Ahhh, so not necessarily something an end user would be dealing with?
-
Welcome to the Community Forums at HiveWire 3D! Please note that the user name you choose for our forum will be displayed to the public. Our store was closed as January 4, 2021. You can find HiveWire 3D and Lisa's Botanicals products, as well as many of our Contributing Artists, at Renderosity. This thread lists where many are now selling their products. Renderosity is generously putting products which were purchased at HiveWire 3D and are now sold at their store into customer accounts by gifting them. This is not an overnight process so please be patient, if you have already emailed them about this. If you have NOT emailed them, please see the 2nd post in this thread for instructions on what you need to do
Dawn 2.0 Underway
- Thread starter Chris
- Start date
Only if the outfit doesn't already support the body shape in Poser. In DS it happens automatically, but how well it fits a new shape depends on many things, like how different the shape is, and how well it was made. In the past, SMS had [inexplicably] decided to remove this feature from the standard Poser version, but Rendo promptly rectified this mistake, not only adding this feature back, but also eliminating the "Pro" and "Game Dev" versions from Poser, so everybody gets all features.
unreal
Noteworthy
Also, it depends on what the outfit is. (loose, tight, sleeveless, legless, skirt, gloves, shoes...)Only if the outfit doesn't already support the body shape in Poser. In DS it happens automatically, but how well it fits a new shape depends on many things, like how different the shape is, and how well it was made. In the past, SMS had [inexplicably] decided to remove this feature from the standard Poser version, but Rendo promptly rectified this mistake, not only adding this feature back, but also eliminating the "Pro" and "Game Dev" versions from Poser, so everybody gets all features.
I've found the "fit room" works best if I do a rough sculpt first. Obviously, automation has to make assumptions about things that are maybe incorrect for what's going on. You can get a lot of mileage from magnets and the morph brush. And then a touch of the brush after.
The biggest hassle is when the figure contains actors not in the clothing (waist, etc.). Auto group is the only thing, short of using the group editing tool (which is not Poser's best implemented tool)
Or when the clothing contains extra bones. Sometimes the extra handle for skirts or hair or ties, belts.
I wish there was a way to make poser do a kind of "merge replace". That is, replace the same named bones and transfer the extra bones, adjusting those extra bones so as makes sense in relation to the new "same named" bones.
I suppose a script could do it. If only there was someone good with Python that had track record of writing gee-whiz utilities.
Totally true!
Agreed. I use a mix of the Morph Tool and Magnets to put things into place and shape before delving into the Fitting Room. I never allow it to shrink-wrap anything, because that's what it will do.
Auto-group should NEVER-EVER be used, at least until the upcoming unimesh Poser comes out. If you do, you will destroy the mesh integrity, so grouping elsewhere is the only option. I recommend MarkDCs Auto-Group Editor, which is standalone, so it doesn't depend on Poser versions to work. When unimesh Poser comes out, none of this will matter anymore, but until then, no Auto-Group in Poser should never happen.
Those bones have to be created in Poser using the Setup Room. Unless we have body handles on the mesh, the Fitting Room won't make them for us.
Ghost bones need to be created manually. I think there is no way around that. There is no way to tell Poser what a ghost bone is supposed to do, or where it should be placed. It only makes sense to us.
unreal
Noteworthy
Can the old rig, be read? There would be the bones and weight maps. Surely there's some way to map that onto a new skeleton? So long as we know what heuristics are involved, and there's some order to their application, then it can be automated.Totally true!
Agreed. I use a mix of the Morph Tool and Magnets to put things into place and shape before delving into the Fitting Room. I never allow it to shrink-wrap anything, because that's what it will do.
Auto-group should NEVER-EVER be used, at least until the upcoming unimesh Poser comes out. If you do, you will destroy the mesh integrity, so grouping elsewhere is the only option. I recommend MarkDCs Auto-Group Editor, which is standalone, so it doesn't depend on Poser versions to work. When unimesh Poser comes out, none of this will matter anymore, but until then, no Auto-Group in Poser should never happen.
Those bones have to be created in Poser using the Setup Room. Unless we have body handles on the mesh, the Fitting Room won't make them for us.
Ghost bones need to be created manually. I think there is no way around that. There is no way to tell Poser what a ghost bone is supposed to do, or where it should be placed. It only makes sense to us.
In machine learning, we'd just take a bunch of meshes, do the process manually, then let an ML algorithm derive the heuristics for us. No?
I was creating more clothing for Dawn2 this week, and I noticed her skeleton has ghost bones that can cause some confusion in the Fitting Room, and later when rigging the outfit. Silly me, since I was the one suggesting these should be ghost bones in the first place. Those are bones that don't have a group on the figure, but that doesn't mean the clothing can't have the missing groups - but I never tried that before, so this is new with Dawn2. This is amusing - I have added the groups to clothing, groups Dawn2 doesn't have, and it works all the same. That's the beauty of ghost bones - it would had worked whether or not I added the groups to the clothing. It just makes things easier if I add the groups beforehand, or else Poser might not understand what I was trying to do.
I am getting this kind of out of context. Could you show me an example of what you want? I can already advance that Python doesn't have access to how Poser projects morphs and weights over models. In a way, it wouldn't be practical for the dev team to provide that functionality to Python now, because all of that is about to change with the new unimesh workflow. Nobody knows how that's going to work, or how long it will take. We only know Bondware's first time estimate didn't work out. The devs told me they are not adding anything else to Python until unimesh Poser is finished, which to me sounds like Python additions in Poser are suspended indefinitely. My guestimate is that we should be seeing the first glimpses of unimesh Poser somewhere in 2022's first quarter. But my guess is as good as yours.
I am getting this kind of out of context. Could you show me an example of what you want? I can already advance that Python doesn't have access to how Poser projects morphs and weights over models. In a way, it wouldn't be practical for the dev team to provide that functionality to Python now, because all of that is about to change with the new unimesh workflow. Nobody knows how that's going to work, or how long it will take. We only know Bondware's first time estimate didn't work out. The devs told me they are not adding anything else to Python until unimesh Poser is finished, which to me sounds like Python additions in Poser are suspended indefinitely. My guestimate is that we should be seeing the first glimpses of unimesh Poser somewhere in 2022's first quarter. But my guess is as good as yours.
This perked my interest. Are you doing mirror bit in Poser? If so how does one mirror a morph to the other side (unless your speaking about the powerful Morph Brush tool set that I use all the time)
Speaking to your convo about groupings. I recently ran into an issue for my old Supersuit for Dusk and really liked how I have that set up with the material zones so thought I'd resize it and make it fit Dawn SE in Poser. Well when I got it all fitted I needed to use the Morph Brush to tighten the fit and found that the groups (probably fitted in Poser a good long while back) left me with very undesirable effects, the zones as I was using the smoothing brush from the Morph Brush tool set was SPLITTING the suite up as I smoothed and tightened things up.
I MUST remember to use the Auto Group Editor from now on but I'm always forgetting to do that! My fault but a reminder NOT to let Poser do it because it screws it all up.
On an aside it gave me an opportunity to fire up Blender 3.0 and weld and smooth out the suit and it worked great without destroying all the material zones like ZBrush would (which I have no idea why they don't have that fixed yet in ZBrush, very off putting) so looks like my relationship with Blender will mature a little more as I need to really get in there and learn what else I can do since ZB is going to be acquired by Maxon and who knows what it's future is going to be in the long run.
I MUST remember to use the Auto Group Editor from now on but I'm always forgetting to do that! My fault but a reminder NOT to let Poser do it because it screws it all up.
On an aside it gave me an opportunity to fire up Blender 3.0 and weld and smooth out the suit and it worked great without destroying all the material zones like ZBrush would (which I have no idea why they don't have that fixed yet in ZBrush, very off putting) so looks like my relationship with Blender will mature a little more as I need to really get in there and learn what else I can do since ZB is going to be acquired by Maxon and who knows what it's future is going to be in the long run.
You have to use the Morph Tool to mirror morphs in Poser.
I can't emphasize this strongly enough: never-ever let Poser auto-group anything, or bad things (like this) will happen. Sometimes we have to learn things the hard way, for then we won't forget anymore. But once unimesh Poser is out, none of this will matter anymore, and we can auto-group in Poser with no issues. But until then... beware! ^____^
unreal
Noteworthy
No. In python. Posers mirror does not work the way I need for this.
For each actor, I determine if the actors cross the center or have a mirror actor. Crossing center is easy enough, If there is at least one neg and pos x, then the mesh crosses the center. For actors, I check the names. For a center crossing mesh, I just make a lookup of each + x, and check for a corresponding - x (with same y, z). This tries at a high accuracy, then gradually reduces it until either it's symmetric or the accuracy is clearly into the visible range. For 2 actor mirroring, it does the same thing only with 2 actors. It builds a list of the verts (an array, since there can be multipoles at the same point). It goes through those and reads the morphs and creates mirrored morphs on the other side. It's smart enough the recreate underlying morphs, plus any wirrd into masters. This comes in handy when I'm doing JCMs. I just do them for one side, wire them into the joints motions, and it creates mirrored versions, named correctly, for the other side. I can create a set of blank JCMs for one side (plus wiring), then I actually define the morphs ( morph brush). When one side is set, I run the script and it mirrors. I can also delete one side, then modify the remaining, then mirror again.
I did this specifically to speed up JCM creation for clothing for La Femme. Heaps of JCMs. I figured out how to quickly make clothing then sculpt, then use quad remesher to make super clean modest poly meshes in blender. But rigging in poser (for me, specifically rigging close fitting La Femme stuff) was brutal. Scripting was necessary. La Femme's topo was not helpful
I have hopes for D2. All the people who know way more than me (that would be nearly everyone) and who have or are playing with D2, hav r said (and made0 some cool stuff. So I expect my hopes are not in vain ^.^
Good things take time. Remember the disaster when SMS rushed out Pauline when she wasn't ready? Her male counterpart, Paul, went into the demo with an incomplete rig, and he couldn't pose. Oh, and don't get me started with LF topology. I have disabled subdivision, and she couldn't pose anymore. Her shape became fuzzy because the edge flow just fights the poses. She can't be sculpted at base resolution, and with subdivision, sculpting in Poser becomes impractical. That's how much poor topology can mess up a figure.
I am getting to know D2 better now that I am making contents for her. The figure design improvements are so many, making life so much easier. She's a step up from D1 in every aspect. Mec4D has already mentioned how easier and faster it is to create contents for her (no JCM nightmare!), and I keep coming to the same conclusions. The improved topology and edge-flow from Chris work nicely with Paul's new weight maps. Those 2 things require each other for best results. The new optimized grouping makes it quicker to complete clothing rigging. The additional bones extend her posability. I have rigged and morphed D2 contents in both Poser and DS, and I give her 2 thumbs up.
I am getting to know D2 better now that I am making contents for her. The figure design improvements are so many, making life so much easier. She's a step up from D1 in every aspect. Mec4D has already mentioned how easier and faster it is to create contents for her (no JCM nightmare!), and I keep coming to the same conclusions. The improved topology and edge-flow from Chris work nicely with Paul's new weight maps. Those 2 things require each other for best results. The new optimized grouping makes it quicker to complete clothing rigging. The additional bones extend her posability. I have rigged and morphed D2 contents in both Poser and DS, and I give her 2 thumbs up.
I like both LF and LH I'm not finding too much of a struggle with their bases. I am very much looking forward to Dawn 2 and Dusk 2 (more so) I know you can't rush perfection but I was hoping that Dawn 2 would hit before XMas. Oh well. I have other projects and still trying to get Gino for LH pushed through at 'Rosity. Everything, right now for me, seems to be hurry up but WAIT! lol
unreal
Noteworthy
So, way less JCMs?
That would be awesome. JCMs make me cry, when making stuff o.o
I don't expect unimesh Poser until maybe this time next year. I sure hope D2 is sooner. Aren't we largely waiting for all the icing, now? Releasing a base character is all well and good. But morph packs, a solid starting wardrobe, and dev stuff kind of make the figure, no?
test renders! throw us a bone (or the whole rig)!
That would be awesome. JCMs make me cry, when making stuff o.o
I don't expect unimesh Poser until maybe this time next year. I sure hope D2 is sooner. Aren't we largely waiting for all the icing, now? Releasing a base character is all well and good. But morph packs, a solid starting wardrobe, and dev stuff kind of make the figure, no?
test renders! throw us a bone (or the whole rig)!
D2 has the usual 3 JCMs on the thighs as D1 has, and also 1 on the shoulders I wish D1 would also had. That's all! To put things in perspective, most other figures in both Poser and DS have JCMs by the HUNDREDS. That is a sign of poor figure design, because JCMs should only be needed on the thighs and shoulders (the ball-joints), and even then, in D2 they do much LESS work than they did with D1, thanks to the better rigging and topology.
One of the things that have visibly improved in D2 is the clavicle\armpit posing, which now produces a more natural, realistic shape than D1 was capable of. Another one is the thigh joint that poses well even without JCMs - that was perhaps the major issue with D1. It poses so much better that Paul has included both front and side leg split poses to D2 to show off her rigging quality. You can see that in the D2 dev thread, where Paul has showed some renders of the leg posing (because I wanted to see it). And remember, those poses were from *before* the JCMs were added!
unimesh Poser ? what is that?
Wow, that question has no simple answer.
By the time Poser was designed, back in 1995, 3D models used to be split into separate parts due to limitations on computers of that era, like lack of enough memory to handle the whole thing as a single piece. So to this date, Poser keeps breaking figures into smaller parts every time we load them, even if the original geometry is a single piece, and lack of enough memory is no longer a limitation with modern computers.
This causes all sorts of problems, because Poser keeps having to weld the figures back into 1-piece to use them. It affects performance, causes morphing and conforming issues, and makes it impossible to export figure geometry back to how it originally was. According to the Poser dev team, the code to handle this has become extremely complex, and it is totally for nothing, since it has been over a decade since the need for models to be split no longer exist. Hence the term "unimesh": meaning a single mesh, not to be confused with DAZ3D Unimesh figures, which means something completely different.
The upcoming unimesh Poser version will no longer split figures when loading them, which is how DAZ Studio works. The benefits spread across a wide range of tasks, but the major benefits will be on better performance, better morphing, better conforming, and Poser exported geometries will finally become compatible with DS and all other programs that require unimesh geometry, such as Blender and iClone. We will be able to export a morph from Poser, and successfully load it in DS, which was never possible before.
Since Poser was designed from ground up to split figures when loaded, changing this means redesigning how Poser works internally. The original Poser creator, Larry Weinberg, told me back in 2016 that this would require rewriting the Poser core. It could break everything else, like all the other "Rooms" in Poser, so that would take a long time. Bondware has been working on it since earlier this year, and I am hoping we should have a working version, perhaps with partial unimesh functionality, somewhere in 2022.