Stop Staring 3. What do you want?

[theflash]

First of all I love your second edition book. Thanks for putting such nice stuff together.

For next version, I think it would be really cool to start off with some advance system using blendshapes. e.g. based on FACS. I think one of the most challenging things is to make individual action unit work together properly to create good shapes.

I would also like to see more on cartoony facial rigs.

I am really curious to see your take on a mixed system with joints being main skin deformers and blendshapes as a corrective or a parallel system.

I will highly prefer less repeatation from your second book. That way people who have second book won't hesitate to buy the third book.

[dunkelzahn]

A chapter about FACS and its practical implementation and additional stuff about Facial muscles will be much appreciated.

Though it will be hard to top the 2nd edition which was one of the best 3D books I bought ;)

[Geuse]

Hi. I have only read the first edition so I wouldn't know what the 2nd is like.
Why not bother with expression anymore?

One thing I'd presonally would like to understand is the approach to deal with displacement mapping and how to effectively set this up for an animated mesh. Dealing with intersection, disp maps for different part of the face all connected to ride on top of the "base displacement" much like what was answered concerning bump mapping in the 1st edition.

Also as theflash mentioned a more indepth on deformable cartoony rigs would be fun.

Perhaps I should go have a look in that second edition.
Keep up all the good work.

[NolanSW]

I want a good cup of joe with my muffin.
I want to be able to go back in time to correct all my mistakes.
I want to know how to make money fall from the sky instead of burgers.

How's that for starters :P ?

Yo Jason, good luck on your next installment.

-Sean

[eek]

First of all I love your second edition book. Thanks for putting such nice stuff together.

For next version, I think it would be really cool to start off with some advance system using blendshapes. e.g. based on FACS. I think one of the most challenging things is to make individual action unit work together properly to create good shapes.

I would also like to see more on cartoony facial rigs.

I am really curious to see your take on a mixed system with joints being main skin deformers and blendshapes as a corrective or a parallel system.

I will highly prefer less repeatation from your second book. That way people who have second book won't hesitate to buy the third book.
'


Id be wary as to introducing FACs, its a massive manual around 560 pages, with 60 action units and all there combination at intensities going from a (slight) to e (extreme) - its not for the faint hearted. Plus the fact that its copyrighted material entitling you to one hand copy only.


What would be awesome is correctives - firstly how to build them. How they get fired by other shapes, (linearly, quadratic, slow-in.out etc) and how there order is crucial in rebuilding them. eg.

A + B (order 0) - base shapes
= C (order 1) - first corrective
C +A - base + corrective
=D (order 2) second corrective

The rule being you have to rebuild them low to high order, this is crucial in face pipelines. As the base shapes may change at any time be it via bones or morphs. Plus storing the absolute of the corrective is very important.

Driving normal/disp maps is a good one too.

Possibly techniques for on face controls, and how there can drive other on face controls - eg. left smile control, right smile control, overalSmile control (drives the other two)

- c

[theflash]

I guess if some theory and examples of how action units work together are explained it would be nice. Not exactly Paul Ekman's system, but something simpler and similar to that idea would be good enough for this book.

[DanHaffner]

I would love a section that helps explain how to get secondary controls to work with blendshapes, so you can push a pose even further.

When explaining your tools, I was honestly lost, not sure if its just me or the way the stuff is written, but your scripts scare the hell out of me and I don't know how to use half of them. I stick to the control maker and the prep shapes ones, the others I honestly have no idea what they do, even after reading the sections that discusses them.

And what the heck is/are FACS?

[yolao]

Hi Jason

1-I would love to see a chapter dedicated to secondary controls that works and follows the blendshapes deformations..Like in this demo of Denislav Georgiev

http://www.youtube.com/watch?v=llDSg5JKHas

2-Also a chapter of automatic secondary motion like mass jiggling result of fast and hard movements, the wind, etc.

3-And i also will like to suggest a chapter dedicated to FACS.

4-Skind sliding over bones....like in this example made in XSI by rmawatson from XSIBASE.
www.rmawatson.com/skinSlide.rar (http://www.rmawatson.com/skinSlide.rar)

5-Cornea bulge sliding under the eyelids.

6-Some way to reaply all the work to a different head with diferent shape and topology, like in this tutorial for XSI by Paul Smith.

http://forums.cgsociety.org/showthread.php?f=24&t=746813&highlight=face+rigging

Thanks.

[DanHaffner]

A section of how to hook up corrective blendshapes would be awesome. A script I was using has you hook it into joint rotations, but I found out very quickly why that is a bad idea.

[uiron]

Hi Jason,

first of all, thanks for an unbelievable resource for facial rigging, this once opened my eyes on how facial setup could really be done! still using quite a lot of concepts from your book, it was extremely helpful.

i would love to see all the modeling-part stripped; not many riggers will model themselves these days; instead, need more focus on : 1. major topology requirements; 2. areas that must be doublechecked when preparing for rigging -e.g., my list would start with checking lip corners/inner shape to make sure it's rounded, not flat, and eyelid topology - is closing the eye possible? also, more general guidelines for tooth size/position, topology between eyebrows, summary of a must-have topology for creases of smile fold, squint wrinkles, etc.. should be like a checklist for a modeler before submitting a model/ rigger upon receiving it and requesting for changes.

what we find in our studio most challenging is making lips and eyes believable - the rest of the face doesn't even matter if those do not work well.

lips still lack some fake for softiness effect (it never feels like upper lip is touching lower lip). animators play quite a lot with second-layer controls (face points) to adjust them in closeups.

our fleshy eye setup is still pretty mechanic, would like to see some more ideas on that. we use a complicated blend of forward-up-down-inward-outward shapes for each stage of the eye - neutral, neutral+squint, closed, closed+squint, wide open. Controlling long eyelashes is an issue too (making them look good when eyes are closed and not penetrate eyebrow mass with eyes wide open/view direction up). we also had to throw away upper eyelid/lower eyelid concept and work with eye open-eye closed as a base, adding overlay of upper/lower lid: this allowed achieving easier watertight blinks/closed eyes at any eye pose, even with a squint blended in.

other "latest fashion" tricks would be nice - switching between closed mouth/open mouth when jaw is moving (chewing vs talking); better sticky lips solution (soft mod one just doesn't cut it). "face points" is mandatory these days as well; two-layer face points system would even be greater (smaller detail face points act on top of bigger detail face points).


Good luck with our book, i'm definitely buying that one!

[ShaderOp]

What do you want? Lay it on me!
I want to pre-order. NOW!

But seriously, it would be great if the content on the CD was made more accessible to those not using Maya. I was able to work through the 2nd edition just fine using Softimage, but it would have been easier if all the blendshapes were available as .OBJ files on the CD instead of just Maya scenes. A video of the final rig in action, showing which sliders affected which shapes would have also helped.

[JasonOsipa]

This stuff is all great, thank you! Anyone else, keep posting!

The corrective stuff is great feedback because, well, that's already in there and the solution is really solid and effective! I refer to them in the book as fix shapes in the text - I will have to revisit the wording to make that clearer. Facs will unfortunately not be a good idea, as I can't cover enough to make sense without violating copyright, but only touching on it as a reference point won't leave the book standalone, which is a big requirement.

On the modeling comments; I need to keep that in there, at least in spirit, because it does actually go on and on about why topology is what it is (which is kind of the second part of the modeling-for-rigging comment), but I can certainly revisit the wording to make it feel more like a discussion on what is and is not "stuff to watch for". Also, I'm figuring out how many/what type of models to add; that could help that area out.

Secondary controls are a definite.

Sticky lips. Damn, you. Wanted to keep the simple-but-not-great solution, as it can turn into a real nightmare of writing to go farther. Alright, you *probably* win. ;)

[blackseraphim]

hi Jason! I bought your 2nd Edition and I surely learned a lot! :D
I was thinking could you tackle about joint facial rigging? or a combination of
joint-blendshape technique?

[woadiestyol]

Hey Jason! Like everyone else, I love the book. In fact, I picked it up somewhat recently and am finishing up working through it at the moment.

I admit I got a little hung up during the interfaces section in understanding the control network that's generated. I understand how to use it with the other scripts in the book, but I really would love more information on the nodes that are being generated and what their purposes are. Maybe it was just that the particular language used to describe that section didn't quite click with me, but if I needed to customize that setup I don't think I could pull it off quickly as I'm not totally clear on what everything is doing and how to hook controls into it manually. If you have any tips in the meantime on how to wrap my head around it a little more, I'm all ears ;).

I'll also put my vote for secondary controls. I've seen some great lip and brow controls layered over blendshapes that I'd love to learn how to implement.

BTW, I'm really excited to hear that the scripts will be in Python soon! I made the switch about 6 months ago and now find myself getting really frustrated every time I have to use or read MEL, so I'm excited to browse through the new goodies.

So chalk me up for a preorder. Thanks again for all the great info thus far, and looking forward to even more!

[FabioMSilva]

it would be nice if you could do some scripting section for 3dsmax like u did with the automation with maya(like automatic scripting to create the ui and linking all the stuff).

looking forward to see version3. i got version 1 and totally love it . i think ive read it throught about 8 times

[GeoffClark]

Jason,
Reading the first edition (yes, it probably bugs you that I'm looking thru old material, but it was 10 bucks).
Thanks for writing it! I am learning a lot from it! thank you!

Geoff

[Leffler]

I would like to see something that would help the face to feel more fleshy/muscular with blendshapes.
I use the half-shapes you introduced in the book and thats a good thing, but not perfect. Lets say a jawOpen-shape, a way to make that look a bit more fleshy and opening in kind of an arc ... don´t really know HOW that would be setup considering how blendshapes work but that´s something for you Jason ;-)

Also maybe a tiny bit more about techniques for creating shapes in an easy way. Would be helpful, but its really not too important so maybe 1-3 pages max.

Other than that ... can´t really think of too much. The second edition is just awesome, pure gold.

Oh ... one more thing for 3rd editon. A new shelf with lots of good stuff :-P

Otto

[mtucker2]

I don't know if it's been mentioned here or not but I would enjoy a little section on how to make "sticky" lips.

[woadiestyol]

I don't know if it's been mentioned here or not but I would enjoy a little section on how to make "sticky" lips.

Check out the current edition (2) - it's in there :thumbsup:

[Coyote12]

Hey Jason,

My opinion would be to not try and rewrite or revise Stop Staring 2 but to write a sequel/part 2 of it. You have a lot of great information that does not need repeating and does not need revision. It just needs expanding and this way Stop Staring 2 is still relevant and part 3 expands the information further.

I would like to see alternative methods and methodology to specific tasks like Sticky lips done with Blendshapes and then done with joints. How secondary controls can add detail to previous broad movements. Why a cartoon character would need one way of building something that a realistic character would need built in a different manor. What a game character needs vs. a film character rig. I think thinks like that would be very helpful.

I look forward to picking it up when it is done. Good Luck.

[JasonOsipa]

XYZ shapes i the simplest solution, and that shipped with ss2.

That way your "start" shape is broken into 3 shapes, each with only the X,Y or Z movement of the original. It automatically sets up driven keys for each so you can handle each of their accelerations/speeds independently. For the open jaw, pretty much make the Z curve ease-in accelerate, and you've got an arc.

Hmmmm. This + the fix/correctives question earlier makes me think I really messed up the wording in the section where I discuss them both: Chapter 11 inside the subsection labeled "Interface concepts"(p269-273) I cover both as well as limiter or "dominant" shapes. Definitely revisiting that verbage; it is clearly lacking!

Feedback like this is great; often I'll get questions about not understanding something I wrote, and that is quite easy to target, but it is sections that *really* went largely misunderstood, that I'll get almost nothing on, as at some point I either buried it or made it prohibitive, and so to most there was nothing to even ask about (and the ones who got it, well, got it and don't have questions). This thread has been and continues to be SUPER helpful, guys! Thank you!

Update of Ch1 being submitted in 8 days. Eep!

I would like to see something that would help the face to feel more fleshy/muscular with blendshapes.
I use the half-shapes you introduced in the book and thats a good thing, but not perfect. Lets say a jawOpen-shape, a way to make that look a bit more fleshy and opening in kind of an arc ... don´t really know HOW that would be setup considering how blendshapes work but that´s something for you Jason ;-)

Also maybe a tiny bit more about techniques for creating shapes in an easy way. Would be helpful, but its really not too important so maybe 1-3 pages max.

Other than that ... can´t really think of too much. The second edition is just awesome, pure gold.

Oh ... one more thing for 3rd editon. A new shelf with lots of good stuff :-P

Otto

[JasonOsipa]

It's tricky. I need to "touch" the book again every few years, as just enough changes under-the-hood in Maya that certain tools/systems break down, and just for technical reasons it requires a spit-and-polish. The thing is, it's not worth anyone's investment to do something that light, so we go in and re-noodle a bunch of stuff. What was really cool from 1st ed to 2nd ed was that the first one actually changed the baseline language of facial animation and interfaces enough that I was able to tear out tons of what at first release was "weird and different" and required explanation, and by time of second release was "just stuff people all know now". Totally unexpected, but thrilling nonetheless. (And humbling!) Reclaiming those pages let me squeeze in just enough more to make the 2nd worthwhile to pick up without disappointing (In most cases!). This time, we're in the same boat again - if we're going to update it at all, we want to try and put together enough of a package that even if you have the 2nd, the 3rd is still interesting enough to pick up.

Making a separate book on the same topic, significantly risks mutual-cannibalizing. I'm not saying that might not be where things go in the future, but right now it just doesn't make enough sense.

Based on amount of discussion and where it is generated, though, I am more and more thinking that the biggest chunk of pages I need to think about reclaiming is the last chapter. Nobody is really screaming for more/different animations so much as more groundwork...

Processing...



Hey Jason,

My opinion would be to not try and rewrite or revise Stop Staring 2 but to write a sequel/part 2 of it. You have a lot of great information that does not need repeating and does not need revision. It just needs expanding and this way Stop Staring 2 is still relevant and part 3 expands the information further.

I would like to see alternative methods and methodology to specific tasks like Sticky lips done with Blendshapes and then done with joints. How secondary controls can add detail to previous broad movements. Why a cartoon character would need one way of building something that a realistic character would need built in a different manor. What a game character needs vs. a film character rig. I think thinks like that would be very helpful.

I look forward to picking it up when it is done. Good Luck.

[underearth]

hey jason,

i would love to see some joint based solution for second layer of control over facial blendshape. I've used hyper real ctr and many joint based solution which does the job right but still there is plenty of room over there.

check out http://www.characteroutsource.com/ this guy did some pretty intresting facial stuff with his Paxton rig in maya.

Apart from that i think max guys have done some pretty nasty stuff with facial Bone rig specially Brad Nobel, paul neale .( check this thread http://forums.cgsociety.org/showthread.php?f=54&t=260019 for his facial rig).

that would be great if you could throw up some light over this.!

thanks
SunnyC

[BubaBrown]

3rd edition is coming? This is great.
It would be great to have all the stuff others have mentioned (combination of shapes and joint rigs especially).
I would like to deepen the cartoon part.

Most interesting would be the tips regarding some procedural solutions for fast facial rigs. You know when you have to rig dozens of faces in a limited time.

[Darksuit]

Jason you may want to take a look at Making Comics by Scott McCloud. Scott has a great chapter that I used to use in teaching along with some of your techniques. in the Book Pages 83-86 specifically, show a really good drawn expample of what happens with blend shapes. You may want to take an idea from this to help explain blend shapes a bit more.

You may also want to talk about using the Wire deformer to control parts of the face.

[joshfilms]

I'm quite new to the world of rigging and I use Cactus Dan's tools but I've heard a lot about your book. I suppose I will wait to buy the 3rd edition. I'm not sure whats in the current books but I would love to see some stuff about muscle systems and working with zbrush displacement maps on rigged characters.

thanks,
Josh

[joel3d]

Hey Jason. Thanks for writing your book, I learned a lot. I'm looking forward to getting the 3rd.

One question, I have a character I setup using your tools. I have all the blendshapes prepped and connected to the sliders...Works great!

But... I want to add some more blendshapes and ssControls to the rig. Seems that If I add new blendshapes and try to prep them it breaks all the previous connections.

Do I have to have all the blendshapes made and prepped before I begin to connect them? Wish there was a way to work in a less liner way.

Thanks. Joel

[tonytouch]

hi jason ,

i would like to see some "joint_based" facial_rigging , including a simple and smooth facial_skinning of these joints and maybe to see techniques , how to create skinclusters for a facial joint_setup with lots of joints - for both lowres and highres models .

that would be awesome :)

[JasonOsipa]

1st Edition v 2nd Edition v 3rd Edition (http://www.facebook.com/photo.php?pid=11798932&l=4d93389f75&id=853475526)

[woadiestyol]

Cooool....are you gonna cover hair/fur? Are you using Maya tools for this or is it Shave?

[JasonOsipa]

This is straight-up Maya stuff. I'm not going into almost any detail about it. For what I needed to do, and what I did do, there's incredibly little to discuss, actually. I show shots of the maps, and have about a para on it, but that's about the extent of it. With my limited page count, I always need to steer stuff back to specific face tech and art. Anything spent on topic A comes out of topic B, you know?

[woadiestyol]

I hear ya, and I'm glad that the focus of the book is in the right areas. I'm one of the folks who was asking for more details on the face tech :thumbsup:

That being said, I have yet to get decent results out of Maya hair or fur for actually creating hair or fur. The blame for that definitely lies between the chair and the keyboard, but I'm sure even a quick half page or so on general tips/tricks/ideas for working with that stuff would be appreciated by a lot of folks.

Looking forward to the new book! Put me on the pre-order list :buttrock:

[JasonOsipa]

I actually had one last chance to affect this ( I am actually doing the author-review-edit on the exact section you asked about right now ). I puffed it out a little bit to include a couple of tips. Again, there really isn't *too much* to even talk about, but I threw in some pointers. Good timing on the request. When I send this file off in an hour, it's locked down but for spell check and layout.

[Coyote12]

It was suspect Three Officer! He was the one who robbed me!

I love the evolution of the "head" over time. Nice progression. I am looking forward to what you have written on secondary controls and any joint based facial set ups. Not that I don't like blend shape based just that no one writes about joint based.

So how long till it should hit the shelves? Inquiring minds want to know.

[JasonOsipa]

Well, I'm quite a bit behind ;). That happens every time. I procrastinate a little, then other work comes up, then I fall behind more because of that, then I jump into it full strength. Once I do, I realize that I can do 10x the quality improvement, in either ideas or visuals, but it'll take 2x the time, and then I spend a ton of time going down that road.

Right now, it is uncertain. I'm hoping we're on shelves before the summer is up.

As far as joint setups, there isn't going to be *too* terribly much, in terms of JOINTS-ONLY setups. I definitely am going into really strong secondary controls, which are done through joints (done in a way that I think a lot of people will be excited about... it's a method that seems like it should fall victim to cycle problems, but doesn't. Controls affect each other as well as the mesh). That could be almost entirely it's own rig, I just rarely use it that way. I think you'll be happy with what is in there.

[woadiestyol]

Sweeet...the secondary controls/cycling issues are something I'd LOVE to learn more about. Looking forward to it!!

[Stellios]

less modelling.

cornea bulge?

[uiron]

hey Jason, I just wanted to say - congratulations on finishing the book! I hope you understand how much impatient we all are already:]

[mccollom73]

Hey Jason,

I have been reading your book, more on the facial rigging side of the book. It's a fantastic book with more information that I can handle. I feel like you do a great job explaining steps and it also gives you a way to figure things our for yourself. With that being said, I would like to request for video tutorials. For me, I would like from the start product of rigging, to the end product, and all the steps in between. You would appeal to more than one/two different learning styles, and throw another one out there, visual. I do like how you explain everything, but I may not be doing this right, and it would be nice to see someone that knows what they are doing to show each step. I am a visual learner, and I am sure there a lot of people out there that are also.

Please consider this, and even if you do not put this in your third edition. Maybe there is a chance that I could convince you to make a video tutorial of the facial rigging and post it on cgsociety so people could use it for reference. Or anyone for that matter. Thank you for your time and reading my post.

Josh

[JasonOsipa]

http://www.youtube.com/watch?v=jMwzC126xc8

[JasonOsipa]

cover image (http://www.flickr.com/photos/22988200@N02/4839781065/)

[theflash]

Hi Jason. The links says its a private photo.

[Kaleidoscope]

I can't view the image either. But I can SAY that I'm super stoked for this book :) Loved the 2'nd edition, and I will probably love this even more.

Thank you for your hard work Jason

[danb]

I really look forward to this book. I bought the first addition a while ago.

I agree with needing more info for toon face rigs. Particularly using clusters/joints for editing of morph driven, or mocap driven face animation.

Just want to say that even though its maya oriented, you've found some fans in the cinema 4d community. :)

[FabioMSilva]

regarding 1st edition,2nd and thir edition images...why do you keep getting uglier Jason? :p:p:

1st edition was the book i probably re-read the most times in my life. so i guess thimes time ill give some thinking of buying the 3rd edition one :)

cheers

[MarkD]

How about using a sculpting app to create wrinkle maps?

A lot of what you cover crosses over to other apps, but there are parts that just don't jump the gap, mostly when you get into very specific tools and scripts. Any chance of exploring the other apps or at least pointing users toward some resources that could help bridge the gap?

Awesome stuff as always!

[j3st3r]

I wish to see several topology solutions as well, there are alternate solutions (avatar has a new way also) that would be useful, including different styles of modelling (point by point, edge extrude, boxmodelling, and retopo)

[player1]

... What do you want? Lay it on me!
muscle-joint rigs

examples below
http://physbam.stanford.edu/~fedkiw/animations/storytelling_muscles_sound_divx.avi (http://physbam.stanford.edu/%7Efedkiw/animations/storytelling_muscles_sound_divx.avi)
http://physbam.stanford.edu/~fedkiw/animations/storytelling_markers_sound_divx.avi (http://physbam.stanford.edu/%7Efedkiw/animations/storytelling_markers_sound_divx.avi)
http://www.behind-universe.org/images/M2/FacialMuscles-w-Ctrls-02.jpg

[JasonOsipa]

Text and imagery has all been locked in for a month now.

[Polimeno]

'
What would be awesome is correctives - firstly how to build them. How they get fired by other shapes, (linearly, quadratic, slow-in.out etc) and how there order is crucial in rebuilding them. eg.

A + B (order 0) - base shapes
= C (order 1) - first corrective
C +A - base + corrective
=D (order 2) second corrective

The rule being you have to rebuild them low to high order, this is crucial in face pipelines. As the base shapes may change at any time be it via bones or morphs. Plus storing the absolute of the corrective is very important.

Driving normal/disp maps is a good one too.

Possibly techniques for on face controls, and how there can drive other on face controls - eg. left smile control, right smile control, overalSmile control (drives the other two)



Totally agree !
Correctives is what LuisSanJuan (http://www.di-o-matic.com/press/success/Nexus-LuisSanJuan.html#page=three) call as Binomial (http://mathworld.wolfram.com/BinomialDistribution.html) Morphs, by the way ?

[eek]

Totally agree !
Correctives is what LuisSanJuan (http://www.di-o-matic.com/press/success/Nexus-LuisSanJuan.html#page=three) call as Binomial (http://mathworld.wolfram.com/BinomialDistribution.html) Morphs, by the way ?

True weight space corrections would be better imo. It supports placing shapes at any dimension [0:1] at any position, and crucially its non linear. Storing order is as simple as assigning each target at creation and its full transform. Check against the orders order of weight and find the highest and set it one higher.

E.g

TargetA - order 0
TargetB - order 0
TargetC == (TargetA & TargetB) - order 1
TargetD == (TargetC & TargetB) - order 2

The reason this works because a blendshape value is the product of weights [1..n] @ 0 > t > 1 i.e a blendshape is not bound to a single value.

[yolao]

So the new book is out now, can you give a list of what`s new in the third edition?

Thanks

[TheRazorsEdge]

A list of what's new would be great! I much enjoyed the first and second editions, but am admittedly a little reluctant to buy the third one on top of the other two. :)
Cheers!

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.