[zdeno]

this is a completely wrong statement, point. what about stacked layers, new (much improved!) UI, color multilight, mxm web integration, new scripting, support of color spaces and other dozens of not-so-small new features? not to speak of a substancial speed improvement under certain scene conditions.

good point , there are many more improvements, but it is funny when NL shows in they newsletter " look how bad was bump and normalmaping in 1.7.1, not mentioning fail in roughness model" but it won't be fixed, instead of service pack we will charge You again.

So now I stack with "not so perfect" 1.7.1 version

I am an architect and if I made mistake in documentation I have to fix this for free, praing I wouldn't have to give some $ to disapointed client.
If software developer admit bump mapping sucks-ok it happens- so lets give me good one for free ! This is not NL invented a new better wheel, bump mapping works better even in scanline in my autodesk VIZ. in 2.0 maxwell it looks just great, but I just paid for that once.

I don,t need "UI, color multilight, mxm web integration, new scripting, support of color spaces" (mxm web integration is the funniest one (95% od this mxm is just useless)
I just want to force brick wall looks like wall constructed with bricks without displacement which took 3 days to render to resonable SL.
IES and leaf-SSS would be most welcome, but only for this .... no thanks.

That's just the point. Maxwell users typically do not want to sacrifice quality for speed. That's what Maxwell is. Uncompromising Quality.

ROTFL. and when there would show again bounce control slider, or sampling filter to run few speed test renders. THEY ALL will lost their mind theirs eyes would explode, and brains turn to rotten tomatoes. wake up Brain. Stand up from your knees.

[dmeyer]

[
I am an architect and if I made mistake in documentation I have to fix this for free, praing I wouldn't have to give some $ to disapointed client.
If software developer admit bump mapping sucks-ok it happens- so lets give me good one for free !

Do you go back and upgrade clients buildings for free as your skill improves as an architect over the years? Or as building materials and processes become more efficient?

[Bubbaloo]

[numerobis]

I just want to force brick wall looks like wall constructed with bricks without displacement which took 3 days to render to resonable SL.

you could switch back to mxw alpha 1.1.33. bump was working great in alpha/beta days as far as i remember...

[Michael Betke]

I use Octane for a while now and really like it.
But I have to say if you have more complex scenes its getting slower too and I have a GTX480 which isn't the slowest thing.

So all I can do is to do rought stuff which needs to get whipped up fast with octane and sent it to clients.
when they are confident with the results I do the final render and setup with Maxwell...

If you ask for a galery just take a look at the octane galery. Its very good arch-viz stuff in there and I doubt clients will see a difference between two unbiased renderings....

I wish Maxwell would get additional GPU support but it wont happen.

[Bubbaloo]

[Tok_Tok]

But I have to say if you have more complex scenes its getting slower too and I have a GTX480 which isn't the slowest thing.

Exactly. GPU is not to replace the CPU, combined it would be perfect. Something like the GPU for realtime preview and DOF, and the CPU for long term rendering (+ GPU for faster results).
GPU rendering would also be perfect for making very realistic materials, especially SSS mats and complex IOR.

[Mihai]

This is 4 pages of empty talk to me, except Michael Betkes post to give some info and opinions about using a GPU renderer. Besides having watched a video showing some objects on a plane lit with IBL, has anybody tried any of them on their regular scenes and say here what they think is good, what is bad/missing, what limitations are there? There are still many questions about the GPU approach which you don't seem to care about. It seems your opinion is - look, another GPU renderer, just like all the other renderers only faster. If so many others can do it, why not Maxwell?

When you have your final scene with all your objects, textures, instances, displacement......how quickly does it do 'real time'? What type of hardware and cost will it take to make it 10x faster than CPU in these cases and not plane/IBL/object scenario?
Can you render at the resolution your clients usually ask from you?
What type of lighting can you use? IBL, emitters, Sky, IES
Does it do caustics? What kind? What limitations? Do they ever clear up?
Sub surface scattering? True volumetric or fake approach?
Single sided translucency for your trees and vegetation?
Instances?
Motion blur?
Displacement?
Any limitations to number of polygons or texture resolution?

Aren't these some of the questions you would like answers to, instead of just asking - why doesn't Maxwell do this "GPU stuff"? Why does NL REFUSE to make Maxwell faster when so many others can obviously do it. If it was that simple...besides the possible technical limitations and not knowing where GPUs and CPUs will be in 1 year from now and are GPUs mature enough to invest a lot of time and effort developing for them...it's always a matter of compromise. How much image fidelity are you willing to give up? The things mentioned in this thread show how much time is invested in perfecting the quality of this render engine. If you simply see it as bug fixes, then I invite you to check some of those things in other renderers and you may be surprised. Do they matter for your business? Maybe not when efficiency and getting your client off your back matters most. But you are free to use any tool at your disposal. I'd like to simply know more about the details of any new tech, instead of this waste of time for me reading this thread ... and irritation from you because of Next Limits intolerable arrogance when they choose not to reveal in public their exact plans for the next 6-12 months...

[Gary]

Thank you Mihai, and I am in full agreement.

I appreciate the hard work, new features, reduced bugs, improved stability and greatly improved plug-ins for Maxwell. I trust Next Limit is diligently working on improving Maxwell in the best ways they see fit.

I expect and hope that Next Limit runs their business like I run mine— evaluating the options, making smart decisions, and putting nothing in the product until it's time (and not talking about new features before it is time). As a result they can be rewarded by offering a quality product that some (not all) people choose to buy.

Much like there are sports cars & mini-vans, not one rendering program can do it all. I for one am glad Next Limit has remained focused on quality and ease of use, making speed gains along the way. If there really were cheaper/faster/better alternatives I doubt we would all still be here reading this. Do I monitor the new technologies? Sure, but GPU rendering is currently not the magic bullet for my work.

Gary

[JTB]

I just want to force brick wall looks like wall constructed with bricks without displacement which took 3 days to render to resonable SL.

you could switch back to mxw alpha 1.1.33. bump was working great in alpha/beta days as far as i remember...

Actually, when I first bought MW beta, I thought there was almost nothing to improve ..... the quality was unbelievable with slower machines.

Exactly. GPU is not to replace the CPU, combined it would be perfect. Something like the GPU for realtime preview and DOF, and the CPU for long term rendering (+ GPU for faster results).
GPU rendering would also be perfect for making very realistic materials, especially SSS mats and complex IOR.

That would be nice, having GPU and CPU together but it won't easily happen with Maxwell... First, we don't know what GPU will be like, in 2 years... Maybe there will be GPU cards with 8Gb RAM and cheap... So no reason to say what's best, when everything changes so fast.
There are many differenct approaches, Arion uses both CPU and GPU which is best I think... Octane has both biased and unbiased rendering which can speed up things... Thea Render has both biased and unbiased rendering and has combined some tools like instances brush which is a great addition for arch-viz renderings (trees, grass, stones etc)...
Almost every other company has something to say for future releases, except NL... They have every right to do so, but I have every right to say that I don't like it...

[choo-chee]

I agree with all but still... about 3-4 years ago I sat down at my home PC and found an interesting software for my work so I d-l it and tried on my crappy home PC. 10 minutes later I called my partner and told him "tomorrow we buy Maxwell Render". Till today I enjoy great images and curse the time they take to render. I think that after few years it's time to improve the worse part of the software-rendering speed- and playing with stuff that doesn't really matter for many users is wrong. No one thinks that MR isn't no. 1 or at least in the first 3 rendering engines with quality. So why try to improve the field that is already almost 100%?! give use speed. the only thing that really changed the speed in maxwell is advancing hardware solutions, the software is almost the same. who cares about caustics clearing faster? is relatively faster to ver. 1 but incredibly slower in terms of development time passed from ver.1. The only thing that keeps ppl away form MR is the speed issue and that's where NL should concentrate if they want new customers and upgrading customers - by the tons.

[dmeyer]

Football/soccer just became interesting.

http://www.maxwellrender.com/news/world_cup_promo.php

[Fernando Tella]

¡¡¡¡ESPAÑA, ESPAÑA!!!!

Edit: Nicole, the price of render node update hasn't changed as the rest (euros); I think it is wrong according to the promotion.

[tom]

Does anybody here own or have access to one of these GPU based engines, yet? Because, I'd like to share something small to test on it.

[wagurto]

Calico is using octane render I think. It would really help to do some comparison.