Advice with secondary forms

DustyShinigami

Hi

I need some advice from fellow character artists. One aspect I've not had enough practise with and that I'm struggling with is the mid-frequency/secondary forms. For the most part, I understand what they are. The issue I'm having is a breakup layer/pass. I'm not sure if other artists incorporate it as well...? I'm presuming so. I did a mentorship with Henning Sanden from FlippedNormals and he showed me his approach to doing it. However, I'm not confident I'm going about it correctly. I could be wrong, but because he comes from a VFX background, maybe he's been overcomplicating things...? Especially as I'm aiming for the games industry. I went back through Laura Gallagher's videos on Outgang, but she only briefly touches on it, sadly. But it seems more simplified...?


One tip I did take away from a FlippedNormals video on the topic is to blur a reference image and make it black and white in order to disguise all the distractions. For the Doom Hell Knight I'm working on, I've gone and done that.

Originals:


Blurred:


And of course I've been attempting it on a Layer. I was also provided with a demonstration from Henning...


I just always feel conflicted. That I'm not quite 'getting it'...? I know that there needs to be breakup and directionality, but feel that something is 'off'. Or that I'm perhaps not capturing some of the secondary forms into the breakup layer...? Or can/should those be done afterwards? After the directionality? I've been trying to do this attempt asymmetrically as well.


This is one of my earlier attempts, which looks wrong, flat, and is too soft.

Any advice/insight would be super appreciated. Thanks.

Eric Chadwick

I think it would help to examine more photographic reference of human musculature, and learning more about the underlying anatomical structures.

Quick web search...

Noren

It looks a bit like you are scratching relatively even valleys of the same depth into your main forms instead of getting the look of actual volumes meeting below the skin. Don't be afraid to leave your main forms were necessary as long as they still read correctly. An area were two muscles meet can be sharp(ish), but the curvature should be fairly continuous across the muscle (at least for those smaller ones) if you know what I mean. However, in the reference, there aren't that many sharp secondary forms to begin with. That's assuming we are talking about the smaller muscle and fat lumps and not details like scratches or folds.

DustyShinigami

Eric Chadwick said:

I think it would help to examine more photographic reference of human musculature, and learning more about the underlying anatomical structures.

Quick web search...

Oh, I know. However, using the official model and looking at the reference images I have, there are some additional shapes in there as well. I did manage to speak to Jason Martin, the original artist, and he mentioned how with creatures, so long as it's based in reality and based on some form of real anatomy, you can fudge it a bit to make something look cool. I kinda mixed in shapes I could see from the Doom Eternal version of the Hell Knight along with the one from Doom 2016.

DustyShinigami

Noren said:

It looks a bit like you are scratching relatively even valleys of the same depth into your main forms instead of getting the look of actual volumes meeting below the skin. Don't be afraid to leave your main forms were necessary as long as they still read correctly. An area were two muscles meet can be sharp(ish), but the curvature should be fairly continuous across the muscle (at least for those smaller ones) if you know what I mean. However, in the reference, there aren't that many sharp secondary forms to begin with. That's assuming we are talking about the smaller muscle and fat lumps and not details like scratches or folds.

Hmm. I see. So I need to add more variation of where forms rise and fall? Some folds, like on the backs of the arms, I have added, but on another layer, which I'll probably incorporate at some point. Scratches I'll likely save for the tertiary. Although I think I have been sketching in some lines with DamStandard.

I attached the reference above from the version in Doom 2016, which has some additional shapes in there, too.

Noren

Yes, more variations (e.g. wider in one spot and more pinched in another), but I was mainly referring to the "cross section" so to speak.

DustyShinigami

Noren said:

Yes, more variations (e.g. wider in one spot and more pinched in another), but I was mainly referring to the "cross section" so to speak.

Sorry, could you paintover where you mean exactly? I'm not quite sure what you mean by 'cross section'. Thanks.

Noren

I was exaggerating, of course, but something like this:



Doesn't need to be actual anatomy, just needs to read like it could be. I took the screenshot as a starting point and smudged it a bit, without trying to replicate the reference model, which is much softer overall (edit: In this area and regarding the secondary forms).

Additionally, I tried to introduce some more variance and to get the impression of some of the muscles going across each other instead of just forms sitting next to each other, which is the result of too even and broad valleys between and too little continuous curvature across the muscles (but I kept a flatter surface for the lat like one). Broader, flat valleys you'd normally find at bone ridges/attachments or tendons.
 

DustyShinigami

Noren said:

I was exaggerating, of course, but something like this:



Doesn't need to be actual anatomy, just needs to read like it could be. I took the screenshot as a starting point and smudged it a bit, without trying to replicate the reference model, which is much softer overall

Additionally, I tried to introduce some more variance and to get the impression of some of the muscles going across each other instead of just forms sitting next to each other, which is the result of too even and broad valleys between and too little continuous curvature across the muscles (but I kept a flatter surface for the lat like one). Broader, flat valleys you'd normally find at bone ridges/attachments or tendons.
 

Ahhh, I see what you mean. Thanks. I'll have another go soon and try implementing it. I'll post the WIP results after.

DustyShinigami

I decided to start the layer over from scratch and use symmetry this time. I've tried adjusting some of the plane changes based on this reference as it has more interesting and visible shapes. Still haven't quite got a good balance with the transitions between shapes though. :-\


Before:


After:

DustyShinigami

@Noren Tried focusing more on the softer transitions and individual shapes...

Noren

Just quickly, and I'm not sure if I'm very good at explaining this:
The "after" variant looks already a bit better, but still has the same underlying problem that's also visible in the latest exercise. It's a good idea to blur the source, but especially with your blue lines, you almost seem to follow a kind of "2D" edge finding approach and outline forms that often aren't really there (at least to this extent) or are parts of tertiary details (minor surface bumps and dents) or much more subtle. And those resulting shapes still sit a bit next to each other on the final mesh.

To put it negatively, it looks a bit like useung a turbulence noise in 3ds max (http://www.neilblevins.com/art_lessons/turbulence/turb_max.jpg) as a bump and I feel partly responsible for that since that's basically the curvature I am suggesting, but it can be much softer for the most part and more subtle in places, and the shapes themselves need to feel less random (in their directionality) and regular. I'm exaggerating, of course, but I hope it helps to get the point across. E.g. if you simply smudged/blurred some of your details and the shapes that have basically an outline going around them, this would probably look better already.

Perhaps you could try to really ignore the tertiary details (blurring alone can't fully accomplish that) for a further exercise and try to find the defining secondary shapes and perhaps try to analyze which shapes the muscles under the skin would have.
And then some of the more random details you are seeing right now will come back organically and more naturally looking once you do your tertiary layer.

DustyShinigami

Noren said:

Just quickly, and I'm not sure if I'm very good at explaining this:
The "after" variant looks already a bit better, but still has the same underlying problem that's also visible in the latest exercise. It's a good idea to blur the source, but especially with your blue lines, you almost seem to follow a kind of "2D" edge finding approach and outline forms that often aren't really there (at least to this extent) or are parts of tertiary details (minor surface bumps and dents) or much more subtle. And those resulting shapes still sit a bit next to each other on the final mesh.

To put it negatively, it looks a bit like useung a turbulence noise in 3ds max (http://www.neilblevins.com/art_lessons/turbulence/turb_max.jpg) as a bump and I feel partly responsible for that since that's basically the curvature I am suggesting, but it can be much softer for the most part and more subtle in places, and the shapes themselves need to feel less random (in their directionality) and regular. I'm exaggerating, of course, but I hope it helps to get the point across. E.g. if you simply smudged/blurred some of your details and the shapes that have basically an outline going around them, this would probably look better already.

Perhaps you could try to really ignore the tertiary details (blurring alone can't fully accomplish that) for a further exercise and try to find the defining secondary shapes and perhaps try to analyze which shapes the muscles under the skin would have.
And then some of the more random details you are seeing right now will come back organically and more naturally looking once you do your tertiary layer.

Okay, awesome. Thanks for the suggestions. I spoke to someone else over on the Experience Points Discord server and they echoed your sentiments. They also advised I focus on the bigger shapes, like these:


...and get those to transition better, whilst keeping in mind the directionality of the lats muscles. Those additional shapes I added were adding more 'noise' and making things look too crowded.
But good point about those details I've highlighted in blue coming through via the tertiary details. I've also been advised to do the tertiary details in Substance Painter instead as there's more control with scaling, or better scaling rather.

DustyShinigami

Noren said:

Just quickly, and I'm not sure if I'm very good at explaining this:
The "after" variant looks already a bit better, but still has the same underlying problem that's also visible in the latest exercise. It's a good idea to blur the source, but especially with your blue lines, you almost seem to follow a kind of "2D" edge finding approach and outline forms that often aren't really there (at least to this extent) or are parts of tertiary details (minor surface bumps and dents) or much more subtle. And those resulting shapes still sit a bit next to each other on the final mesh.

To put it negatively, it looks a bit like useung a turbulence noise in 3ds max (http://www.neilblevins.com/art_lessons/turbulence/turb_max.jpg) as a bump and I feel partly responsible for that since that's basically the curvature I am suggesting, but it can be much softer for the most part and more subtle in places, and the shapes themselves need to feel less random (in their directionality) and regular. I'm exaggerating, of course, but I hope it helps to get the point across. E.g. if you simply smudged/blurred some of your details and the shapes that have basically an outline going around them, this would probably look better already.

Perhaps you could try to really ignore the tertiary details (blurring alone can't fully accomplish that) for a further exercise and try to find the defining secondary shapes and perhaps try to analyze which shapes the muscles under the skin would have.
And then some of the more random details you are seeing right now will come back organically and more naturally looking once you do your tertiary layer.

Here's this evening's attempt:

Noren

Much better!

Perhaps it make sense to move this to 3D Art, though, if you plan to continue to work on this specific model.

DustyShinigami

Noren said:

Much better!

Perhaps it make sense to move this to 3D Art, though, if you plan to continue to work on this specific model.

Sure. Makes sense. I plan on asking for feedback as I go with it. Although should I just create a new thread or will a mod move it?

Neox

@Eric Chadwick for vis

@DustyShinigami
https://www.instagram.com/reel/DG8K5eWAZgv/?igsh=MWk5eGE0cmR6dXliZA==
This one reminded me a fair bit of your goals here. I feel a bit like the structure is smoothed out too much after its applied, but i think the direction is quite similar

Eric Chadwick

moved!

DustyShinigami

Neox said:

@Eric Chadwick for vis

@DustyShinigami
https://www.instagram.com/reel/DG8K5eWAZgv/?igsh=MWk5eGE0cmR6dXliZA==
This one reminded me a fair bit of your goals here. I feel a bit like the structure is smoothed out too much after its applied, but i think the direction is quite similar

Thanks for the link; I'll give that watch. I think the next stage is adding some volume and directionality to the shapes. That way things won't hopefully look as smoothed down.

DustyShinigami

Eric Chadwick said:

moved!

Thanks

DustyShinigami

Okay, I've tried experimenting with adding volume and some noise. Though I'm not sure I've approached it correctly. I did add some directionality and then smoothed it down.

DustyShinigami

I'll probably juggle and post this torso along with a human one I'm working on as part of an arm anatomy course. It will be good for additional practise and studying.

DustyShinigami

Just out of curiosity, does anyone know of any anatomy courses that cover secondary forms? And goes into it with a good amount of coverage?

pxgeek

I really like how Gio Nakpil approaches hierarchy of forms in that regardless of how you delineate between secondary/tertiary or whatever; tertiary forms support the secondary and secondary support primary forms and that they aren’t added haphazardly, but built on top of one another.

https://youtu.be/ZGbBrWRRN7o?si=JaOzrm6hiDsRkQL3&t=87

https://www.youtube.com/live/qAq294tcuDM?si=mjRNMk-jd50Eenql&t=1328

I feel like in your last image those secondary forms seem like they are placed with no rhyme or reason and consequently detract from underlying structure of the primary forms…they also look very similar shape and size and spaced almost evenly that gives it that “averaged cg” (please pardon the term) look.

While I don’t know any courses solely focused on secondary forms, I’d offer that any videos you can find of Gio Nakpil and Glauco Longhi would be worth your time. Of course, Zuccarello is great too!

DustyShinigami

pxgeek said:

I really like how Gio Napkil approaches hierarchy of forms in that regardless of how you delineate between secondary/tertiary or whatever; tertiary forms support the secondary and secondary support primary forms and that they aren’t added haphazardly, but built on top of one another.

https://youtu.be/ZGbBrWRRN7o?si=JaOzrm6hiDsRkQL3&t=87

https://www.youtube.com/live/qAq294tcuDM?si=mjRNMk-jd50Eenql&t=1328

I feel like in your last image those secondary forms seem like they are placed with no rhyme or reason and consequently detract from underlying structure of the primary forms…they also look very similar shape and size and spaced almost evenly that gives it that “averaged cg” (please pardon the term) look.

While I don’t know any courses solely focused on secondary forms, I’d offer that any videos you can find of Gio Napkil and Glauco Longhi would be worth your time. Of course, Zuccarello is great too!

Hey, thanks for the links and suggestions. I'll give those a watch and try looking into their videos. See what I can find.

I'll go back to an earlier save and start over. Hopefully I'll get more of an idea how to approach things as I'm a bit unsure at this point. ^^;

DustyShinigami

Okay, I've pretty much gone through both of those linked videos. There's still under 50 minutes left of Gio Nakpil's video; he's currently on tertiary details. His approach to doing secondary forms on a head certainly makes sense and cleared up some doubt. When I was doing my mentorship with Henning Sanden, and I was working on heads, I think I pretty much got it after he showed me in a video. It's just for a creature's torso, like I'm working on, it doesn't seem as obvious, if that makes sense...? I'm not sure if it's because Jason Martin implemented some additional shapes to make things look more interesting.

So I'm not sure if my way of thinking for this next stage is on point (feel free to correct me if I'm wrong), but now that those transitions of those main shapes are done, I'm seeing some additional more subtle shapes under the surface. And I guess I was trying to implement those last time, but got carried away and overdid it. But I think I need to try and add them, but to ensure they're more subtle and flow with the muscles.

pxgeek

DustyShinigami said:

So I'm not sure if my way of thinking for this next stage is on point (feel free to correct me if I'm wrong), but now that those transitions of those main shapes are done, I'm seeing some additional more subtle shapes under the surface. And I guess I was trying to implement those last time, but got carried away and overdid it. But I think I need to try and add them, but to ensure they're more subtle and flow with the muscles.

Yeah, that’s about the right idea. You want those secondary forms to serve some sort of purpose whether it be as a function of anatomy or just something there to add to the visual design- It’s all there to support those bigger shapes.

Maybe what might help in your specific case with the torso is to take a closer look at those shapes your seeing and decide if it's an underlying muscle, or fat deposit, or bony landmark, or tendon, or just something there you see that makes it look cool for aesthetics; and have that inform your sculpting insofar as needing to integrate the form, or softening/hardening transitions, adding/taking away volume, or removing it all together. My opinion was that I wasn’t getting any of those reads and thus made the overall design of the torso less readable.

What could also help (apart from zooming out and looking at the model as a whole) is to work the sculpt around the full character at once rather than a small part in isolation as it can be easy to get myopic and over develop one area and not realize until the end that the arm doesn’t really relate to the shoulder and the shoulder doesn’t relate to the torso, etc. (hey, which is another kind of hierarchy of form…Ha!)

DustyShinigami

pxgeek said:

DustyShinigami said:

So I'm not sure if my way of thinking for this next stage is on point (feel free to correct me if I'm wrong), but now that those transitions of those main shapes are done, I'm seeing some additional more subtle shapes under the surface. And I guess I was trying to implement those last time, but got carried away and overdid it. But I think I need to try and add them, but to ensure they're more subtle and flow with the muscles.

Yeah, that’s about the right idea. You want those secondary forms to serve some sort of purpose whether it be as a function of anatomy or just something there to add to the visual design- It’s all there to support those bigger shapes.

Maybe what might help in your specific case with the torso is to take a closer look at those shapes your seeing and decide if it's an underlying muscle, or fat deposit, or bony landmark, or tendon, or just something there you see that makes it look cool for aesthetics; and have that inform your sculpting insofar as needing to integrate the form, or softening/hardening transitions, adding/taking away volume, or removing it all together. My opinion was that I wasn’t getting any of those reads and thus made the overall design of the torso less readable.

What could also help (apart from zooming out and looking at the model as a whole) is to work the sculpt around the full character at once rather than a small part in isolation as it can be easy to get myopic and over develop one area and not realize until the end that the arm doesn’t really relate to the shoulder and the shoulder doesn’t relate to the torso, etc. (hey, which is another kind of hierarchy of form…Ha!)

Thanks for the advice. ^_^ One thing I'm going to find quite difficult, and it will certainly take some time to get to grips with, is knowing what and when to add things to the design to improve the visual appeal. Especially if it's something like this where there's more freedom to do that compared to a regular human character. 

DustyShinigami

Haven't made a tonne of progress, though I have spent quite some time around the centre of the back and taking it slow.

Before:

After:

DustyShinigami

A bit more progress with it. Still unsure if it's going in the right direction though. If it's 'readable' and makes sense.

pxgeek

DustyShinigami said:

One thing I'm going to find quite difficult, and it will certainly take some time to get to grips with, is knowing what and when to add things to the design to improve the visual appeal. Especially if it's something like this where there's more freedom to do that compared to a regular human character.

Yep. That’s kinda our whole thing: developing a strong visual library and design sense. And without those, we have references to help us out.

From what I can see of Gio: he is strongly looking to nature and experimenting through his studies to try to find ways of incorporating those things into a more fantastical realm.

_____

I think you’re making some good improvements from your previous iterations.

While my own personal tastes tells me it’s still vague in some areas, I can definitely see where you are making much more intentional choices about what some of the secondary forms are and how they fit into the larger shapes. Keep at it!

DustyShinigami

pxgeek said:

DustyShinigami said:

One thing I'm going to find quite difficult, and it will certainly take some time to get to grips with, is knowing what and when to add things to the design to improve the visual appeal. Especially if it's something like this where there's more freedom to do that compared to a regular human character.

Yep. That’s kinda our whole thing: developing a strong visual library and design sense. And without those, we have references to help us out.

From what I can see of Gio: he is strongly looking to nature and experimenting through his studies to try to find ways of incorporating those things into a more fantastical realm.

_____

I think you’re making some good improvements from your previous iterations.

While my own personal tastes tells me it’s still vague in some areas, I can definitely see where you are making much more intentional choices about what some of the secondary forms are and how they fit into the larger shapes. Keep at it!

Yeah. I’ll eventually have to start gathering similar references for future projects. I recall Gio being drawn to exotic fish and gnarled trees etc. I have been in touch with Jason Martin about his Hell Knight from Doom Eternal, which he said he’d get back to me when he’s less busy. Presumably, he’s working on the new Doom game. I’m curious what sorts of references he used whilst making it.

But that’s good to know that I’m making good improvements; that’s reassuring. Thanks.

DustyShinigami

pxgeek said:

DustyShinigami said:

One thing I'm going to find quite difficult, and it will certainly take some time to get to grips with, is knowing what and when to add things to the design to improve the visual appeal. Especially if it's something like this where there's more freedom to do that compared to a regular human character.

Yep. That’s kinda our whole thing: developing a strong visual library and design sense. And without those, we have references to help us out.

From what I can see of Gio: he is strongly looking to nature and experimenting through his studies to try to find ways of incorporating those things into a more fantastical realm.

_____

I think you’re making some good improvements from your previous iterations.

While my own personal tastes tells me it’s still vague in some areas, I can definitely see where you are making much more intentional choices about what some of the secondary forms are and how they fit into the larger shapes. Keep at it!

Which areas in particular are still looking vague to you?

pxgeek

the red section.

I like the treatment you did for the scapula muscles.

DustyShinigami

pxgeek said:

the red section.

I like the treatment you did for the scapula muscles.

Okay. Hmm, interesting. I mean, it could be why it was approached differently for the Doom Eternal version. I'm very curious to hear from Jason and find out a bit more about these details and his approach. Those shapes in red are what I'm seeing in the Doom 2016 one. Though I guess the bottom row of stacked shapes aren't quite right.


And then the Doom Eternal version, which I've flipped the image:

Do you reckon the shapes on the Eternal model make more sense?

Thanks

DustyShinigami

Done a bit of observation and paintover with the Eternal model and tried to find all the shapes.

pxgeek

I guess I’m not really seeing it the same way you are.

What I like about the originals is that they are still simplified into just a few groups of large muscle masses and has a strong sense of shape design, which gives it a kind of clarity and impact that I feel is a little lost on yours.

Or maybe to put it another way: those secondary shapes you’re adding are a bit over-emphasized.

The differences between the two Doom versions are pretty nuanced to me…they both read well.