Like everyone else, I'm excited about all the new features, but I have not heard mention of whether or not the interface will now take advantage of 3D acceleration (GeForce3, etc.). I'm NOT talking about the rendering, but rather the INTERFACE itself. Currently with Poser4, even on my smoking 2GHZ machine with a GIG-O'-RAM, the interface became almost unuseable when enough models were on screen. I'm hoping (no, praying) that this will be eleveated with new version. Kup, anyone?

Most likely not, although at the rate the hardware is improving, by the time Poser 6 comes out I wouldn't be surprised if a typical 3D graphics chip could accelerate preview mode easily. The geometry might have to be converted to triangles first.

SO it has to do with how Poser renders the preview geometry, eh?

Well, it's more a matter of how consumer-level 3D hardware is optimized. 3D games use triangles, so the hardware is designed to handle triangles only. Poser meshes are typically built with quads. So for the geometry to work with a GeForce4, for instance, you'd need to convert it to triangles first. For now, this conversion would have to be done with the CPU, so any performance gains would be offset. But a few years down the road, who knows?

Thanks for the explanation ... but still a bummer. Here's to hoping your wrong! :)

I've noticed that in most 3D applications, OpenGL or DirectX 3D acceleration works great with only a few items in the scene. But as you add more items, or high-polygons items, the hardware quickly loses all of it's advantage. When you run out of on board RAM on these cards, then the card gives up entirely and you get reverted to software rendering. Victoria with a pair of 3000x3000 texture maps, and you're already exceeding what most gaming cards can handle. Game 3D cards are totally optimized for getting high framerates in high-speed, low-polycount games. I doubt very much if any of the current nVidia or Radeon cards would help speed up rotation of a millennium figure scene much, even after triangulation.

True. Today's cards are pretty robust (as much as 128MB of onboard memory) and they're only going to go up from here. But game characters are typically 2000 polys or lower, and an entire scene with multiple players might be 10000-15000 polys, tops. Like I said, wait a few years. The hardware will catch up to Poser; it's evolving much faster than software.

Umm, not to risk being toasted, but that's not entirely accurate. Most consumer cards are certainly optimized for gaming (and tris) they can certainly be used for applications and quads. If Poser were to support OpenGL, cards like geForce would provide signifant hardware acceleration for the preview/display. Lightwave (and UVMapper Pro, and most 3D apps for that matter) will take advantage of the cards hardware via the OpenGL layer, and in Lightwave, quads are almost a religion. This can make a huge difference in speed and quality of the interface and previews (try rotating a 100,000 polygon model with software vs with a GeForce hardware accelerated OpenGL, you'll see what I mean). The down side is you cut out a lot of machines that can't do OpenGL in hardware. Posers preview/gui routines work on just about any machine, and custom routines can be made faster than doing OpenGL in software, Open GL requires a decent 3D card with good drivers with OpenGL support, such as nVidia's geForce line. Add to the above the fact that Adding an OpenGL support isn't trivial, then you'll probably see why Curious Labs chose to avoid this. When NewTek switched to full OpenGL, it took them several updates to get the speed back to where it was before. I'm wondering if features like this, multiple CPU and network rendering aren't being held back for a new Poser Pro pack? Regards- Lemurtek

Really? Cool!

I knew that Nvidia's professional-level Quadro products can do some amazing stuff on workstations (Nvidia once demoed a lower-res Final Fantasy scene in realtime), and that today's game software isn't yet taking full advantage of today's game hardware, but I didn't think a GeForce could handle quite that many polys.

I think that kupa shot down the possibility of multiple-CPU support in an earlier thread, but he may have hinted something about network rendering in the near future.

Thanks Lemure ... I KNEW that Hardware-based OpenGL apps that I use were definitely sped up an extraoridinary amount (such as 3DS Max, Illusion (particles, not 3D), and even Bryce) ... that's why I couldn't understand why Poser would not take advantage of this awesome power. As far as "cutting out" systems that can't do harware OpenGL, two things: A) Get with 90's people!! If you don't have a system that can do hardware OpenGL, you shouldn't be playing with these kind of apps in the first place. B) All the programs I use that incorporate Harware OpenGL acceleration give you the option of turning it off, and using software mode as well. So why couldn't Poser do this?? -Tim

I have a lowly GeForce 1 (first rev) card and I can confirm that it handles Mil figures quite well (MUCH faster than Poser software rendering). I know this from an OpenGL application that I wrote to play with .obj files. So yes, it's possible and desireable - it's just a matter of programming time (OpenGL is actually relatively trivial to implement, extremely trivial, compared to the existing software renderer, for example). OpenGL also handles the triangulation for you (you can pass in 100-sided polygons if you want, the lower-level parts of the driver convert it to triangles for today's hardware). ...but since there's been no mention of hardware accelerated preview modes, I assume that CL hasn't commited the resources to do the conversion yet (it would be a major line-item on thier spec sheets if it was there).

Not very technical here, but Poser is the only program that switches the fan on to cool my laptops CPU nearly ALL the time. Even Vue DEsprit uses less CPU power and that raytraces.

Interesting.

A question on graphics cards..I have poser 4.03 and will be purchasing 5.0 as soon as I can.. I have a PC with a p4 1.4 ghz chip 640 meg of Rd 800 ram and a 32 meg G force 2 card... What would be the best graphics card to upgrade to? I should be able to get poser 5.0 to work on my machine as is but if you were me what would you upgrade hardware wise?? Keep in mind this is a hobby and my funds are limited! (two kids in collage!! ugh) thanks PE

There is probably a good reason why Poser, unlike other graphics apps, isn't well suited to acceleration. For an app to use any hardware acceleration it needs to repackage its own internal vertex, polygon and texture data into a format usable by the graphics card (including the OpenGL API). Poser, by its nature has a lot of complex meshes (ie. figures) that change geometry (deform) on every frame as the figures animate. At least for animation purposes (rather than just moving the camera in a single frame), that means that a lot of work needs to be redone in every frame. If the overhead of this exceeds the time saving for the accelerated display, then it is simply not worthwhile. On the other hand most other apps are not primarily concentrated on figure animation, therefore it may be expected that a higher proportion of mesh geometries will be static (and don't need to be reconstructed) during an animation, even if the actual objects themselves move. Bill

Bill, there is some relevence to what you say about reformatting the data, but at some point, in just about every 3d pipeline I've ever messed with (hardware or software) the data is reduced to a set of polyons and the various material data etc. anyway, so I really don't see it as a problem. In OpenGL, if (or while) the model remains static, you can build display-lists and just render the lists (which are stored on the video hardware) until/unless something changes. So when you're just rotating, translating or scaling the entire model (or moving the camera around, for example), the app simply re-uses the display list each frame. If the model is animating, it either passes in all the data each frame or rebuilds the display-lists each frame... either way should still be faster than software rendering. While there may be some overhead involved while animating, 90%+ of Poser users are just rendering still shots. And even when they are animating, the time you actually have the animation active is pretty small compared to the time you spend setting up poses, and rotating the view around to get clothing conformed, etc. In short, I don't see any disadvantages to offering hardware rendering in preview mode (aside from the time that the programmers need to initially implement it) and there would be HUGE advantages relative to making the application more useful/productive.

Ok, NOW I'm getting depressed. Now that I know it is possible (which I originally thought) and they're just not implementing it (or have not in the past) is really disconcerting. Is there any chance they have done this and have just not mentioned it? Although I would imagine their silence on this thread is a pretty strong indicator. Why would you not want to make the interface useable?? Anyone who does any real Poser work knows that the main window bogs down and becomes practically unuseable when scenes get more than a few hi-res textured items. HOW COULD THEY OVERLOOK THIS?

They most likely have not overlooked this. They have just chosen not to implement it at this time. This could be for many reasons: 1. Limited development staff 2. Unfamiliarity with OpenGL/DirectX 3. Limited time for development 4. Too much re-development needed that would not allow them to piggy-back off of existing code 5. Upcoming (unannounced 3rd party products) that would also require substantial redesign for any changes to the basic Poser app. For whatever reason, it is not in there. It is not something that can be fixed with a patch at a later date. It would require complete redesign of their software, such as would be in Poser 6. At this point, the best thing is just to suck it up, accept it, and move on. Just an observation.

Yep, I'll take what I can get, but they are REALLY behind the curve on this as other software has utilized it for years. Also, I'd agree with you if this were say, Poser 4.5, but this is a entirely new release. I'd gladly, as I'm sure would a LOT of people, wait for them to fix this cripling interface flaw. I mean, a piece of software can have all the bells and whistles in the world, but its worthless if it becomes unbearable to operate. Oh well, I'm still holding out hope that they just forgot to mention it. :) -Tim

Well, just FYI. In another post, Kupa specifically mentioned that this release will not have any form of hardware acceleration.

WAAAAAAAAAAAAAA!!! [tears streaming down cheek] Thanks for the news Ryamka, even though it's bad news. I'm still stunned they would not address this crucial issue. Oh well, I'm sure it won't be the first (but probably the most) dissapointing facet of the new release.

Spanki said: >OpenGL is actually relatively trivial to implement, extremely trivial. Maybe a silly question: Would it be possible to make an openGL renderer in Phython? I mean, if it's trivial, P5 WILL support Phython..... Now that would be something that would ring some (cash registry) bells (Listen to Pink Floyd: Money)! Isn't there some Phython whizz out there (PC Bos maybe)that can do some magic? So I guess it can't be done that way, huh? Or else someone would have already done it.

Well, the problem is we're not talking about rendering of the image. That is a mute point when it comes to 3D accel. really. We're talking about the window (does it have a name? Workspace, Preview Window, or what?) where you do all of your setup and posing. This is where the acceleration is desperately needed.

I guess I should qualify my statement (a bit more). Everything is relative... Implementing OpnGL 'within' the app should be relatively trivial (ie. it's much easier than writing your own 3d pipeline down to the point where you are rasterizing polygons - so, much easier than the code that's already there/written for example). But I don't know enough about Python (or the given interface/implementation) to know if that would be trivial (or even practical). My guess is that due to the amount of interface you need with the rest of the existing application (dials, mouse and such), it would be much better implemented within the app and not as a plugin/add-on.

great.... and im still waiting for them to fix the stupid pro pack.........................