Cool lightwave plugin/previewer !

[T4D]

Hmm i’m alittle confused ?

1 min to render ?
was thinking Pim was a game/3D engine of sorts ?
now it’s trying to be a fprime ?

sorry to point out there are some issues with that pim render, there’s no light fall off on the lights Look at the darker areas.?
gives me the feeling like a game screen grab not a full 3D render ? ? ( to me anyway )
It just doesn’t look like a render you would wait for ?

what PIM is or will be, is still pretty cloudy for some of us ?

any chance of released date or features lists or a demo’s of what you can get out of it ?

[Wout]

Allow me to clarify, PIM is still very much a realtime plugin. However, it has the ability to render out frames at tremendous speeds with 256+ passes. The lighting in this scene comes from one arealight for the sun, one for the overall blue sky,and pointlights with lots of motionblur to create the effect of light bouncing around. Lights do have falloff; linear, inverse and inverse square (all the lights in this scene have inverse square falloff) I just made this scene to show that you can do a lot of things with Pim, realtime graphics aswell as ‘offline rendering’.

[T4D]

THanks I Understand

Do we have a target for it’s release ?

[John_Gonzales]

Not bad!

We rely heavily on FPrime in spite of the improvements of the native renderer. FPrime is an amazing tool and has paid for it’s self many times over but the later iterations of FPrime does seem to have some stability issues with later versions of Lightwave (I have the latest build). When FPrime works it is amazing however when it doesn’t it’s a bloody nightmare as it seems to be a very delicate balancing act between FPrime and Lightwave’s architecture. It is a very bittersweet situation.

I have seen the earlier videos of PIM and it’s impressive indeed. One question are you guys planing any support for the Mac?? That and nodal support in later versions of LW?

Glad to see this something of the magnitude is coming good luck on your development and do keep us posted.

Cheers

[John_Gonzales]

Allow me to clarify, PIM is still very much a realtime plugin. However, it has the ability to render out frames at tremendous speeds with 256+ passes. The lighting in this scene comes from one arealight for the sun, one for the overall blue sky,and pointlights with lots of motionblur to create the effect of light bouncing around. Lights do have falloff; linear, inverse and inverse square (all the lights in this scene have inverse square falloff) I just made this scene to show that you can do a lot of things with Pim, realtime graphics aswell as ‘offline rendering’.

How does this tool work work with Radiosity??

Cheers

[John_Gonzales]

Hmm i’m alittle confused ?

1 min to render ?
was thinking Pim was a game/3D engine of sorts ?
now it’s trying to be a fprime ?

sorry to point out there are some issues with that pim render, there’s no light fall off on the lights Look at the darker areas.?
gives me the feeling like a game screen grab not a full 3D render ? ? ( to me anyway )
It just doesn’t look like a render you would wait for ?

what PIM is or will be, is still pretty cloudy for some of us ?

any chance of released date or features lists or a demo’s of what you can get out of it ?

Funny you should mention this when FPrime first appeared it was marketed as a previewer. However, it exceed expectations especially for us. We now use FPrime as a rendering solution especially where radiosity is concerned. We can now offer radiosity renderings to our clients at a reasonable timeframe so radiosity IS now a realistic venue for us with FPrime. Worely has done a great job although FPrime 3.1 is very robust it still seems to have issues with LW 9 +. We can render out final images and animations but as a previewer it is still lacking in some aspects of stability… We are hopeful that Worely will make good on this. Till then we use FP 1.5 with 8.5 for previewing and or FPrime 3.1 for final rendering in 9. It’s turned into quite the hodge-podge these days.

Cheers

[richcz3]

…any chance of released date or features lists or a demo’s of what you can get out of it ?

…Tune in again around this time next year :twisted:

[Wout]

How does this tool work work with Radiosity??

Cheers

Radiosity hasn’t been implemented yet, though I believe it is on the to do list Just to make sure everyone knows, I’m a Pim beta user and am not actually developing it. Pimtools is Enki’s brainchild

[softdistortion]

I’m a Pim beta user and am not actually developing it. Pimtools is Enki’s brainchild

Probably your avatar throws people off…I also forgot you were a tester…Either way, maybe you can pass on the message to get a push on with opening this up, maybe some info from the developers directly?..PIM’s had amost 60k views in this thread alone…seeing some direct feedback would seem appropriate if it’s possible at all.

Quite a few of us would like more info when we might get to try it out.

[ztreem]

hi !

Just to let you know PIM is still alive =>
Here’s something I’ve been working on for about half a year now(on and off, after work).
241843 polygons, 13lights, all textures were painted.
About 1 minute to render.

Hmmm… In LW 9.2 you can use radiosity and render somthing like that in around 1 minute. But I’m sure it’s alot of things you can do in PIM that LW can not, like all the realtime things.

[Wout]

SoftDistortion=> I told Enki about the thread ‘reawakening’

Hmmm… In LW 9.2 you can use radiosity and render somthing like that in around 1 minute. But I’m sure it’s alot of things you can do in PIM that LW can not, like all the realtime things.

Maybe, but not with 256 AA passes, right ?

[ztreem]

SoftDistortion=> I told Enki about the thread ‘reawakening’

Maybe, but not with 256 AA passes, right ?

In 9.2 all AA is done in one pass so you don’t need 256 passes.

[TrueArt_Support]

In 9.2 all AA is done in one pass so you don’t need 256 passes.

With any AA you have multiple passes… But in latest LW while using ACT they’re internally done, one by one for specifiec screen location… Without caching ray-trace evaluation 256 passes (no matter if they’re done in pre-LW92 way, or internal way) will be 256 times slower than 1 pass… With internal passes it can be little faster when no refractive, no reflective surface is used, and no GI and ray-trace shadows are off, because in that case CPU cache memory is not reloaded with a lot new data between internal passes…

[ztreem]

With any AA you have multiple passes… But in latest LW while using ACT they’re internally done, one by one for specifiec screen location… Without caching ray-trace evaluation 256 passes (no matter if they’re done in pre-LW92 way, or internal way) will be 256 times slower than 1 pass… With internal passes it can be little faster when no refractive, no reflective surface is used, and no GI and ray-trace shadows are off, because in that case CPU cache memory is not reloaded with a lot new data between internal passes…

Ok, that may be true, but Newtek said the AA is done in one pass, if they do passes internal then thats another thing. What ever I don’t think you ever need 256 passes of AA, seems unnecessary to me.

[TrueArt_Support]

Ok, that may be true, but Newtek said the AA is done in one pass, if they do passes internal then thats another thing.

That’s fact… You set these internal passes in Camera Properties window, when using ACT (non-Classical Camera), by adjusting “Antialiasing” parameter… (In pre-LW92 and in Classic Camera it’s drop-down list of predefined values, but in new LW92 it’s animatable integer value).

What ever I don’t think you ever need 256 passes of AA, seems unnecessary to me.

I think the same… More important/useful would be progressive unfinite rendering…

[Wout]

Well it seems to me that 256 AA/mb/dof (all the same ) passes in less than a minute is something usefull. This fprime render of the exact same scene shows some difference (normal map shader doensn’t render with fprime) but it’s about the AA/MB quality. 1 min fprime render; rendered on a PC, AMD Athlon 64 3400+,1gig ram,ati x800.

[ztreem]

Well it seems to me that 256 AA/mb/dof (all the same ) passes in less than a minute is something usefull. This fprime render of the exact same scene shows some difference (normal map shader doensn’t render with fprime) but it’s about the AA/MB quality. 1 min fprime render; rendered on a PC, AMD Athlon 64 3400+,1gig ram,ati x800.

First of all I have not seen or tested your scene so it’s hard to tell if I can render it in 1 min in LW, but by the look of it in your first post the render looked like it would render in under 1 min and it maybe still is in 9.2. LW 9.2 can be alot faster than F-prime when rendering radiosity.
About the AA I think if you just see to antialiasing I don’t think you need 256 passes, if you talking about MB and DoF then it can be useful. In 9.2 the new MB and DoF works a little different then before so you don’t need that many passes for that either, but it can render the old way too. Instead of measuring number of passes I think it’s more interesting to compare quality of renders instead of how many passes it does.

[Wout]

First of all I have not seen or tested your scene so it’s hard to tell if I can render it in 1 min in LW, but by the look of it in your first post the render looked like it would render in under 1 min and it maybe still is in 9.2. LW 9.2 can be alot faster than F-prime when rendering radiosity.
About the AA I think if you just see to antialiasing I don’t think you need 256 passes, if you talking about MB and DoF then it can be useful. In 9.2 the new MB and DoF works a little different then before so you don’t need that many passes for that either, but it can render the old way too. Instead of measuring number of passes I think it’s more interesting to compare quality of renders instead of how many passes it does.

Okay, just so we understand eachother; there’s no radiosity in this scene, the bounced light effect was created by pointlights moving around subframe combined with 256 passes of motion blur. And about comparing the quality of render, aren’t the screenshots obvious enough ? I have to say I haven’t tried LW9 yet, but are you honestly saying that a LW9 render with radiosity and 35 passes(if that is still the maximum AA value, in other words as close as you’ll get to 256) at 1280*960 will render under one minute ?

[ztreem]

Okay, just so we understand eachother; there’s no radiosity in this scene, the bounced light effect was created by pointlights moving around subframe combined with 256 passes of motion blur. And about comparing the quality of render, aren’t the screenshot obviouss enough ? I have to say I haven’t tried LW9 yet, but are you honestly saying that a LW9 render with radiosity and 35 passes(if that is still the maximum AA value, in other words as close as you’ll get to 256) at 1280*960 will render under one minute ?

I know you don’t use radiosity, I was just comparing the quality of your render and think it can be matched by a render with radiosity in 9.2 in about a minute, that’s all(I can be wrong). I don’t think it will render under a minute with 256 AA passes and so on, so in that case PiM is amazingly fast. LW 9.2 supports how many AA passes as you want, but render something with 256 passes would take forever, so you’re right that’s quite fast… but in the end it’s the result that counts not how many passes you used.

[TrueArt_Support]

Wout, you don’t have even LW9, not to mention LW9.2?!

There is very simple way to check this- share you scene with us, and we will render it under LW v9.0, LW v9.2, K-Ray v1.7 and F-Prime v3.01…