In the one PaperTiger image I clicked on, it said toolchain yada,yada and Luxrender. I'm sure I missed the Octane part but no matter. This is getting beyond subjective and into what might or  migh not have been done in postwork or what it might have looked like with xyz lighting etc. Short of someone working to produce two images as identical as possible, using the best techniques for each tool … and even then, it would probably answer little  i.e. people are probably going to end up pretty much where they started

I do think that the emphasis on Poser models is appropriate - we are talking about rendering Poser scenes no?  All the hard edged arch/viz goodness that any of these renderers can produce doesn't matter if you want to do images where people are a prominent part. In that regard, I haven't seen anything from Octane, Lux etc. that is better than the best Poser images, even going back to what Catharina was doing in Poser 6 or so. The final girl on the couch Octane render is very nice and for the speed, price, esthetic equation if it works for you great. I just don't personally, subjectively see anything superior there on the dimension of (human) realism. What any of these might be if the artist stood on his left foot facing east, I don't know. As always, use what works for you.

Quote - I do think that the emphasis on Poser models is appropriate - we are talking about rendering Poser scenes no?

Sigh.  We should be talking about rendering.  Period.  This was a thread about ALTERNATIVE rendering to Poser firefly.  According to some here, there are no render engines out there that can do better human rendering than Poser's Firefly.  My argument here is that the technology behind engines such as Octane IS far superior for realistic lighting results than Firefly, so why shouldn't the results be better, regardless if you're rendering a Vicky gen 10, or the temple she is naked in.

I could achieve photorealistic results in Poser firefly too, and even pass them off as having been done in say Vray, and very few people would ever doubt that claim.  I know it's possible to get high end results from Firefly, but there are areas where it's impossible to get consistent, predictable results all the time.    I know too, however, that the engine itself, in the end, isn't as robust as something like Vray, MentalRay, or in this case, Octane... and I'd never argue that it is.  It's like saying Hexagon is as good as Maya, just because a few people have modeled some incredibly realistic human characters with it that could have passed for a model made in Maya.

The image I posted above was rendered in Firefly 2012 early last year, not too long after it's release.  I used a single emitter (no actual lights in the scene).  EZskin for the skin.  While the results were very impressive, I know from experience and countless renders that those results are not consistent.  If I changed my skin textures, that render didn't look as realistic, and I needed to make adjustments to the skin shader, etc.  Also, if I wanted to animate the character and render it with that level of realism, it was impossible.  Why?  Well, because Firefly isn't a physically-accurate render engine, and it's IDL etc. isn't robust enough.  There is all sorts of issues to contend with when attempting photorealistic results in Firefly.  The materials, no matter how good, only get you so far.

Anyway, I can't sit here and argue that this pic isn't as good as that one, or that render isn't as realistic as this one, etc.; because it's a particular pic, created by a particular artist.  I can, however,  argue that there is NO reason anyone could not get better results in Octane using Poser figures, because the render engine IS technologically and realistically more robust, with far more options, with a more physically accurate rendering algorithm.

The results I got from Poser 2012 were fantastic, but only after hours of tweaking, test rendering, etc.  In the end, however, there were many limitations to the engine;  there was no way to escape the IDL from object emitters from looking blotchy on white or lighter flat surfaces, and some artifacting in very detailed areas of larger scenes, etc., no matter how you tweaked the somewhat limited render options, and there was no way to escape flicker in animation, or grainy DOF without smoothing it in photoshop, etc.  All these things are facts.  However, I COULD get a completely artifact free GI in Octane, with instant feedback of my scene lighting, and I could get silky smooth DOF, with control over exposure IN the engine, and I could get REAL physically accurate caustics to reflect on the human skin from the outdoor pool scene.   So I could go on and point out the shortcomings of Poser's Firefly, and just because you haven't seen a render you like better than one you saw in PP2012 by a particular artist, doesn't mean the engine can't do a better job, in less or equal time.

I'm not saying it isn't possible - don't think I ever said that. I've seen superior results from Max, C4D etc., Vicky Yeoh's young grl in Max, the silent man in C4D … Most of the human renders I've seen in any program fall short of that leve - IMO. Again, in my subjective opinion, I haven't seen it in the Octane renders either. That is not the same thing as saying it doesn't exist or isn't possible. I hope that's clear.

As for the physical accurace thing, I was never on that bandwagon, even back when I first drooled over nuts and bolts rendered in Lightwave. My only interest is in whether it looks 'right' to me, not whether it meets IEEE specs. I know other people hive different views. Bottom line, again IMO, Max, Maya, C4D, Kerkythea, Lux, Octane et al, can do stunningly realistic inorganics. Skin and hair are tougher challenges for any render engine. I respect your skills and those of anyone who can play in that pool tremendously. All I have is my ageing eyes and personal taste. If any particular program gives you what you feel are superior results, especially in a much shorter time, you've hit paydirt. As I said earlier, Octane is probably more than adequate for what I would want to do, but that's usually not photorealistic portarit type stuff. If I were much more technicaly lettered, I suppose I could look at the technical specs and say yeah, that's the right stuff, but I'm not. Again, I can only judge by what I see. It may be more of an indictment of my vision or taste, it certainly isn't a knock on any program or artist.

I'd love to try out octane, but I'm not ready to invest into a graphic card with the chance I end up with images that a lot of octane user seem to end up with..... they're not so good (putting it midly now). Yes, there are some nice renders coming from octane, but I also see a majority of images that are flawed. I'm not willing to take the gamble. The problem is that even though octane may be as superior as you say it is, lot's of users can't seem to get those results, contrary as to what you seem to tell us.... the poser-octane related images show otherwise and some of them aren't even rendered in octane at all.

I'm having a completely different experience with firefly, I've developed a certain look for my promo images and by now I can achieve those in just a matter of minutes, no matter what textures I do use, even figures or props don't influence that at all. I know there are quite a number of others who can put out the same result time after time. That is not said that firefly doesn't have shortcoming, every render engine does have them and you need to learn to work with them or find something that's better for you. To each his own. Personally I see too many flaws to invest heavily into the hardware and software needed to even go for something like octane, it's not like it's a few dollars to try, it's quite an investment with an open end.

So, I'll be on the fence for a while longer.

Can I put in a word for Hybrid renderers?

They may well be a transition product in the history of rendering but I think most of us with older hardware can use one without upgrading our GPUs.

I'm thinking of something like Indigo (another commercial unbiased renderer).

The advantages of keeping the data in main memory make it worthwhile sacrificing some of that super, ultra-fast 'GPU only' speed. 

It solves your model size and texture size limitations and you still get a considerable speed boost from the GPU.

A quick, very unscientific comparison with a simple test scene and the lighting set up sort of, vaguely, approximately similar. :rolleyes:

Luxrender (via Luxus from DAZ Studio) and Indigo (from C4d). Win7-64, 8GB, an old AMD X6 and an old nvidia 550ti.

(I understand Luxrender's implementation of hybrid mode is not fully working so the Luxrender result is not a fair indication of its performance. Its just here as a baseline.)

I got 200K S/s in Luxrender (hybrid)
Indigo gets up to 2.1M S/s in hybrid mode.

...if you've already taken your stuff into something like C4D or Max you can try the Indigo demo.

(The only actual empirical data here is that Indigo doubled its render speed with CUDA activated.)

Hmmmm, Indigo also supports openCL, which means it can use my Radeon graphic card.

I'll have to give the demo a shot...

(Thanks for the tip!)

Hybrid could be very interesting indeed. I'm assuming it can do huge render sizes as well, but is a lot faster in rendering. It should also deliver the same quality as gpu rendering gives and do away with the flaws pure gpu rendering does have.

What flaws do pure gpu renderers have?

 A more accurate term is 'limitations'. Mainly hardware: your scene -has- to fit into the memory on the video card. System memory isn't relevant there. And you are limited to the texture slots that the GPU supports. a 600 series NVIDIA chip, for example, supports up to 1,344 texture slots.....and before thinking that there is not issue, keep in mind that anything that presents as an image is considered at 'texture'. So if you have bump maps for all your textures, that cuts the available 'slots' down to 672. A bump and a displacement map, it goes to 448 'slots'. Toss in a normal map, and you get '336'. Toss an alpha mask on top of those, and you have all of 268 'slots', assuming each actual texture requires all 4 other 'textures'. And that is the total for your entire scene.

That said, NVIDIA has just fielded a rack mount GPU rendering system for the pro's ($40k is a professional level toy....but it gets you lots of gigabytes of ram to play in and 32 dedicated GPU's), and Otoy, the maker of Octane, is working on a cloud based service for desktop users to pay for to get a lot more distributed render power.  

Ah yes, the texture limitation, that is basically the thing putting me off trying Octane.

But I believe that the systems that use openCL have no such restrictions.

Or have I missed something?

The programming language won't make a difference. Its the hardware limits you're up against.

nVidia were dropping broad hints last year that they had found a way to use the huge bandwidth of PCIe3 to get round this. But I've heard of nothing in the recently released CUDA update addressing this.

If I was cynical I would say they don't want you buying cheap 1GB cards when you could be spending a lot more on 4GB cards. :sneaky:

Ah yes, the texture limitation, that is basically the thing putting me off trying Octane.

The texture limit is a factor for the pre GTX 600 cards, but the 600 and titan cards offer a lot of textures and it's only something you need to manage on very large scenes (for example, Stonemason's city scenes).  The plugin comes with a selection of tools to reduce both the number of texturemaps used and the size of the texturemaps (not really an issue with the latest cards).  For example, you can remove bump maps from a figure, prop, material or all bumps from the whole scene.  In general bumpmap numbers are the most important, and they are the easiest to delete with the plugin.

Paul

Click To enlarge.

OK - here is one.  Render time was about 10 mins on my PC, but would be < 1 mn on a Titan.  Rendered inside Poser.  I used the standard SSS shader, but did include the normal map that came with the figure.  Geometry was only 200k polys.  He needs some eyelashes :-) 

200,000 polys? There's a way lower res version of that head floating around the net. You should try that next time.

I wonder if you can use octane render with low quality settings to make the render look less realistic, similar to firefly render quality but with a much faster render time?

Firefly Render of quality X = 45 minutes

therefore .....**
**

Octane Render of same quality X = 15 minutes

... thus Octane render shaving off 30 minutes of render time...

So the quality of the renders are the same except octane just renders faster. Anyone know? If I could cut my render time in half but keep the less realistic look of Poser I would buy Octane in a heartbeat.

Firefly Render of quality X = 45 minutes

therefore .....
Octane Render of same quality X = 15 minutes... thus Octane render shaving off 30 minutes of render time...

Why not try the plugin demo and see how fast it renders?  Speed it quite dependent on your video card.  Get a Titan, and the speed is staggering (but so is the cost)!

Quote - Firefly Render of quality X = 45 minutestherefore .....
Octane Render of same quality X = 15 minutes... thus Octane render shaving off 30 minutes of render time...

Why not try the plugin demo and see how fast it renders?  Speed it quite dependent on your video card.  Get a Titan, and the speed is staggering (but so is the cost)!

Man, soooo tempting! I definitely have to try the demo, it's been on my todo list for like a month.

Just so ya know how hollywood makes there realism.
On the biggest PC's $$$ could buy.With a crew of experienced CGI Artist.
They model ,rig ,weight there own charaters in highend apps.polycount 20,000.
Then send to zBrush. vextor map = polycount 4 million.
If they needed a forest they would use vue as a plug in there highend apps.
They would render the movie on a very big renderfarm 100's of PC's.
Would have took 3000 years to render transformers on a home PC.

Since we don't have million dollar budgets.
We're not going to = HollyWood realism.
But we can still make killer video's.

check list.
App's tell you what they can do.
they never say what they can't do.

1.HighEnd app's of your choice that can do every thing you need to do.
HighEnd app's that suports the plugs you need.Poser,if you need a forest will it use Vue or a app like Vue ? if needed suports Renderfarms ?

If you want more.
2.zBrush rules for the ultimate vextore map textures.
takes skill and a Cintiq wacom tablet or Intuos wacom tablet.

Quote - So, let's move away from theory and hype and show some reality.

OK - here is one.  Render time was about 10 mins on my PC, but would be < 1 mn on a Titan.  Rendered inside Poser.  I used the standard SSS shader, but did include the normal map that came with the figure.  Geometry was only 200k polys.  He needs some eyelashes :-) 

This is a joke right? Now you're doing exactly the same again. You're telling us you're moving away from theory and hype and yet you throw out theory and hype again. OK, 10 minutes in poser (and the image is looking way better quality from what I've seen the gpu renders do and conjunction with poser), so that's reality. But would be less then <1min on a titan that is pure theory again. Stop comparing the real poser times to theoritically gpu times and start doing some serious timing instead. The titan isn't out at all, we don't know it's costs, all we know is that it may be fast and will be very expensive, you're now again hyping!

So, you're still hyping and being theoritically, you now even tell is what non released gpu's do..... that's pure hype man! Start showing some real numbers with gpu's we can get now or just admit they're not as fast as you claim them to be IF you want the same quality as this poser render you show. Because that's what it's all about same quality and for the most times the gpu rendered images have shown less quality.

The 10 minutes was for the Octane render (and that is the produced image). If I remember correctly, he uses a GTX 570 which is an older card. The titan - which IS released and available - would take less than a minute.

There is no easy way to compare render times. Octane - like Lux - keeps rendering. You have to decide when it is finished. Each pass it increases the samples and gets more accurate. Sometimes 100 samples is enough and sometimes you need 12000 samples. Some people will be satisfied with a 500 sample image and others will go up to a minimum of 6000 samples - even if they don't see the difference. Low light conditions tend to need more samples as bright sunlit images.

Poser's render times are also variable. Low lights conditions also need a higher IDL quality setting which dramatically increases render time.

Another complicating factor is that there are 3 render engines: Direct Light (biassed and very fast), Path tracing (like IDL) and PMC (faster with specular materials such as glass and water). And you have to take the graphics card(s) into account. A GTX590 is the fastest midrange card and is at lot faster as a low range GTX4xx card which has fewer GPU cores and runs at a lower clockspeed. Same is true for Firefly. A 12 core machine will render a lot faster as a dual core machine.

So comparision between Firefly and Octane in terms of render time are only valid for one machine setup. A machine with a low end graphics card and a high end CPU configuration might wel be faster rendering an image with firefly as with Octane

I will post a comparision - which is only valid for my machine - a little later when the renders are finished.

Excellent, looking forward to that. I know they figures will differ from machine to machine, but it will still give an impression.

First image is the straight Firefly Render. IDL, 3 RT bounces, 3 pixel samples, Min Shading rate at 0.2. This is on an i7 990X, 12core, 24GB memory

Render time 38min 52 seconds

All images have been reduced to 1200x800 JPG to make them smaller

Same machine. Graphics card is a GTX680 with 3GB VRAM

Rendertime for this image 23m34sec

This one was at 1000 samples, I could have stopped at 700 samples giving pretty much the same result which would make the render time about 15min

This is after rendering 1 minute (170 samples if I recall correctly). It uses a different render kernel which is less accurate but much faster.

Render time: 1m:0sec

If you want to use this render kernel, you would adjust the light a bit

To be complete:

This is exactly the same scene. Original dimensions were 1670x1108, 32bit color (for octane). I used a preset for the skins which is pretty much identical to EXSkin and adjusted some of the metal on the jewelry. No other changes in the scene - it imported directly as is. The lights I used in the Octane render where the lights in the mesh which were in the ceiling. In poser they were a bunch of spotlights. No other lights in the scene

Thanks. The last one is indeed very fast, but it shows in quality, it's imo pretty bad. The first octance render has good quality and renders in half of the time, that's good. The poser image somehow looks wrong, it's not the quality poser can deliver, somehow the light seems to be off and the image isn't sharp at all. But that can be changed and improved without increasing the render times, not a big deal.

If one has the hardware already and your willing to pay for the licenses, octane could be a good option. If, like me you don't have the hardware, adding a second computer to render is an option as well and may even be cheaper and will cut the render times easily in half as well. For now, I'm not going to run out and spent lots of money on hardware and the octane license, I'm going to what a little while longer and see what reality and lux updates will bring.

I could have improved the quality for the Poser render, but that would have increased the render time dramatically. The lights could have more intensity - but then I would have to adapt the specular settings as well. It is a careful balancing act and takes a lot of time. Sharpness is a question of turning on a filter and is a matter of preference. Sharp images are pretty unrealistic and a dead give away for CG images.

The DL render is something which I did to show off the speed. I mentioned that something had to be done about the lighting - it is a different render technique requiring different lights. I think "pretty bad" is a gross exaggeration.

Adding a second computer will not make one bit of difference in render time since each image is still rendered on one computer. In contrast, you can add a video card in your PC and your render time will be cut in half in Octane.

Reality with Lux will probably be slower as firefly since Lux is unbiased and uses the CPU for rendering. The GPU render facility has the same limitations as Octane

But I did not expect at all that you would change your mind. You never do. I just wanted give other people examples of what Octane actually does and correct the some of the information which was given.

I really hate to break it to people.

There is no way that the Titan is 10 times faster than previous generation cards.

It can do 4.5 teraflops only in thoery.... Thats where the problem lies....

To achieve 4.5 teraflops, every core must be saturated and doing something every step, and not waiting on another core for information, etc... Which doesn't happen in the real world. Some render engines show GPU usage/saturation, and it is normally never maxed out saturation wise.
Nvida also said the Titan is only 3 times faster than a GF110. And only at one thing.

Double precision is not Nvidias strong point.

Will the Titan be fast? Yes it will be faster than previous Nvidia cards.

But it is slow compared to an AMD FirePro S10000, which does all but 6 teraflops in theory.

You can't beat the price of a Titan card, that's one thing Nvidia has going for it. The problem is finding one instock somewhere...

I'm not tech savvy ,but half to say todays PC's n 3D cards are alot faster then the
windows 95 32 bit 32 Ram PC's ,with the 4 RAM 3D cards.

I meant 500 series Fermi and 600 Kepler when I was refering to previous generation cards.

If you read the test reports about the Titan, they state that the GTX 690 is still faster then the Titan is, about 10%, so all the miracles people expect of the Titan will not hapen at all..... that's why I already called it hype. So in speed the Titan will never do these renders in the claimed time in this thread, but it will have more memory.

Now, if you'r like me and do not use poser as a main rendering engine at all, for the price of one titan of gtx 690, I can buy 2 fast i7 machines. Adding that to the one I've got already, I will have 3 machines to render and with Vue, they can alll work on the same scene..... let's see if any titan can beat that.

There is no way that the Titan is 10 times faster than previous generation cards.

You are right - it's only 7 times quicker than my 550Ti (Octane benchmark test scene).  Still pretty impressive.   (550Ti = 1.1millions samples/sec, Tital = 7.51million samples/sec).

If you read the test reports about the Titan, they state that the GTX 690 is still faster then the Titan is, about 10%, so all the miracles people expect of the Titan will not hapen at all

Octane renders fasteer on the Titan than the GTX 690 (which does 6.1million samples/sec).  The Titan also uses around 1/2 the power of the 690 and is substantially quieter.

Now, if you'r like me and do not use poser as a main rendering engine at all, for the price of one titan of gtx 690, I can buy 2 fast i7 machines.

I don't believe a) you can get 2 x i& PCs for that price (add in OS, second copy of Poser, etc, pus the added power usage), and b) 1 x Titan is going to substantially outperform 2 x PC's.  My i7 PC renders in Poser slower than my GTX 550Ti, so even 2 x PC's will be far slower than a single Titan.

Plus, power users are putting 2, 3 or 4 Titans in the one PC, delivering the render speed of a room full of PC's.

Paul

Actually you can buy 2x i7's for that amount, I just did. For the price of 1 titan, I got 2x i7 3770, 8GB Ram, 500Gb, video card, Wireless and W7. No need for extra poser copies since it can network render. I'm mainly using it for Vue. So now I've got 3 fast pc's linked and i'm sure those render as fast as...... It's not 2 pc's against the Titan, it's 3 since the Titan needs a pc as well to run, so it's fair to add that one to the mix as well :-)

I can even throw a few copies of poser 9 (on sale) into the mix if needed, let's not forget that you need the very expensive octance license..... but since I don't need to I can spent that money on other stuff.

That just seems like the most incredible waste of space, resources and electricty :-

Space yes, electricity I don't think it differs much, gpu rendering uses a lot of power. Resources, no, all the things is built to do is render. Of course you can get a different cpu, a workstation with 8 cores / 16 threads or more that would be less space consuming. The thing is, for me it's a trade off. I render with Poser and Vue. Vue is purely cpu, poser can do both now..... but, I render for print, so huge renders and gpu rendering will fail there. So, for me (and many others), building our own small rendering farm where the focus lies on the cpu is the best option.

If I would only use poser, it would perhaps be a whole different story, but still the render size would be a major issue.

Quote - Space yes, electricity I don't think it differs much, gpu rendering uses a lot of power. Resources, no, all the things is built to do is render. Of course you can get a different cpu, a workstation with 8 cores / 16 threads or more that would be less space consuming. The thing is, for me it's a trade off. I render with Poser and Vue. Vue is purely cpu, poser can do both now..... but, I render for print, so huge renders and gpu rendering will fail there. So, for me (and many others), building our own small rendering farm where the focus lies on the cpu is the best option.

If I would only use poser, it would perhaps be a whole different story, but still the render size would be a major issue.

Poser does GPU rendering?  Am I missing something?

All this talk about Octane vs. Vue or any other renderer for that matter is a load of old BS if you ask me. Are there bad renders being done with Octane? Of course. Are there bad renders being done with Poser and Vue? Oh yes!  Getting a new renderer doesn't automatically make your images better. If you're bad at lighting and posing with Poser, Octane, Reality or Vue won't fix that problem for you. I bought Octane after trying out the demo. I like how fast it renders and how it handles lighting and materials.  That being said I still have a lot to learn to get the most out of it.   It also suit me personally since I have a very old system but a new GTX660 graphics card (bought for computer gaming).

News to me... Only gpu assisted for preview as far as I know

Quote - News to me... Only gpu assisted for preview as far as I know

That's what I thought and the OpenGL preview is just that, a preview.   But why would GPU renderers fail at rendering for print?  I fail to see how you can fail to print renders based on the renderer it comes from.

Renders for print generally require high dpi/resolution, perhaps not so good for gpu rendering which is more limited by available RAM?

Quote - Renders for print generally require high dpi/resolution, perhaps not so good for gpu rendering which is more limited by available RAM?

I can render an image that's 3200x2400 px.  That gives me a 10x8 Inch image at 300 dpi.  At 200 dpi it's 16x12 inches.  I haven't tried to render any bigger but don't know of any limitation in that regard.

Quote - Poser does GPU rendering?  Am I missing something?

Depends on where you draw the line between rendering and preview. In Poser 9, you can do an OpenGL preview render and get procedural shaders, ambient occlusion, IBL, gamma correction, shadow maps and motion blur, most of which are things that the Poser 4 renderer was unable to do.

Hello!

I would like to buy octane but my hardware is two years old (Intel i7960 with 12 GO of RAM and a quatro 4000 (256 cores))

I tried the demo plug for poser (pro 2012) , it works well, but I miss some textures in any case and don't render if I have too many élements in the scene.

I can render simple scène (I don't ask more, I'll postwork) but at the size of the demo.

Can't change my hardware now, and I wonder what will happen with larger renders. I've read that the registred plug is more powerfull.

Is it true ? Will my configuration be able to handle the software ?

I think the quadro 4000 is the limiting factor here, not so much in terms of speed, but the video chip architecture. It is a Fermi which is limited to 64 RGB and 32Grayscale textures. The 2GB VRAM should be OK for most purposes.

Once you run out of texture slots there will be textures missing in the render.

As far as I know, there is no difference between the demo and the full version in this respect

Spectacular!

Quote - I just uploaded this one to my gallery.  Poser to Octane.  55min render in Pathtracing mode on my GTX 560, which is lower end of Nvidia cards.  I stopped the Poser Firefly comparison midway through the Indirect Lighting, because although I prob could have eventually come close to matching the quality in Firefly with very high settings, the IDL calculation alone was very slow on the hair, and after half an hour, I felt there was no way it was going to match the render time/quality.  It still had to calc the SSS, and do the render.

Click To enlarge.

Maybe you just need to upgrade your video card?

Quote - Hello!

I would like to buy octane but my hardware is two years old (Intel i7960 with 12 GO of RAM and a quatro 4000 (256 cores))

I tried the demo plug for poser (pro 2012) , it works well, but I miss some textures in any case and don't render if I have too many élements in the scene.

I can render simple scène (I don't ask more, I'll postwork) but at the size of the demo.

Can't change my hardware now, and I wonder what will happen with larger renders. I've read that the registred plug is more powerfull.

Is it true ? Will my configuration be able to handle the software ?

I tried the demo plug for poser (pro 2012) , it works well, but I miss some textures in any case and don't render if I have too many élements in the scene.I can render simple scène (I don't ask more, I'll postwork) but at the size of the demo.Can't change my hardware now, and I wonder what will happen with larger renders. I've read that the registred plug is more powerfull.Is it true ? Will my configuration be able to handle the software ?

The demo version of the Poser plugin is many versions behind the current commercial release, but if it works on the demo, that's a good sign.  The size of the render makes no difference (other than speed).  Whilst there are many users of the plugin on the Quatro 4000, there have been a few users who had trouble with that card due to it's age and needed to upgrade.  I would personally get a low-end 600 series card (GTX 650 or 660) for around $200 to remove any doubt.  The 600 series card will also double your textuemap capacity.

Paul

Hello!

thank your for your answers

I'll consider to buy a new graphic card and the Ge Force GTX 660 Ti is in my budget.

Thank again.

Quote - Hello!

thank your for your answers

I'll consider to buy a new graphic card and the Ge Force GTX 660 Ti is in my budget.

Thank again.

I use the GTX 660 Ti and I'm happy with it. Both for Octane and gaming ;)