Off topic stuff moved from "Refining Our Direction at HiveWire 3D" thread

[Semicharm]

@kobaltkween When I mentioned "unimesh", I wasn't referring to Genesis or compatibility with DS in any manor whatsoever. As I stated above, Poser's geometry format is fragmented, literally broken. That has caused problems and will continue hamper all future development with the app. At this point, there's no escaping the numerous pitfalls without switching to a solid "unimesh" geometry. That has been discussed and documented quite thoroughly at this point. Daz was right to make the switch to solid geometry with Genesis and Poser will eventually need to do something similar. My comment was out of concern of the long term survival of Poser, nothing about Poser supporting Daz products.

 

[Ken1171]

Nothing I said was about morphs. Furthermore, what you said isn't entirely true either. Changing the vertex count on one part, say an arm, does not affect the head or any other part because Poser uses fragmented "grouped" obj files. It's been that way for decades. You've certanly worked with Poser long enough to have figured that out by now.

I am sorry, but unfortunately that is not how it works. Poser indeed splits the geometry into separate, split geometry, but only after you load it from disk. If you compare the original OBJ on disk, the vertex count is different from the one loaded in Poser. That is, the original geometry on disk, and the one being used in Poser are 2 different things. The original OBJ geometry is a single mesh, and the one loaded in Poser is not.

Therefore if you change even a single vertex on any part of the body, it will invalidate all existing morphs, no matter what part of the body it affects. This is like that in both Poser and DAZ Studio, except that DS doesn't split the geometry when loaded. It always works with a single mesh.

But don't take my word on it - you can test this yourself. Export Dawn from Poser to OBJ, and then compare the vertex count with her OBJ in the geometries folder - it won't match. Another test - remove a single vertex from the original OBJ from the geometries folder, and then load Dawn in Poser, and try to use any of her morphs - it won't work. They are now all invalid. It's better if you see it for yourself than to just take my word on it. Just give it a try and find out.

 

[Ken1171]

@kobaltkween When I mentioned "unimesh", I wasn't referring to Genesis or compatibility with DS in any manor whatsoever. As I stated above, Poser's geometry format is fragmented, literally broken.

I would just like to add that the geometry we create for use in Poser is not fragmented. It is a single piece, here referred to as "unimesh". This term is not owned by DAZ or only used in DS, it's the same used in Poser. The issue is that Poser changes this unimesh geometry once loaded, splitting it into separate parts. To make it worse, Poser is unable to put the mesh back into one piece when exporting to OBJ, or any other format (even GoZ), so all geometries exported from Poser no longer match the original. One direct consequence is that morphs exported from Poser will not work in any other program but Poser itself.

You can test this by loading Dawn in Poser, and then exporting her to zBrush with GoZ. Then export the model in zBrush back to OBJ, and try to use it as a morph target in DAZ Studio. It won't work. You will get a "geometry mismatch" error message. If you compare this geometry, or any exported from Poser to the original from the Geometries folder, they won't match in vertex count.

The main point here is that the geometry from Poser products is NOT fragmented- at least the original ones from the geometries folder. They are indeed unimesh. The only exception is when some content creators use the geometry exported from Poser, which will always be fragmented. I used to in the past, but not anymore. Some vendors, like DarkSeal, still uses geometry exported from Poser, but this is rare nowadays, and not quite recommended.

 

[caisson]

AFAIK the reason that Poser breaks meshes goes like this: rigging as a process involves creating a system of bones or joints and then telling the program which part of the mesh is affected by which bone; this is called skinning or binding a mesh. The way Poser has done this since the very first version is to match named polygon groups to bones with the same name. When a grouped mesh is brought into Poser and bound to a rig Poser then internally splits/breaks/fragments that mesh into pieces as defined by those groups so that it can deform the mesh with the rig. Every refinement to the rigging system in Poser, like weight mapping, has been built on top of this core functionality.

It does have disadvantages, and there are other methods of binding, but changing the program that fundamentally would be a significant task (I'm not betting on it at the moment!). Obviously as every Poser figure since Poser 1 has been made this way, it does work, and with Zbrush and the GoZ bridge multi-resolution morphs and reverse deformation morphs are possible, which I know is not an ideal solution by any means, but it is a solution. (Incidentally, this is why La Femme has a grid-like topology - it's quite deliberate as it makes sculpting higher resolution morphs easier.)

@kobaltkween's point about Poser users needing to support new Poser figures is well made. That the Hivewire base figures now ship with Poser is a great selling point, Dawn 2.0 is looking awesome, La Femme is getting support, L' Homme is in progress ... I can see plenty of shiny things, and some of them are here right now

 

[Ken1171]

AFAIK the reason that Poser breaks meshes goes like this: rigging as a process involves creating a system of bones or joints and then telling the program which part of the mesh is affected by which bone; this is called skinning or binding a mesh. The way Poser has done this since the very first version is to match named polygon groups to bones with the same name. When a grouped mesh is brought into Poser and bound to a rig Poser then internally splits/breaks/fragments that mesh into pieces as defined by those groups so that it can deform the mesh with the rig. Every refinement to the rigging system in Poser, like weight mapping, has been built on top of this core functionality.

That is correct, the reason body groups exist in both Poser and DS is to bind the bones to the mesh. However, those body groups do not need to split the geometry to exist. That was a design decision made over a decade ago when Poser was created. For example, DAZ Studio also uses body groups, but it keeps the geometry as a single piece. We have many examples of 3D programs that break the geometry to create groups (Poser, 3DSMAX, Maya, etc), and the ones that don't (DS, Modo, Blender3D, etc).

So there is a distinction between why we need body groups for, and whether or not they need to split the geometry to exist. Poser and DS both use body groups, but one has decided to split the geometry to use them, and the other didn't. One thing doesn't require the other, but there are rather serious consequences if we choose to split the geometry like Poser does.

One interesting thing about groups is that it is possible to rig conforming clothing in Poser using the entire model as a single group. The named groups make the binding easier, but they are NOT required. We can do the binding manually using the Joint Editor without having named groups. To do so, all you need is to edit a joint, and whatever you paint weights on will be bound to the respective bone. I have done this with a shirt, where the entire geometry was the "hip" group. That was the only group in the whole shirt, and it works! ^____^

 

[carmen indorato]

That does sound like a good idea, but unfortunately, it is not possible. Changing even a single vertex would invalidate all existing morphs created for the figure. Once the figure is released and morphs are created for it, its topology is cast on stone.



Just to complement - for many years I used to believe DS was created because Poser was being bought and sold many times in a row, and DAZ wanted to be more independent from it. That was until Chris told me the actual reason was because Poser's owner company had decided to enter the content market with the creation of Content Paradise, which represented direct competition to DAZ. So it comes down that CP was the actual reason why DS was created. If only they knew that CP would never present serious competition to ANYONE, DS could have never existed.

Well the way I understood it DAZ was created when Zygote saw the potential of creating content for Poser user. Zygote at the time created 3d content for higher end 3D users. Poser gave them a great cash cow incentimv eot branch off a subsidiary to capitalize on the app users and the obvious growing market opening up. Content Paradise wa sno the only content creators out there way back then either. There were many otherd who set up shop and closed down for whatever reason. Renderosity and DAZ remained as the biggest out there.

 

[carmen indorato]

I would just like to add that the geometry we create for use in Poser is not fragmented. It is a single piece, here referred to as "unimesh". This term is not owned by DAZ or only used in DS, it's the same used in Poser. The issue is that Poser changes this unimesh geometry once loaded, splitting it into separate parts. To make it worse, Poser is unable to put the mesh back into one piece when exporting to OBJ, or any other format (even GoZ), so all geometries exported from Poser no longer match the original. One direct consequence is that morphs exported from Poser will not work in any other program but Poser itself.

You can test this by loading Dawn in Poser, and then exporting her to zBrush with GoZ. Then export the model in zBrush back to OBJ, and try to use it as a morph target in DAZ Studio. It won't work. You will get a "geometry mismatch" error message. If you compare this geometry, or any exported from Poser to the original from the Geometries folder, they won't match in vertex count.

The main point here is that the geometry from Poser products is NOT fragmented- at least the original ones from the geometries folder. They are indeed unimesh. The only exception is when some content creators use the geometry exported from Poser, which will always be fragmented. I used to in the past, but not anymore. Some vendors, like DarkSeal, still uses geometry exported from Poser, but this is rare nowadays, and not quite recommended.

Hm......thus explaining why I have always had problems with content from creators like DarkSeal.

 

[carmen indorato]

@kobaltkween When I mentioned "unimesh", I wasn't referring to Genesis or compatibility with DS in any manor whatsoever. As I stated above, Poser's geometry format is fragmented, literally broken. That has caused problems and will continue hamper all future development with the app. At this point, there's no escaping the numerous pitfalls without switching to a solid "unimesh" geometry. That has been discussed and documented quite thoroughly at this point. Daz was right to make the switch to solid geometry with Genesis and Poser will eventually need to do something similar. My comment was out of concern of the long term survival of Poser, nothing about Poser supporting Daz products.

Unimesh was also used as a selling point starting with DAZ Mil1 and Mil2 figures but never really followed up on. Though their use of the word apparently was totally differnt from what is being discussed here.

 

[carmen indorato]

I apologize for upsetting you. I was responding largely to this:



And more generally to the preceding back and forth about Rendo and DS products. Or more widely, a content community that supports two (or more) pieces of very, very different software. I'm just saying that I think we should look forward, and let those platforms naturally diverge without shaming anyone for working in one and not the other. And that Poser customers need to support modern Poser figures, like the HiveWire3D family, La Femme, and Ero, rather than wait for Poser development to follow behind DS development.

I only brought up the past as a reason to look forward, wishing all the developers good luck on their divergent paths.


What determines content is demand and sales. Because once most vendors have an established workflow on a platform, the core of their work is (for models) in their modeler and texturing tool of choice. Most content creators' primary love is working in their modeling, sculpting, or texturing tool(s), not the final software platform or figure they publish to. Which means that every hour spent working on a version that sells less rather than a version that will sell the most loses money.

People often say, "Why not make version X supporting [insert figure or platfom here] and make more money?" The problem with this logic is it ignores any other products the vendor makes. And it implicitly assumes that porting content is simple, easy, and fast, when it in fact doubles every aspect of your work except the part you love most. A new version will still need publishing to a platform, testing yourself, overseeing beta testing, doing promo renders and writing promo text, promoting the new version, packaging everything, etc. The only thing you _don't_ have to do is the part that makes all that other stuff worthwhile.

It's a whole lot easier and more lucrative for most vendors to just pick a single workflow and spend that time on a new product rather than a lower-selling version of an existing product. More importantly, it's more _fun_.

Demand drives the what vendors publish for. This is why many of today's DS-only vendors were yesterday's Poser-only vendors.

Sorry, I guess I wasn't clear in this, either. I was (and am) saying that Poser customers need to invest in new Poser figures if they want new Poser content. Invest in them and show them off.

I genuinely get not being able to afford much in the way of indulgences. I really, truly get that. I also get wanting to stick with a figure you've invested _so_ much in. My own V4 library is huge. I also get wanting to stick with the certainty of popular DAZ figures. I get why the community still clings to both older figures and DAZ figures. But for the Poser content community, and Poser itself, to survive, Poser content customers must invest in new content for modern figures. Holding back might feel safer, but en masse (not putting this on any one person _at all_; you do what you need to), it's slowly strangling the community.

Ken1171 can't do it for all of us.

And now I'll get back to Blender's #Nodevember and working on a set of materials for Poser.

Figures are NOT he only need we Poser users have. Content does not start and end with figures. we also need newer and better sets and environments, props and hair that works on older and newer figures, and of course animals and creatures that render better in newer versions of Poser and ook more realistic standing next to both older human characters and newer ones. I don't have to be reinvesting on Human figures when my old er one still work great as as I can expect. But I do want the other stuff mentioned above and will invest in it if properly made and supported.
I may not buy into Dawn or Dusk but I will spend money on the animals and creatures and environments and props and hair here and elsewhere.

 

[Ken1171]

Well the way I understood it DAZ was created when Zygote saw the potential of creating content for Poser user.

That was referring to why the "DAZ Studio" software was created, not the DAZ3D company.

 

[Ken1171]

Unimesh was also used as a selling point starting with DAZ Mil1 and Mil2 figures but never really followed up on. Though their use of the word apparently was totally differnt from what is being discussed here.

Oh yes, good point. Indeed "Unimesh" in that context means something completely different. That was instead referring to how the Millennium figures all used the same base mesh between the same family generations. That was revolutionary at the time, because it has changed the meaning of "figure". For example, Millennium 4 figures were the first to release Aiko as a morph to Victoria 4. Before this, Aiko 3 and all the other Mil3 bases were independent figures.

Hivewire3D uses the same "Unimesh" approach, where Dawn, Dusk, Baby Luna, and all other HW figures share the same base mesh. This is what allows us to morph Dawn into Dusk, Baby Luna, or even the HW Gorilla with ease. I like this concept a lot. Genesis in DS uses that too.

 

[Semicharm]

I am sorry, but unfortunately that is not how it works. Poser indeed splits the geometry into separate, split geometry, but only after you load it from disk. If you compare the original OBJ on disk, the vertex count is different from the one loaded in Poser. That is, the original geometry on disk, and the one being used in Poser are 2 different things. The original OBJ geometry is a single mesh, and the one loaded in Poser is not.

Therefore if you change even a single vertex on any part of the body, it will invalidate all existing morphs, no matter what part of the body it affects. This is like that in both Poser and DAZ Studio, except that DS doesn't split the geometry when loaded. It always works with a single mesh.

But don't take my word on it - you can test this yourself. Export Dawn from Poser to OBJ, and then compare the vertex count with her OBJ in the geometries folder - it won't match. Another test - remove a single vertex from the original OBJ from the geometries folder, and then load Dawn in Poser, and try to use any of her morphs - it won't work. They are now all invalid. It's better if you see it for yourself than to just take my word on it. Just give it a try and find out.

I've already done that when trying to fix issues with various figures. Of course, that requires understanding the how the vertex indexes in the OBJ file format work and how Poser uses them, how they they are grouped internally AND how that relates to the references used in morphs. I completely understand all of that and even done edits of obj files and morph data by hand. Well, I sometimes cheat and import the obj data into a spreadsheet to recalculate the indexes. So yeah, I know exactly what I'm talking about. When done correctly by someone who knows what they hell they're doing, inserting a vertex into a model only affects the group its added to. In fact, it can even be done without breaking the morphs within the group, but that can be considerably more complected in practice. Hell, SM added entire geometry to Roxie's head when they updated her with transparency mapped lashes, without having to rebuild all of the morphs. Actually they missed a few, such as BrowUp, which apparently still use the original morph. It works perfectly fine, except the new geometry for the lashes don't move with the eyelid. I was also able to convert a morph for the original Roxie to Roxie1.1 with only a minor edit.

 

[Semicharm]

@Ken1171 A simpler way to put it is that you're saying the game of Jenga in impossible, merely because you only know how to play it with a sledgehammer.

 

[Semicharm]

It's damn well is possible to alter the geometry without having to rebuild any morphs when you actually comprehend how Poser works.

 

[LisaB]

You are stating this as absolute truth where it is not. There were many factors that led to Daz Studio being developed, not just the creation of CP. Chris talks about some of it in the spotlight interview published here in the forums.

Edited to add - perhaps we could step out of the box of the past a bit. The past has wisdom hidden in it that we can pack our pockets with as we step into the future. But we don't have to pull that wisdom out of our pockets every time something changes. If our hands are full of the past, how will we pick up the shiny objects that may appear on the path to the future which is yet unboxed and undefined?

I am just stating what Chris has told me when I said I thought it was because Poser was being bought and sold so often. That's what I believed for many years, until Chris corrected me. So will we be censored for talking about the history of the hobby we love so much?

I talked to Chris about this and you are correct. It was when CP was brought to the table with an invitation to join in that made the point for the creation of Daz Studio.

No Ken, you will not be censored for talking about the history of what we do. All we ask is that you are aware of how some of what you state as your opinion can be seen as an invitation to argue the point which can quickly move into a situation that requires moderation.

My statement on full pockets, the path to the future and shiny objects wanting attention still stands.

I apologize for upsetting you. I was responding largely to this:

And more generally to the preceding back and forth about Rendo and DS products. Or more widely, a content community that supports two (or more) pieces of very, very different software. I'm just saying that I think we should look forward, and let those platforms naturally diverge without shaming anyone for working in one and not the other. And that Poser customers need to support modern Poser figures, like the HiveWire3D family, La Femme, and Ero, rather than wait for Poser development to follow behind DS development.

I only brought up the past as a reason to look forward, wishing all the developers good luck on their divergent paths.

I am not upset at all. I said something because I do not want this thread to devolve into negativity and arguments. I have no issue with what you have shared here, however, I do see things that have potential to turn from sparks to fires. There are far more creative ways to invest time that putting out needless fires.

 

[KageRyu]

For clarification regarding Unimesh for those interested, as I remember Daz3D's introduction and Hype they had their own interpretation vs that of the general 3D world.
unimesh(general 3D) - as kobaltkween pointed out in general 3D application terms any figure that is one solid piece of polygons and welded vertexes is a unimesh. This is not exactly What Daz3D meant. This is opposed to an object made of several unconnected polygonal structures that may appear to be touching at joints or seams but actually have unwelded vertexes overlapping one another in those locations (for those non-modelers among us).
Unimesh(Daz3D) - When Daz3D began bandying about it's Unimesh terms (I believe with the Generation 2 millenials) the major hype and focus as explained at that time was that the Figures that carried the Unimesh designation were all based on the same basic mesh figure. They shared the same polycount and vertices. They used similar UV mapping. In theory this allowed them to share textures (which was the major selling point) but also to a limited extent some of the morphs and would make it easier for creators to transport morphs between the figures to make identical morph packs between the figure family. This was built upon more with Generation 3 Millenials and you might almost say this was when the groundwork ideas for what became Genesis was being laid. I have lost track of how many forum reboots agio this was, but the forums were much more open and friendly back then, with more behind the scenes looks at the technology, and Daz3D itself had different management and was a much more personable company IMHO back then (we are talking 2004-2206 era). I have used some of the various Generatio 3 morphs on a Generation 3 figure not directly intended, and have swapped textures on them. It is easiest with certain figures (Freak 3, David 3, and M3...but some of the morphs will work on the Gen 3 teens). My point is, though, that Daz's original intended meaning of Unimesh was a single base mesh for all of the figurees - and this was exactly how they defined it at that time in the forums. Their intended definition and usage may have changed since then, but was never the broader meaning used by the wider 3D community.
Oh... edit to add, direct from one of the older members at Daz3d...
"With the third generation, they used the same mesh for everything. This doesn't mean that the SHAPE is the same but that they have the exact same vertices in the same position. This includes Victoria 3, Aiko 3, Stephanie Petite, David 3, Michael 3 & Freak 3. Compatibility was not 100% because the way the mesh was grouped for M3 & F3 was different from the others. All these figures also had very similar UV mapping. This was the first unimesh.


The 4th generation characters share another, newer unimesh. I think that the grouping for the 4th generation may be more consistent, but the UV mapping is less. (DS can switch UV maps it its necessary). Note that the Girl 4 is a unimesh figure while the original girl (Kim Goosens') was not."

 

[Ken1171]

No Ken, you will not be censored for talking about the history of what we do. All we ask is that you are aware of how some of what you state as your opinion can be seen as an invitation to argue the point which can quickly move into a situation that requires moderation.

For goodness sake, Lisa - I have already told you that was NOT my opinion, but something that Chris told me on occasion. I was just passing the info to someone who was thinking the same as I did before Chris had clarified that for me. Not once, but you have called me a liar TWICE before you've decided to double-check with Chris, just to find out I was telling the truth. I am personally offended by these remarks. You are NOT doing me any favors by harassing me in the forums like this. Are you trying to publicly demoralize me or something? This is ridiculous and it has to stop.

 

[Hornet3d]

Figures are NOT he only need we Poser users have. Content does not start and end with figures. we also need newer and better sets and environments, props and hair that works on older and newer figures, and of course animals and creatures that render better in newer versions of Poser and ook more realistic standing next to both older human characters and newer ones. I don't have to be reinvesting on Human figures when my old er one still work great as as I can expect. But I do want the other stuff mentioned above and will invest in it if properly made and supported.
I may not buy into Dawn or Dusk but I will spend money on the animals and creatures and environments and props and hair here and elsewhere.

Now that is something I fully agree with, most of my money spent on the hobby for the past few years has been props, environments and other stuff like HDRs and materials. I work with Dawn, I like Dawn and the only other figures I work with are M4 and occasionally V4. Further more I am not in the market for characters but I do need scenes I can use the figures in, rooms, buildings, landscapes space ships the list is a long one. Two of my favourite artists over at Daz are Petipet and Jack Tomalin, Petipet for modern and futuristic housing and transport, Jack Tomalin for the fantastic graveyard scenes.. At Renderosity it is Truform and Blue Tree Studios much for the same reason as Petipet. Here I have purchased almost all of the the dynamic clothing by Lully, the flowers from Lisa with the textures from Napalm Arsenal. Backgrounds from Rowan 54 and Photo GG. Clothes for Dawn from Hunter 3D. Superfly Materials from Phd and so many other artists I could go on for a long while. None of this is for figures in the way the term is normally used and almost all of it makes use of some of the new features in Poser 11 such as Superfly. If I try an use a scene created for an earlier version on Poser I am generally dissatisfied with the results unless I have played heavily with materials and set the scene to mask some of the short comings.

I love the way Hivewire is going and the only other figure I am really interested at the moment is Dawn 2.

 

[Semicharm]

@Ken1171 Being dismissed and not taken seriously by someone who you'd think would know better... Now you know how I feel.

 

[LisaB]

This thread is locked from further replies for the time being. It will be unlocked when I am ready to unlock it.