I've been using Redshift for quite some time and I like the tight integration with Houdini, but I find it more and more difficult to ignore the RenderMan call given their pretty aggressive development schedule and future integration with Solaris.

However, I don't see it mentioned very often in render engine discussions, with most of the users mentioning Mantra, Redshift, V-Ray and Arnold.

I'm curious as to who around here uses RenderMan and why. I am also curious about who has looked into it and decided not to use it, and why not?

Renderman is outright rejected by anyone not up to date with rendering tech. It's seen as very old skool and slow, which is what RM 22+ and XPU is trying to fix.

Non-biased renderers produce a nicer look.

Is RenderMan biased or unbiased?

Other RenderMan A&SF 2019 presentations:
https://renderman.pixar.com/news/renderman-art-and-science-fair-2019—presentations [renderman.pixar.com]

RenderMan offers a combination of unbiased and biased rendering techniques which provide both accuracy and technical efficiency

https://renderman.pixar.com/tech-specs [renderman.pixar.com]

Solarisse
I think I'm biased, despite I didn't try H+Rman in production yet (used Clarisse before but don't want the export/import, frequent crash etc nightmares again).

Have you seen this?

Pixar guys make an effort to make a good impression, but starting with H18 and Karma, i can not say i am awfully impressed with it. Also since I was using Mantra I just can't give up on the procedural material networks I am so used to creating, which eats up all kinds of funny business.
I couldn't find a fake bevel shader out of the box in renderman too, did i miss it? Read through docs, seem to be absent there as well.
Also Karma promises a big speedup, let alone it's own version of GPU some time later. Kays, you've been using RS for a while now, i cant seem to see any reason to go prman in your situation. I was entertaining this idea (of considering renderman as a main render) since I haven't switched to GPU rendering, but now i don't even think about it and will just wait for Karma. Mind you, I don't render hollywood VFX blockbusters, so I can't claim to have knowledge of how it will fare against karma in a typical alien invasion on New York scenario.

Apart from digital doubles (realistic human rendering) I am not sure where it will outshine karma. Thats my 2 cents.
Let alone Karma will be as native as it gets.

Yeah, I agree, Karma is promising, its main downside being the timeframe (i.e. how soon before it's production-ready, and how soon before a GPU, or at least GPU-enhanced, version is out?).

I did watch the Houdini Hive presentation and that's what prompted me to ask the question. I still do like Redshift, as a matter of fact I'll be doing a presentation on using Redshift within my most recent project, in Los Angeles in the Fall.

It's a music video, I dunno if you've seen it, but if you're interested here it is:


Anyway, I did toy around with RenderMan a bit and I have to say I do like the responsiveness of the IPR and fairly straightforward node network. Oh, and the material browser is so much nicer than anything else at the moment. I hope Karma makes improvements in that field.

^great vid

Generally speaking, you can always build your HDA for a material, so it's kinda tomato-potato.
Karma seems to be same~ish level of responsiveness, based on my totally scientific youtube recording evaluations.

As for you my friend, going forward would be a 1x 2020 RTX Titan + 6x 3080ti via nvidia link - the speed would be unbeatable for sure in RS against anything, - that is if you are perfectly fine with RS and what it has in store.

Pixar goes hard with presenting it's tech, in fact it has the most coverage out of all competitors, yet I'll repeat myself - after Karma is out and stable - what's the point?

glitchi
after Karma is out and stable - what's the point?

I think you're right. Karma is poised to take the Houdini users by storm, and hopefully it will be just as cool as the initial sneak peeks hint at!

Again, that's just my personal take - I don't render 140,000 lights at once and other stuff that pixar emphasizes in it's presentation. Possibly someone from big studios who are vastly smarter than I am will laugh at me, so take it with a grain of salt.

My 2 cents:
A combo package for indies would be the best: Houdini Indie + RMan Indie (exactly same as the non-commercial, just same usage constraints as H: $100k/year max) .
Yes, it seems that Karma would make RMan pointless, but Pixar is ahead in time. You can use it right now, and XPU is coming, I guess way before the Karma GPU. Dylan Sisson mentioned in the general pesentation that RMan 23 will be available from the release day of H18:
https://vimeo.com/353632662#t=867s [vimeo.com]

They work closely, so if Karma would be better for every kind of project, it would be a strange situation…

For me the stability/roboustness (1.) and interactivity/memory efficiency(2.) are the most important things. Raytracers have basically the same featureset from artistic viewpoint, the rest can be done by shading node trickery, every one has some missing features like the fake bevel shader… The question is that after work you go home with the feeling that you hate computers and regret because you chose this profession, or you think that this is one of the best professions that humans ever invented. I experienced both situations, and as I see it highly depends on the number of crashes, issues, slowdowns, restart, doitagain etc / day = the MCAS factor: Machine Crashed Again Shhhhhhhh

I have tested Renderman 22.5 a little and really like the responsiveness and general ease of use. Especially interesting on Mac sice there currently is a lack of GPU renderers for production. Maybe 22.6 is better but 22.5 was quite unstable and missed some features I needed like fulldome rendering (180 degree fisheye, sciviz!). Also, to slow compared to RS on a few RTX cards.(I have recently switched to RS since I needed something to render a project fast enough for delivery.)
On a similar note, I have also beta tested AMD Pro Render for a year or so and it is very similar in thinking compared to Renderman (for obvious reasons). RPR is quite shaky as well and missing features but is progressing quite nicely and since it exist as a Hydra delegate it will be interesting to test out later this year when all the stars finally align

Renderman 23 release will land on H18 day one (if goes according to plan), so will Karma beta.
I've used prman, the latest 22.6, and it's faster/more responsive than Mantra. As filipw mentioned, if it's the speed you are after, RS will be an option. RS is VERY responsive, almost magically so - you gotta try it to believe it.
I am talking here about a non-Hollywood level of visuals mind you, which will require a renderfarm whichever render you chose.

PS - if it's the speed what you are really after, UE4 is an option too. I'm not joking. Lots of tradeoffs, but 60fps with instances, autolods, realtime vertex-paint effects, deferred decals, foliage, amazing materials system, GPU particles (Niagara), Houdini engine, the list goes on.

edit: going forward, RS (and octane) will be having a realtime version of their render engine - currently named RS RT. Their presentation makes bold claims, most importantly - it will use same scene, lights and shader setups, so you can build your scene in realtime, and then switch to “full fat” Redshift if needed without additional tweaks.

I don't envy for a second the development teams of render engines. The competition is fierce, and that's good for the end users.

Strange situation.
Mantra (soon to be replaced) + Karma (still beta) + Renderman (hopfully XPU).

It seems that Karma is just a nominal renderer.

Sitting and waiting for prman XPU is a strange bet. If it arrives in 2020 that'd be the best case scenario.

What I find truly intriguing is watching SideFX promoting so much Renderman while disclosing so few details about Karma. As you said, Renderman is ahead on time and available now, but I don't know in what position that puts Karma on.

RM22.6 feels like 3Delight - which is a very good thing™

goat
RM22.6 feels like 3Delight - which is a very good thing™

Well 3Delight for Houdini is coming soon, so get ready to have fun rendering again.

tinyhawkus

goat
RM22.6 feels like 3Delight - which is a very good thing™

Well 3Delight for Houdini is coming soon, so get ready to have fun rendering again.

I look forward to it so we can finally ditch Maya as the antiquated rendering husk that it is.

Wren

tinyhawkus

goat
RM22.6 feels like 3Delight - which is a very good thing™

Well 3Delight for Houdini is coming soon, so get ready to have fun rendering again.

I look forward to it so we can finally ditch Maya as the antiquated rendering husk that it is.

Have you tried 3Delight with Gaffer [www.gafferhq.org] ?
Free, open source and Chappie likes it.