Saekano Girls (Megumi Ponytail WiP) (1 Viewer)

[Mineur]

I think I finally figured out a way how. It actually hit me while I was sleeping. What do you think now? Shading's a bit iffy, but I wrestled with it for too long to care anymore.

And then I noticed the ear. I'll get around fixing that somehow.

Lol so in a dream haha? But ya it looks better and better, and like stuntcock explained the ear thingy is rly easy to fix, especially if you are aware of it b4 starting to color it ^^

 

[4dascience]

If you'd like, we could tackle a dynamic hairstyle project collaboratively. You'd supply the reference artwork, and then we'd walk through the steps of tracing it and animating it. You could then use the completed FLA file (and the walkthrough documentation+screenshots) as a guide for future solo projects - and it could serve as a tutorial resource for others.

@skymastolf66 and I worked through a similar collaborative project a while ago, but he wasn't interested in creating dynamic hair mods. So we provided some discussion of the design decisions, but we didn't actually produce a step-by-step guide to hair-mod creation.

The key factors are:

• the hairstyle must be suitable for dynamic animation
  • you can find a brief overview of dynamic-hair attributes here, and specific details here
  • there's a "sweet spot" to aim for. If the hairstyle is too simple then it isn't worth the effort of animating; if it's too complex then the FPS penalty will make the hairstyle unusable during normal gameplay.
  • I prefer hairstyles with weird shapes, because they give me an opportunity to test the limits of the physics code. But I'm not going to insist on a goofy hairstyle; you're free to propose a realistic hairstyle from one of your favorite anime or videogame characters.
• the hairstyle must be drawn with redundant/overlapping details
  • when a strand of hair swings aside, we should be able to see another layer of hair underneath it. We shouldn't see a blank patch of scalp.
  • therefore we need to make assumptions about underlying hair layers (which usually aren't visible in the reference image). We can draw in these details by relying on secondary reference images, or just make them up with imagination.
    
  • we need to maintain lots of separate layers while working on the image (in GIMP, Photoshop, InkScape, or whatever). The "Merge Layers" command is taboo.
• the hair image must be in vector format before it can be properly animated
  • ideally, it would be drawn/traced in vector form
  • if you're more comfortable with raster tools, then you could hand off a raster image and I could trace it before proceeding with the project

• overall impression: nicely done! You've shown a lot of improvement in just a few days :)
  
• the outline is too thick. Zooming in shows it to be wider than the outline of the girl's face, and it's noticeably "fuzzy."
  • a thinner and sharper outline would blend in more nicely.
• the ahoge strands tend to disappear off the top of the screen at normal zoom levels. This isn't an "artistic flaw"; you've accurately recreated the original design. But it's an example where it might be appropriate to tweak the original design so that it fits better into the gameplay scene.
  • this is especially important with dynamic-hair projects. There's no sense in spending an hour carefully animating the ahoge strands if they're going to get clipped off by the edge of the game window.
  • sby's animtools mod allows users to "pan" the camera upwards, or zoom out so that the entire hairstyle remains visible. So ... if clipping becomes a problem then we might choose to fiddle with the camera instead of adjusting the actual hair.
• the "sidelock" strand should be moved to the top-left quadrant so that it passes over the ear
  • 
• the bangs can be reshaped slightly to avoid eclipsing the eye. Eye movement and eye contact are very important in SDT because they inform the player about the girl's emotional state. We shouldn't block the eye unless it's absolutely necessary; we should also leave the eyebrow visible when possible.
  • 
• the bangs could be tweaked further so that they pass behind the nose (as shown in the reference images)
  • I haven't bothered to do so here (because laziness)

Before and after:

I think in general, her hair seems rather big, after trying her for a little. I did fix the bangs, though. I think I'm getting a hang of this more or less. Though, it can take a bit of time.

I would definitely love to do a collab work! I can't really seem to think of anybody off the top of my head at the moment though.

 

[4dascience]

Actually, scratch that, how about this?

 

[stuntcock]

Actually, scratch that, how about this?

Too complex for an initial project.

If you click here, you can see a crude "wireframe" sketch of the animation plan for a previous project. That project rendered most of the "scalp" hair as a single static element. Minor animation was provided for the hairbow and moderate animation for the bangs, but effort was focused on the most visible and distinctive element: a long S-shaped ponytail.

This design lacks a single focal element which we could animate in isolation. We'd pretty much need to animate the whole thing; otherwise the static-vs-animated boundary would be very obvious and unsightly.



Unfortunately - there are just too many independently-moving strands in this design. It would take a long time to draw and animate, and the complexity is so high that a step-by-step walkthrough would not serve as a good tutorial for novice modders.

You'll notice that the sketch includes several Y-shaped and X-shaped convergence points. The hair physics system can support such details, but they're somewhat "advanced." I wouldn't mind including one or two of them in a tutorial project, but this design contains too many.

The design also includes at least two anatomical constraints (locations in which the hair rests against the shoulder and collarbone). They're technically feasible, but not ideal for a beginner/tutorial project because of complexity (we need to balance the collision-repulsive force against the mass and inertia of the hair ... which involves math, testing, and fine-tuned parameters).

Please don't be discouraged! Other suggestions are welcome, and we could resurrect this project at a later date (when we're less worried about "tutorial" stuff).

 

[4dascience]

Too complex for an initial project.

If you click here, you can see a crude "wireframe" sketch of the animation plan for a previous project. That project rendered most of the "scalp" hair as a single static element. Minor animation was provided for the hairbow and moderate animation for the bangs, but effort was focused on the most visible and distinctive element: a long S-shaped ponytail.

This design lacks a single focal element which we could animate in isolation. We'd pretty much need to animate the whole thing; otherwise the static-vs-animated boundary would be very obvious and unsightly.



Unfortunately - there are just too many independently-moving strands in this design. It would take a long time to draw and animate, and the complexity is so high that a step-by-step walkthrough would not serve as a good tutorial for novice modders.

You'll notice that the sketch includes several Y-shaped and X-shaped convergence points. The hair physics system can support such details, but they're somewhat "advanced." I wouldn't mind including one or two of them in a tutorial project, but this design contains too many.

The design also includes at least two anatomical constraints (locations in which the hair rests against the shoulder and collarbone). They're technically feasible, but not ideal for a beginner/tutorial project because of complexity (we need to balance the collision-repulsive force against the mass and inertia of the hair ... which involves math, testing, and fine-tuned parameters).

Please don't be discouraged! Other suggestions are welcome, and we could resurrect this project at a later date (when we're less worried about "tutorial" stuff).

Jesus, that's more intensive than I thought it'd be. Trying to think of others, but dunno.

Ever find yourself not being able to enjoy your stuff, if only because you did it yourself? Feels different after a few hours of work on it.

 

[stuntcock]

Ever find yourself not being able to enjoy your stuff, if only because you did it yourself? Feels different after a few hours of work on it.

Hahaha hell yes. That's why I prefer to do collaborative animation projects. If I draw the initial lines then I notice all of the flaws and shortcuts: "the shading patterns don't make any sense" ... "the ponytail is implausibly thick; hair doesn't work like that" ... "why should I spend time animating this half-assed garbage?"

If someone else draws the lines then I can just assume that they're correct, and proceed with the actual work. I'll still notice flaws in the animation, but I can accept those (because I know that additional fidelity would have incurred an unacceptable FPS penalty).

 

[4dascience]

Yea, I'm getting a little bit of that. A case of wanting to make more sexy characters, but running out of ideas, and then in the end I don't enjoy them as much.

How much harder is it to do clothes?

 

[stuntcock]

How much harder is it to do clothes?

• Projects can be smaller in scope
  • You don't need to start with a full-body costume
  • You can start with a pair of shoes, or gloves, or even an earring
• Requires Flash
  • Doing the work in a raster-drawing program is possible but not really worth the effort.
  • Unlike the girl's head (and hair), the body size is on a slider. Raster images suffer aliasing and tend to look bad.
• More layers to worry about
  • Each element must be drawn on a specific layer, which is rigged to a specific bodypart.
  • For example: a skirt would consist of four to five pieces (rigged to her back, abdomen, right thigh, left thigh).
  • Because you're working with separate shapes, and because the SDT girl's anatomy is slightly wonky, it's difficult to directly trace a reference image. The work involves more adaptation and imagination.
  • Elements can be drawn in fancier ways (so that they stretch between different bodyparts instead of being rigged to one part) but this involves more work. Example: Huit's bikini tops include bra straps which stretch across the torso and collarbone.
• Breast slider is a thing
  • You can ignore it completely
  • If you support it, then you need to either exclude some clothing designs (which radically change shape depending on breast size) or do much more work in order to support them.
  • Some clothing designs (such as a full-body skintight suit, or a long dress) are relatively easy to build with breast slider support.
• RGB support is a thing
  • You can ignore it completely
  • If you decide to support it, then you need to do a bit more planning. Shading and highlight details must be placed on a separate layer.
• Mod compatibility is a thing
  • There's a limited set of "slots" available for clothing mods by default. Vanilla "gloves" can't do handjob mode, and vanilla "tops" can't have long sleeves. Vanilla imports lack versatility, but they're very reliable and rarely generate conflicts.
  • The set of available slots can be expanded by using mods (such as sby's template extension) or by using a different modding technique (such as Faceless' FLA template).
  • Either of the aforementioned approaches would make your mod dependent on the SDT Loader; vanilla users would be unable to play with your new clothing.
  • If your mod becomes popular then people may complain about compatibility - your layers might overlap with those of a popular bra or jacket. It's up to you whether to tweak your work for compatibility, or just leave it as-is.
  • If your mod becomes popular then people may ask you to make it work with sby's moreClothing mod. This might be a very straightforward task or a very difficult one. Of course, you can always say "no."

 

[4dascience]

• Projects can be smaller in scope
  • You don't need to start with a full-body costume
  • You can start with a pair of shoes, or gloves, or even an earring
• Requires Flash
  • Doing the work in a raster-drawing program is possible but not really worth the effort.
  • Unlike the girl's head (and hair), the body size is on a slider. Raster images suffer aliasing and tend to look bad.
• More layers to worry about
  • Each element must be drawn on a specific layer, which is rigged to a specific bodypart.
  • For example: a skirt would consist of four to five pieces (rigged to her back, abdomen, right thigh, left thigh).
  • Because you're working with separate shapes, and because the SDT girl's anatomy is slightly wonky, it's difficult to directly trace a reference image. The work involves more adaptation and imagination.
  • Elements can be drawn in fancier ways (so that they stretch between different bodyparts instead of being rigged to one part) but this involves more work. Example: Huit's bikini tops include bra straps which stretch across the torso and collarbone.
• Breast slider is a thing
  • You can ignore it completely
  • If you support it, then you need to either exclude some clothing designs (which radically change shape depending on breast size) or do much more work in order to support them.
  • Some clothing designs (such as a full-body skintight suit, or a long dress) are relatively easy to build with breast slider support.
• RGB support is a thing
  • You can ignore it completely
  • If you decide to support it, then you need to do a bit more planning. Shading and highlight details must be placed on a separate layer.
• Mod compatibility is a thing
  • There's a limited set of "slots" available for clothing mods by default. Vanilla "gloves" can't do handjob mode, and vanilla "tops" can't have long sleeves. Vanilla imports lack versatility, but they're very reliable and rarely generate conflicts.
  • The set of available slots can be expanded by using mods (such as sby's template extension) or by using a different modding technique (such as Faceless' FLA template).
  • Either of the aforementioned approaches would make your mod dependent on the SDT Loader; vanilla users would be unable to play with your new clothing.
  • If your mod becomes popular then people may complain about compatibility - your layers might overlap with those of a popular bra or jacket. It's up to you whether to tweak your work for compatibility, or just leave it as-is.
  • If your mod becomes popular then people may ask you to make it work with sby's moreClothing mod. This might be a very straightforward task or a very difficult one. Of course, you can always say "no."

Damn. Yea, I was wondering why there aren't more bras, negligees and so on, but now I think I know why.