PDA

View Full Version : What can be done to improve character animation workflow in Lightwave?



Ryan Roye
08-26-2013, 01:12 PM
I felt this question by Matt:


Let's turn this into something more constructive! ;)

Of course, we have plans for CA tools in LightWave, but I have some questions.

What do you feel is missing that would aid character animation in LW?

Which existing areas do you feel need fixing / enhancing?

What do you feel is currently possible, but difficult / time-consuming in LW?

Cheers
Matt

... needed its own thread. So, let's get the ball rolling and get as much useful information in here as clearly and concisely as possible so the devs have something on the table to look at eh? Some say things have been said a thousand times, but I think it would be incredibly helpful if we could aggregate that information HERE.


I'll start off with the things I feel would improve character animation workflow:

- Interactive weight painting in Layout. This alone would fill a very, very big hole in the rigging workflow. People wouldn't need to avoid weighting if it didn't involve jumping between modeler and layout constantly.

-Genoma: This is a system designed to make rigging a bit more of a user friendly process, but it needs to preserve existing elements in a scene when running the "update rig" command because people often pair facial rigs or controls that cannot be created right within genoma. Things like expressions, morph assignments, motion modifiers, etc all get deleted when updating the rig. Obviously, this is incredibly problematic for practical usage even for novice users and introduces a lot of added work and complication that nullifies the convenience factor almost entirely.

-Overly Long Dropdown lists: We all hate them. If any dropdown list is more than 10 items long, it needs to have a way to type in something to find an entry (or, better yet, how about a "use layout selected" button?). Perfect examples are scattered all over Lightwave's interface... for example, in motion options any dropdown menu that uses the scene's items are too long to be practical to use.

-Allow Multiple Targets/assignments for falloff controls. This concerns goals, targets, motion controllers, etc. These things all have one thing in common: Only ONE item can be used with them. This is why I have always recommended baking instead of even bothering with IK/FK switching and falloffs. If multiple items could be assigned as targets in a timeline, there would be no need for baking this stuff.

-Easier keyframing of envelopes: You know that little key icon in Morph mixer? and how you can keyframe stuff without having to open up a graph editor? Why can't we have it work this way for the entire interface?? Why do we, for instance, have to open the graph editor or assign a null in order work with these values? It would add a lot of convenience if we could manipulate keyframes for enveloped items the same way we do in morph mixer.

I think that might get the ball rolling, I will pitch in more to this thread as I have more time to point out issues.


GUIDELINES

- Try to be detailed with your descriptions of issues/problems and why they impact the animator/rigger's workflow. Be as concise and detailed as possible.

- This thread is all about helping the Lightwave developers. So, let's keep that the focus of attention. Criticism of tools/processes are ok, but if you post that, you should also post what an optimal solution should be.

nickdigital
08-26-2013, 01:24 PM
My list off the top of my head. I'm sure I'll have more as I continue to think about it.


A channel box like Maya. There's too much of a need to navigate through multiple items to get a single task done.
Better performance when heavy or multiple rigs are in a scene.
A lattice deformer.
A deformations controlled by a curve. Or a way to get bone chains to stretch along the SplineControl tool. Messiah has this.

UnCommonGrafx
08-26-2013, 03:10 PM
- Lattice Deformer;
- Ability to animate in VPR, at a quick clip. That is: bones work, render works, everything follows my mouse in this condition;
- The forum scoured for ALL the old stuff, and fixed based there upon;
- People who have shown expertise to be absorbed by NT for a year as "Artist in Residence" so that we can all benefit from their knowledge. Rebelhill (Craig) comes to mind in this context.
As of late, NT has only absorbed. I'd rather see the visiting scholar scheme come back.

LW_Will
08-26-2013, 08:19 PM
My list off the top of my head. I'm sure I'll have more as I continue to think about it.


A channel box like Maya. There's too much of a need to navigate through multiple items to get a single task done.



I was thinking that the search box could be a part of every list. The Items list for Objects, Bones, and Lights... every place there is a list, there could be one of those boxes.

nickdigital
08-26-2013, 08:22 PM
I was thinking more about tieing channels to a single item to make animating easier. Not searching.

Dexter2999
08-26-2013, 11:16 PM
This doesn't address "more than 10 items" but I have wished there was a way that bone lists were auto collapsed/expanded following the currently selected object. So, the only rig displayed in the scene editor/drop down would be the one you are currently working on. (Selection of Multiple objects would expand multiple trees.)

Also, it has crossed my mind that instead of just a color coded icon, the actual line in the scene editor could/would be color coded by object, so different objects rigs could be quickly identified while scrolling faster than users can actually read.

These are just some quick ideas. I'm sure better minds will present better alternatives.

Ernest
08-27-2013, 12:18 AM
Chris Jones was asking for a more artist-friendly way to rig. Rebelhill replied that, in order to offer that, you'd need to create a rig for every possible scenario. Yes! He was right, but then he went on to conclude that the task was impossible in the short/mid term because of that.

I think he actually pointed the real and only solution. A solution that is proven to work in other areas of 3D graphics and that isn't, in any way, outworldly difficult to implement.

It's the solution that DAZ Studio used to take all the technical complexity out of character modeling and making it 100% artist friendly. It is the solution that Vue used to take all the technical complexity out of sky and vegetation creation and making it 100% artist friendly. It's what Mixamo is trying to (eventually?) do with game animation... You need a store, like DAZ/Renderosity/Cornucopia where people can upload their rigs and/or rig building blocks. That is the only way to get a massive library of rigs so that you can find the one you need in any scenario.

I can't remember the last time we modeled a human (or horse) from scratch. There's always something close enough available that just needs to be retopo'ed. Likewise, I can't remember the last time we modeled a plant. If it's not in Cornucopia, it surely must be in XFrog. If there was a massive library of $20 genoma rig elements from riggers of all sorts of different rigging philosophies, I think I'd just as readily forget the last time we bothered to rig a character ourselves. And, if we had to, then we'd do what we do whenever we actually have to model a plant or a character: We'd put it up in the store so that nobody else has to bother with that one again.

Of course, Genoma would need quite a few changes before any 3rd party could create truly custom rig elements that could still be mixed and matched like the current genoma pieces, but it's all within the realm of current tech.

RebelHill
08-27-2013, 02:23 AM
get a massive library of rigs so that you can find the one you need in any scenario.

Its an interesting idea for sure... and probably quite a plausible one. The only bar I can see is that which I mentioned in the other thread... having a "general" piece of code that's able to deconstruct the rigs properly into the component parts that need to be repositioned and which can somehow "interpret" the intentions of the rigger. A single preset may well contain multiple parts that must not be touched by the "refitting" user, and others that need values/attributes modified slightly for different contexts/fittings... that's TOUGH.

I did have one interesting (and laregly stupid) idea appear in my mind though...

A rig can be thought of as a series of parts which share spatial connections/relations (hierarchy) and which commit influence upon one another (IK, constraints, etc). You could therefore think of any given rigs design as being described as something known as a constraint satisfaction problem (in this case constraint meaning "permissable outcome", not motion constraint)... Such as... when in IK, limb must go to goal, when in FK goal must be unmovable, changes to right arm must not affect left arm, etc, etc. Now CSP solvers are widely and well known, and very general, it only takes a few task (input) specific tweaks to get them solve any properly described problem. Thus... if you could somehow allow a user to "tell" the computer what a rig must and must not contain, what each part/mechanics must and must not do... then you could let the comp chew away until it found a solution (a rig) which satisfied all of those requirements and constrained conditions. That would give you a properly "intelligent" autorigger.

There are, however... 2 GLARING problems with this.

First, how to describe the "requirements" in the first instance. Rigs, and the ways they are intended to work, can be quite "abstract"... more easily described as design ideas rather than technical specifics. But even allowing for that, the user would have to be able to give the computer a set of constraints that were actually resolvable... if they asked for stuff that didnt make any real sense, or which had internal structural conflicts... then the system would either never reach a solution, or would have to return a "min conflicts" solution (which would still be of no use)... so there'd still be quite a burden of knowing what you were doing placed upon the user.

Secondly... the "description" of a character rig in that regard would be a VERY complex "problem" with lots of intricacies, interconnections and very, very tight and narrow available solutions... Its not a simple "vista", like timetabling, or roadway organisation problems are... And since the solve time for CSPs is exponential in the size of the problem... then I can forsee that even with a perfectly delivered set of constraints, the computer could potentially take a long time to reach a solution... Im not sure Id want to wait a week while I the autorigger figured out how to do what I had asked of it!

Sigh... thats what I get for eating cheese and drinking pepsi before bedtime... I wake up with this kinda nonsense in my head.

safetyman
08-27-2013, 08:50 AM
I agree with Earnest on a DAZ-type solution...

1. Give away for free a fully working rig (genoma or otherwise) somewhere on the LW website. Maybe multiple rigs; biped quadraped, whatever.
2. A 1- or 2- click solution to apply the above rigs to your model with little or no extra work.

I know it's probably impossible, but that's what I want. I don't have the resources to hire RH or buy his wonderful rigging tools as much as I'd like to.

VonBon
08-27-2013, 08:50 AM
Just to get user created Rigs into Genoma.

Could be as simple as having a "Export Interface" for the person exporting the
Rig to Genoma. The information has to be in the scene file. Could have the script
cycle through each item in the scene and ask the user to "Label/Tag" each part
for the corresponding parts inside Genoma. The scene file should already retain
the information on how these parts are connected and how they work together.
There should also be a separate file that contains any "motion modifying" data that
would be loaded into layout when the Rig is opened or when requested by the user.

The Rig or Rig parts should work as the creator intended so the computer doesn't
have to figure out what it needs to do, it already knows since its in the scene file.

RebelHill
08-27-2013, 09:19 AM
I don't have the resources to hire RH or buy his wonderful rigging tools.

The free version is for YOU!

robertoortiz
08-27-2013, 10:00 AM
The first thing I would do.
(And I mean this)
I would sponsor an animated project from upcoming talent.

But instead of asking him to do it in Lightwave, ask him point blank, to imagine the best tools to do his dream project.
And I talking from the GROUND UP total rewrite.
And I do think the LW community already has someone like that.
To not pick the brain of an artist like
Chris Jones (of the passenger fame) is frankly a waste of a good opportunity.
http://www.youtube.com/watch?v=OGW0aQSgyxQ



And scope out a brand new toolset based on what he feels are his needs.
We need fresh ideas and people like him can bring them.
Maya was developed this way with the Bingo the Clown short.
http://www.youtube.com/watch?v=Iu9L23A6yDU

I have more ideas based on a traditional animation pipeline, that I will post later.

I know that the new team is open to different view points, after all the hiring of Lino and Matt have proven this..

chikega
08-27-2013, 10:42 AM
Being able to adjust bones/joints and watch the deformations in realtime within the animation itself in addition to setup mode. Like this:

http://www.eggington.net/6Promo/video/Bonagami_Skelevision.mp4

Also having access to modeling tools to adjust geometry (vertices) within Layout.

jburford
08-27-2013, 10:46 AM
The first thing I would do.

And I talking from the GROUND UP total rewrite.





I think that a "Ground Up" rewrite of the CA Tools/Toolset will be impossible/unlikely without a Complete Overhaul/rewrite of Lightwave as we know it. And this will not be happenning anytime soon.

More than likely perhaps something like ChronoSculpt which then could be bolted into the new Lightwave. . . .. but talking about a hefty amount of work, after first accepting and understanding what the users are seeking.

jburford
08-27-2013, 10:49 AM
Being able to adjust bones/joints and watch the deformations in realtime within the animation itself in addition to setup mode. Like this:

http://www.eggington.net/6Promo/video/Bonagami_Skelevision.mp4

Also having access to modeling tools to adjust geometry (vertices) within Layout.



. . . . . . . meaning a Rewrite and Unification. . . .

Greenlaw
08-27-2013, 10:53 AM
For me, making the joints system compatible with third party apps would be a big improvement for mocap users, and I think it would be a 'low hanging fruit' for LW3DG to accomplish.

The current Joints system isn't directly compatible with other applications because of the weight map offset issue that forces users to take extra steps to correct the problem. In fact, the weight map offset isn't even consistent with Lightwave's bones system, and it prevents some rigging tools from working correctly LW Joints.

I would love to be able to easily transfer a joints based rig in LightWave to Maya or Motion Builder but the way it currently exists, LW Joints require a special 'alternative' setup just to transfer the rig, otherwise you will need to redo the weight association in the other program.

Yes, I understand that changing this in LightWave will break existing rigs that use joints but IMO that's a small price to pay to get a joints system that behaves like everybody else's joints system and makes LightWave more compatible with programs like Motion Builder, Maya, etc. And I understand that this adjustment may break the 'twist' feature but, really, I don't care. I would much rather have consistency and third party compatibility.

According to RH, LightWave's Joints system actually worked correctly during the 9.6.x public beta but it got changed just before release. I'm not a programmer but it seems like all that's required is that weight maps be moved back to be associated with its intended Joint (like how it works with LightWave Z Bones and third party 3D programs) and not to the parented Child Joint.

Could we please have LightWave Joints set back to the way it worked back then?

Thanks in advance LW3DG. :)

G.

Greenlaw
08-27-2013, 11:02 AM
And +1 for Weight Painting and Mesh/Endomorph editing in Layout. These capabilities for Layout were proven in 3DCel's little x32 tech demos a few years ago, so I'm hopeful that these feature will be coming to Layout in a robust production-worthy x64 version.

G.

robertoortiz
08-27-2013, 11:16 AM
I

More than likely perhaps something like ChronoSculpt which then could be bolted into the new Lightwave. . . .. but talking about a hefty amount of work, after first accepting and understanding what the users are seeking.

We are on the same page.
I would dare to say that you could build a better character animation pipleline on top of something like Chronosculpt (AKA the Hydra Engine) than classic LW.

There are too many legacy issues with the workflow right now.

For starters:
The lack of an universal undo. This is a deal breaker for complex animations.
Cross Panel Jumping.
Drop down menus from hell.
The isolation of plugins- as in a lot of plug ins dont talk to each other.

And there are way more.

Greenlaw
08-27-2013, 11:29 AM
I wish to see Schematic View improved so that it allows you to hide/filter nodes and the ability to nest nodes in Groups. We also need the ability to save and load Schematic Layouts for the full scene and for selected items.

Schematic View Tools could be much improved too. The current version works okay with very simple scenes but it gets easily confused with larger scenes that have many items with similar names and clones (like many typical 'production' scenes,) and running Tree View or Skeleton View will generate a big confusing mess.

Besides character rigs, I used to like using Skeleton View to lay out scene maps for environments but it's just too unreliable now. Even Tree View, which should be pretty straightforward, will fail when selected items have names similar to unselected items.

Occasionally, I still try using Schematic View Tools because this tool used to be very useful, but nowadays this usually ends in frustration.

I imagine the entire Schematic View system (Layout and Tools) will need a complete overhaul to address the above issues but it would be a worthwhile effort.

G.

Ryan Roye
08-27-2013, 12:25 PM
Graph Editor Enhancements:

- I'm not 100% sure, but it doesn't seem possible to move around the playhead in Graph Editor without using the mouse. I see no command to move the playhead in the keyboard shortcuts menu. Specifically, I mean like moving the playhead back/fourth with the left/right arrow keys. The Graph editor keys should be controlled identically to how the timeline works outside the graph editor. This would save the user having to swap in/out of graph editor when quickly keying things. Currently, it seems only possible to jump between existing keyframes, not move the playhead freely with the keyboard.

- Create Key dialogue: Adding the incoming curve and TCB values would save the user from having to repeatedly perform redundant actions to change these. Optimally, the values should be retained so that they are persistent. I understand you can set the default incoming curve via options, but this really needs to be in the create key dialogue and there is no way to set the tension as keyframes are created. (HINT: this would be nice to have in LW's normal create key dialogue as well).

- Commands like Numeric Scale and Numeric Move should be interactive so that one can play the timeline whilst making changes. Obviously, the user should have the option to toggle interactivity off in case the keyframes are really dense. This would make adjusting existing animation with these tools an easier and more intuitive process.

- Ripple editing option. If I have 100 keyframes, and I want to stretch out keyframes 40 thru 60, the keyframes past that really need to get moved forward in the timeline appropriately instead of being overwritten. This would greatly improve the user's ability to manipulate the timing of their animations without a ton of work or 3rd party plugins.

- More than 1 undo level would be nice. I know this one is problematic but it is worth mentioning. Leave/backtrack footprint only does so much to alleviate this.

- A more user friendly channel filtering method. The built-in filter is very poorly implemented and isn't straightforward to use. The user should not have to, for example, figure out that they have to type *ion.* in order to filter out position AND rotation channels. I am not exactly sure how this would be implemented... but perhaps checkboxes could be added beneath the channel list that the user can click on? (XYZ/HPB/XYZscale/Other)

- Searchable Channels for the lower-left part of Graph editor. Navigating the dropdown list is a real pain in many cases.

bobakabob
08-27-2013, 12:49 PM
I mainly use mocap files to animate in Lightwave, which are pretty straightforward, dare I say a pleasure, to set up in Layout. I also enjoy creating basic rigs in Lightwave though Genoma has made things much more immediate and effective for CA projects. Thankyou Lino :)
http://vimeo.com/68946684

As a one man band who is mainly a modeler / Zbrush artist (and very basic animator), I really appreciate the craft of experts like Lino and RebelHill in working towards sophisticated rigs. It can only be good for LW that debates like this flare up from time to time (though unfortunate in this case it was initiated by an extremely rude op (troll?) who flouted forum etiquette big style in his initial rant).

Some great suggestions in this thread and earlier discussions. Agree 100% with:
* Point manipulation in Layout (though Chronosculpt may well be the way forward)
* Making joints compatible with other programs as RebelHill and Greenlaw suggest.
* Auto weight map generation in Layout a la XSI + weight painting.
* Dumping the cumbersome bones - follow - null - goal IK setup. In xsi again it's much less painstaking to set up.
* Colour coding the mile long item selection lists.
* More Genoma tuts specifically rigging bipeds and facial rigging in LW.

LW is where it is regarding animation, though does have a hardworking talented Dev team. I for one am looking forward to future developments. Must say it's hard to comprehend the original 'complaint'. The great examples of completed projects such as 'Workout' and 'Passenger' show that vision and talent combined with LW's capable animation toolset can create truly startling results. If users feel LW has shortcomings there are great alternatives / plugins such as Messiah and Blender or more specialised apps such as Xsi.

robertoortiz
08-27-2013, 03:45 PM
One thing I wish developers of 3d programs did is to go back to the core notion of what is an effective character animation workflow.
When I see Maya, or Softimage or even Blender I see great tools, but they follow toolset paradigms that come straight of the 90's, when computing power and memory were in short supply.

There has to be a way to re-think from the ground up how to do character animation into something more manageable and less time consuming.
Right now the biggest character animation sins that Lightwave commits is the lack of a clear workflow for the tools you need for animation and the time to take to track down elements in a scene. As people have mentioned before a more unified workflow would be in order.

Hell I would even look into another areas of animation, LIKE STOP MOTION to see how they do it.

For starter of what the heck I am talking about check out the CORE 12 Principles of Character animation and lets try imagine how LW could implement them. (good Luck)
http://en.wikipedia.org/wiki/12_basic_principles_of_animation

hazmat777
08-27-2013, 04:52 PM
Not to go too off-topic, but if there are CA's out there who don't have this book, (Illusion of Life) find one. As far as I know, there are two different versions. BUY THE ORIGINAL VERSION, NOT THE "REVISED EDITION". The originals graphics were taken from original slides (not dupes) and it has more content. Just for bragging rights, I got a Mint Condition 1st Edition for $49.00 ! Must have not known what he had !

Ryan Roye
08-31-2013, 12:22 PM
Some other suggestions related to rigging:

- A method to achieve mirror bone editing in Layout. While one can take advantage of mirror tools with modeler's skelegons, when perfecting deforms it is often faster to adjust bones in Layout. Deleting 1/2 of the rig and mirroring it often causes problems with many tools that depend on bone ID#'s being the same, so it is not a very viable solution in productions that demand a solid rig template meant to be used on many characters.

-Axis adjustment for joint compensation. Currently, joint compensation only works with the pitch axis and if one has a library of models that don't have the character's hands positioned in a certain way, some crazy workarounds are required to take advantage of it. It would be incredibly helpful if one could adjust the axis in which this operates so the original position of the mesh is irrelevant.

erikals
08-31-2013, 12:31 PM
* Auto weight map generation in Layout a la XSI + weight painting.

we kind of have it > http://youtu.be/wjm52B1Oqkg

Kryslin
09-01-2013, 01:04 PM
Hmm... Some ideas from a Noob in CA...
1) Fix some of the long standing bugs first.
2) Fix the interactivity bug re: nodal channel editing.
3) Document, Document, Document!
4) Weight painting in Layout
5) Lattice Deformer
6) Improved Performance in Layout Display
7) The ability to add channels to items without a master plugin to an any item.
8) The ability to use items to control blending without node editing or expressions (an "I" Icon, select it and it brings up a dialog for Item, Channel, Invert, and a range to clamp to). If a plugin, give it a com ring interface so it can be controlled by lScript/Python...
9) Add some more UI elements to the Sliders toolkit- 0/1 push buttons, radio buttons, etc. Also, be able to attach a Sliders UI to an object.
10) Document, Document, Document!
11) Filtering based on object name... If you have a bone selected (Named, say, Char_Rt_Leg_Upper), sort the weight maps names by how close a match they are - Since several people have pushed the "logical and consistent naming scheme" around here, the item you are looking for will be at the top of the list... Also, have an option to disable it, for those who think it would be a pain, or whose workflows don't require a strict naming scheme.
12) A Genoma Construction Kit : Consists of the various bits that make up a genoma sub rig - FK Bones, Controllers, Poles, Re_Parents, IK Goals, etc. Would allow a 3rd Party to build a sub rig. Genoma should be able to support this, since the creation of the controllers, targets, poles, etc. Is all handled by sub scripts called from a main script.
13) Improved Autoweighting. The Vertex Paint plugin gives a good start, it leaves a lot to be desired in some cases.
14) Merge D-Storms Skelegon Editor functionality into the Skelegon Tree. Also, when a skelegon is highlighted in the tree, select it in the viewport.
15) Document, Document, Document!

bobakabob
09-01-2013, 01:43 PM
Genoma - request for Lino - Make it impossible to delete key point and poly elements in Modeler which would break the rigging once sent to Layout. In other words 'idiot proof' Genoma :)

Davewriter
09-02-2013, 07:57 PM
we kind of have it > http://youtu.be/wjm52B1Oqkg

One more of those in my very long "I didn't know that!" list.
Thank You SIR!

Dexter2999
09-02-2013, 08:57 PM
Would it be beneficial to have drop-down-menus (or pop-up-windows) that behave like nested menus?
So all bones in your item are present, other objects are present but their bones and parented associations hidden, you mouse over an object and it's root bone/child items appear in the next pop up... mouse over the root bone, the root tree appears, spine and hips, select left_hip and the left leg tree appears, etc...

Or is this rediculous?

Ryan Roye
09-02-2013, 09:18 PM
Would it be beneficial to have drop-down-menus (or pop-up-windows) that behave like nested menus?
So all bones in your item are present, other objects are present but their bones and parented associations hidden, you mouse over an object and it's root bone/child items appear in the next pop up... mouse over the root bone, the root tree appears, spine and hips, select left_hip and the left leg tree appears, etc...

Or is this rediculous?

If the dropdown menus can't be eliminated entirely, I think a good alternative would be to simply be able to click an item in 3d space and have that be the selection (kind of like how the "eyedropper" works in most 2d programs). Some of Lightwave's more handy tools like cyclist are *CRIPPLED* by these dropdown menus.

Mr Rid
09-02-2013, 11:27 PM
Boy, I'd LOVE to advise on the interface, and see LW character tools re-written from the ground up to be more friendly to even non-CG artists who intuitively want to stick their hands in CG and play around, without having to read a lot of manuals and plod thru hours of tutorials. I always see ways of simplifying the interface, without compromising control. But more experienced character animators should be intimately listened to. The UI should be as intuitive for the widest range of user experience, and not be left to whatever a more tech-minded programmer thinks is convenient. I think there should be more R-click pop-up menus, and click+drag-able graphic displays, to minimize the need for opening clunky panels with convoluted value arrays. IKB function started down this path. There should be a beginner friendly layer of functionality that allows any user to jump right into rigging and quickly see characters move, so its FUN. Then there is another layer of complexity for more advanced expression.

-Since v5.6, I wondered why LW doesnt come with a decent, versatile human rig and base figure mesh, that can be easily edited to fit other human figures you may make or acquire. So that beginners can get figures moving quickly without having to learn rigging from the ground up.

-Get all rigging tools the hell out of Modeler. You need real time deform feedback in Layout. And I find weight maps an awfully inefficient way of controlling deforms...

-The long-forgotten 'limited range' tool could be upgraded to reduce need for weight maps- namely, range and falloff values independent for each axis, like with textures- +X,+Y,+Z, -X, -Y, -Z. These values could be represented as click+drag icons in the display that do not require the opening of optional value panels.

-LW in general needs the 'preselection' that I find so useful in Daz. This means that as the cursor passes over items, they each highlight so you know exactly what you are going to select before you click to select (the click-select in Layout can be awfully inaccurate, and unusable at times). And selection bias should favor items closer to POV.

-Preset rigs and common joint types that have automatic settings appropriate to their typical purpose (Genoma sorta, but IN Layout).

I am just thinking out loud here, but I wish I could 'draw' preset 'joint types' and 'joint links' that auto-follow the contours and depth of a mesh. Roughly, I could run the cursor over say a human figure mesh, as polys are 'preselected', while Layout 'sees' thru to the opposite poly normals, and calculates the thickness of the mesh. Select 'Apply Joint' mode, select a 'shoulder/Ball-&-socket' preset, select an axis, click on a poly in the mesh shoulder, and a shoulder joint is added halfway deep into the shoulder, with preset deforms and min/max values typical for a shoulder joint. Adjust handles to suit, and min/max values are graphic inclinations I can click+drag. No need to open panels.

Select 'Elbow' preset (or it is auto-selected if you opt to apply bones in a natural order of hierarchy), click on the elbow of the mesh, adjust. Move to 'wrist', repeat. Then 'Knuckle', or 'Standard', etc. Joints may 'auto-link,' or be post-linked in 'Link edit' mode, in different ways. A 'link' between joints can simply be set to 'bend, stretch, rigid, etc' type, so you dont necessarily need to add each of a string of bones for say a spine. A single link may exist between say the neck and hip, with 'bend' selected. A string of finer controls exist 'under the hood' that can be accessed only when needed.

LW_Will
09-03-2013, 12:00 AM
If the dropdown menus can't be eliminated entirely, I think a good alternative would be to simply be able to click an item in 3d space and have that be the selection (kind of like how the "eyedropper" works in most 2d programs). Some of Lightwave's more handy tools like cyclist are *CRIPPLED* by these dropdown menus.

MAYBE you could change a class of commands? If you select a object (bone, null, etc) and then the next thing you select could be Cyclist, or whatever. Open a dialog box that is part of the Cyclist, or whatever, or a dialog box that will ask you what program you needed to open up. Maybe a shift-RMB, or control, or RMB-KEY-ALT... could work... :shrug:

erikals
09-03-2013, 03:33 AM
One more of those in my very long "I didn't know that!" list.
Thank You SIR!
http://erikalstad.com/backup/misc.php_files/wink.gif http://erikalstad.com/backup/misc.php_files/047.gif

stiff paper
09-03-2013, 05:02 AM
Some of Lightwave's more handy tools like cyclist are *CRIPPLED* by these dropdown menus.
You mean the endlessly scrolling vertical lists? I suggested to somebody in the 'old' (i.e. 13 years ago) LW dev team that the pain of endlessly scrolling pop-up lists could be at least improved by simply making the pop-up pay attention to screen resolution, and then the lists could variably have as many columns sideways as sensible, 'cause, yeah, you know... the whole screen. It would prevent scrolling in anything but the craziest cases. But, hey, 13 years ago. And here we are. Still.

chikega
09-03-2013, 05:30 AM
It would be nice if LW Layout had a Setup Tab that would allow an artist to rig a character without thinking about things as "Record Pivot Rotation" and such. I'm thinking of messiah here ... much easier to rig in.

safetyman
09-03-2013, 06:38 AM
Boy, I'd LOVE to advise on the interface, and see LW character tools re-written from the ground up to be more friendly to even non-CG artists who intuitively want to stick their hands in CG and play around, without having to read a lot of manuals and plod thru hours of tutorials. I always see ways of simplifying the interface, without compromising control. But more experienced character animators should be intimately listened to. The UI should be as intuitive for the widest range of user experience, and not be left to whatever a more tech-minded programmer thinks is convenient. I think there should be more R-click pop-up menus, and click+drag-able graphic displays, to minimize the need for opening clunky panels with convoluted value arrays. IKB function started down this path. There should be a beginner friendly layer of functionality that allows any user to jump right into rigging and quickly see characters move, so its FUN. Then there is another layer of complexity for more advanced expression.

-Since v5.6, I wondered why LW doesnt come with a decent, versatile human rig and base figure mesh, that can be easily edited to fit other human figures you may make or acquire. So that beginners can get figures moving quickly without having to learn rigging from the ground up.

-Get all rigging tools the hell out of Modeler. You need real time deform feedback in Layout. And I find weight maps an awfully inefficient way of controlling deforms...

-The long-forgotten 'limited range' tool could be upgraded to reduce need for weight maps- namely, range and falloff values independent for each axis, like with textures- +X,+Y,+Z, -X, -Y, -Z. These values could be represented as click+drag icons in the display that do not require the opening of optional value panels.

-LW in general needs the 'preselection' that I find so useful in Daz. This means that as the cursor passes over items, they each highlight so you know exactly what you are going to select before you click to select (the click-select in Layout can be awfully inaccurate, and unusable at times). And selection bias should favor items closer to POV.

-Preset rigs and common joint types that have automatic settings appropriate to their typical purpose (Genoma sorta, but IN Layout).

I am just thinking out loud here, but I wish I could 'draw' preset 'joint types' and 'joint links' that auto-follow the contours and depth of a mesh. Roughly, I could run the cursor over say a human figure mesh, as polys are 'preselected', while Layout 'sees' thru to the opposite poly normals, and calculates the thickness of the mesh. Select 'Apply Joint' mode, select a 'shoulder/Ball-&-socket' preset, select an axis, click on a poly in the mesh shoulder, and a shoulder joint is added halfway deep into the shoulder, with preset deforms and min/max values typical for a shoulder joint. Adjust handles to suit, and min/max values are graphic inclinations I can click+drag. No need to open panels.

Select 'Elbow' preset (or it is auto-selected if you opt to apply bones in a natural order of hierarchy), click on the elbow of the mesh, adjust. Move to 'wrist', repeat. Then 'Knuckle', or 'Standard', etc. Joints may 'auto-link,' or be post-linked in 'Link edit' mode, in different ways. A 'link' between joints can simply be set to 'bend, stretch, rigid, etc' type, so you dont necessarily need to add each of a string of bones for say a spine. A single link may exist between say the neck and hip, with 'bend' selected. A string of finer controls exist 'under the hood' that can be accessed only when needed.

All good suggestions and I agree 110%..!!

robertoortiz
09-03-2013, 03:28 PM
I agree with Mr Rid. After seeing how CLEAN the workflow on Chronosculpt is, I am convinced that what is keeping the LW character animatioon toolset bogged down are its legacy issues.


Mr Rid said
"...see LW character tools re-written from the ground up to be more friendly to even non-CG artists who intuitively want to stick their hands in CG and play around, without having to read a lot of manuals and plod thru hours of tutorials. I always see ways of simplifying the interface, without compromising control. But more experienced character animators should be intimately listened to. The UI should be as intuitive for the widest range of user experience, and not be left to whatever a more tech-minded programmer thinks is convenient. "


and I think he is SPOT ON.
I would encourage people to sketch some workflows and post them on this thread. I know I will as soon as I finish a killer project for work.

-R

djwaterman
09-04-2013, 02:33 AM
we kind of have it > http://youtu.be/wjm52B1Oqkg

Erikals, what is that view setting you have at the beginning of the video where the geometry is blue and semi transparent?

DogBoy
09-04-2013, 03:26 AM
Erikals, what is that view setting you have at the beginning of the video where the geometry is blue and semi transparent?

I think it is just a semi transparent surface.

erikals
09-04-2013, 04:24 AM
yep, you gotta have OpenGL transparency on though, click "d" and go to "GL"...

stevegraymusic
09-05-2013, 10:37 AM
Forgive me if some of this stuff is already in place and I just haven't discovered it.
I consider myself more of a creative person than a tech person - so anything that would make the process less technical would be great!


1. Biped presets that include face rigging in Genoma !

2. "Create rig" function in Layout also creates the sliders for body and face controls

3. Dynamic soft body rigging system. Fat, breasts, hair, accesories etc move interactively for more natural movement instead of a calculation
ex.
http://www.youtube.com/watch?v=LLb_Ck3orPg
http://www.youtube.com/watch?v=WNZw57nC7Cs

4. Preset hand systems for common hand poses (closed fist , spread, curl, clench, point, etc)

5. Deformation system - In addition to bones while rigging you have markers to place for deformation (hold bones etc) - i.e. Rhiggit

6. Onion skinning function

7. Built in lip sync plug - i.e. TAFA, 3rd wish, magpie etc

8. Easy switching back and forth from animating to rig edit mode to make rig adjustments in Layout
and back to animating as often as necessary - i.e. Rhiggit
(How about just acquiring Rhiggit!! and combining with Genoma !! :)

9. (perhaps there is already a way to do this but...Key Smoothing function ? -
If you start out your animation by blocking poses (linear incoming curves)
A simple way to convert to splines - tweening without having to adjust every single keyframe of every single object or bone
(heading pitch and bank X - Y and Z in the editor to see the effect of a different curve

11. Genoma - simpler bone markers for rig setup - what should be moved or should not be moved (color coded etc)
while placing bones during rigging - (some of the connectors, rotation bones etc are not clearly marked or are so small you could move them inadvertenly - but they have a significant effect on the outcome when the rig is created)

12. A walk cycle builder creator

13. Ability to adjust and apply stepped linear etc in the timeline instead of having to open the editor

14. set of commonly used expressions that can be applied to a rig

15. User friendly morph mixer with sliders - ( TAFA)

16. ability to rename bones and objects in the drop down list in layout

geo_n
09-06-2013, 11:08 AM
Pre highlighting and items that you can just drag, move, rotate without the need to click once on it before manipulating it.
Ikboost can do it.

Surrealist.
09-06-2013, 12:47 PM
Heck just give me items I can actually click on and select without some voodoo, summoning up spirits from the great beyond or magic cantation to invoke a spell on the interface.

CaptainMarlowe
09-06-2013, 11:20 PM
we kind of have it > http://youtu.be/wjm52B1Oqkg

Except that if you try to do it with a genoma rig, it asks to rename the bones.
If you say yes, the weight painting will be OK, but Genoma won't work (and crashes layout on my Mac) because of the name changing.
If you say no, the weight painting will be screwed.
Another exemple of a tool implemented without correct liason to the others...

Greenlaw
09-07-2013, 06:07 AM
Except that if you try to do it with a genoma rig, it asks to rename the bones.
If you say yes, the weight painting will be OK, but Genoma won't work (and crashes layout on my Mac) because of the name changing.
If you say no, the weight painting will be screwed.
Another exemple of a tool implemented without correct liason to the others...

I haven't had that experience but I agree with you and Mr. Rid--the functionality that Vertex Paint offers needs to be moved to Layout and it needs to work with modern LightWave tools and features. And I'm not asking for a Vertex Paint update--what we need is a completely new set of tools to create/assign/paint weights directly on a rigged character within Layout's viewports, and it needs to work with both bones and joints.

To be honest, it's never made sense to me that we have these features in Modeler and not in Layout. I assume Vertex Paint was designed this way to work around limitations of LightWave from the version 6 and 7 era...but that was a really long time ago wasn't it? :)

G.

tyrot
09-07-2013, 11:55 AM
I just ask a plugin from our plugin writers.

BVH to HMOT ...

Hmot with MotionMixer support. I dont think it should be so hard to code.
Advantages:

We do not need to manually save HMOT and remap the each bone every time. So plugin may work like this
1- Select BVH
2- Create Actor Name
3- Export HMOT

Then we can open motionmixer and use all those BVH's without any problem since they have same actor and bone setup.

Dodgy or XchrisX can make it work i HOPE!

jasonwestmas
09-07-2013, 12:15 PM
To be honest, it's never made sense to me that we have these features in Modeler and not in Layout. I assume Vertex Paint was designed this way to work around limitations of LightWave from the version 6 and 7 era...but that was a really long time ago wasn't it? :)

G.

Very long time ago. I could say that about a lot of tools. Vertex map editing in general should be in layout too.

Surrealist.
09-07-2013, 03:07 PM
But are these things really practical to ask for I wonder? I mean they are going to keep making improvements with this old LW. But I think a lot of the things we would want to see require a rewrite of LW to be in place first. At least that is what it seems like. And so I am curious as to what NewTek as in mind in that regard. And certainly I don't expect them to come out and say anything definitive. So to me it just seems a little like groping in the dark. I know what I'd like to see. I have outlined it many times as have others.

I think it would be wise to wait and see what they do with LW 12 and 13 to have a better idea of what LightWave will be capable of. Maybe it will be another new tech like is in CS that is able to do things entirely differently. That would be interesting. And I'd like to see something more innovative personally. We already have lots of choices on the conventional side of things. (Messiah Blender) As they are today. Maybe more interesting and powerful approaches are in store for the future.

That would be the surest way to set LW apart from the pack. CS is going in the right direction I think.

Here's to hoping anyway.

jasonwestmas
09-07-2013, 04:03 PM
Like you Richard, I just use whatever works, I stopped waiting for NT a long time ago. If the animation tools that are in modeler aren't in layout by LW 12. I simply won't ask anymore.

CaptainMarlowe
09-08-2013, 01:33 AM
I haven't had that experience but I agree with you and Mr. Rid--the functionality that Vertex Paint offers needs to be moved to Layout and it needs to work with modern LightWave tools and features. And I'm not asking for a Vertex Paint update--what we need is a completely new set of tools to create/assign/paint weights directly on a rigged character within Layout's viewports, and it needs to work with both bones and joints.

To be honest, it's never made sense to me that we have these features in Modeler and not in Layout. I assume Vertex Paint was designed this way to work around limitations of LightWave from the version 6 and 7 era...but that was a really long time ago wasn't it? :)

G.

You bet ! ;)

erikals
09-08-2013, 02:03 AM
urmf... :/

exporting a Genoma rig to skelegons with D-Storm "Load Bones" doesn't work... http://erikalstad.com/backup/misc.php_files/confused_002.gif

Dexter2999
09-08-2013, 02:08 AM
Sorry if this is a stupid comment/question. (I don't actually do much 3D anymore.)

But Vertex isn't in Layout because Layout isn't geared to deal with point/edge/poly info. per se, is it? Attributes like this get defined in Modeler, then Layout deals with objects as a whole and takes the detailed attribute info. the users generate from Modeler and apply it to the model. Correct?

So, this doesn't sound like a trivial matter but a substantial rewrite issue, doesn't it?

Or am I just straight up wrong in my understanding of this?

jeric_synergy
09-08-2013, 10:27 AM
Pretty much. It's like (and I'm a little behind so maybe we can do this now) how weight painting SHOULD be in Layout, where it'd be useful, but isn't. Jumping back and forth just KILLS interactivity.

erikals
09-08-2013, 11:29 AM
exporting a Genoma rig to skelegons with D-Storm "Load Bones" doesn't work... http://erikalstad.com/backup/misc.php_files/confused_002.gif

yes it does... >
http://www.youtube.com/watch?v=jHhTauxOvsw

jasonwestmas
09-08-2013, 11:45 AM
Sorry if this is a stupid comment/question. (I don't actually do much 3D anymore.)

But Vertex isn't in Layout because Layout isn't geared to deal with point/edge/poly info. per se, is it? Attributes like this get defined in Modeler, then Layout deals with objects as a whole and takes the detailed attribute info. the users generate from Modeler and apply it to the model. Correct?

So, this doesn't sound like a trivial matter but a substantial rewrite issue, doesn't it?

Or am I just straight up wrong in my understanding of this?

Well, it's been like this ( and I think I've been in the loop for 90% of what NT has communicated to us users ) Some of us, (mostly animators) point out specific workflow issues that slow things down drastically because of the lack of interactivity between tool-sets. Such as the vertex mapping editing issue where we have animation tools in modeler and not in layout. In the past we were told by some people from NT that certain things just can't be done and we needed a rewrite. Then years later we get the opposite message and are told that if LW is broken up in a modular fashion and reorganized then a brand new mesh and animation system can be inserted without totally sacrificing the Lightwave familiarity. . .which will open up new capabilities and keep existing users. Whether this latest direction is only about performance improvements or also includes changing the workflow for the better, that's still unknown to us. Quite often we are told these days that anything is possible with LW, but it's just a matter of time and that this current path LW3DG is on is faster than a complete rewrite of the interface.

As for getting specifics for animation workflow improvements. I have not seen many announcements on a fundamental level after LW 9.5. Only adding onto an extremely old deformation and keyframing system fundamentally speaking.