Mudbox 2009
polycounter lvl 19
http://area.autodesk.com/mudbox_preview#performance_preview
Saw this link on Tweak. Interview with Dave Cardwell and a small video demo of new features. GPU powered interactive HDRI and AmbOc on a 25mil mesh interactive in viewport. Crazy stuff.
Photobucket link
http://vid258.photobucket.com/albums/hh257/robioto/performance.flv
Interview with Dave Cardwell
The Area: What kind of performance enhancements can we expect to see in the upcoming version of Mudbox?
Performance was one of the big things that people wanted to see improved in Mudbox, from Mudbox 1 and we've completely re-architected the rendering engine and the brush engine to increase the performance several times in the application. What you're going to see in Mudbox 2 is a completely new rendering and brush engine that's going to give you the ability to visualize the entire model with all the high resolution details on screen at once, and not just small local subdivision regions. And so this will allow you to, as an artist, make better decisions about the characters or the environments you're actually working on. But it's also going to give you the ability to present that asset to your art director, client or the director of the project that you're working on.
The performance is geared towards the rendering speed, camera interactivity and also brush speed. Also the amount of polygons you can subdivide too. And how fast you can subdivide...actually changing level speed. That was really slow in Mudbox 1, we've actually improved that a lot. File size is another thing we have been able to reduce. The amount of memory that the files take up on disk has been reduced. So you're seeing all those performance enhancements and you're going to get that in Mudbox 2.
The Area: Does that mean that the architecture has been completely rewritten?
The architecture for the rendering and the brush engine has been completely rewritten. We've actually implemented a stamp-smoothing algorithm that was in development for ten years - the same algorithm that was used in Studio Paint. So you're going to see much smoother brushes. If you remember in Mudbox 1, if you do really tight circles and curves, you end up with a really faceted line - you won't see any of that anymore. And also have explicit control over stamp spacing. That's going to help you in applying textures and patterns in a very controlled way. And for sculpting, it's just going to mean a lot smoother strokes, and very little lumpiness. Performance improvements all around. In some cases, we've had to enhance a significant portion of the code to be able to accomplish this. A lot of code has been enhanced from Mudbox 1.
The Area: Are you taking advantage of hardware and multi-threading?
The entire application is GPU accelerated and of course, the graphics card has multiple threads, the main CPU processors in some instances, we are using multi-threading for certain things, but we're currently not using threading across the entire application. What that means is that the application still could get a hell of a lot faster in the future. One of the nice things about being at Autodesk is that you get a heads-up on the technology that's going to come down the pipeline from various hardware manufacturers, be it graphics cards or processors. And so based on this, we're making sure Mudbox is in a good position to take advantage of all that stuff as it comes out. And soon as it's out there, we're going to be able to take advantage of those technologies with Mudbox because of the code changes that we're making now.
The Area: What about hardware needs - do you need a hi-spec'd machine?
Do you need a supercomputer? No. There's going to be a recommended hardware list, just like all the Autodesk products, but you can run it on significantly less powerful hardware. But there will be a qualifications chart. It will basically qualify in a very similar hardware configuration as Maya. Just as an example, we're running Mudbox 2009 - that's what we're calling it, Mudbox 2009. I'm running it here on my laptop, a Dell 1520 Inspiron with an 8600 mobile graphics card and I can work with very very high polygon counts at extremely interactive speeds. Even if you had a super computer to run Mudbox 1, it's much faster than Mudbox 1 ever was. So it's a huge improvement, like several times, not just like 10, 20, 30 percent, but we're talking like a few hundreds of percent faster than Mudbox 1.
The Area: Is this what you used to record the video?
No. To record the video, the video that's shown on AREA now, I used my home workstation. I'm running a 64-bit, Windows 64. Actually I've got two graphics cards I use. I use an 8800 GTX which runs extremely well, but for the video I used QuadroFX 4600...it's got a lot of ram on it. It's not necessarily significantly faster than the 8800 GTX cause that runs equally as well for Mudbox 2009. But yeah I'm running Windows 64, with 8GB ram, QuadroFX 4600. It's a computer you'll want to use if you're doing production work. You don't want to limit yourself based on hardware and computers are pretty cheap nowadays anyway.
Saw this link on Tweak. Interview with Dave Cardwell and a small video demo of new features. GPU powered interactive HDRI and AmbOc on a 25mil mesh interactive in viewport. Crazy stuff.
Photobucket link
http://vid258.photobucket.com/albums/hh257/robioto/performance.flv
Interview with Dave Cardwell
The Area: What kind of performance enhancements can we expect to see in the upcoming version of Mudbox?
Performance was one of the big things that people wanted to see improved in Mudbox, from Mudbox 1 and we've completely re-architected the rendering engine and the brush engine to increase the performance several times in the application. What you're going to see in Mudbox 2 is a completely new rendering and brush engine that's going to give you the ability to visualize the entire model with all the high resolution details on screen at once, and not just small local subdivision regions. And so this will allow you to, as an artist, make better decisions about the characters or the environments you're actually working on. But it's also going to give you the ability to present that asset to your art director, client or the director of the project that you're working on.
The performance is geared towards the rendering speed, camera interactivity and also brush speed. Also the amount of polygons you can subdivide too. And how fast you can subdivide...actually changing level speed. That was really slow in Mudbox 1, we've actually improved that a lot. File size is another thing we have been able to reduce. The amount of memory that the files take up on disk has been reduced. So you're seeing all those performance enhancements and you're going to get that in Mudbox 2.
The Area: Does that mean that the architecture has been completely rewritten?
The architecture for the rendering and the brush engine has been completely rewritten. We've actually implemented a stamp-smoothing algorithm that was in development for ten years - the same algorithm that was used in Studio Paint. So you're going to see much smoother brushes. If you remember in Mudbox 1, if you do really tight circles and curves, you end up with a really faceted line - you won't see any of that anymore. And also have explicit control over stamp spacing. That's going to help you in applying textures and patterns in a very controlled way. And for sculpting, it's just going to mean a lot smoother strokes, and very little lumpiness. Performance improvements all around. In some cases, we've had to enhance a significant portion of the code to be able to accomplish this. A lot of code has been enhanced from Mudbox 1.
The Area: Are you taking advantage of hardware and multi-threading?
The entire application is GPU accelerated and of course, the graphics card has multiple threads, the main CPU processors in some instances, we are using multi-threading for certain things, but we're currently not using threading across the entire application. What that means is that the application still could get a hell of a lot faster in the future. One of the nice things about being at Autodesk is that you get a heads-up on the technology that's going to come down the pipeline from various hardware manufacturers, be it graphics cards or processors. And so based on this, we're making sure Mudbox is in a good position to take advantage of all that stuff as it comes out. And soon as it's out there, we're going to be able to take advantage of those technologies with Mudbox because of the code changes that we're making now.
The Area: What about hardware needs - do you need a hi-spec'd machine?
Do you need a supercomputer? No. There's going to be a recommended hardware list, just like all the Autodesk products, but you can run it on significantly less powerful hardware. But there will be a qualifications chart. It will basically qualify in a very similar hardware configuration as Maya. Just as an example, we're running Mudbox 2009 - that's what we're calling it, Mudbox 2009. I'm running it here on my laptop, a Dell 1520 Inspiron with an 8600 mobile graphics card and I can work with very very high polygon counts at extremely interactive speeds. Even if you had a super computer to run Mudbox 1, it's much faster than Mudbox 1 ever was. So it's a huge improvement, like several times, not just like 10, 20, 30 percent, but we're talking like a few hundreds of percent faster than Mudbox 1.
The Area: Is this what you used to record the video?
No. To record the video, the video that's shown on AREA now, I used my home workstation. I'm running a 64-bit, Windows 64. Actually I've got two graphics cards I use. I use an 8800 GTX which runs extremely well, but for the video I used QuadroFX 4600...it's got a lot of ram on it. It's not necessarily significantly faster than the 8800 GTX cause that runs equally as well for Mudbox 2009. But yeah I'm running Windows 64, with 8GB ram, QuadroFX 4600. It's a computer you'll want to use if you're doing production work. You don't want to limit yourself based on hardware and computers are pretty cheap nowadays anyway.
Replies
I really can't tell if you're being serious or not
Jonathan: sorry to say, but fact is Autodesk really has not innovated. Nearly every addition to 3dsmax can be linked to xsi, houdini, maya, etc having it a release or two before. Autodesk looks to be going more and more the route of EA sports. Slight updates every year.
I see nothing. Which should now be expected from Autodesk updates.
(seriously, the page is blank. Are they reinstalling it?)
If mudbox is handling such levels of detail it'll be nice, but I hope they're adding more to it than more polies on the screen with hdr and ambient occlusion.
Was there any information on when Mudbox 2009 would be released?
Since it has ambient occlusion in the viewport, I wonder if there is an option to bake it out.
If you don't mind my asking how did you manage to capture the video?
I used a download manager that has an option for downloading flvs. There are some addons for Firefox that do the same thing.
He just saved the flv video. there's extensions for firefox that let you do this.
tsk 2009 =/...so lame.
there might be a firefox extension that allows you to do it.
Always eager to hear about new tools! ty
Still though, I can't really see anything there that would make me want to ditch Zbrush. So far, I have never wanted a 25-million-quad mesh (although even that when they were showing it, had a triangle/poly count of 16xxxxxx in the viewport counter so I'm not sure what they were talking about), and there's nothing demonstrated in that video which I can't already do in ZBrush.
Anyways, can't wait.
That could of come in handy for me a few months back, Scuplting a 21 million poly wall for work, Zbrush was crying alot...
Mudbox is the simpliest app i have ever used, but zbrush is pretty good although is pain in the ass to learn.
A bigger canvas size in zbrush makes editing slower. One thing i hate of mudbox, is the filesize, crazy! like we had teras of HDs... 500mb 300mb... damn
I was hoping for more then just performance tweaks, hopefully they have some more announcements coming soon. I've never really pushed Mudbox to the point it would melt so performance was never really an issue. Sure Mudbox is easy to learn because there's 7 tools and only 3 anyone really uses.
I don't mean to sound like I'm bagging on Mudbox, I like, we use it at work, it works great for what it does. I was hoping for it to compete instead of tow the Autodesk line of standard version "un-grades".
I agree with per about the ZBrush learning curve and general unfriendliness, but it has a lot to offer and it would be great if Mudbox stepped up to that level and was able to retained its approach to user friendliness. But I'm not sure anyone can do that without convolution of the interface.
Bottom line, I'm getting tired of buying the same copy of software when the only difference is a 09 sticker over the 08 label. Thanks for spending all your time fixing something I never really had an issue with... Hopefully they have more news coming.
Features I'm still holding out hope for:
(I'm sure there are some work around for some of this, but for a program that "just works" they have a few things that need to "just be fixed" so they can "just work".)
- Retopo...
- Selective sub-d that doesn't break the mesh into pieces.
- Get rid of the weird layer rules that prevent you from sculpting on lower layers. Or organize layers in sub folders according to the level of sub-d they belong to.
- Mirror axis that works on its own axis not on the objects axis (it can start there but should be movable) and it shouldn't use some crazy "select two polys and hope it works" voodoo.
I have a lot more but I have stuff to do...
Hahaha, awesome visual.
ZB has always had performance - fake or otherwise, it gets results and that's all that counts for me. MB2 getting a performance boost isn't enough for me to upgrade either. I've had that for so long I take it for granted. It's the other stuff I need more now. Like awesome masking and grouping/visibility features, mesh extraction, transpose and deformation tools, which when combined with masks simply fucking rock. That shit speeds up my process more than performance does. I rarely work with a single mesh that pushes more than 3 million polygons anyway.
3d is fake so I don't get it when people complain about zbrush being fake at 3d. I really don't give a shit about what kind of monkey math is happening under the hood, so long as I get my results.
Painting is the only reason I'm interested in mudbox 2 at this point. But I want to see it and try it for myself before I upgrade. Still waiting for modo to remove the suck from its painting tools after 2 years so I'm not going to drop money on another program that does it half assed (I put way more faith in the skymatter guys getting it right than lux for what it's worth).
I don't care who wins the best software award though. I'll happily switch apps like a whore with six virginias if it benefits my work.
Spend a few hours with Zbrush and you can get good results. You're still grappling with a camera that makes no sense, bizarre canvas/subtool systems, and brushes that, while you can get some results from once you get used to, really don't do things in a very effective way.
Compared to mudbox even the simplest elements of zbrush, like the standard brush and smooth, are awkward and joyless. I spend more time grappling around the smooth tool trying to get it to work how i want it to in ZB than i do sculpting, whereas in mudbox it's perfect right from the start.
with mesh extraction , transpose and other thingies you can do 3d concepting on the fly with little or no restrictions. You can explore ideas, you can let your imagination run wild, without worrying too much about technical stuff. I kinda prefer mudbox, but ZB ain't the scariest 3d boogyman in town
Most people are using the clay, or claytubes, now instead of just the standard and it has a nice feel. I've not really had a problem with the smooth though so I might be easily pleased.
___
Per: Just on that distinct look point, is that referring to the red wax shader and the exaggerated forms it gives?
I tend to prefer the fast shader, or one of the simple clay ones. The material system is mildly interesting, but it feels superfluous. It's handy to get a change in read (like flipping something in PS) or making a high poly look more interesting, but I don't find it a big change in workflow.
Otherwise ZB is more natural for concepting work as iatriki said. If you're coming in with tight meshes and know exactly what you want to get out there wouldn't be many advantages to ZB.
The ZB UI just doesn't seem that distractingly unnatural to me now. MB is still more natural, but (contrary to Suprore) I don't find ZB is pulling me out of what I'm doing any more -- which it did before 3.1, and some of the plugins they've released.
Subtools for example.. Why do I have to append a 'polymesh3D1' before I can import another obj? All work arounds and hacks. Yes, the software *CAN* do allot but it's fundamentally flawed and it's gonna come back and bite when the shit hits the fan and it needs to move to true 3D.
Zbrush is freedom, plain and simple. I'm the first to admit the GUI isn't hot shit, but it ultimately has little influence over my decision to use it or not. And some of the actual UI elements, like masking and working with polygroups are pure fucking genius. Seems to me like those who pan it solely for it's UI issues are the ones that haven't used the program to its fullest extent. I got over that a long time ago. I like the fact that it isn't just a stripped down set of brushes that only cater to a game or film modeling grunt's needs because I need it to do more than just 1 or 2 things.
I can build an entire character in z without ever leaving except for the final, low poly mesh and I can do it a lot faster than I can with wasting time in a polygon modeler. There is freedom in that that makes 'old school' polygon modeling much less important. It doesn't eliminate it by any stretch, but it definitely cuts down on the bs and modeling with polygons still involves a lot of bs, no matter what software you use or how long you've been doing it.
The problem most people have with zbrush is that they are trying to fit it into such a tiny, constricted, standardized process that it's no wonder they get frustrated with it. I wouldn't want to use it either if that were the case, much like I don't like using a big bloated studio app for something as simple and non-pipeline critical as modeling. Why not build a bust or something simple like that in mudox? I would. I like simple tools too and I go out of my way to research anything that might make me faster, or just able to enjoy my work that much more while I'm doing it. I'm just glad I was given the option of using both ZB and MB at work because I've definitely had a need to be able to do more than just sculpt some details on a single mesh since I got here.
With zbrush I really don't give a shit what my art director tells me to change. I can do it, at any stage, which makes it invaluable for iterative work and I can do it on the fly while he's standing over my shoulder - especially at the start of a project when nothing is nailed down and it's more important to throw as much shit at the wall hoping something will stick. But even moreso, jeebus forbid, at a later stage in production when you really don't want to be doing that kind of thing. It shouldn't happen, but it does, and when it does I'm definitely glad I have zbrush to work with.
Anyway, like I said, would be nice if mud would expand the toolset to cover other areas that fall under sculpting, for those of us who need it. I don't care who comes up with an idea first. I only care about who implements it better in their software. If that involves a little copying and leapfrogging each other's tech, that's cool too.
http://www.zbrushcentral.com/zbc/showthread.php?t=057154 they released that a few months ago, makes multi obj imports as painless as any app.
I don't have much of a the problem with the camera, if my persp fov is set to 70 in both zb and max it's reasonably close.
To me mudbox was the bigger pain to use with this keyboard shortcut stuff.
I could never do without zrush, it absolutely rocks.
if I need to change the base mesh I can reproject the mesh easily on to the new base.
subtools and layers are great also.
polypainting is great for blocking in colour
the only thing I like about mudbox is the display, looks more like a traditional 3d app display,.
I have messed around with mudbox a bit , but when it comes to the crunch there is too much
stuff it can't do and will always be playing catchy up with zbrush.
it s like some people like lightwave but I always found that a bit odd to work with.
i find mudbox in that category, but probably would learn it if I had to.
What is it about digital artists that gets them so attached and worked up about their weapons of choice ?