If you ICE experts out there had to recommend the top 5 (first) things to learn about ICE what would they be? I'm looking for cornerstone/pedestal type foundation items. I'm going to put some time into really learning ICE and I'd like to know what the most valuable concepts are.
1. Context 2. Locations, getting data at locations, etc. 3. Reference frames, i.e. knowing which data is local to the object with the tree and which data is not 4. Arrays and array manipulation 5. Transformations and 3D math in general, e.g., knowing the difference between multiplying a vector or a position by a matrix. 6. Simulated versus unsimulated trees and how it all fits with the operator stack
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Simon Van de Lagemaat Sent: Thursday, December 15, 2011 01:08 PM To: softim...@listproc.autodesk.com Subject: Top 5 things to learn about ICE?
If you ICE experts out there had to recommend the top 5 (first) things to learn about ICE what would they be? I'm looking for cornerstone/pedestal type foundation items. I'm going to put some time into really learning ICE and I'd like to know what the most valuable concepts are.
- ICE trees, the operator stack, and evaluation (eg simulation versus modeling)
- Attributes
- Transforms, matrices, vectors, and coordinate systems
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Simon Van de Lagemaat Sent: December-15-11 1:08 PM To: softim...@listproc.autodesk.com Subject: Top 5 things to learn about ICE?
If you ICE experts out there had to recommend the top 5 (first) things to learn about ICE what would they be? I'm looking for cornerstone/pedestal type foundation items. I'm going to put some time into really learning ICE and I'd like to know what the most valuable concepts are.
Graham is right, locations is good thing to understand. I'd replace "attributes" in my list with "locations"...
For arrays, I guess it depends on whether or not you do any programming. If you do, then arrays aren't that hard to grasp.
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Stephen Blair Sent: December-15-11 1:33 PM To: softim...@listproc.autodesk.com Subject: RE: Top 5 things to learn about ICE?
Here's 5 things off the top of my head...
- Data flow in an ICE tree
- Data context, structure, and type
- ICE trees, the operator stack, and evaluation (eg simulation versus modeling)
- Attributes
- Transforms, matrices, vectors, and coordinate systems
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Simon Van de Lagemaat Sent: December-15-11 1:08 PM To: softim...@listproc.autodesk.com Subject: Top 5 things to learn about ICE?
If you ICE experts out there had to recommend the top 5 (first) things to learn about ICE what would they be? I'm looking for cornerstone/pedestal type foundation items. I'm going to put some time into really learning ICE and I'd like to know what the most valuable concepts are.
I don't program at all. But I'm sure if I bang my head up against that wall enough it will crumble!
Thanks for all these tips guys, this will really help.
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Stephen Blair Sent: December-15-11 10:56 AM To: softim...@listproc.autodesk.com Subject: RE: Top 5 things to learn about ICE?
Graham is right, locations is good thing to understand. I'd replace "attributes" in my list with "locations"...
For arrays, I guess it depends on whether or not you do any programming. If you do, then arrays aren't that hard to grasp.
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Stephen Blair Sent: December-15-11 1:33 PM To: softim...@listproc.autodesk.com Subject: RE: Top 5 things to learn about ICE?
Here's 5 things off the top of my head...
- Data flow in an ICE tree
- Data context, structure, and type
- ICE trees, the operator stack, and evaluation (eg simulation versus modeling)
- Attributes
- Transforms, matrices, vectors, and coordinate systems
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Simon Van de Lagemaat Sent: December-15-11 1:08 PM To: softim...@listproc.autodesk.com Subject: Top 5 things to learn about ICE?
If you ICE experts out there had to recommend the top 5 (first) things to learn about ICE what would they be? I'm looking for cornerstone/pedestal type foundation items. I'm going to put some time into really learning ICE and I'd like to know what the most valuable concepts are.
1 - context, locations, operator stack 2 - strong foundation on vectors, because that's what you are going to manipulate about 80% of the time (at least for me) 3 - encapsulation/modules. Know how to break a problem into small tasks, know when to compound nodes so that they can be used elsewhere. (that's the beauty of ICE) 4 - get used to node naming conventions. They can be a good shortcut to find what you want in the big list of nodes. (Turbulize/Randomize/Test/Modulate/Filter by/etc) 5 - medium to long term: focus on performance! Avoid Repeat nodes, know how to optimize geometry queries, build an intelligent graph, skip unnecessary tests, know that getting the squared length of a vector is faster than getting its length (if all you need is comparision).
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Fabricio Chamon Sent: Thursday, December 15, 2011 9:33 PM To: softim...@listproc.autodesk.com Subject: Re: Top 5 things to learn about ICE?
1 - context, locations, operator stack 2 - strong foundation on vectors, because that's what you are going to manipulate about 80% of the time (at least for me) 3 - encapsulation/modules. Know how to break a problem into small tasks, know when to compound nodes so that they can be used elsewhere. (that's the beauty of ICE) 4 - get used to node naming conventions. They can be a good shortcut to find what you want in the big list of nodes. (Turbulize/Randomize/Test/Modulate/Filter by/etc) 5 - medium to long term: focus on performance! Avoid Repeat nodes, know how to optimize geometry queries, build an intelligent graph, skip unnecessary tests, know that getting the squared length of a vector is faster than getting its length (if all you need is comparision).
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Matt Lind Sent: December-16-11 10:30 AM To: softim...@listproc.autodesk.com Subject: RE: Top 5 things to learn about ICE?
In other words, learn to be a programmer.
Matt
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Fabricio Chamon Sent: Thursday, December 15, 2011 9:33 PM To: softim...@listproc.autodesk.com Subject: Re: Top 5 things to learn about ICE?
1 - context, locations, operator stack 2 - strong foundation on vectors, because that's what you are going to manipulate about 80% of the time (at least for me) 3 - encapsulation/modules. Know how to break a problem into small tasks, know when to compound nodes so that they can be used elsewhere. (that's the beauty of ICE) 4 - get used to node naming conventions. They can be a good shortcut to find what you want in the big list of nodes. (Turbulize/Randomize/Test/Modulate/Filter by/etc) 5 - medium to long term: focus on performance! Avoid Repeat nodes, know how to optimize geometry queries, build an intelligent graph, skip unnecessary tests, know that getting the squared length of a vector is faster than getting its length (if all you need is comparision).
On Sat, Dec 17, 2011 at 5:30 AM, Matt Lind <ml...@carbinestudios.com> wrote: > In other words, learn to be a programmer.****
> ** **
> ** **
> Matt****
> ** **
> ** **
> ** **
> *From:* softimage-boun...@listproc.autodesk.com [mailto: > softimage-boun...@listproc.autodesk.com] *On Behalf Of *Fabricio Chamon > *Sent:* Thursday, December 15, 2011 9:33 PM > *To:* softim...@listproc.autodesk.com > *Subject:* Re: Top 5 things to learn about ICE?****
> ** **
> 1 - context, locations, operator stack****
> 2 - strong foundation on vectors, because that's what you are going to > manipulate about 80% of the time (at least for me)****
> 3 - encapsulation/modules. Know how to break a problem into small tasks, > know when to compound nodes so that they can be used elsewhere. (that's the > beauty of ICE)****
> 4 - get used to node naming conventions. They can be a good shortcut to > find what you want in the big list of nodes. > (Turbulize/Randomize/Test/Modulate/Filter by/etc)****
> 5 - medium to long term: focus on performance! Avoid Repeat nodes, know > how to optimize geometry queries, build an intelligent graph, skip > unnecessary tests, know that getting the squared length of a vector is > faster than getting its length (if all you need is comparision). ****
1. Context 2. Operator stack effects on ICE trees (and the power therein) 3. The 'location' concept 4. Changing your thinking to view ICE as a means to manipulate data in general, which can be used in many creative and surprising ways in XSI (ie: ICE is not just for procedural modeling and simulation/particles.) 5. Translating pseudocode and core math/programming concepts into ICE, which is a lifetime's worth of work (arrays, math/linear algebra/matrices, algorithms and approaches to problems in CG both new and old, etc etc)
To have some meaningful assistance, it would be worth knowing whether you've ever coded before and whether that's the route you're expecting to go with ICE. If you're more of an artist, the list of 5 would be very different than if you're expecting to run it like a coder. I'm confused by the "Context" advice as I have no idea how this is going to help as a start point. As a non coder I would suggest these:-
1. Start with poking around inside some of the demo scenes. Basic Fire, Spawning Fireworks, Particle Forces Surface etc. I'd open these up, dig around in there, remove/replace nodes and generally see what's going on until you're happy you understand what's doing what. 2. Try going into compounds, even the factory ones and see what's happening and even change them. Quite a lot of them are very simple and are there to just package up regular tasks in a neat way. That makes it easy to see what's going on and possibly show how you can build your own. Even poke around in the emit compounds and figure out what's doing what. You'll have a much better understanding of the general structure of ICE workflows that way. 3. Experiment Experiment Experiment. Try to create simple effects and make sure you understand why things are happening. Some of my early tests were all about trying stuff and just seeing what happens. The main thing is to understand how to move particles or points around in a controlled way. ICE isn't just a simulator, so try moving points around yourself manually. 4. Don't be afraid to use loops (Repeat and Repeat with Counter). It's usually frowned on by coders, but if you're a non coder, just have a go. The concept of a loop is very simple and can help in all sorts of situations. There are ways around using them, but like I said, for a non coder it's the concept of a loop that is straight forward to understand. 5. Read and understand the Getting and Setting Data in ICE Trees in the help docs, including the section on Setting Custom Attributes. This is something you'll need to do a lot of, so have a good read of that and try out some of the examples.
That's it. Good luck!! Chris
On 17 December 2011 13:55, Andy Moorer <andymoo...@gmail.com> wrote:
> 1. Context > 2. Operator stack effects on ICE trees (and the power therein) > 3. The 'location' concept > 4. Changing your thinking to view ICE as a means to manipulate data in > general, which can be used in many creative and surprising ways in XSI (ie: > ICE is not just for procedural modeling and simulation/particles.) > 5. Translating pseudocode and core math/programming concepts into ICE, > which is a lifetime's worth of work (arrays, math/linear > algebra/matrices, algorithms and approaches to problems in CG both new and > old, etc etc)
Not really. Programming just happens, as a discipline, to have been at these problems and models for a long enough time to have found patterns, tools, and methodologies to categorise and solve most of them in the most efficient way possible.
Learning to be a problem solver doesn't necessarily require that you become a proficient computer scientist and accrue knowledge of the specifics of several languages and APIs. In fact ICE has made it extremely easy to experiment with and teach about paradigms and models with a visual and immediate feedback that before required a major entry fee in terms of knowing language and an API.
The problems and models don't change, nor do the fundamentals you should learn, just you now get exposed to things that before would have been removed from most users because you don't pay for the entry ticket anymore before bumping into that wall. It's a good thing ;)
On Sat, Dec 17, 2011 at 5:54 AM, Simon Van de Lagemaat <
si...@theembassyvfx.com> wrote: > Ya that's the sticky part :-) ****
> ** **
> *From:* softimage-boun...@listproc.autodesk.com [mailto: > softimage-boun...@listproc.autodesk.com] *On Behalf Of *Matt Lind > *Sent:* December-16-11 10:30 AM > *To:* softim...@listproc.autodesk.com > *Subject:* RE: Top 5 things to learn about ICE?****
In fact, it's a very very good thing! ICE is nothing like programming, it's visual fun!
On Tuesday, 20 December 2011, Raffaele Fragapane <
raffsxsil...@googlemail.com> wrote: > Not really. > Programming just happens, as a discipline, to have been at these problems
and models for a long enough time to have found patterns, tools, and methodologies to categorise and solve most of them in the most efficient way possible.
> Learning to be a problem solver doesn't necessarily require that you
become a proficient computer scientist and accrue knowledge of the specifics of several languages and APIs. In fact ICE has made it extremely easy to experiment with and teach about paradigms and models with a visual and immediate feedback that before required a major entry fee in terms of knowing language and an API.
> The problems and models don't change, nor do the fundamentals you should
learn, just you now get exposed to things that before would have been removed from most users because you don't pay for the entry ticket anymore before bumping into that wall. It's a good thing ;)
> On Sat, Dec 17, 2011 at 5:54 AM, Simon Van de Lagemaat < si...@theembassyvfx.com> wrote:
nah, he has a point, its fun (for some) in comparison to trying to learn Python or C++ or whatever :) instant gratification of the red node means it don't work before you execute your script!
On 20 December 2011 15:07, Andy Moorer <andymoo...@gmail.com> wrote:
In my experience ICE has provided that ease in step that was the missing link before, and more and more people are picking up some sort of programming fundamentals, or even full blown C++. I know it was hugely useful in my workshop for the maths weeks. It's sort of a gateway drug :)
On Tue, Dec 20, 2011 at 9:06 PM, Chris Marshall <chrismarshal...@gmail.com>wrote:
I'll confirm what Raf said, being that I took the workshop and having ICE handy to wrangle together the concepts of the math portion was a huge asset and great visual learning tool. My vague understanding of what vectors were and my obliviousness of how a matrix broke down was quickly remedied.
I've found that prototyping or troubleshooting my Python vector / matrix maths with ICE has been such a life saver. I'm not hesitant any longer on diving into more computational focused tools.
It is a gateway drug and Raf has us hooked up to IVs pumping it into our veins at the office. :)
-------------------------------------------- Eric Thivierge Currently: Digital Artist, Rigging at Animal Logic http://www.ethivierge.com
On Wed, Dec 21, 2011 at 10:59 AM, Raffaele Fragapane <
raffsxsil...@googlemail.com> wrote: > In my experience ICE has provided that ease in step that was the missing > link before, and more and more people are picking up some sort of > programming fundamentals, or even full blown C++. I know it was hugely > useful in my workshop for the maths weeks. > It's sort of a gateway drug :)
> On Tue, Dec 20, 2011 at 9:06 PM, Chris Marshall <chrismarshal...@gmail.com > > wrote:
>> In fact, it's a very very good thing! ICE is nothing like programming, >> it's visual fun!
That is true, with me it was more a gateway drug to Maths & Physics Algorithms and not programming, Several books & repeated exposure and some of it is starting to make sense!
ICE, expanding horizons since XSI 7
On 21 December 2011 00:14, Eric Thivierge <ethivie...@gmail.com> wrote:
> I'll confirm what Raf said, being that I took the workshop and having ICE > handy to wrangle together the concepts of the math portion was a huge asset > and great visual learning tool. My vague understanding of what vectors were > and my obliviousness of how a matrix broke down was quickly remedied.
> I've found that prototyping or troubleshooting my Python vector / matrix > maths with ICE has been such a life saver. I'm not hesitant any longer on > diving into more computational focused tools.
> It is a gateway drug and Raf has us hooked up to IVs pumping it into our > veins at the office. :)
> -------------------------------------------- > Eric Thivierge > Currently: Digital Artist, Rigging at Animal Logic > http://www.ethivierge.com
> On Wed, Dec 21, 2011 at 10:59 AM, Raffaele Fragapane < > raffsxsil...@googlemail.com> wrote:
>> In my experience ICE has provided that ease in step that was the missing >> link before, and more and more people are picking up some sort of >> programming fundamentals, or even full blown C++. I know it was hugely >> useful in my workshop for the maths weeks. >> It's sort of a gateway drug :)
>> On Tue, Dec 20, 2011 at 9:06 PM, Chris Marshall < >> chrismarshal...@gmail.com> wrote:
>>> In fact, it's a very very good thing! ICE is nothing like programming, >>> it's visual fun!
No. ICE is a framework for creating operators which abstracts the low level details away from the user so he/she can focus on the task at hand as it relates to 3D. In order to make effective operators, the user needs to be knowledgeable in the area of what ICE is doing at a high level so he/she can use the available resources properly. Just like with scripting you can be functional without that knowledge, but you'll be far from efficient - or as a friend of mine would always say, "know just enough to be dangerous".
To use an analogy, just because you know how to speak French doesn't make you a French poet. Knowing the language is not enough. You have to understand the inner workings of all the pieces to know how to put them together in a cohesive manner. In the case of ICE, that means getting more versed in topics of computer science, and possibly mathematics and/or physics depending on what you're doing.
Matt
From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Eric Thivierge Sent: Friday, December 16, 2011 11:16 PM To: softim...@listproc.autodesk.com Subject: Re: Top 5 things to learn about ICE?
ICE == a programming language no?
-------------------------------------------- Eric Thivierge Currently: Digital Artist, Rigging at Animal Logic http://www.ethivierge.com
On Sat, Dec 17, 2011 at 5:30 AM, Matt Lind <ml...@carbinestudios.com<mailto:ml...@carbinestudios.com>> wrote: In other words, learn to be a programmer.
Matt
From: softimage-boun...@listproc.autodesk.com<mailto:softimage-boun...@listproc.autodesk.com> [mailto:softimage-boun...@listproc.autodesk.com<mailto:softimage-boun...@listproc.autodesk.com>] On Behalf Of Fabricio Chamon Sent: Thursday, December 15, 2011 9:33 PM To: softim...@listproc.autodesk.com<mailto:softim...@listproc.autodesk.com> Subject: Re: Top 5 things to learn about ICE?
1 - context, locations, operator stack 2 - strong foundation on vectors, because that's what you are going to manipulate about 80% of the time (at least for me) 3 - encapsulation/modules. Know how to break a problem into small tasks, know when to compound nodes so that they can be used elsewhere. (that's the beauty of ICE) 4 - get used to node naming conventions. They can be a good shortcut to find what you want in the big list of nodes. (Turbulize/Randomize/Test/Modulate/Filter by/etc) 5 - medium to long term: focus on performance! Avoid Repeat nodes, know how to optimize geometry queries, build an intelligent graph, skip unnecessary tests, know that getting the squared length of a vector is faster than getting its length (if all you need is comparision).
If ever there was a way of scaring animators away from ICE, Matt, your comment is it! What we need is for everyone to play with ice and not even think about the maths side of it. How far they get will be down to their own ability. I really don't regard myself as someone who understands the maths side of things, but I get by because of the way ice works. It bridges the gap between artist and coder, but also allows coders to dig deeper. Just because I'm not a French Poet doesn't mean I can't kick some serious ICE butt!! ;-)
> To use an analogy, just because you know how to speak French doesn’t make
you a French poet. Knowing the language is not enough. You have to understand the inner workings of all the pieces to know how to put them together in a cohesive manner. In the case of ICE, that means getting more versed in topics of computer science, and possibly mathematics and/or physics depending on what you’re doing.
> Matt
--
Chris Marshall Mint Motion Limited 029 2002 5762 07730 533 115 www.mintmotion.co.uk
> If ever there was a way of scaring animators away from ICE, Matt, > your comment is it! What we need is for everyone to play with ice and > not even think about the maths side of it. How far they get will be > down to their own ability. I really don't regard myself as someone who > understands the maths side of things, but I get by because of the way > ice works. It bridges the gap between artist and coder, but also > allows coders to dig deeper. > Just because I'm not a French Poet doesn't mean I can't kick some > serious ICE butt!! ;-)
> > To use an analogy, just because you know how to speak French doesn�t > make you a French poet. Knowing the language is not enough. You have > to understand the inner workings of all the pieces to know how to put > them together in a cohesive manner. In the case of ICE, that means > getting more versed in topics of computer science, and possibly > mathematics and/or physics depending on what you�re doing.
I don't think Matts comments are scary at all. Seems like a fair breakdown of two possible approaches to ICE. My experience has shown that although both camps, those with and without programming knowledge, can usually get the job done. There is, however, a pretty big difference in how the trees are put together. Also, the time taken to reach the desired result is exponentially proportional to the complexity of the job.
All that's being said is that - yes, you can play with ICE all you want, but having some technical knowledge is only going to increase the bounds of your playground.
On 22 December 2011 09:25, Sebastian Kowalski <l...@sekow.com> wrote:
> If ever there was a way of scaring animators away from ICE, Matt, your > comment is it! What we need is for everyone to play with ice and not even > think about the maths side of it. How far they get will be down to their > own ability. I really don't regard myself as someone who understands the > maths side of things, but I get by because of the way ice works. It bridges > the gap between artist and coder, but also allows coders to dig deeper. > Just because I'm not a French Poet doesn't mean I can't kick some serious > ICE butt!! ;-)
> > To use an analogy, just because you know how to speak French doesn’t > make you a French poet. Knowing the language is not enough. You have to > understand the inner workings of all the pieces to know how to put them > together in a cohesive manner. In the case of ICE, that means getting more > versed in topics of computer science, and possibly mathematics and/or > physics depending on what you’re doing.
> > Matt
> --
> Chris Marshall > Mint Motion Limited > 029 2002 5762 > 07730 533 115 > www.mintmotion.co.uk
--
----- "Consistency is the last resort of the unimaginative" - O. Wilde
Always when learning something, attack to your weak spot. In case youre well grounded with programming, learn to do something visual/practical with it. In case youre artist that can`t program, try to make system that eases the your everyday tasks. ect.
Famous quote by Einstein "The definition of insanity is doing the same thing over and over again and expecting different results"
> I don't think Matts comments are scary at all. Seems like a fair breakdown > of two possible approaches to ICE. My experience has shown that although > both camps, those with and without programming knowledge, can usually get > the job done. There is, however, a pretty big difference in how the trees > are put together. Also, the time taken to reach the desired result is > exponentially proportional to the complexity of the job.
> All that's being said is that - yes, you can play with ICE all you want, > but having some technical knowledge is only going to increase the bounds of > your playground.
> On 22 December 2011 09:25, Sebastian Kowalski <l...@sekow.com> wrote:
>> ** >> AMEN!
>> Am 22/12/2011 09:48, schrieb Chris Marshall:
>> If ever there was a way of scaring animators away from ICE, Matt, your >> comment is it! What we need is for everyone to play with ice and not even >> think about the maths side of it. How far they get will be down to their >> own ability. I really don't regard myself as someone who understands the >> maths side of things, but I get by because of the way ice works. It bridges >> the gap between artist and coder, but also allows coders to dig deeper. >> Just because I'm not a French Poet doesn't mean I can't kick some serious >> ICE butt!! ;-)
>> > To use an analogy, just because you know how to speak French doesn’t >> make you a French poet. Knowing the language is not enough. You have to >> understand the inner workings of all the pieces to know how to put them >> together in a cohesive manner. In the case of ICE, that means getting more >> versed in topics of computer science, and possibly mathematics and/or >> physics depending on what you’re doing.
>> > Matt
>> --
>> Chris Marshall >> Mint Motion Limited >> 029 2002 5762 >> 07730 533 115 >> www.mintmotion.co.uk
> --
> ----- > "Consistency is the last resort of the unimaginative" - O. Wilde
On a side note (pun not intended), I've recently realized that my skills as a coder have increased substantially because of my experience with ICE. I had to write a script to plot out a simulation and do all the math within the script (It involved a lot of converting between different coordinate spaces for different parts of the calculations).
Because of experience interacting with math in ICE, I had the vector and transform manipulations down so well, I was able to do all of it in my head, and the simulation worked on the first run of the script.
It was a nice moment for me, like unexpectedly recognizing that the training wheels were off the bicycle.
