New Render Pass Plugin!

[gaetan]

Wow! look nice! :thumbsup:
http://www.lwpassport.com/

[RedBull]

Is it a C based plugin or Lscript?

[Phil]

LScript, I think, given Jezza's iDof plugin.

[RedBull]

LScript, I think, given Jezza's iDof plugin.

Thanks considering the amount of scripts breaking over 9.x, and the price given not being a C based plugin, it turns me off a little actually.

I may have to give it a miss for the moment, but LW should already have something like it, i guess I'll wait for a NT developed C plugin.
Thanks for the info.

[jeremyhardin]

Interesting take RedBull,

Considering this was written during 9.x, it's unlikely that this will be breaking for quite a while. And I've written around frequently changing areas of LightWave and Lscript. It's using 'safer' methods, not the newest/latest/greated/volatile.

The price is based on functionality and time spent, not the language used to create it.

But if you can live without render passes for the day(s)/month(s)/year(s) until NT implement it, then I'd understand you're lack of interest. I personally can't do that, and have no intention of creating copies of my objects for every custom data pass I create, then creating scene copies that I have to keep updated with the lastest animation/camera/lighting information.

[trick]

...it's unlikely that this will be breaking for quite a while...

I have been waiting ages for an update on HD_Instance. The same for TequilaScream. Surpasses, which had the same functionality as Passport and cost me FAR less, is likely to NEVER appear again. If you have spent that many time for creating this plugin, I'm afraid that when NT releases a new version, you will need to spent that time again and I will have to wait again. So forgive me that the last thing I will do is buy new plugins for a long while until proven otherwise...

[Lightwolf]

I have been waiting ages for an update on HD_Instance. The same for TequilaScream. Surpasses, which had the same functionality as Passport and cost me FAR less, is likely to NEVER appear again. If you have spent that many time for creating this plugin, I'm afraid that when NT releases a new version, you will need to spent that time again and I will have to wait again. So forgive me that the last thing I will do is buy new plugins for a long while until proven otherwise...
It's a chicken and egg thing, isn't it? I fully understand your point. As a third party (puts on third party hat) you rely on people actually purchasing your stuff otherwise there is no point in on-going development (unless you do it for fun in your spare time). Which is why Surpasses died... little interest and the developer moved on to greener pastures.

On the other hand... Jeremy has a good track record with iDof, so that should tell you something as well.

Cheers,
Mike

[archijam]

I have been waiting ages for an update on HD_Instance.

Interesting that you mention that, his last post was from siggraph, when he said we was going to hang out at the Autodesk stand (?) .. very criptic.

Is HD Instance working fine with 9.3? I had not heard anything to the contrary ..

Sounds like a great plugin Jeremy. :)

j.

[trick]

Interesting that you mention that, his last post was from siggraph, when he said we was going to hang out at the Autodesk stand (?) .. very criptic...

Well, since I already use Max with VRay, fRender and Maxwell, I don't see what he could offer there, since all the render layer and instance functionality that I could desire is in one of these combinations. It would be really hard to develop a FAST volumetric instancing engine in combination with VRay GI !! I still prefer LW (still 7.5 with HDInstance and Surpasses ;)) for certain tasks and wish some more solid and integrated developments in these areas could be concentrated on.

[Snosrap]

Thanks considering the amount of scripts breaking over 9.x, and the price given not being a C based plugin, it turns me off a little actually.



You might just as well dump LW then, the whole darn thing is nothing but a bolted on plugin. .ls .lsc .p who cares. If it works - it works.

Cheers
Snosrap

[duke]

It instills a great sense of confidence to see all the demo videos done with a ground plane, 3 lights and a few primitives. We once bought Surpasses and tried using it on a few large scenes and it crapped it pants every time.

[Titus]

Watching PassPort working makes me think that a render pass plugin seems a very boring task to program but at the same time not really daunting. I don't get why this "basic" feature is so difficult to achieve?

[mike_b]

I dont get why this plugin hasnt come out sooner:)
It looks like an excellent time saver, and adresses a serious lack in LW's setup. So I am very keen to get my hands on it, its going to save me a lot of time, and I suspect will pay for itself very quickly.
There are a lot of great free plugins out there, but you cant have every plugin for free

[jeremyhardin]

It instills a great sense of confidence to see all the demo videos done with a ground plane, 3 lights and a few primitives. We once bought Surpasses and tried using it on a few large scenes and it crapped it pants every time.

It instills a great deal of incentive to put more time into projects like this when demos are met with sarcasm and comparisons to a completely different product. :D ;)

Seriously though duke, here are some things to keep in mind with this...

1. This is not surpasses, and tries to avoid layout altogether for interaction. This is a render layer plugin, so the results are present in the render. This frees it resource-wise quite a bit (until you go to render). That's not to say it isn't slower on larger scenes. So long as PassPort exists in LW, it'll act the same way LW does with large scenes. :D
2. The reason for using smaller content with the demo is for speed. Let's face it, large scenes render slowly, and I didn't want to crossface or leave the video running for an hour to show a large scene rendering a single frame. You can see a tester comment about a large production scene here (http://forums.cgsociety.org/showpost.php?p=4695052&postcount=53) if you like. And I encourage you to try the demo when it's released if you need this type of thing.



For users that don't need this type of thing, that's fine. I'm not evangelizing a workflow. I'm merely providing a tool for users that have a similar workflow to mine.

[jeremyhardin]

Watching PassPort working makes me think that a render pass plugin seems a very boring task to program but at the same time not really daunting. I don't get why this "basic" feature is so difficult to achieve?

I would beg to differ as to the 'not that daunting' part of your comment. If it's so easy, why aren't there so many 3rd party options for this right now?

(Ok, partially because a lot of LW users don't work this way and don't see the value. But still. ;) :thumbsup: )

[animotion]

Funny, once I found out that you wrote it, it gave me the urge to want to buy it.
Good work man. :thumbsup:

[jeremyhardin]

In case anyone missed this, PassPort (http://www.lwpassport.com) is for sale now at the introductory price (for a limited time). A 5-day fully functional demo is up too.

[Matt]

Which is why Surpasses died... little interest and the developer moved on to greener pastures.

I think Surpasses also suffered from being just too complex, from what I remember of watching the demo videos it looked like it did almost everything, but it seemed quite convoluted to use.

[Lightwolf]

I think Surpasses also suffered from being just too complex, from what I remember of watching the demo videos it looked like it did almost everything, but it seemed quite convoluted to use.
Complex yes, convoluted no. Layers are complexto handle properly to start with, as are passes. Passport itself is quite complex as well, but (a lot) less convoluted than doing it manually ;)

Cheers,
Mike

[Paul Brunson]

Jeremy, excited for the plugin. I assume the introductory price is $299. Just curious, what will the price go up to after introduction? (trying to figure out how urgent it is for me to scrounge up the $299 :) )

I'm very interested in multipass rendering using Mike's EXR trader (buffers) and now Passport for passes. If you don't mind I'm curious about your multipass workflow. Lightwave has thrown me a couple curve balls that I've found messy workarounds for, but hope for something better.

1) Shadows in the diffuse buffer (also beauty pass buffer), this is driving me crazy, I'd love control over shadows in the composite. But having them burned into the diffuse buffer makes it difficult.

2) Lightwave light color(s) not taken into account in the raw RGB buffer. I can always use the beauty pass which does include the coloring from lights. But then I give up what little control I have over the shadows.
(I've found I can manipulate the shadows via adding a inverted shadow buffer back into the shadows. But this only works if I am using the diffuse buffer, if I use the beauty pass instead then the technique falls apart.)

3) Coverage buffer, all my multipass composites loose their anti-aliasing and I have to look to fake it with edge blurring. Know of any better way? I've seen Mike comment on this. I wish newtek would open the SDK so Mike could include coverage with EXR Trader.

I'm assuming passport at least gives me the option now to easily setup a "shadowless" render pass that I could then combine with a second pass that includes shadows.

But I'm curious if you have any insights you could give me from your experience?

[jeremyhardin]

Jeremy, excited for the plugin. I assume the introductory price is $299. Just curious, what will the price go up to after introduction? (trying to figure out how urgent it is for me to scrounge up the $299 :) )
Heya Paul,
Introductory price was $175. $299 is the normal price, and it shouldn't be going up any time soon.

I'm very interested in multipass rendering using Mike's EXR trader (buffers) and now Passport for passes. If you don't mind I'm curious about your multipass workflow. Lightwave has thrown me a couple curve balls that I've found messy workarounds for, but hope for something better.
Hmm, this is a complex question, and one I'll try to answer soon. Reason being because it varies so much from project to project, and between my 'work' pipeline and my 'freelance' pipeline.

1) Shadows in the diffuse buffer (also beauty pass buffer), this is driving me crazy, I'd love control over shadows in the composite. But having them burned into the diffuse buffer makes it difficult.
The age old problem! You can disable raytracing per pass or control shadows per object per pass with overrides, but I can't fix LW's buffer I'm afraid.

2) Lightwave light color(s) not taken into account in the raw RGB buffer. I can always use the beauty pass which does include the coloring from lights. But then I give up what little control I have over the shadows.
(I've found I can manipulate the shadows via adding a inverted shadow buffer back into the shadows. But this only works if I am using the diffuse buffer, if I use the beauty pass instead then the technique falls apart.)
For this I personally would rather have the raw rgb buffer without light contribution, then do a light colour pass (by overriding the surfaces to white for that pass). As well, I sometimes override my lights to be red/green/and blue for key/fill/rim so that I can raise and lower light contribution amount in post.

3) Coverage buffer, all my multipass composites loose their anti-aliasing and I have to look to fake it with edge blurring. Know of any better way? I've seen Mike comment on this. I wish newtek would open the SDK so Mike could include coverage with EXR Trader.
Again, I can't fix the lack of this buffer, but what I did include is a resolution multiplier override. So while you don't have coverage, you can render your data passes at double resolution and scale down in post. Your edges are a bit better, and data is still preserved.

I'm assuming passport at least gives me the option now to easily setup a "shadowless" render pass that I could then combine with a second pass that includes shadows.
Yes, shadows can be controlled per-pass by changing ray recursion to 0 per pass, or by changing object's shadow-casting properties per-pass, or by using light inclusion/exclusion for shadow-casting lights per-pass.

[adamredwoods]

LScipt usually craps out due to poor coding. no different than C. C's only advantage is speed and access to a few features.

I'm confident Jeremy did a great job since he is very knowledgeable in Lscript's shortcomings. But I haven't tried PassPort, so I cannot comment on it specifically.

[Digital Hermit]

Jeremy,

Can Passport render a shadow pass with out the objects and if so would it be faster?

Reason:

Say I want to render scene with a Hi-Rez “static” shot of a building; rendering it once as a background shot. Could I then go back and render an animated element, but only capture the shadow that falls on the shape of the structure? I could then take it to composite.

I am mainly thinking that this would be a great time saver when dealing with shadow falling on certain static shots.

Thanx,

[jeremyhardin]

Can passport do it? Absolutely. That's exactly the sort of thing it was intended for. As a matter of fact, you could set up both of those passes (the static shot and the shadow pass) in the same scene, then click render all passes to render them both at once from that one scene file.

As to rendering faster, it is in some cases where you can eliminate objects from some of your passes, but in this case you want the shadow-casting objects in the scene but invisible. So in this case it'd render in the same amount of time.

The difference is that you don't need 2 scenes (the beauty and the shadow scene) and you don't need to sit and render one scene after the other since you can just choose to render both at once from one scene in PassPort.

[Cageman]

It instills a great sense of confidence to see all the demo videos done with a ground plane, 3 lights and a few primitives. We once bought Surpasses and tried using it on a few large scenes and it crapped it pants every time.

As a beta-tester of PassPort I can assure you that it works very well with complex/heavy scenes. I've tested it with scenes that, frankly speaking, almost breaks my computer apart (see my computer-spec below). With PassPort I was able to chunk them up into smaller, manageable pieces. Having scenes that takes up little over 1.5GB ram when loading them (on my machine) will take forever to render (if they mange to render at all). PassPort is well worth its price.