Jump to content
3DCoat Forums

Leigh's Sketchbook


LJB
 Share

Recommended Posts

  • Advanced Member
Taros - Thank you, Glad someone found it useful I hoped it might prove so I read many askings for extra increase features I even put in a request for Localised voxel increase myself.

At the moment I, as are some others here on the forum, am struggling with brushes and expected effect to try and see just how far we can push detail in 3D coat. This effort took a while as the brushes are not current suited to fine detail like this, in ZBrush it would have taken minutes but I tire of that workflow, now i have experienced the freedom of voxels. I will continue to bring up the detail only in beginnings at the moment.

I believe more so than resolution its more to do with the stroke and brush behaviour. I work it like this in Zbrush everything get converted to quads at high subdivision and that show the poly count as quads where as we read Voxel in Tri so Double the amount of Tris would be expected for same level of detail. This is 10mil tris so roughly the same resoltion as i would achive with 6 level of SD in ZBrush. I know we can achieve it i feel we are getting closer but its the brushes and stroke that are holding us back more than the resolution. To that end i will request a drag rectangle type brush similar to ZBrush that will give us a better capabillity. Its not only being able to achieve the detail its being able to achieve the detail in a comfortable manner that will garentee a big future for 3D coat.

Thats my two pence on this resolution vs detail subject.

- leigh

- idea removed for rewrite and different approach -

3dioot

Link to comment
Share on other sites

  • Advanced Member
Now something interesting has happened. What would happen if you would start working in true size? (its not uncommon for this to be demanded for the proper functionality of dynamics systems for example). If you would work in true size you could start linking resolutions to object detail. (this is why i will request a unit system in 3dcoat) You could say that, for truesize objects, 1 is a really low resolution. 3 is a medium resolution which allows you to define nice forms. 6 is the resolution that allows you to sculpt extreme surface details like pores (numbers are fictional for now offcourse).

These levels will be true for all objects that you sculpt if you sculpt them lifesize.

I see what you mean, but I don't want to always be thinking in terms of lifelike sizes when sculpting but more like ZBrush where you can slowly go up and down in resolution as needed. For example: If I import an head and I slowly add detail and then increase the resolution and add more detail and then all of a sudden I realize I want to do a major shape change to the head like add horns for example. In zBrush you simply lower the resolution and then sculpt the horns in at the low resolution and then increase the resolution back up very high and continue modeling. Then when you are finished with the sculpt you export a displacement map at the high resolution and then lower the resolution to it's lowest level and export the .obj file out at the lowest resolution. What you are left with is a low poly mesh with a very detailed displacement map, therefore it renders fast due to the fact that it is a very low poly mesh but it still looks incredibly detailed because it has a very detailed displacement map.

This approach has many advantages. Lets say you want to import a mesh object for sculpting. Currently you have to guess at what resolution it is imported. You have to experiment by scaling your mesh up and down in resolution "1" and trying out what fits best with the detail or your mesh. If you try to import on a layer which already had resolution changes (either through scaling or by the use of "increase resolution) its completely random since there is no way to relate the resolution to the root layer.

If you know the size of your object (or estimated size) with this system you can pick the appropriate resolution. If you have a mesh thats finished and highly detailed you will import at level 6 and it will be turned into voxels with all the details preserved. If you have the equivalent of a basecage you would use in zb or mb you would import it at level 1 or 2. Completely predictable results (for the first time in the release of voxel sculpting.

I don't see why you would need to guess what the resolution of the layer is in the first place? If you were in Voxels, then a simple number somewhere could tell you what the current resolution of the object was, but if you imported an object, then why couldn't it be a different resolution if it's already on a separate layer just look at the resolution count (assuming Andrew adds this) and then increase the resolution if you need to.

If Andrew could allow a you to see the resolution somewhere in the UI, then this would help. I don't think there should be a problem with having different resolutions for different objects. For example if you had a warrior with armour, maybe the detail would only be added where the fine sculpting was added. So for example you have a basic armour shape and the back part is smooth so it would be low poly, but the front would automatically be more detailed as finer detail was added with the sculpting brushes. Just an idea but probably too hard to implement. :)

Link to comment
Share on other sites

  • Advanced Member
I see what you mean, but I don't want to always be thinking in terms of lifelike sizes when sculpting but more like ZBrush where you can slowly go up and down in resolution as needed.

"As far as i can think it through its still technically impossible to create multivoxelresolution (ive spend hours thinking about it, making notes and sketching things out). I feel its important to accept this. Voxel sculpting brings many, many major advantages. Features that will never be fluently possible with mesh sculpting. But multilevel sculpting is something it simply cannot do."

"There are lots of challenges with this approach. Not the least one will be to make people understand why this would work since its such a deviation from the perspective of a mesh based sculpting."

And the first thing you do is go straight back to saying you want multiresolution just like Zbrush. You are fully entitled to your opinion. I would like multiresolution myself. But its very unlikely to happen. Its time to see what is possible besides multiresolution sculpting. The next big obvious step is resolutionless sculpting. Thats a major challenge but it is doable. That was the intent of this wall of text. To express some idea's on how to achieve this. If all you have to do is sculpt lifesize to gain this then truly, how much of a problem would this be? Id dare say thats a limitation any human can overcome.

As for your example on how easy zbrush works. I think thats well known by most people, including myself. I dare to state however that if you would try to sculpt horns from a basemesh that doesnt support it (like a sphere) you would be in trouble. Both at the high end where you need to subdivide into oblivion and will most likely end up with extreme rectangular polygons and at the low end where you will have a basemesh which is so stretched its hard for the renderer to calculate correct displacement.

I don't see why you would need to guess what the resolution of the layer is in the first place? If you were in Voxels, then a simple number somewhere could tell you what the current resolution of the object was, but if you imported an object, then why couldn't it be a different resolution if it's already on a separate layer just look at the resolution count (assuming Andrew adds this) and then increase the resolution if you need to.

My point was that resolution only means something in relation to size. Really the answer is in the text i wrote.

If Andrew could allow a you to see the resolution somewhere in the UI, then this would help. I don't think there should be a problem with having different resolutions for different objects. For example if you had a warrior with armour, maybe the detail would only be added where the fine sculpting was added. So for example you have a basic armour shape and the back part is smooth so it would be low poly, but the front would automatically be more detailed as finer detail was added with the sculpting brushes. Just an idea but probably too hard to implement. :)

I never said there was a problem to have different resolutions for different objects. As for your suggestions on resolution being dependant on sculpted detail i have given alot of thought to that as well. In my above post i had actually started describing that under the name of "adaptive resolution". There are however big problems with the implementation of that which i found out while i was typing it all out so i removed it. Ill recap them quickly.

The first problem is that of an upper limit. A cube could always be displayed within the resolution a single voxel offers. A sphere would, in theory, require an infinite resolution. The solution came to me while i was typing out my "voxel resolution definition". If you work in truesize you can state that at a certain level you would begin to start sculpting detail which cannot be seen (lets say even finer then pores). This means that you can set your upper limit at that resolution. Nomatter how much detail you add, it will never need a higher resolution then this one. This is a very good reason to introduce a truesize sculpting workflow.

Then there is the problem of at what resolution sculptural changes are made. You only know wether a form can be optimized after it has been formed. Does this mean all brushes should perform at the maximum resolution by default? This would create an unworkable situation. Lets say that gets solved one way or the other. The optimization process would still need to take place. Think of the surface tools in the voxel room for a second. The "only" thing that needs to happen there after each stroke is the conversion from surface to voxel. Yet this gives a noticable feeling of lag after every stroke even on the most powerfull systems. To have an "optimize" routine run after each and every brush stroke would create the same effect.

-----

I still think that the concept of brush resolution (as i guess is done with smooth) is the most promising to increase voxel sculpting performance.

3dioot

Link to comment
Share on other sites

  • Contributor
"As far as i can think it through its still technically impossible to create multivoxelresolution (ive spend hours thinking about it, making notes and sketching things out). I feel its important to accept this. Voxel sculpting brings many, many major advantages. Features that will never be fluently possible with mesh sculpting. But multilevel sculpting is something it simply cannot do."

"There are lots of challenges with this approach. Not the least one will be to make people understand why this would work since its such a deviation from the perspective of a mesh based sculpting."

And the first thing you do is go straight back to saying you want multiresolution just like Zbrush. You are fully entitled to your opinion. I would like multiresolution myself. But its very unlikely to happen. Its time to see what is possible besides multiresolution sculpting. The next big obvious step is resolutionless sculpting. Thats a major challenge but it is doable. That was the intent of this wall of text. To express some idea's on how to achieve this. If all you have to do is sculpt lifesize to gain this then truly, how much of a problem would this be? Id dare say thats a limitation any human can overcome.

As for your example on how easy zbrush works. I think thats well known by most people, including myself. I dare to state however that if you would try to sculpt horns from a basemesh that doesnt support it (like a sphere) you would be in trouble. Both at the high end where you need to subdivide into oblivion and will most likely end up with extreme rectangular polygons and at the low end where you will have a basemesh which is so stretched its hard for the renderer to calculate correct displacement.

My point was that resolution only means something in relation to size. Really the answer is in the text i wrote.

I never said there was a problem to have different resolutions for different objects. As for your suggestions on resolution being dependant on sculpted detail i have given alot of thought to that as well. In my above post i had actually started describing that under the name of "adaptive resolution". There are however big problems with the implementation of that which i found out while i was typing it all out so i removed it. Ill recap them quickly.

The first problem is that of an upper limit. A cube could always be displayed within the resolution a single voxel offers. A sphere would, in theory, require an infinite resolution. The solution came to me while i was typing out my "voxel resolution definition". If you work in truesize you can state that at a certain level you would begin to start sculpting detail which cannot be seen (lets say even finer then pores). This means that you can set your upper limit at that resolution. Nomatter how much detail you add, it will never need a higher resolution then this one. This is a very good reason to introduce a truesize sculpting workflow.

Then there is the problem of at what resolution sculptural changes are made. You only know wether a form can be optimized after it has been formed. Does this mean all brushes should perform at the maximum resolution by default? This would create an unworkable situation. Lets say that gets solved one way or the other. The optimization process would still need to take place. Think of the surface tools in the voxel room for a second. The "only" thing that needs to happen there after each stroke is the conversion from surface to voxel. Yet this gives a noticable feeling of lag after every stroke even on the most powerfull systems. To have an "optimize" routine run after each and every brush stroke would create the same effect.

-----

I still think that the concept of brush resolution (as i guess is done with smooth) is the most promising to increase voxel sculpting performance.

3dioot

Can Voxel resolution issues be helped by having a base res in the parent volume and higher res in additional volumes? Not having tested this, can you modify the low res volume and see the changes in the 'outer' hires volumes?

Link to comment
Share on other sites

  • Contributor

Firstly an updated detail head -

Closer.jpg

Still not finished but Im getting more comfortable with the detail capabillities. Straated making my own detailing set of pens and will share them with all once complete.

Secondly I Like what you guys are saying some nice ideas batting back and forth.

3Dioot, yes brush/pen independant resolution sounds good. and possible i think, What about something like a brush that could paint, marks voxel area then a slider that could be used to increase density dependant on area masked/selceted?

I kind of take for granted that currently leveling up/down is a impossibillity so I would push both Localised voxel density and Optimisation of large brushes performance. The move brush can work very well at high resoltion areas but the larger it is the worse it gets. Optimise this and like i said before Resolution would not be so important.

- leigh

post-1266-1246839601_thumb.jpg

  • Like 1
Link to comment
Share on other sites

  • Contributor

Couple more updates -

Head_Nearly_Complete_01.jpg

Head_Nearly_Complete_02.jpg

Getting there. One big problem is the Stamp will only function at to a certain size. as you drag too far it dissappears and no stamp is applied.

  • Like 1
Link to comment
Share on other sites

  • Advanced Member

Great! I was studying the skull today , making it in layers .

You can change the color of the default material for every seperate layer when you switch to paint room for a second.

That way every layer gets its own color. Nice to see the muscles apart from the bones etc.

Link to comment
Share on other sites

  • Advanced Member

Very nice work Leigh. He has a warm genuine smile which is not easy to achieve in a sculpt. :)

I look forward to your brushes gift as well.

Yes opening the mouth is a beautifull example of where mesh based sculpting would have faltered.

I have an idea on how to fix brush resolution seperate from min and max levels and truesize sculpting. Dont feel like putting another wall of text in your thread though. Perhaps ill save it for the beta thread.

Keep up the great sculpting.

3dioot

PS

Edited for positivism.

Link to comment
Share on other sites

  • Contributor
Great! I was studying the skull today , making it in layers .

You can change the color of the default material for every seperate layer when you switch to paint room for a second.

That way every layer gets its own color. Nice to see the muscles apart from the bones etc.

Try painting a copy layer of the Jaw also that way you can rotate it open translate it into place and acheive a nice pose. You can find a layered skull and mandible from me HERE!

post-1266-1246909219_thumb.jpg

post-1266-1246909224_thumb.jpg

Link to comment
Share on other sites

  • Advanced Member
"As far as i can think it through its still technically impossible to create multivoxelresolution (ive spend hours thinking about it, making notes and sketching things out). I feel its important to accept this. Voxel sculpting brings many, many major advantages. Features that will never be fluently possible with mesh sculpting. But multilevel sculpting is something it simply cannot do."

3dioot

You've obviously given this a lot of thought (more than me):) and I think we are on the same page in the end. I just want the ability to sculpt in High detail and also the ability to be able to export a low poly or low resolution object out of 3DCoat to be used in other 3D applications. Without this ability 3DCoat is useless to me.

For example if I were to create an extremely detailed dinosaur, I would want the ability to sculpt that in Voxels and then export that out of 3DC as a low poly mesh to be used in other 3D applications. We agree the voxel sculpt needs to be converted to poly's to be useful in a 3D application so however this was done would not matter to me, but I want to export a low poly mesh out of 3DCoat and not a high poly one as it would be far too cumbersome. Thanks for the banter. :)

leigh, excellent head sculpts. I see you have a lot of talent for sculpting and anatomy. :)

Link to comment
Share on other sites

  • Contributor

SpacePainter - Thank you Again

3Dioot - Thanks, Please share your thoughts on brushes PM maybe we should get our heads together on that!

Paint Guy - Thank you, Please keep up the banter I will add to it As and when i come up with something new.

Got a little carried away on the open mouth sculpt and he turned into the devil? - leigh

post-1266-1246920870_thumb.jpg

Link to comment
Share on other sites

  • Contributor
SpacePainter - Thank you Again

3Dioot - Thanks, Please share your thoughts on brushes PM maybe we should get our heads together on that!

Paint Guy - Thank you, Please keep up the banter I will add to it As and when i come up with something new.

Got a little carried away on the open mouth sculpt and he turned into the devil? - leigh

I think we all would enjoy watching you play. ;)

Link to comment
Share on other sites

  • Contributor

Playing around with the pose tool and lasso, here is my first test pose.

Test_Pose.jpg

lots of cleanup required (well it aint a substitute for rigging thats for sure) but I still prefer it to transpose in ZBrush.

Link to comment
Share on other sites

  • Contributor

Taros - Thank you, I want to try and get the weight transfer right here, or it will ruin the pose! Using the pose tool is way too tricky at high resolution of voxels, and this make it only viable depending on how you want to use the result. Let me explain my thoughts. Topological symetry will neve be possible with Voxels so posing here will only really work in early stages if you intend to try and keep things symetrical. Also the fact that things need a whole bunch of clean up means that posing is best left to early stages. Here with this character he is as close as i would like to go b4 posing. any higher and detail will be lost.

On another note regarding functionallity of pose tools rotation, view based rotation does not function correctly, it should when view based is selected no longer lock the axis of rotation because this makes it tricky to judge which axis to manipulate meaning it is not based on view at all. What i would like to see is when using view based rotation a single double ended arrow perpendicular to the viewport this would illustrate proper view based pose rotation.

Link to comment
Share on other sites

Taros - Thank you, I want to try and get the weight transfer right here, or it will ruin the pose! Using the pose tool is way too tricky at high resolution of voxels, and this make it only viable depending on how you want to use the result. Let me explain my thoughts. Topological symetry will neve be possible with Voxels so posing here will only really work in early stages if you intend to try and keep things symetrical. Also the fact that things need a whole bunch of clean up means that posing is best left to early stages. Here with this character he is as close as i would like to go b4 posing. any higher and detail will be lost.

On another note regarding functionallity of pose tools rotation, view based rotation does not function correctly, it should when view based is selected no longer lock the axis of rotation because this makes it tricky to judge which axis to manipulate meaning it is not based on view at all. What i would like to see is when using view based rotation a single double ended arrow perpendicular to the viewport this would illustrate proper view based pose rotation.

I made the same experience and am not happy with the current status. I see the same problems: Posing is possible only in a early state - but early posing costs the possibility for symmetry skulpting... a tricky situation. The solution is only to export the ready skulpted object and rig it in a animation software. Then, if necessary, reimport and make corrections. A science for itself... :)

Regards

Chris

Link to comment
Share on other sites

  • Contributor

Finally a bit of actual work in here!

Im modelling a character for Discovery Channel and did some of the primary texturing in 3d Coat so I'll share him here.

Discovery_Final.jpg

Dicovery_OL.jpg

Discovery_HL.jpg

Unfortunately i did not finish texturing in 3D Coat but this was more due to my inexperience using it to texture instead i finished in Mudbox. I will brush up 3D Coat texturing and try to intergrate further into my work.

Rendered in XSI BTW, incase your wondering.

[Edit] Whoooooops..... I mean Softimage 8P

Link to comment
Share on other sites

Nice.

I am interested, how well it works to simple quadrangulate the voxel-object, save it with uvws in obj format. Then bake voxel information to a normalmap and import it to softimage. Maybe I will make some tests too.

Link to comment
Share on other sites

  • Contributor

Made a start on that test here Is 3dcoat views, I dont have softimage installed here but im heading to the office in 20 or so i will render it out and see how it look. Im not really very happy on the quadrangulation result but if it look good in xsi then it s sure a substitute to decimation master. We will see!!

Quad_HighPolyPaint.jpg

Quad_WireFrame.jpg

Quad_LowPoly.jpg

BTW I really havent tried on the detailing just roughed in ares is all, I feel i get slightly better results with voxel tools than painting tools on detail so eventually maybe i will take this fellow further detail wise, I dont see the point till i know what is happening when i get it into render. I have licenses to Maya and Max also so i will run off some comparisons when i have the time.

Some of you may think that is a little high poly wise but let me remind you that this would just be for still image as a wip shot to client befor retop and painting. to give suggestion of finished result. so high poly in XSI not a problem for static images.

[Edit] For best results using quadrangulation on model with higher detial I think is best to use little or no normal smoothing, which keeps the poly count higher in areas of detail like hands and mouth.

- leigh

Link to comment
Share on other sites

  • Contributor

Quadrangulation is only very effective from the way I see it using microverts,because you can export mid-hires(or higres) level displaced mesh but you can't from perpixelpainting.It's gonna have a slightly higher polycount than the quadrangulated mesh but if you combine it with a displacement map you'll get a far less messy result than decimation master.

But really if you are using only for illustration directly from voxesl-to decimation master way is faster.

character rocks btw :)

Are you planning to do a full composited shot with background and all?

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...