What if Softimage died?

[luceric]

I'll keep thinking there is always room for redisigning thing (I know i know you have to deal with toons of constraints), improving and yes I think i need to keep in my mind this "Utopia" to go further and produce good work ;-)

EDIT: Just looking at an example, Yeti from peregrine in maya, integrated in another node editor.... why is it not integrated in the maya node editor ? is it because maya don't permits that ? (I point here why it's not just like Melena on softimage...)

Melena is a bunch of ICE operators. SOuP, shaders, and hundreds of other examples are Maya nodes. Yeti procedural graph is not made of Maya nodes. Their node editor edits an expression tree that they will execute themselves. The view in Maya called "Node Editor" (or the hypergraph) is a view of the Maya nodes (what we would call classic operators in XSI, combined with the Schematic view) that make up the Maya scene graph. If it's not a Maya node, you're not going to see it in that view, just like you wouldn't see something other than ICE operators in the ICE tree. "Move point"? 3dobjects? shaders? nope, different planes of existence.

Something would chose to not be Maya nodes when it needs to be executed completely standalone, for example in a shader on the renderfarm. You do not want to bring in the whole of Maya in your shader just to evaluate some basic mathematics. The whole point was to have lightweight processing at render time for greater scalability, not bake out the results out of the 3D app before hand. So the result of that is that Yeti has its own procedural graph.

[NNois]

"Move point"? 3dobjects? shaders? nope, different planes of existence.

Something would chose to not be Maya nodes when it needs to be executed completely standalone, for example in a shader on the renderfarm. You do not want to bring in the whole of Maya in your shader just to evaluate some basic mathematics. The whole point was to have lightweight processing at render time for greater scalability, not bake out the results out of the 3D app before hand. So the result of that is that Yeti has its own procedural graph.

Humm, thanks for the explanation, so it's all about speed and scalability... for me this kind of things is a dev/design problem, and in my "utopia" world the user should not be limited by these technical problems ;-)

All i'm saying is that today for my point of view what's make the force for softimage and ice is this possibility to control in a single graph(or multiples organized graphs) the generation of the geo, the kinematics, the shaders (with custom attributes), this permits to devellop and imagine things better.

If today devs start saying "hey it's not great for scalability, it's not enough fast. let's restrict things" for me it's simply making things easier for devs and complicated for users.

Anyway, it's a real pleasure to discuss and understand things from you, thank you very much luc-eric

[luceric]

All i'm saying is that today for my point of view what's make the force for softimage and ice is this possibility to control in a single graph(or multiples organized graphs) the generation of the geo, the kinematics, the shaders (with custom attributes), this permits to devellop and imagine things better.

I understand what you're saying in the rest of your post, and it's fine. This paragraph bit is written with Softimage-colored glasses, however. There are various worlds in XSI that are in their individual bubbles and that makes things more simple.

Bare with me until the last sentence, I am not saying one app is better: Unlike XSI, Maya that has a unified node graph when you can edit and connect together all operators, fcurves, shaders, transform (what you'd call ice kinematics) in the one graph. It's a core concept, fundamental and what got it where it is today. And you've probably seen the threads where people here have said that the node editor was confusing and ICE was cleaner. That's probably the main reason.

So we have that example that suggests it's not easier on the user that way. It's not unified XSI, there are separate worlds for scene hierarchy and transforms, operator stack, ice tree and render trees. And that clarifies things. But you can't reconnect an extrude operator, if it wan't an ICE compound.

ICE is a separate system on the side of the "real" architecture and can pokes in some places in the XSI scene with get/set data. You can "generate all your geometry" if you create a primitive with the UI (which creates a native primitive), put a legacy generator, and then add an ice tree on the operator stack where you will use ice modeling. You can drive ice kinematics but it's a separate system parallel to constraints, expressions, and you can't use the local transform, etc. Because those are separate systems.

The reason why ICE is easier and cleaner is because it is specialized. you might think it's super general because you can access "everything" in the scene, but in fact it's a sandbox, and what you can get/set is carefully selected and converted to what ICE needs as necessary, and then converted back on the way back to the native XSI scene objects. I have not used ShaderFX myself, but it's designed to be the right "sandbox" for the users creating realtime shaders for game.

But yeah, that's going to look like a group of specialized modules in one app. And that sounds wrong, at least for a freelancer/consumer level app. So those discussions are going on all the time, even though I sound as categorical on this subject as I always do on anything.

Ultimately I think the solution is for things to "feel" unified, not and necessarily burden the workflow with a technical unification the way the nodes were in Maya. So you got nodes and node grapher UI that are the same for different modules, but there are workflow-specific difference when appropriate. At this price point you want each module to be the best at what it offers. The freelancer simplification will probably happen more at the Maya LT level.

[MauricioPC]

All i'm saying is that today for my point of view what's make the force for softimage and ice is this possibility to control in a single graph(or multiples organized graphs) the generation of the geo, the kinematics, the shaders (with custom attributes), this permits to devellop and imagine things better.

I understand what you're saying in the rest of your post, and it's fine. This paragraph bit is written with Softimage-colored glasses, however. There are various worlds in XSI that are in their individual bubbles and that makes things more simple.

Bare with me until the last sentence, I am not saying one app is better: Unlike XSI, Maya that has a unified node graph when you can edit and connect together all operators, fcurves, shaders, transform (what you'd call ice kinematics) in the one graph. It's a core concept, fundamental and what got it where it is today. And you've probably seen the threads where people here have said that the node editor was confusing and ICE was cleaner. That's probably the main reason.

So we have that example that suggests it's not easier on the user that way. It's not unified XSI, there are separate worlds for scene hierarchy and transforms, operator stack, ice tree and render trees. And that clarifies things. But you can't reconnect an extrude operator, if it wan't an ICE compound.

ICE is a separate system on the side of the "real" architecture and can pokes in some places in the XSI scene with get/set data. You can "generate all your geometry" if you create a primitive with the UI (which creates a native primitive), put a legacy generator, and then add an ice tree on the operator stack where you will use ice modeling. You can drive ice kinematics but it's a separate system parallel to constraints, expressions, and you can't use the local transform, etc. Because those are separate systems.

The reason why ICE is easier and cleaner is because it is specialized. you might think it's super general because you can access "everything" in the scene, but in fact it's a sandbox, and what you can get/set is carefully selected and converted to what ICE needs as necessary, and then converted back on the way back to the native XSI scene objects. I have not used ShaderFX myself, but it's designed to be the right "sandbox" for the users creating realtime shaders for game.

But yeah, that's going to look like a group of specialized modules in one app. And that sounds wrong, at least for a freelancer/consumer level app. So those discussions are going on all the time, even though I sound as categorical on this subject as I always do on anything.

Ultimately I think the solution is for things to "feel" unified, not and necessarily burden the workflow with a technical unification the way the nodes were in Maya. So you got nodes and node grapher UI that are the same for different modules, but there are workflow-specific difference when appropriate. At this price point you want each module to be the best at what it offers. The freelancer simplification will probably happen more at the Maya LT level.

Wow ... very nice post. Thanks. I wasn't part of this discussion but it's always nice to learn more about the inner workings of Softimage and Maya.

[NNois]

So those discussions are going on all the time, even though I sound as categorical on this subject as I always do on anything.

Ultimately I think the solution is for things to "feel" unified ...

Yes, obviously we agree on this point, unifing things was my first note here.

And if it's manageable the pair would be also a unified system/mechanism for thoses sandbox to communicate each other, even with yeti etc (it's your node editor right ?)... damn i'm falling again in my "utopia" sorry

[Falam]

A Subscription for Softimage is $850, for Maya $625, that has increased. Everyone crapped on Softimage 2014, I understand why some of you cancelled your Softimage subscription. It's true you don't have to stay on Subscription, people do use up to three year old versions of their software.

[MauricioPC]

A Subscription for Softimage is $850, for Maya $625, that has increased. Everyone crapped on Softimage 2014, I understand why some of you cancelled your Softimage subscription. It's true you don't have to stay on Subscription, people do use up to three year old versions of their software.

What's the reason for Softimage subscription to be more expensive?

[xsisupport]

A Subscription for Softimage is $850, for Maya $625, that has increased. Everyone crapped on Softimage 2014, I understand why some of you cancelled your Softimage subscription. It's true you don't have to stay on Subscription, people do use up to three year old versions of their software.

What's the reason for Softimage subscription to be more expensive?

Because Softimage is Subscription with Advanced Support, and the Maya price is for Subscription only (aka Basic).
If you add Advanced Support to your Maya Subscription, then the price is about the same, or a little higher, than 850.

[MauricioPC]

Because Softimage is Subscription with Advanced Support, and the Maya price is for Subscription only (aka Basic). If you add Advanced Support to your Maya Subscription, then the price is about the same, or a little higher, than 850.

Thanks for that info. I always heard complains about the price difference but in that case it makes sense.

[Falam]

Phone Support, that is what you pay $200 extra for, would be nice if you had an option. Really there is no other alternative, the 3D world is really at either a cross roads or standstill.

[azurecgi]

Because Softimage is Subscription with Advanced Support, and the Maya price is for Subscription only (aka Basic). If you add Advanced Support to your Maya Subscription, then the price is about the same, or a little higher, than 850.

Thanks for that info. I always heard complains about the price difference but in that case it makes sense.

What doesn't make sense is that Softimage users aren't even given the option of a lower cost subscription!

I wonder how many users have actually used this additional support we are forced to pay for. I've been using Softimage for 6 years and have never needed it (and I'm not sure it would help if I did). When I need help with a problem I normally turn to the community first.

[Falam]

There are alot of great talented people creating ICE compounds, those compounds get 40% finished then they cease to be developed upon, in other words no more time is spend on them. Depending on the community has it's pros and cons.

Suppose someone needs a muscle system, today. He is fairly knowledgeable in ICE, he knows he would need two or three weeks to turn out something workable for production, he doesn't have that time, someone or a studio wants to see a prototype, today. He sees the alternatives, he is impressed, he can accomplish the muscle system with the alternatives with very nice results. What do you think they are going to use ?

As it's been mentioned two programs are better then one, until one program ceases to be used at all. Softimage, Maya, Houdini are all not perfect apps, go to any one of their users and they'll all praise there program. Suppose NewTek bought Softimage, NewTek has it's own vision.

[MauricioPC]

I like the idea of a NewTek Softimage. Don't quite know how to explain, but I like Rob Powers. I think he's a good CEO/Marketer of cool stuff.

[angus_davidson]

I like the idea of a NewTek Softimage. Don't quite know how to explain, but I like Rob Powers. I think he's a good CEO/Marketer of cool stuff.

Would be happier if the foundry folks took it over myself. Always wanted to love Lightwave , but could never get past the modeller.
Very similar to my feelings on blender too ;) A hybrid of Softimage and Modo would be really interesting.

[MauricioPC]

Would be happier if the foundry folks took it over myself. Always wanted to love Lightwave , but could never get past the modeller.
Very similar to my feelings on blender too ;) A hybrid of Softimage and Modo would be really interesting.

Just hopes the prices of The Foundry stays the same. NukeX is very expensive.

[El Burritoh]

Speaking of the Foundry... I'm interested to see what turn the toolsets and job titles take in the next few years as Comp and 3D become even more blended than they are now. With Katana and Nuke, the Foundry is pretty much the only entity seriously able to start merging these two stages. AD doesn't have anything here.