Poll

3 features you want the most:

Tiles map
31 (6.7%)
Further imrpovements to Corona Image Editor
8 (1.7%)
Significantly faster DOF (Depth of Field) rendering
25 (5.4%)
Sketch/Toon/Stylized shader
29 (6.2%)
Dedicated CarPaint Shader
7 (1.5%)
Dedicated fabrics shader
27 (5.8%)
Lightmix extended to materials, textures, ...
27 (5.8%)
Interactive rendering in 3ds max viewport (with gizmos, object selection, manipulation, ...)
16 (3.4%)
Rendering memory usage improvements
16 (3.4%)
Speed of rendering improvements
42 (9%)
Speed of interactive rendering improvements specifically
31 (6.7%)
GPU/Hybrid rendering
88 (18.9%)
Stability improvements (bug fixes)
12 (2.6%)
Improvements to caustics
23 (4.9%)
Thin film/coating shader
5 (1.1%)
Parsing performance optimization (e.g. for animations)
40 (8.6%)
New and better frame buffer (docked and floating)
23 (4.9%)
Further improvements to Chaos Scatter
15 (3.2%)

Total Members Voted: 178

Author Topic: The most wanted feature?  (Read 524854 times)

2019-01-07, 19:48:30
Reply #615

Jpjapers

  • Active Users
  • **
  • Posts: 1644
    • View Profile
Also lupaz not all dev is user driven. There's likely alot more at play.
But it is somewhat disappointing when something that is on a user poll and the trello gets pushed back but its beyond anyone's control.

2019-01-07, 19:53:12
Reply #616

bluebox

  • Active Users
  • **
  • Posts: 268
    • View Profile
I always thought that speed comes (partially) also with leaner code. Personally Corona is fast enough for me. On the other hand I'm under an impression that there is some kind of consensus that there is a lot of room for improvement on the field of tone mapping etc.

Apart from memory optimizations is this being considered as "the big feature" for the upcoming release ?

2019-01-07, 20:05:44
Reply #617

lupaz

  • Active Users
  • **
  • Posts: 951
    • View Profile
Of course everyone will say yes to the question "do you want to render faster?"
But how fast can Corona get by just improving the code? I can't imagine the difference being significant to match a GPU render engine and a couple of GPUs.
I'd think Corona is at the mercy of CPU development, and I think the user base knows this and it's ok with it.

And I'd say the same with RAM. Get more RAM if you need it.
But we can't get a better BRDF elsewhere and plug it in to Corona.

2019-01-07, 20:34:41
Reply #618

Jpjapers

  • Active Users
  • **
  • Posts: 1644
    • View Profile
Of course everyone will say yes to the question "do you want to render faster?"
But how fast can Corona get by just improving the code? I can't imagine the difference being significant to match a GPU render engine and a couple of GPUs.
I'd think Corona is at the mercy of CPU development, and I think the user base knows this and it's ok with it.

And I'd say the same with RAM. Get more RAM if you need it.
But we can't get a better BRDF elsewhere and plug it in to Corona.

Pretty sure siger has a brdf shader as dub at did some tests with it

2019-01-07, 20:38:48
Reply #619

lupaz

  • Active Users
  • **
  • Posts: 951
    • View Profile
Of course everyone will say yes to the question "do you want to render faster?"
But how fast can Corona get by just improving the code? I can't imagine the difference being significant to match a GPU render engine and a couple of GPUs.
I'd think Corona is at the mercy of CPU development, and I think the user base knows this and it's ok with it.

And I'd say the same with RAM. Get more RAM if you need it.
But we can't get a better BRDF elsewhere and plug it in to Corona.

Pretty sure siger has a brdf shader as dub at did some tests with it

Thanks. It was just an example.
Edit: Are you sure it's a BRDF or is it maps?
« Last Edit: 2019-01-07, 20:45:55 by lupaz »

2019-01-07, 20:44:52
Reply #620

zokni

  • Active Users
  • **
  • Posts: 9
    • View Profile
Allright, I will try to explain myseft.

First of all, I am pretty sure nobody wanted to hurt or offend you guys, including myself as well.
I had/have an opinion what I've shared. It looks like for the majority of your customers different things in the engine are (more) important. Noted.

Sorry, I am not going to play with the quote feature I just want to reflect on some things.

...there was an opinion here that the NVIDIA Denoiser was really nothing major, just something generic.

It was me ofcourse and after all I apologize a little bit about that. It wasn't completely nice to say like that and I didn't exactly meant like that. As I said I like that feature, its a game changer, thats why I'll renew my subsciption as well. I was dissapointed a little bit, becuse I expected something more besides that as well. It's not needed to tell me again, I understand.

You can't please all of the people all of the time, as the saying goes :)

It's true, and true. We only would like to see a little bit more exact things, or more accurate things in the development roadmap it that is possible. There was a missunderstanding here, (some) people (including me) took the development roadmap more seriously than we should have had to. Now we know those are not exact things, just planned possible things.

but I every time I deleted my post because it was too impolite.

If I was impolite I apologize again. We are just talking not arguing. Nobody accuse anybody with anything.

most people want better performance in huge scenes (RAM, startup and shutdown times)

I would love to see/have those. We work on large scenes as well.

Only fraction of our users post here. In fact, many of the biggest powerusers simply do not have time to post here.

OK, It's a little bit hard to understand what you meant by that. I mean, I understand a little bit, but let me not reflect on that.




Corona is definitely a large company now. I am not talking about the number of people who work there I am talking about how they see the world businesswise.
You know what, I understand that one as well. they want to keep an eye on the companies future, I would do the same what they did/do.

Everybody needs to understand that, it's not the same company what it was a couple of years ago. I am not saying it becomes better or worst, it just becomes a real company.
The merging with Chaos Group, the focusing more on money what big sharks want made some people dissapointed. Nobody blames you guys. Every company has to make that decision who wants to grow above a certain level.

We love Corona, we love the family, some people just started to feel that the other family members not listening with full focus anymore. 4-5 (who knows) years ago every opinion was important but it can't be like that anymore because there are thousands of opinions out there. Understood.

We are archviz guys, you need to understand that. Ofcourse we ask for those things what we would need or love to see in the next releases. If the company focusing more on different areas, (maybe because they want to break into other parts of the industry as well) we understand that as well, companywise, but we not celebreate that.

I didn't even want to write this comment, not because I am dissapointed or mad with anybody. I didn't want to do it because I think I understand your point. I am not a company owner ofcourse but I've managed to live the past 40 years on this world and I try to keep my eyes and my mind open as much as I can. That doesn't mean I am right or wrong. Some people will think I am right (in a few things) the others will think I am the stupidest person who ever lived. That is completely fine.

As an end note:
We are all looking forward to the next releases. You guys do your best, we will try to do the same as well.

I forgot to mention something and I think thats the real point of this whole thing.
We are artist, we have different needs. Large companies (those are the important clients, thats understandable, I buy 1 licence, they buy 1000 licences) want more speed, less RAM usage. The quality of the corona renders are already superb, ofcourse they want to speed up their renders to make money more quicker with a less expensive render farm.

We - artists - on the other side prefer those "special" things (as well), what makes that f* vase on that table more photorealistic. :)
« Last Edit: 2019-01-07, 21:38:02 by zokni »

2019-01-07, 20:49:31
Reply #621

Jpjapers

  • Active Users
  • **
  • Posts: 1644
    • View Profile
Of course everyone will say yes to the question "do you want to render faster?"
But how fast can Corona get by just improving the code? I can't imagine the difference being significant to match a GPU render engine and a couple of GPUs.
I'd think Corona is at the mercy of CPU development, and I think the user base knows this and it's ok with it.

And I'd say the same with RAM. Get more RAM if you need it.
But we can't get a better BRDF elsewhere and plug it in to Corona.

Pretty sure siger has a brdf shader as dub at did some tests with it

Thanks. It was just an example.
Edit: Are you sure it's a BRDF or is it maps?

It's actually complex fresnel, skimmed the post and i thought i saw bdrf , apologies.
« Last Edit: 2019-01-07, 21:32:51 by jpjapers »

2019-01-07, 23:10:12
Reply #622

Ondra

  • Administrator
  • Active Users
  • *****
  • Posts: 9048
  • Turning coffee to features since 2009
    • View Profile
few notes:

since I see chaos group thrown here or there: in the 1.5 years we are together we never were ordered to implement or not implement something in Corona. All decisions were ultimately our own.

speed: it is true that for new renderer/renderer with not many features, lot of speed comes form the "lean-ness". But as the renderer gets complex and supports many features, that is no longer true and you have to invest lot of time to improve/keep the speed.

Those corporate fatcats that we cater ;) to are studios of 6 to 15 people. We were in contact with them since the beginning and they are still archviz users. I brough it up just as illustration that there are other sources of feedback, but this does not mean we are now doing exclusively that. In fact, we want to attempt (cannot promise anything) most of the features mentioned here in v4.

cpus being dead: 2 years ago if you were building a machine, you would buy a 4core i7 with maximum being 8. Today, you can build 12core threadripper for the same money, with options to scale up to 32core in single affordable desktop CPU. I am not saying we will get 128 core CPUs in next 2 years... but say what you want, CPUs were anything but dead in the last 2 years.

BTW: Even majority of the forum IMHO agrees that the new features are useful, since when I remove something from the most wanted list here in thread as done, most times it was in the TOP3 choices. Too bad I dont have historical poll data.
Rendering is magic.How to get minidumps for crashed/frozen 3ds Max | Sorry for short replies, brief responses = more time to develop Corona ;)

2019-01-08, 00:08:54
Reply #623

lupaz

  • Active Users
  • **
  • Posts: 951
    • View Profile
cpus being dead

If anyone here thinks this then they got the wrong forum to talk about it.




2019-01-08, 04:45:34
Reply #624

zokni

  • Active Users
  • **
  • Posts: 9
    • View Profile
few notes:

Thank you for your answer Ondra.

Offtopic more or less

Talking about CPUs/GPUs is a pretty interesting thing. I am planning to push my career towards freelancing because of different reasons. It is a longer term plan, lets say in a year.

Ofcourse money (I mean the amount of money what I actually have or what I would like to spend on start later) is an important aspect. I am planning to buy a powerful workstation in the next couple of months (what I would be able to use as a render machine as well) and I am confused a little bit.

It looks like the only smart solution atm are those Threadrippers with 24/32 cores.

I worked in three different offices in the past 6 years. In all of them they used Vray when I started and especially 5-6 years ago (be honest) that was the right decision. In 2 of the 3 offices we were able to persuade our boss(es) to switch to Corona from Vray.

In my opinion the end result of Corona is nicer, especially the light distribution. The rendering speed of Corona has been always an issue (in my actual office we use 24 (a little bit older, but still decent dual xeons as a farm, the approx speed of one of them (with 2 xeons) is like a modern 12 core i9, so lets say 24 of those). Even with those machines (in distributed, split render) sometimes we have crazy render times (4 hours in 4-5k but most of our scenes are heavy).
Thats why I said it depends where we look the case from. As an artist, I prefer more photrealism as a wannabe freelancer I prefer speed and less RAM usage. I am fighting inside with myself. :)
If I would be able to make decisions which way Corona would follow, according to my brain I would choose speed, according to my heart I would choose more photorealism.

Personally I prefer to keep the noise level of my renders pretty low, I use 2% what can be an overkill to some people. I like the real time denoiser but I don't like to use the denoiser in the final render (that will change no later than on that day when I will need to provide the necessary computer power for the renders :)) I know, probably its a huge mistake to use 2% of noise without any denoising at the end but I like to keep the end result of my renders as less affected as possible.

From this aspect I think some kind of Hybrid solution is important. Let's say if it would be possible to speed up the scene parsing process with a GPU or the displacement calculation, that would be a big plus. To be honest I know nothing about computer programming I have never been interested in that thing. I have no idea what can be possible and what is not.

According to those pretty expensive CPUs (lets talk about gaming cards because those are the things what most mortal can afford) even a powerful GPU (2080Ti) is cheap.

CPU wise, what is your opinion, which would be better as a start? To buy a more powerful workstation (let's say the the 32 core AMD) and buy separate render machines later or buy a less powerful workstation (let's say an 8-12 core intel) and buy separate render machine(s) the same time as well?

I think, these days, when GPUs are more and more powerful for less and less money you can't avoid (at least you have to think about it as a future plan) to go towards some kind of hybrid solution. If you can archive 50% more speed with the cost of €5000 CPUs or the same with €2500 GPUs, probably nobody would say: don't do it.

What is your opinion about those things above?

Another interesting thing is (I am not going to ask your opinion about that :)), which render engine to chose. Atm, Vray (especially according to speed and reliability in distributed) is more safe, more predictible. But Corona, according to quality, "friendlyness" and my heart is a better solution. If you make money with it and not "just" your boss (if he reads this, I love him ofcourse), speed is important, or more than important and "safety" as well.
« Last Edit: 2019-01-08, 04:50:51 by zokni »

2019-01-08, 14:01:09
Reply #625

sirio76

  • Active Users
  • **
  • Posts: 146
    • View Profile

I think, these days, when GPUs are more and more powerful for less and less money

To be honest its quite the opposite, GPU speed increment are becoming marginal and price are going up. Once you forget about all the RTX mumbo jumbo(no renderer are able to utilize RT core right now and even when it will be possible we should see the real benefits) as a matter of facts a 2080Ti offers barely 30% more render power when used in GPU renderer compared to previous generation(1080Ti is more than two years old so it's just 15% a year) but costs almost twice as much.
I totally agree with Ondra, CPU renderer are here to stay(BTW people were telling CPUs were dead since much longer than 2 years).
My 2 cents :)

2019-01-08, 14:27:37
Reply #626

Ondra

  • Administrator
  • Active Users
  • *****
  • Posts: 9048
  • Turning coffee to features since 2009
    • View Profile
BTW people were telling CPUs were dead since much longer than 2 years
small note: CPUs were quite "dead" (or in non-internet-discussion words, their development was slow  and performance gains minimal) approx between ~Haswell and Ryzen. All that thankfully changed with a bit of competition from AMD
 
Rendering is magic.How to get minidumps for crashed/frozen 3ds Max | Sorry for short replies, brief responses = more time to develop Corona ;)

2019-01-08, 14:38:26
Reply #627

sirio76

  • Active Users
  • **
  • Posts: 146
    • View Profile
That's true but as noted it was only because of the lack of competition, not because the technology was dead itself. And that's why it's important to support AMD on the GPU side too, unless you don't strictly need CUDA they offer competitive cards for less money.

2019-01-08, 19:01:07
Reply #628

bluebox

  • Active Users
  • **
  • Posts: 268
    • View Profile
In fact, we want to attempt (cannot promise anything) most of the features mentioned here in v4.


Thanks for answering Ondra. Looking forward to the new release then. It would be a major thing if all those features would get implemented.

2019-01-08, 19:45:48
Reply #629

Jpjapers

  • Active Users
  • **
  • Posts: 1644
    • View Profile
I guess in summary thanks for the responses from the team. Myself and my colleagues are looking forward to seeing what other cool stuff you guys pull out of the bag.