the past (Unless I've over looked some thing), one thing that is clear TG2 if it is going to be out there in the commercial arena is for the sake of commercial viability at the vary least is going to have a feature set that is at least on par with the market leaders, or at least have the features from them that are considered industry standards, I am not advocating that TG2 or its successors become bloatware but there are things it dosen't have and this post is not about them.

Think of the software industry as a war (which it is) and in war there are basically two kinds of people, those who lead and those who are lead and please bear in mind what Manfred Von Richthofen said about fighter pilots (But is relevant to the software industry, after a fashion) he said "A fighter pilot is ether vary good or is vary dead". At the end of the day it will be industry reaction to TG2 as a commercial product that will be a deciding factor, if you try to Gage current industry reaction to Terragen by using Google or any other search tool, it seems as if its not on any bodies radar at this time, and using some thing like Google Groups the overall reaction to it is mixed.

As to my own experience with TG2, what limited time I've had to play with it I have managed to get some basic images out of it but was frustrated by my old system with its limited memory and hardware bottlenecks these should now be solved now that I have a Dell Precision T7400 Workstation.

My thinking is this if TG2 is aimed at least say production artists then give them the tool set they need to accomplish there job quickly, to a high level of proficiency in the extreme deadline environment which they have to do their jobs, give Technical Directors the tools and access that they'd require and make life for compositors easier, and least but not least make integrating Terragen assets with other pipeline assets as painless and straight forward as possible, the last thing that you want worry about is software that doesn't talk nicely with other software assets in and existing pipeline.

Regards to you.

Cyber-Angel                                

[gregsandor]

.....

[rcallicotte]

I'll be waiting for your answer, Matt.  Thank you.

Hi Calico,

I'll be answering your question shortly. Please consider this a placeholder

Matt

[Matt]

Hi Calico,

I think I understand your questions. I'm not sure how well I can answer though, but I'll try

First, a few things you probably already know but which don't fully answer your question.

One of Terragen 2's main goals is to fill a particular niche which has been open for a long time, that of rendering landscapes and environments cost effectively without having to compromise on photorealism. With our limited resources we've tried to stay ahead in certain key areas, particularly atmospheric realism, displacement-heavy rendering (even if there are many problems still to solve), and enough procedural tools to allow it to be flexible enough to be used in production situations where other solutions would be inadequate or take longer to get the necessary results. I don't mean to say that the procedural tools are its greatest asset - I would say that they are merely adequate at this point - but they support Terragen's other strengths. There's still a lot of ways that we need to improve communication with other tools so that Terragen can fulfill its potential in that niche.

Lots of other things that will affect Terragen 2's wider appeal, like the ability to render objects from other sources, quickly place them in the scene and texture them with ease; faster and more direct methods of modelling and surfacing the landscape or getting that data from outside modelling and paint packages. Those are things that we will improve in future updates. But our strategy hasn't been, and is unlikely ever to be, to create a general-purpose 3D modelling and rendering package, so that still leaves us with the problem of how to create scenes where the landscape is only part of the whole composition. For users coming from certain backgrounds this will be less of a problem as we further develop the tools needed to composite with elements from other renderers. At the same time, we'll be continuing to improve Terragen's capabilities to do a lot of the non-landscape rendering and texturing within Terragen itself. We'll never be able to do all of these things as well as Blender, Cinema 4D, Photoshop, etc. but how important this becomes to the individual user will depend on what their aims are and the relative importance of the different themes and elements of their art, I suppose. If the landscape is only a minor part of the composition, or if the other aspects have to be the very best they can be, the artist will have to break out the other tools and make them work together. By improving the import/export capabilities of Terragen we can make this process easier.

If I'm not mistaken, the crux of your question comes down to the fact that you see at least two possible ways that Terragen can be incorporated into larger, high quality works, but there is not enough clear advice on how to actually achieve that. My personal (3D-CG-biased) opinion is that to achieve the most photorealistic results, generally, most things will need to be done in 3D, with Terragen only being used where it excels, and that 2D work will often form a big part of the work (as textures, 2D backgrounds, 2D backgrounds mapped into 3D, whatever is best for the job at hand). Terragen could fit into that pipeline in various ways. How it fits will usually depend on many more external factors than just Terragen itself. The loftier the goals, the less central Terragen becomes, and therefore the problems to be solved and the techniques to be employed become less about Terragen and more general to digital art.

If the goals are simpler, Terragen alone may be perfect for the job. And then of course there are jobs that Terragen is just not suited to. Unfortunately the ones in between tend to be the most difficult

After all that, there are improvements to be made which will help everybody - no matter what their abilities and experience. Terragen needs to get easier to use, faster and more reliable. We'll  tackle all those things.

I know this isn't really answering your question, but we do want to improve the tools, provide more resources for learning how to incorporate Terragen into workflows and pipelines. We'll need to do that to sell this software and to allow existing customers to get the best out of it.

Let me know whether I've gone any way towards addressing your questions or completely missed the point   It's been a long day!

Matt

[EDIT: I noticed that I emphasise photorealism a lot. I won't apologise for that - it's what I strive for But if you exchange 'photorealistic' with 'high quality', I think a lot of what I said still applies]

[rcallicotte]

Thanks so much, Matt.  Your explanation is plenty.  I realize you couldn't write the book(s) or do the tutorial(s) I need to study.     Really, to not joke about it, this has helped very much.

The biggest help in what you are saying is that Terragen might improve the import and especially the export of its system.  The improving import system will perhaps help with smaller jobs that will sometimes require some sort of 2D digital work to touch up the end result of work inside Terragen.  The work that will need export from Terragen will probably fit "The Golden Compass" or "Star Trek: Nemesis" sort of venue and that is fascinating. 

Thank you for helping me to put Terragen in its place without belittling it and yet staying true to its clearer definition of purpose.  I appreciate your humility.

I think I'll stay tuned.   

[Matt]

Thanks. I'm glad that something I said is still connected to reality

Matt

[JimB]

Can I pipe up a bit as well?

A VFX shot breaks down into multiple elements composited together, as a rule, and what Terragen does is allow for a realistic, mostly procedural element to create a complete environment background, or part of, that does away with the need for an expensive live action or miniature shoot. It can usually be done to the Director's specification, and very importantly, it allows the Director and VFX Supervisor (and sometimes Director of Photography even) to 'tweak' the environment. That's where the procedural aspect comes in useful - you don't need to repaint a whole area, or find or shoot photos you don't have, or remodel a 3D object or practical miniature.

The ideal scenario for a shot using Terragen is when there is no live action or man-made parts that need to intermingle with the landscape, like an establishing shot, all nature, shot in the form of a flyover as an example. But I've not had a single shot yet that does that. I think that's were the big difference lies between the beautiful stills that we see (and the occasional animation) created as their own self-contained piece of artwork, and the so-called "pro" film or broadcast work (I hate that "P" word - it's a pretension - nobody dies or goes to jail if it goes wrong) where it's simply an aid to tell a more complete story, and as such needs to be combined with other storytelling elements which have to integrate temporally with other shots and their elements.

Fuzzy bit over, the practical side to exporting a mesh, and very important one, is the guesswork goes out of that integration of elements, and speeds up the pipeline. The exported terrain mesh allows for accurate shadows to be made in the other 3D app of other 3D elements, as an example, or the limits of the action and animation can be set to make sure the relationship between camera, animated elements and environment are known and set. The exported terrain mesh can also be used in most decent compositing packages (the ones that can import 3D objects) as a cue for that type of  software to hide and reveal 2D elements, amongst other things. The renderer in Terragen, unfortunately, can't shade modelled 3D objects the way other 3D renderers can, simply because that's not what it's supposed to do so far. If you look at Iron Man and get your hands on the latest Cinefex, you'll realise that the surface of his armour shading isn't just reflection, displacement, colour and specular (and the other usual bits), it's also a whole new set of shading methods that had to be developed because it was realised just the above didn't pass muster any more. TG2's strength lies elsewhere and does many things that the other renderers would find difficult to do, and take up a lot of human resources to make happen which would make a shot more expensive and prohibitive. I personally think it's gonna be three years at least before Terragen2 will be able to do everything the likes of Mental Ray can do, and I'm fine with that. Any work you might see in a film that apparently does everything TG2 can currently do probably took a huge effort and is proprietary.

If you're doing work in visual effects it's normal for mulitple apps to be used for any one shot; nobody has a problem with it and everyone realises (or they should realise) that no single app can cover everything. It's also possible for a single app to have different versions tailored to different markets with differing needs, and be no worse off for it.

[rcallicotte]

Gosh, Jim. Thanks!  This is very helpful.  Do you tend to use a 3D application like XSI to import TG2 meshes?  If so, how easy / hard is it to match a good texture in XSI for the Terragen mesh?

[JimB]

Gosh, Jim. Thanks!  This is very helpful.  Do you tend to use a 3D application like XSI to import TG2 meshes?  If so, how easy / hard is it to match a good texture in XSI for the Terragen mesh?

The Lightwave meshes can be imported directly into XSI usually (although there seems to be a few glitches on XSI's side sometimes). If I have a known camera in XSI, which I then export into TG2 (via .chan import), I can render the terrain in TG2 (as well as an ouput mesh if I want a light mesh) from that camera's POV, and bring them all back into XSI using the XSI version of the same camera as a Texture Projection Camera to project the TG2 render onto the XSI mesh. You just have to be methodical and know what you want from the outset. Alternatively, use an Ortho camera in TG2 to create an XZ texture render and slap it onto a standard (high rez) mesh of the terrain in XSI.

The thing to bear in mind is that an imported terrain mesh 1000 units square in XSI is 1000 units square in TG2. You can actually do without a special importer/exporter for locked off cameras by simply typing the camera values (rotation and translation) from one to the other, bearing in mind TG2 uses a different co-ordinate system which I can't remember offhand (Z+ becomes Z- in one or the other, etc). This applies to pretty much all 3D apps - I've done the same between Maya and XSI and even the old Softimage3D in the past; it's a breeze. It's a pain to do for animation, though, although I have done it, spending hours and hours and hours typing values per frame when no converter was available between Softimage3D and Maya (not fun at all). A unit is a unit is a unit. If a model measures 25 units in XSI, when obj'd and put into Maya or TG2 it's 25 units in those as well, and if you move the camera 12 units above it in XSI you can do the same in Maya and get the same result. The confusion starts when we begin talking metres, centimetres, inches, etc.

At least, that's how I remember it.....