PDA

View Full Version : Node reference material



CAClark
03-22-2008, 06:03 AM
Does anyone know if there are any genuinely GOOD materials out there explaining in ENGLISH what the various nodes do, and why you would use them etc...

I don't want math geek mumbo jumbo, I just need to know why and how :D

Cheers!

Mitja
03-22-2008, 08:34 AM
I don't want math geek mumbo jumbo...
I just wanted to say that these few words made me laugh for about ten mins :D ...
Anyway I dunno the answer to your question.

cheers!



...mumbo jumbo... :lol:

Matt
03-22-2008, 09:30 AM
Does anyone know if there are any genuinely GOOD materials out there explaining in ENGLISH what the various nodes do, and why you would use them etc...

I don't want math geek mumbo jumbo, I just need to know why and how :D

Cheers!

This is something I've been after for a while too!

In a perfect world I'd love a well presented library of small video tutorials that cover EVERY node, what it does, where you'd use it, how it works with other nodes.

Would be invaluable!

Weepul
03-22-2008, 03:53 PM
I've been thinking about making something like this... The #1 problem with making clear and concise mini-tutorials regarding what each node is used for is that they can be used in any way you want that gets the result you'd like! :p

Maybe I could give it a test run...does anyone have a specific sort of node usage topic they'd like covered?

CAClark
03-22-2008, 05:03 PM
True, but each node has to have had a base function in mind at the beginning :D

Cheers!

CAClark
03-22-2008, 05:47 PM
does anyone have a specific sort of node usage topic they'd like covered?

All of them really. Its amazing that NT haven't already got this under way. I haven't got time to just connect nodes together and see what happens in the hope that i suddently stumble across the effect I'm looking for. Time is money, and more importantly, time is finite.

Cheers!

multipass
03-22-2008, 10:06 PM
im curious as to what the heck the function input of a node does?

RedBull
03-22-2008, 11:15 PM
im curious as to what the heck the function input of a node does?

Taken from the LW9 Surface and Render Manual.

"Functions are the only type of connection
that are actually bidirectional. This bidirectionality is
transparent to the user but Function nodes receive
information from the node they are modifying, alter the
information in some way and then send the results back
to the node they are connecting to. A good general rule
of thumb to remember about function type nodes is just
to not intermix them with dissimilar types. Keep your
Functions in the yellow and all will be mellow."

Page 89, goes into more detail.

Steamthrower
03-24-2008, 01:40 PM
Does anyone know if there are any genuinely GOOD materials out there explaining in ENGLISH what the various nodes do, and why you would use them etc...

Easy Craig, let me give you a quick run through of nodes.

The following singularly perturbed Dirichlet problem ε[2]Δu - u + |u|[p-]u = 0, in Ω, u = 0 on ∂Ω has a nodal solution u[ε] which has the least energy among all nodal solutions. Moreover, it is shown that u[ε] has exactly one local maximum point P[ε][1] with a positive value and one local minimum point P[ε][2] with a negative value, and as ε → 0, formula math where φ(P[1], P[2]) = min(}P[1] - P[2]}/2,d(P[1], ∂Ω ), d(P[2], ∂Ω)). The following question naturally arises: where is the nodal surface {u[ε](x) = 0}? In this paper, we give an answer in the case of the unit ball Ω = B[1] (0). In particular, we show that for ε sufficiently small, Pf, P[ε][2] and the origin must lie on a line. Without loss of generality, we may assume that this line is the x[1]-axis. Then u[ε] must be even in Xj, j = 2,..., N, and odd in x[1]. As a consequence, we show that {u[ε](x) = 0} = {x ∈ B[1](0) x[1] = 0}. Our proof is divided into two steps: first, by using the method of moving planes, we show that P[ε][1], P[ε][2] and the origin must lie on the x[1]-axis and u[ε] must be even in Xj, j = 2,..., N. Then, using the Liapunov-Schmidt reduction method, we prove the uniqueness of u[ε] (which implies the odd symmetry of u[ε] in x[1]). Similar results are also proved for the problem with Neumann boundary conditions.

Matt
03-24-2008, 02:03 PM
LOL! Ahhhhh, it all makes perfect sense now!

:D

warrenwc
03-24-2008, 03:22 PM
Easy Craig, let me give you a quick run through of nodes.

The following singularly perturbed Dirichlet problem ε[2]Δu - u + |u|[p-]u = 0, in Ω, u = 0 on ∂Ω has a nodal solution u[ε] which has the least energy among all nodal solutions. Moreover, it is shown that u[ε] has exactly one local maximum point P[ε][1] with a positive value and one local minimum point P[ε][2] with a negative value, ...(blahblah)

Hee Hee!
You forgot to account for the Shmendrick singularity.:D

PhantomPhish
03-24-2008, 06:15 PM
im curious as to what the heck the function input of a node does?

Functions allow you to map the output of a node in various ways. For example, you might have a Turbulence texture, ranging from 0 ( black ) to 1 ( white) with shades in the middle.

Using a Box Step node, you can tighten up the contrast of the Turbulence to make like a cow-print type of pattern, with mostly black and white blotches, and just a thin sliver of grey in between. Like adjusting curves in Photoshop.

PhantomPhish
03-24-2008, 06:17 PM
I've been thinking about making something like this... The #1 problem with making clear and concise mini-tutorials regarding what each node is used for is that they can be used in any way you want that gets the result you'd like! :p

I agree, a tutorial 'per node' makes little sense for most of them. Kurv studios has a node DVD for sale and there are some other free mini tutes around...

Hopper
03-24-2008, 08:22 PM
fprime is the best tool to learn from
It sure sounds like it. And from what I've seen of some of the tutorial videos that use it, it's a must for some. Too bad I'm just a hobbiest and my wife would beat me with my own arm if I bought it.

I still haven't rendered her "new livingroom" yet, so maybe I can use that as an excuse! :D

She has my nuts in the freezer and she lets me keep my Man Card just for show.

JBT27
03-27-2008, 03:57 PM
All of them really. Its amazing that NT haven't already got this under way. I haven't got time to just connect nodes together and see what happens in the hope that i suddently stumble across the effect I'm looking for. Time is money, and more importantly, time is finite.

Cheers!

I don't know what the latest docs are, but the ones I have don't include Item Info, and some nodes are not fully explained, like Vector Scalar - the modes are not explained. Are there newer ones? The links were all broken earlier when I checked.

Video tuts would be good, and surely someone can concoct reasons for using all of them to show off specifically what they are for, even though I understand that it's open season on how you use each one.

Things like Dot, which is cosine something or other - sounds lame but unless I see a reason to use that, ie. a sample of it in action making a surface and therefore telling me what it does and why it's useful, I won't touch it.

You can experiment for hours and get nowhere with some of these nodes, mostly because there is no clear explanation of when and why to use them.

This afternoon, after about five hours of experimenting, I managed to create a setup to have the colour of one object controlled by the geometry of another below it, for terrains. I kind of get how it works, but I'll admit I stumbled on it through trying things out that I thought sounded like they might be what I wanted. But it uses Vector Scalar, and the Length mode works in the setup, but the YChannel does not - frankly I have absolutely no idea why and the docs ain't helping :thumbsdow

Julian.

RedBull
03-27-2008, 05:39 PM
I don't know what the latest docs are, but the ones I have don't include Item Info, and some nodes are not fully explained, like Vector Scalar - the modes are not explained. Are there newer ones? The links were all broken earlier when I checked.


You can experiment for hours and get nowhere with some of these nodes, mostly because there is no clear explanation of when and why to use them.



Sorry i disagree with everyone here, the LW9 Surface and Render manual, has quite well written info, as well as several tutorials explaining in depth the exact reason to and not to use them. It's one of the few well written parts of the LW9 manual.

The Item Node functions are explained between pages: 268 and 270.

On top of that playing, trial and error etc, i mean it really shouldn't be up to NT to give a video tutorial of every node, try and learn some yourselves you lazy bastards! The manual is a good start and most of seem quite self-explanatory anyway and i don't claim to be any expert, but i think people must be bloody lazy not to read two pages in a manual....

JBT27
03-28-2008, 05:45 AM
Though it's surprising how quickly you understand something when it's shown in context alongside a real-life example.

That's my oversight on the Item Info docs.....I was using the Web Help, where it still has not been included.....I redownloaded the Web Help yesterday and installed it - still the same and most of the files that comprise it are almost a year old. As the words are written for the PDF, you wonder why no-one has included them by now in the Web Help.....

Mimicking most of the equivalent layered surfacing operations with nodal is pretty easy I reckon - there were stumbling blocks like Texture Displacement, how to mimic layers set to alpha, the Distance To modes of gradients. Those were covered by forum posts, and William Vaughan did a video on Distance To using nodes, but only the basic one object to another.....the more complex and alot less obvious node setup for doing X, Y and Z Distance To have not been covered anywhere by NT, that I know of.....doubtless I'll be proved wrong :) That was also ultimately covered by a forum thread.

I'd agree that NT don't have a responsibility to provide us with tutorials for everything.....'everything' does not require a tutorial, there's alot that's well within reach of every user and there are great chunks which should be common-sense.

The bit that I find frustrating are the nodes that are maths functions, like Normalize - from the Web Help:

'Vector normalization involves taking a vector and finding its length. All three elements of the vector are then divided by this length. This has the effect of creating a normalized vector, that is, a vector with the same direction as the input vector but with a unit length.'

OK.....fine.....so what?? I don't think it would be out of order to suggest that NT, in time, provide some tutorials as to why this stuff is useful in the creation of cgi, and where you might use it. Back to context again - you can't and shouldn't cover everything in tuts, but some relevant hints and examples would not go amiss. For those of us who do bother to sit down and learn what we work with, it could save alot of time and push us in the right direction.

Julian.

Captain Obvious
03-28-2008, 08:03 AM
Vector normalization and stuff like that is only really useful if you're creating a node tree that already relies on heavy use of mathematics, and unless you understand the concept of normalization, I see no point in using it.

JBT27
03-28-2008, 09:10 AM
Vector normalization and stuff like that is only really useful if you're creating a node tree that already relies on heavy use of mathematics, and unless you understand the concept of normalization, I see no point in using it.

Thanks.....fair point.....I need to be more hard-headed about the time spent on some of this stuff - big difference between learning the 'regular' nodes, which is not all that hard for the most part, versus becoming fluent enough in the relevant math to be able to use the rest. There's not enough hours in the day already :)

But still, there is one thing I'm stuck with - thought I'd solved it, but apparently not - so I'll post a thread about that.

Julian.

CAClark
03-29-2008, 04:30 PM
try and learn some yourselves you lazy bastards!

That's the kind of response that quite frankly pisses me off. Well done for you, you understand nodes, I'm made up for you.

CAClark
03-29-2008, 04:33 PM
I'd agree that NT don't have a responsibility to provide us with tutorials for everything.

Very true, all it needs is a clear concise explanation of what it does, and what affect it has on key nodes it's intended to be used for. IMO that isn't a tutorial, it's reference material that provides the user with the ability to know which nodes are most likely to do what they are looking for.

I'm happy to pay for tutorial material if it's good, but the Kurv nodes DVD was a big let down, when I had hoped it would be the key to grasping nodes.

Cheers!

Jim M
03-29-2008, 08:11 PM
I think people find it difficult coming from a different model of surfacing.
There is a general difficulty in getting to the point where nodes can be used fluidly and without feeling like you are going through treacle trying to 'work it out'.
Often people don't have a specific question in mind, but a general sense of difficulty, and this is natural with such a system as it is very different in its operation and depth.
I am no expert but my advice would be... learn slowly until the penny drops. You cant expect to learn lightwave in a day and due to the depth of nodal surfacing I would take a similar attitude. Start with what you know within the node system, and build slowly from there.

One difficulty or obstacle to fluid use is that there is a variety of complexity with the nodal system that people miss. Some of it easy and some of it obscure. Define what is familiar, find the easy stuff !

I understand probably 90% of what it does, I understand 80% percent of the inputs, and I understand 50% of the actual real life use... now what do I do with a preprocess output .. etc etc ...

A lot I figured out myself but simple examples are all I require to build upon, and be creative. This is the same for most people, probably. There is no mystery to nodes, though some are esoteric it seems. You just have to use them at certain phases of the moon.

Having an example of something we've never used will give us ideas to incorporate into our workflow. Rather than knowing somethings function, and trying to work out how to use it.

the end

RedBull
03-30-2008, 02:08 AM
That's the kind of response that quite frankly pisses me off. Well done for you, you understand nodes, I'm made up for you.

"I'm made up for you" Obviously you write as well as you read.
Oh, no we don't wanna piss off the person who cannot even read a manual.
Thanks, yeah reading is really hard and the manual that NT employees made for retards like yourself, should really have paint by number to teach brainless farts like yourself how to RTFM....

And it's clear from your initial comments you have not read the manual, but had the nerve to attack NT for not teaching you EVERY nodes use, when it's explained in detail for idiots just like yourself.

If you had made an attempt to read or understand perhaps, other would help you.... But with your attitude, I imagine surfacing with any system is going to be to complicated for your small little brain to cope with, perhaps try a different industry.... Or stick with just one discipline like Modeling... :)

Jim M
03-30-2008, 05:44 AM
Well it seems another thread has degenerated... someones been drinking too much caffeine ...


when it's explained in detail for idiots just like yourself.
Sometimes detail isn't the full picture, just lots of separate pieces.

Having seen CAClarks work, I am not alone in thinking he shouldn't find another industry to work in. And moreover he is not alone in his experience with nodes.

People who try to belittle the ignorant, by calling them idiots ?? Ooh the irony...

JBT27
03-30-2008, 05:44 AM
I think people find it difficult coming from a different model of surfacing.
There is a general difficulty in getting to the point where nodes can be used fluidly and without feeling like you are going through treacle trying to 'work it out'.
Often people don't have a specific question in mind, but a general sense of difficulty, and this is natural with such a system as it is very different in its operation and depth.
I am no expert but my advice would be... learn slowly until the penny drops. You cant expect to learn lightwave in a day and due to the depth of nodal surfacing I would take a similar attitude. Start with what you know within the node system, and build slowly from there.

One difficulty or obstacle to fluid use is that there is a variety of complexity with the nodal system that people miss. Some of it easy and some of it obscure. Define what is familiar, find the easy stuff !

I understand probably 90% of what it does, I understand 80% percent of the inputs, and I understand 50% of the actual real life use... now what do I do with a preprocess output .. etc etc ...

A lot I figured out myself but simple examples are all I require to build upon, and be creative. This is the same for most people, probably. There is no mystery to nodes, though some are esoteric it seems. You just have to use them at certain phases of the moon.

Having an example of something we've never used will give us ideas to incorporate into our workflow. Rather than knowing somethings function, and trying to work out how to use it.

the end

I think that's useful comment.

The best approach is to reach the point with nodal that entirely allows you to match what you can do with layers. There is a tendency to look at all those nodes, read all the descriptions of them, get overwhelmed and then frustrated, and ultimately get nowhere with it.

To be honest, matching basic layered surfacing is very easy. Once you hit that point, it's not a giant leap to understand just how powerful nodes really are when you start adding Mixer, say, and some basic math nodes for tighter control - that may not be rocket-science, nor win you an Oscar, but it will give you alot more flexibility than layered. Arguably, that level of nodal use is 'all' many artists will ever need.

As Captain Obvious pointed out in his reply to my post above, though not in so many words, the old addage: "Don't play with stuff you don't understand!"

There are swathes of nodes that apparently do require reasonable knowledge and experience of the relevant math. This is one of those cases where if you don't know what they are, you probably don't need to worry about it. Other side of that coin is wanting to know what they are and how to use them, but that's a personal choice to go and learn the math you need to do that; that's not NT's responsibility, in so far as those nodes are concerned. They are there for the people who do know and can do stuff with them.

I'd really like to know this stuff and work with those nodes, but I struggle with maths. I may try to get into that at some point, but I also know that I don't need to in order to expand my surfacing with nodes and do the work that I do. Again, it's not NT's responsibility to teach me the maths.

That said, and as I said earlier, I think more tutorial coverage on project driven nodal tasks wouldn't go amiss. Like William Vaughan's video on doing 'distance to' gradients with nodal, which is very different to layered. Or what I started my thread on about colouring one surface according to the distance of another, along a single axis. After wasting the best part of three days last week, getting very frustrated, I posted the thread, and had the answer within a couple of hours, in one of Denis's nodes from DP Kit.

There are going to be many occasions when RTFM just isn't enough, and that's where more tutorials and these forums come into play.

Julian.

colkai
03-30-2008, 05:47 AM
"I'm made up for you" Obviously you write as well as you read.
Not wanting to get into anything but in case you aren't UK based. ;)

Dunno if you know, but actually, that's a common turn of phrase, meaning, in a less than sincere manner, that the person is happy you have the thing in question, be it an item or otherwise.

As in "oh, so we're all fired but you get a pay rise, I'm made up for you!". :D

colkai
03-30-2008, 05:55 AM
IAs Captain Obvious pointed out in his reply to my post above, though not in so many words, the old addage: "Don't play with stuff you don't understand!"
Julian.

You know, I heartily disagree with this because if you don't tackle the stuff you don't understand, how can you progress?

Nodes, like anything else, just takes a shed load of time and energy to wrap your head around everything. Now, if one were to say, start with the basics and work your way up, that's a different story.

For me, it would be, muck about, try to replicate a "std" surface, then see how other nodes impact that. There are one or two useful nodes tuts out there which give a taste of things and help one gain a foothold on the ladder, but I think most folks are put off simply because sometimes, it seems a heck of a lot of work to get what std surfacing and layers can produce quickly.

Of course, nodes then add all sorts of toys on top of that, but as with learning other packages, sometimes, it's hard to see why it should be done or why it is so (seemingly) complex.

When I was learning Motionbuilder, I knew Jack about it, but banging my head against brick walls and fiddling meant I ended up where now folks continually ask me how to do some stuff as it, like nodes, has some very "grey" areas and sometimes, there really is no substitute for going beyond ones comfort zone.

Even if you end up with a few more grey hairs / less hairs (delete as applicable ;) )

Jim M
03-30-2008, 06:13 AM
Agreed Colkai!
+I might PM you later when I get stuck in MB :)

Captain Obvious
03-30-2008, 06:48 AM
As Captain Obvious pointed out in his reply to my post above, though not in so many words, the old addage: "Don't play with stuff you don't understand!"
You couldn't have misunderstood me more if you tried. My point was actually more along the lines of don't break your brain trying to figure out the really esoteric stuff first -- start with the basics. You only need to worry about dot products and vector normalization and stuff like that if you're building your own lambert shaders and stuff like that inside Nodal. You can do it and it's a lot of fun, but it's not the sort of thing I would recommend doing until you're very comfortable with the concept of nodes.

Start with something simple. Then do something a bit more complicated. Then something even more complicated. Then re-do all of them with the lessons learned.

Once you're able to use Nodal to do anything you could do in the old system, THEN you're ready to start doing the really freaky stuff that you simply CANNOT do in Layers.

JBT27
03-30-2008, 06:57 AM
You know, I heartily disagree with this because if you don't tackle the stuff you don't understand, how can you progress?

Nodes, like anything else, just takes a shed load of time and energy to wrap your head around everything. Now, if one were to say, start with the basics and work your way up, that's a different story.

<snip to save space!!>

Even if you end up with a few more grey hairs / less hairs (delete as applicable ;) )

Oh, actually.....quite right.....I heartily disagree with myself now you point out how I said what I said :D

So basically, yeah.....what you said!

.....except that I guess I am specifically aiming at the maths nodes - the ones that do a math thing, like sine, cosine, tangent and so on. For the most part I am assuming that they really are small cogs in a 'big wheel', providing you know what that 'big wheel' is going to be. Figuring through surfacing using the majority of nodes is one thing, that you can work through and experiment with and get somewhere. Knowing the maths to wield those other nodes in useful trees is a whole other thing, and a whole lot of learning and understanding, and indeed aptitude, if you can't already.

I know what sine, cosine and tangent are, by the way, but as yet I cannot see how they might be useful in a nodal surface tree - I don't know what normalize is or does - I need to find out and understand, and importantly want to, even though I do struggle with maths :)

I'm hoping and guessing that it's not as hard as I convince myself it might be, nor will take too long to learn.....I have quite a few grey hairs and a fair few will be to do with age :D

Julian.

JBT27
03-30-2008, 07:02 AM
You couldn't have misunderstood me more if you tried. My point was actually more along the lines of don't break your brain trying to figure out the really esoteric stuff first -- start with the basics. You only need to worry about dot products and vector normalization and stuff like that if you're building your own lambert shaders and stuff like that inside Nodal. You can do it and it's a lot of fun, but it's not the sort of thing I would recommend doing until you're very comfortable with the concept of nodes.

Start with something simple. Then do something a bit more complicated. Then something even more complicated. Then re-do all of them with the lessons learned.

Once you're able to use Nodal to do anything you could do in the old system, THEN you're ready to start doing the really freaky stuff that you simply CANNOT do in Layers.

Oops - I did understand but replied very poorly - apologies.

Yes, good and sound advice - thanks - that's what I will do.

I am getting much more comfortable with building basic stuff and actually love using nodal - understand what's going on with what I can do, and want to know more, as I replied to colkai.

One step at a time..... :)

Thanks again!

Julian.

colkai
03-30-2008, 08:26 AM
Oh, actually.....quite right.....I heartily disagree with myself now you point out how I said what I said :D

So basically, yeah.....what you said!
Hey, no worries. :)


.....except that I guess I am specifically aiming at the maths nodes - the ones that do a math thing, like sine, cosine, tangent and so on.
...
I know what sine, cosine and tangent are, by the way, but as yet I cannot see how they might be useful in a nodal surface tree -

Julian.

Yeah, here I tend to do wierd things like google, not on nodes, but on trying to find what sort of thing such functions could be used for. A bizzare approach I know and one that can take a lot of digging.
Even then, some functions are like "say whu?" so where possible, I try to look at node presets and reverse engineer them, mucking about with function to see the impact. Even stripping it down so I can see just that section.

Sometimes it works, other times it can get very ugly and I end up twice as confused, but hey, at my age, I can always write that off as being "because I'm old". ;)

As for trying to convince yourself about these things, right there with ya. :D

jameswillmott
03-30-2008, 06:07 PM
I'm happy to pay for tutorial material if it's good, but the Kurv nodes DVD was a big let down, when I had hoped it would be the key to grasping nodes.


Wow, this is very disappointing to hear. I authored that DVD and I'm sorry you didn't get much out of it. Please private message me so we can go through what you expected to learn, I can figure out why the material wasn't up to your expectations, and if you have any questions about nodes I can hopefully answer them in a way that gives you some value.

My apologies.

kopperdrake
03-31-2008, 04:22 AM
I feel your pain Craig. I freaked at Vue's node-based texturing, probably down to infamiliarity, so when LightWave introduced them I freaked again. Figuring I'd best knuckle down and get to grips with them I bought the Kurv DVD but it really didn't help things click into place. I've just run a 2-day project where I really wanted to use the glass node, but try as I might I could not tweak the results as quickly and easily as I could in standard surfacing so I ended up using the old gradient technique. I know it takes time to learn, my point is that often you don't know which bit you need to learn until the project happens along, and then a deadline negates any learning time you might have had.

Not all of us have weeks on end for trial and error. We're not lazy, heaven knows I literally haven't had a full day off since last year, it just boils down to the fact it would be great for a good handful of tutorials that held your hand through some node-based surfacing projects which introduced each nodein as clear to real English as possible.

If someone made a comprehensive set of tutorials on nodes I could watch in bed I'd buy them. Sad but true...gawd knows why my wife puts up with me (phwoar...nodes...phwoar...) ;)

PhantomPhish
03-31-2008, 05:36 AM
Not all of us have weeks on end for trial and error. We're not lazy, heaven knows I literally haven't had a full day off since last year, it just boils down to the fact it would be great for a good handful of tutorials that held your hand through some node-based surfacing projects which introduced each nodein as clear to real English as possible.

The problem is, nodes aren't that useful by themselves, once you learn how to use them together is when they start becoming really powerful. Thats the whole point of node based surfacing.

JBT27
03-31-2008, 06:02 AM
I feel your pain Craig. I freaked at Vue's node-based texturing, probably down to infamiliarity, so when LightWave introduced them I freaked again. Figuring I'd best knuckle down and get to grips with them I bought the Kurv DVD but it really didn't help things click into place. I've just run a 2-day project where I really wanted to use the glass node, but try as I might I could not tweak the results as quickly and easily as I could in standard surfacing so I ended up using the old gradient technique. I know it takes time to learn, my point is that often you don't know which bit you need to learn until the project happens along, and then a deadline negates any learning time you might have had.

Not all of us have weeks on end for trial and error. We're not lazy, heaven knows I literally haven't had a full day off since last year, it just boils down to the fact it would be great for a good handful of tutorials that held your hand through some node-based surfacing projects which introduced each nodein as clear to real English as possible.

If someone made a comprehensive set of tutorials on nodes I could watch in bed I'd buy them. Sad but true...gawd knows why my wife puts up with me (phwoar...nodes...phwoar...) ;)

Project-based tutorials would be especially useful for nodes.....probably go so far as to say essential - I'd pay for such things. I say again that NT ought to consider some video training, hopefully by William Vaughan, that covers some intermediate stuff.

As PhantomPhish says, by themselves they are not useful, abstract.....but then they are not intended to be useful on their own for the most part.

But then what projects would be universally useful?

Perhaps a list of suggestions? What do people do the most, and are most confounded by nodally?

What about that fabulous lava someone did back in the early days - tough to do with layered - that would be a good nodal project, blow by blow.

Julian.

Chuck
03-31-2008, 05:46 PM
The format in our documentation of Nodes is that each Node is covered with:

A basic description,
An explanation of the inputs and outputs on the node,
An explanation of the edit panel for the node,
A usage example "mini-tutorial."


The best way that I can see for us to improve on the documentation is to have very specific and detailed feedback on the existing documentation.

Are there nodes for which documentation is not present in the original text or the "What's New" documents for the v9.x series?
For which nodes do you find the documentation not clear enough?
Which specific elements of the documentation for a given node are confusing to you?
What are the questions you have about those elements?
Are the explanations too technical, too hard to follow?

For which nodes are the Usage Examples not sufficient?

Is it the case that the given example itself does not assist you in understanding how to use the node?
Is the usage example not clearly explained?
Or is it the case that the given example is useful and clearly explained, but the node really needs additional usage examples for a more thorough understanding?


I've passed along some of the comments here on items not being present in the HTML docs that are covered in the PDF manual. Our docs staff will work on getting the HTML docs updated.

For those of you who can take the time to assist in providing detailed feedback, we'll do our best to improve the text and to add supporting usage examples as needed and requested, and you'll very certainly have our appreciation for your contribution to our efforts.

Chuck
03-31-2008, 05:50 PM
Project-based tutorials would be especially useful for nodes.....probably go so far as to say essential - I'd pay for such things. I say again that NT ought to consider some video training, hopefully by William Vaughan, that covers some intermediate stuff.


William's first 24 hour run of tutorials includes 13 Node tutorials, and he will be resuming tutorial production soon (he's in the midst of a major production project at the moment).

JBT27
04-01-2008, 03:18 AM
William's first 24 hour run of tutorials includes 13 Node tutorials, and he will be resuming tutorial production soon (he's in the midst of a major production project at the moment).

This is all very good news!

I knew about the existing tutorials (geez.....I even carry them around on my iPod - so does my business partner!), but more tutorials are always very welcome.

I for one will be happy to contribute to these discussions.

Julian.

Glendalough
04-01-2008, 07:37 AM
... Time is money, and more importantly, time is finite.

Cheers!

Well this is one of the best threads here since LW 9 and nodes have come out!

Think it is about time something was done in this area. Have been doing LW9.x for over 6 months and still barely understand the nodes. Just use presets and examples on the forum to build on in a haphazard manner.

What we don't want and there is far too much of:

"you can take the Whatever Node and plug this into another Whatever or Entity and then this can be Combined with any Animation Parameter.....this is just a small glimpse of the unlimited possibilities ..."

Badly explained, but for another example, say:

You can take the garden hose into the kitchen and have a 3rd source of water (after hot and cold taps) perform water displays, the possibilities are endless, and don't forget you can also take the hose into any room in the house, you are only limited by it's length and the water pressure...

But starting from the very beginning, first inconsistency (confusion) that I noticed :

1. When you use nodes, ie check the node box on the original surface editor, none of the original surfacing attributes work EXCEPT bump???

2. If you hook a material node up to the final surface tab, all other inputs on final tab don't seem to do anything. See image, this is a real bummer, as the possibilities here are endless for nothing happening at all.

3. If material is set up correctly, how does one augment the surface by using the shader nodes? Where do these figure?

Need specific (1st rate, no compromise) examples of exact situations and precisely how they work. What we don't need are generalizations.

JBT27
04-01-2008, 10:20 AM
OK, I'm going back to this whole issue of just how much maths you need to work with nodes.....which is obviously like opening a debate on the length of a piece of string :)

This goes to some nodes, as has been pointed out already, like Normalize and Dot, where you really do have to understand what the math is behind them and how they work together.....and therefore why and where you would use them.

It may sound daft, but I see no harm in NT providing some paragraphs as to what these are mostly intended for, and that they require a certain proficiency in maths to be able to use them. So as Captain Obvious said, if you are building your own shaders, like Lambert, or a raytracer (and there's a thread on that somewhere I think), then these are for you. I suspect the vast majority of artists will not be doing this.

There will be some who would like to know more, but as I said earlier, I don't believe it's NTs responsibility to offer maths tutoring at that level.

That said, now and then on the forums, someone will remind someone that such and such an output will have to be normalised to get a result, for instance. In this kind of case, I think that some explanation would be useful. If the nodes are there, ie. as NT have seen the worth of developing and including them, it would be equally useful to provide enough primer info for people to get their heads around what the node does, enough to make sense of how and when to use it.

A similar analogy is being able to make good use of a car without knowing all the mechanics of how it works, and why. OK, maybe that's not a great analogy, but that's the idea.

This page on Wikipedia about vectors:

http://en.wikipedia.org/wiki/Vector_(spatial)

is useful.....I understand some of it - there are diagrams, which help alot.

Now I'm not talking about all the maths nodes - use of Add, Multiply, and so on, should be fairly obvious. Simple examples would be enough for the real newbie. But even something simple like a constant being controlled by a multiply node gives clues to the power of nodal versus layered.

I don't think anyone serious is looking for an easy ride on this, and NT need not get into the business of teaching math, but we need alot more than descriptions which, though accurate and concise, are still abstract and out of context.

Julian.

JBT27
04-01-2008, 10:44 AM
This must be one of the most useful nodes in general use, and yet there is little explanation about what each output is and what you can do with it.

Before I get flamed over that, here's what I mean:

One of the output descriptions:

*****
Ray length (Type: Scalar):
Ray length is the length of the ray from the source to the spot
under evaluation.
*****
OK - that's simpler than others. But 'source' needs expanding.....presumably that's the light? But what if there are multiple lights in the scene affecting that spot? Are they averaged? Can you confine it to just one of them, and if so how? What could you use it for? Just some basic examples giving users the clue to run with it themselves.

The normals outputs likewise - I think I know what normals are, but in this context how useful are these outputs and again, how do we make use of them?

And I'm sorry, but the Incoming ETA is a prime example of what I'm asking:

*****
Incoming ETA or Incoming Refraction Index. ETA is the name
of the Greek letter commonly used in 3D graphics to specify
refraction Index. When the Greek character eta is unavailable,
the letter ?n? is often used to represent the refractive index in
many online discussions and examples.
*****

The vast majority of this description tells me about nomenclature - if it's not called eta, as in the Greek letter, it will be shown as 'n'. I'm not criticising that as a piece of info - it is useful - but exactly how is this output useful? Presumably Spot Info is outputting the IOR of the surface?

It needs more explanation and examples and hopefully some diagrams.

There are no usage examples for this node currently.

Julian.

colkai
04-01-2008, 01:34 PM
Yep, it's stuff like that that leaves you scratching your head.
I'd wager that stuff was written by a techie who knows fully what it does and so assumes everyone else must.
A failing, alas, all us techies suffer from at times.

spiroz
04-01-2008, 01:44 PM
One other factor to easier understand nodes is more feedback
in the node editor. For example, I think it would be a lot easier
to understand the concept of shader/material override if other
inputs in the surface node turns ghosted when a material or a
shader gets connected to it.

Lightwolf
04-01-2008, 02:13 PM
The funny thing is, some of the descriptions in the SDK are better than in the manual (and yes, most of these outputs correspond to variables previously only visible through the SDK).

Examples (LW9.2 SDK):
rayLength
The distance the viewing ray traveled in free space to reach this spot (ordinarily the distance between raySource and wPos).

incomingEta
Incoming refraction index. (note: Hey, that's all that's really needed to describe it).

56970
A definition for a spot in general (just copied from the SDK as well):
For each pixel in the rendered image, the renderer finds the spot in the scene that the camera sees at that pixel. If the spot is on an object, its appearance depends on the suface assigned to the polygon it lies in.

So a lot of the descriptions are already available. They might need some polishing, but they're available.

@JBT27: If you know how Lamber works (or shading in general) you'll have little trouble using the more advanced nodes... and you'll know why you need to normalize a vector (The main reasion is to find the angle between two vectors, that you can only do with normalized vectors... i.e. vectors with a length of 1.0. They are also used to define directions... you use normalized vectors so you can multiply them by a distance(scalar) to define a position in space - since they have a length of 1, after the multiplication they will have a length of distance). And yes, that's high school math ;)

Cheers,
Mike

JBT27
04-01-2008, 02:37 PM
The funny thing is, some of the descriptions in the SDK are better than in the manual (and yes, most of these outputs correspond to variables previously only visible through the SDK).

Examples (LW9.2 SDK):
rayLength
The distance the viewing ray traveled in free space to reach this spot (ordinarily the distance between raySource and wPos).

incomingEta
Incoming refraction index. (note: Hey, that's all that's really needed to describe it).

56970
A definition for a spot in general (just copied from the SDK as well):
For each pixel in the rendered image, the renderer finds the spot in the scene that the camera sees at that pixel. If the spot is on an object, its appearance depends on the suface assigned to the polygon it lies in.

So a lot of the descriptions are already available. They might need some polishing, but they're available.

@JBT27: If you know how Lamber works (or shading in general) you'll have little trouble using the more advanced nodes... and you'll know why you need to normalize a vector (The main reasion is to find the angle between two vectors, that you can only do with normalized vectors... i.e. vectors with a length of 1.0. They are also used to define directions... you use normalized vectors so you can multiply them by a distance(scalar) to define a position in space - since they have a length of 1, after the multiplication they will have a length of distance). And yes, that's high school math ;)

Cheers,
Mike

Ah right.....so I'll still admit I struggle with maths and invariably have to think about it, but I pretty much get now what normalise is, in principle.

So an explanation like that in the docs ..... maybe laced with some nice diagrams for people like me :D ..... and that's really 'all' that is needed. What I consider the more regular surfacing nodes, like turbulence, grids and bricks and so on, are probably fine as is and can be easily experimented with. But with these math nodes, you need to have some clue of what they and you are doing to be able to experiment.....randomly connecting maths nodes on assumptions is not experimenting in my books - well, it is, but I've spent hours doing that and got nowhere!

That's quite bizarre about those SDK docs - those are exactly the kind of explanations that should be in the regular docs for the rest of us.....maybe they accidentally switched doc writers.....:D

Julian.

Lightwolf
04-01-2008, 06:13 PM
But with these math nodes, you need to have some clue of what they and you are doing to be able to experiment.....randomly connecting maths nodes on assumptions is not experimenting in my books - well, it is, but I've spent hours doing that and got nowhere!
Yup, that's like plugging in symbols until you get pythagoras theorem, os the defintition of sines/cosines etc.
The weird thing is... I suppose if you really sat down and went through some of the maths now, having a CG background, it would probably make a lot more sense than it did in school.


That's quite bizarre about those SDK docs - those are exactly the kind of explanations that should be in the regular docs for the rest of us.....maybe they accidentally switched doc writers.....:D
Well, the SDK has clear, dry and concise descriptions because that's all that is really needed. The user docs assume less (knowledge), but don't really provide the basics even though they try to.
The SDK docs are also a lot older (written to the best part by Ernie Wright, between LW 6.0 and 7.0).

Cheers,
Mike

JBT27
04-03-2008, 04:02 AM
Yup, that's like plugging in symbols until you get pythagoras theorem, os the defintition of sines/cosines etc.
The weird thing is... I suppose if you really sat down and went through some of the maths now, having a CG background, it would probably make a lot more sense than it did in school.

<snip>

Cheers,
Mike

Yep - context - it's all about context versus abstract.

Despite mithering about my lack of second-nature maths skills, truth is I reckon alot of us do use some maths everyday in this work and actually alot of this does make alot more sense now than it did as abstract concepts in school, because now there is some point to it.

That's true of alot of learning for practical use though - hence my hopes for alot more project-based tutorials from NT.

But it also peeks my interest enough to dust-off the trig and start learning again.....:)

Julian.

08-19-2008, 08:53 PM
LoL this is the funniest thread i ever read in my whole life,i mean in the first pages....
that looked like UK war hahaha anyways, i completly understand everyone's point of view besides the fact im still too noob for this(its been only a few months i got into LW's world and CG society ,but i got the same idea bout nodes, soon i'll be posting my novice projects but think its still too early for that,kinda feeling embaressed lol,this is too pro for me yet hehe :D
really great discussion

btw welcome me guys,just joinned newteks forum ^^ forgot to introduce myself
take care folks

starbase1
08-26-2008, 04:27 AM
I've passed along some of the comments here on items not being present in the HTML docs that are covered in the PDF manual. Our docs staff will work on getting the HTML docs updated.

For those of you who can take the time to assist in providing detailed feedback, we'll do our best to improve the text and to add supporting usage examples as needed and requested, and you'll very certainly have our appreciation for your contribution to our efforts.

Chuck, I think a few nice presets would go a LONG way to getting people started - it's always easier to tweak something that works than start with a blank screen!

Nick

starbase1
08-26-2008, 04:32 AM
I must say I was very nervous when I first approached nodes, it did look horrendously complex!

In practice I found things were a lot easier than they looked, not least because any channels etc you don't control with nodes stick with the old system, so you can mix and match with more nodal bits as confidence grows.

The only thing I do think worth noting is that a few areas where I felt something was missing were well covered by the excellent and free DP kit, which I recommend highly.

For example, I reallly would be clueless about controlling specular from a colour channel the old way. (I wanted to do this for a very high res Earth, to make the oceans shiny without another very high res map). A bit of mucking about soon had a colour keyed shinyness, and after that tuning it was easy.

(Still want more presets though!)
Nick

JohnMarchant
08-26-2008, 05:12 AM
Once we are able to preview onscreen in ogl i will be alot happier. But im getting into and liking nodes

08-28-2008, 09:03 AM
Word
hope soon i will able to manage nodes like a second nature...
cheers folks

DaveWhitney
11-28-2008, 11:08 AM
Yep, it's stuff like that that leaves you scratching your head.
I'd wager that stuff was written by a techie who knows fully what it does and so assumes everyone else must.
A failing, alas, all us techies suffer from at times.

So I work at Microsoft and specifically in the group I work in, the docs are written by people who were not involved in the product's development. That sounds idiotic at first, but the end result is they have to keep asking the product team members how a specific feature works until they themselves get it. The product team reviews the doc and sends editorial comment back ("That isn't what it does", etc). The end result is documentation that is more useful to the non-familiar than anything that would have been written by a techie.

I think the solution here is a feedback system where Newtek product people do their best to explain each feature so that a non-techie (or other folk who don't previously "get it") try to write some documentation. Then Newtek product people correct mistakes and fill in gaps and have the non-techie refine the doc based on the revised information. Best mechanism for this is a wiki. Newtek's first step really ought to be to convert the docs into wiki form and keep them online. Regular folks could revise the wiki only to the extent of making comments like "needs more detail," "I don't get this," etc. People with editorial control would then make refinements as requested.

zarti
02-27-2010, 08:04 PM
So I work at Microsoft and specifically in the group I work in, the docs are written by people who were not involved in the product's development. That sounds idiotic at first, but the end result is they have to keep asking the product team members how a specific feature works until they themselves get it. The product team reviews the doc and sends editorial comment back ("That isn't what it does", etc). The end result is documentation that is more useful to the non-familiar than anything that would have been written by a techie.

I think the solution here is a feedback system where Newtek product people do their best to explain each feature so that a non-techie (or other folk who don't previously "get it") try to write some documentation. Then Newtek product people correct mistakes and fill in gaps and have the non-techie refine the doc based on the revised information. Best mechanism for this is a wiki. Newtek's first step really ought to be to convert the docs into wiki form and keep them online. Regular folks could revise the wiki only to the extent of making comments like "needs more detail," "I don't get this," etc. People with editorial control would then make refinements as requested.

:agree:

Lightwolf
02-27-2010, 08:20 PM
So I work at Microsoft and specifically in the group I work in, the docs are written by people who were not involved in the product's development.
It does make absolute sense (Having said that, I've never explicitly read docs from MS so I don't really know if the effort pays off - and the help texts I've read were o.k.-ish).
Another approach (which is should actually be complementary) is to write the documentation as part of the design process before code is written.
You'd still need somebody now familiar with the development to ask the seemingly stupid questions at one point in time though, no doubt.

Cheers,
Mike

JBT27
02-28-2010, 04:12 AM
In principle it is a very good idea, but it needs provisos to make it work well, in the case of nodes.

This seemingly quite modern convention of bringing people in who don't know what they're talking about, on the basis that ignorance equals fresh thinking, is also deeply flawed. In some fields, it means that each successive generation wants to reinvent the wheel, rather than build on what went before and genuinely progress whatever it is they are dealing with.

In this case, the idea works well providing that programmers are relating the concepts to advanced users, TDs maybe, who in turn relate their interpretation to intermediates and newbies. In addition, some aspects of nodal, as we are told, that only relate to building your own raytracer, for instance, should be flagged up - when I asked about some of these functions a couple of years back, I was told that 'If you don't know what they are, you don't need to worry about them.' In the case of some of the more, to me, esoteric maths nodes, that may be true. But in turn it would help enormously if someone would step-up and actually bother to explain, simply and with diagrams, what each of these things do. Maybe you can't, in all cases, but then maybe NT ought to rethink how these functions are made available.

It's been said before that it is not NT's job to teach math, but if they are going to include something as powerful, and by necessity, complex, as nodes, then there is a duty to develop a tech writing procedure that starts with the programmers and ends with writing by those who have both 'got' what something does, and importantly can communicate it clearly.

Julian.

zarti
03-07-2010, 05:13 PM
ok.
here you have something guys; easily translatable in lw's nodal trees.
look down at Lesson.7., there is a downloadable video. it is nice and clean!
@that time, lw was used to illustrate polies'n'normals ... now that lw has nodes, nothing has changed @lw's home. so pity ...


anyway, this is @least a good example how NT should explain nodes
Lesson 7: A video to watch - practical implemention via low-level shading nodes in Mental Ray with XSI (http://www.warpedspace.org/tutorials/shading_theory/)


p.s.: someone here (@least) could make some money for sure if he/she/they could fill a dvd with simple tutes like that.
why this still doesn't happen here ?!

p.p.s.: if someone finds similar things, please post/collect those here ...

share people, share!

zarti
03-07-2010, 05:24 PM
Light Sourced Lambert Shading (http://www.exaflop.org/docs/waynelight/index.html)