Author Topic: Tonemapping - Plz Halp  (Read 79813 times)

2020-05-04, 11:32:31
Reply #150

Designerman77

  • Active Users
  • **
  • Posts: 507
    • View Profile
May I ask what exactly is your point? You claim that Corona is a perfect render engine and devs. could go home and stop optimizing it?
It is NOT perfect and still needs a good amount of work until it can be called really photo-realistic. PERIOD !

Calling other peoples` thoughts "garbage" is an offense to all the guys here that take their precious time, energy and experience, willing to
make this great software even better.

I am sure the devs at Corona are not offended that users make suggestions. Actually, the nowadays quality of this software is based on the open minded attitude of the Corona dev team, not on an attitude of ignorance and "defending the existing limitations".

I'm just asking for a pragmatic approach. If you follow my post history, you'll actually see that I'm trying to push for some improvements as well. But come on, comparing a photograph of a kitchen with a render of a completely different scene, Making a poll with random renders... what's the point of this? This is useless as this proves nothing -> garbage. I'm not defending anything, you do.

I'm sorry for you but @BardhylM was right.

disclaimer : just to clarify, when I said no controlled environment, I was talking about the process of comparing real photos to renders. That part was not about lupaz test which is great even if it has some flaws I pointed out. So your test lupaz was not included in the garbage stuff.


You split hairs in a problem that is visible even to a half blind. MY opinion.
The goal of this software ist to work photo real. Nothing less.

Improvements are not done by neglecting the flaws.
Do you also react like this if your clients ask for improvements in your renders? Like "shut up with your garbage, they are perfect" ???

2020-05-04, 11:43:25
Reply #151

Fluss

  • Active Users
  • **
  • Posts: 553
    • View Profile
I did not ask anyone to shut up.

I'm not neglecting anything.

We are trying to answer one question. Not "did you see it ?", not "Do you like it ?", but "Why?". That's the point.
« Last Edit: 2020-05-04, 11:47:26 by Fluss »

2020-05-04, 11:46:23
Reply #152

cjwidd

  • Active Users
  • **
  • Posts: 1069
    • View Profile
    • Artstation
Eh, I feel like some of these comments are unnecessarily hostile and counterproductive...

2020-05-04, 12:02:03
Reply #153

ynotsop

  • Active Users
  • **
  • Posts: 39
    • View Profile
That's fine, I'll take the heat for posting those two pictures, but to me it illustrates how Fstorms renders mimic a camera. And I don't have to have a render based on a real photo to see it. The subjects are different but the overall materialism is very similar. To me the pictures look like they were taken with the same camera which goes back to my first point. Maybe I'm the only one seeing it and that's fine too.

The only thing garbage here is the attitude some of you have, and according to some only the ultra knowledgeable should be allowed to participate in discussions and the others can observe from their caves. If you don't agree with someone you can at least express it in a polite manner and we can keep the conversation on a level. I think it's fair to say that everyone participating in these forums are doing so with good intentions.

Eh, I feel like some of these comments are unnecessarily hostile and counterproductive...

I agree.

2020-05-04, 13:05:05
Reply #154

Fluss

  • Active Users
  • **
  • Posts: 553
    • View Profile
The thing is plenty of threads like this exist on the forum and they're all the same -> "Look how it's good", "Are you blind", "We want the same". Basically people moaning they want photo-real results without doing their part.
People keep posting example pictures from the internet, which they have no clue how they were made, how much post-production there is, how much effort the artist put in to look like that etc etc... Basically attributing the fact that the image looks good because of the renderer and not because of the artist.

So I tried to orient the discussion on a break it down approach. If it is so simple and Fstorm is so magical, show us examples you made and compare the results with what you can produce in Corona, 1:1. It's easy to make claims, it's less to prove them. The side by side comparison I've seen so far are not as mindblowing as some of you claim it to be. I never denied there is something common to some of the Fstorm renders we've seen, I'm just inclined to understand why rather than asking for something blindly...

Everybody is allowed to express an opinion, including me! So if I find these ever returning comments useless, I am allowed to express it. The thing is those are just making the thread going in circles like the 99 others before. Now if I offended some of you by calling those garbage, I apologize for that. I'm just tired of seeing the same stuff over and over again.

« Last Edit: 2020-05-04, 14:04:52 by Fluss »

2020-05-04, 13:22:56
Reply #155

bluebox

  • Active Users
  • **
  • Posts: 268
    • View Profile
I think we should wrap this up at this point to a conclusion that everyone is trying to contribute towards the same goal - improvements in the area that as we can all see is really really wanted by many.

Current tone mapping is not working as it should in many areas that were discussed back and forth many times.
Original contrast destroys blacks, oversaturates colours to a point that you are no longer able to see the original texture used which is really annoying when working with product design or anything that requires more than just a punchy-candy picture.
LUTs are also not the way to go, they should be used to get a certain degree of mood imho, not the base acceptable picture.

What concerns me the most is the fact that this in one of many topics abut all those issues and we still can't get noone from the Corona team to participate.
We still get all those "key new features" like faster color picker and the important stuff as we can see and extrapolate from the most wanted features poll gets neglected.

Those discussions basicaly IMHO end up in a vacuum and we get this BS policy of a roadmap that is all but "no promisses".
If the community wants certain features and they end up not being developed and we can't get any promisses then I vote for bringing back the box licence so I can upgrade when and only when the features that community wants get implemented.

My few cents.

2020-05-04, 14:37:20
Reply #156

agentdark45

  • Active Users
  • **
  • Posts: 581
    • View Profile
Regarding that image of my earlier test, as others noted, I would have to use the exact same HDRI and closest as possible materials, not a simple conversion.

So I repeated the test. Same HDRI, Same LUT (KimAmlan02), Matched white balance. very simple materials.

Do you really see a 50% bump in realism in the Fstorm one? Maybe I don't have your eyes, it is entirely possible that I actually don't have the perceptual capacity.



Anyway, now that I've conducted the test in a more proper manner, my point stands. I will write it in bold so you don't miss it and you stop arguing with something I'm not saying:

Fstorm has better tonemapping.   I wish Corona had a DSLR like tonemapping

However, I don't think that is the main reason for the Fstorm community producing seemingly more realistic images more consistently than the corona render community.
I think there is a small number of users (Agentdark45 included) who have superior understanding and perceptual abilities and actually notice the super slight differences between the two renderers, which are real and measurable.

But I believe what MOST people notice is a vibrant community that attracts great artists who are learning faster and feeling comfortable sooner with Fstorm, and thus, becoming good enough that, overall, the Fstorm community is producing better renders, more artistic, more realistic.

Just to be super clear, FSTORM HAS BETTER TONEMAPPING, 100% agree. And it has a real impact in realism! I wish Corona implements this soon
ps. I also wish corona becomes a lot more beginner and artist friendly by making it MUCH EASIER to produce the best possible images it already can, which, BTW would inevitably be WORST than Fstorm regarding tonemapping. At least until corona implements a better tonemapping, which I will continue to push and wish for.

Apologies if I missed the spirit of your post, it seemed contradictory at the time of posting. I'm glad to see you are supporting the call for improved tonemapping in Corona.

My main contention here is that some posters are claiming:

1. "Everything is fine with Corona as is, zero perceptual difference in tone-mapping between Fstorm and Corona; the issue is noobish user error being fixated on stock settings".
2. While other posters are stating that there is a clear and obvious limitation with Corona's tonemapping (despite Bertrand Benoit levels of Corona mastery and tweaking), and we want improved colour space/ACES like tonemapping implemented as a priority over other seemingly trivial features that eat up precious dev time.

Regarding your recent comparison, yes there is a difference between the two (perhaps not 50%) but I can clearly see the black crush in the shadows issue I have mentioned before in trying to achieve a contrasty punchy image whilst retaining not overly clamped highlights (look at how much detail is lost in the lower right shadow on the floor in the Corona version, along with the shadow at the base of the exterior sphere, the soft shadow gradient on the left hand interior wall, and the tendancy for Corona to go to pure black in corner gaps - the dynamic range of the shadows have been butchered). Now scale this issue up to images that mostly deal with subtle tonal variations (i.e moody or brightly lit product shots).

As a side note outside of tonemapping, you can also see the filtering issue I mentioned previously in the Corona example. Look at how much detail has been smoothed out of the concrete walls and wood material in the Corona version.
« Last Edit: 2020-05-04, 14:55:43 by agentdark45 »
Vray who?

2020-05-04, 14:52:58
Reply #157

Juraj

  • Active Users
  • **
  • Posts: 4324
    • View Profile
    • studio website
Less strawmen and attacks or devs will lock this epic thread 😀

Everyone really wants the same, there is just various level of allergy to hyperbolic expectations and conclusions.
This might be tough topic even for devs, I've only ever seen Pixar doing indepth reaserch into realism that goes beyond the technical part, esp. because so much of it strays into subjective.

I personally love comparisons, I know they are flawed but lit can be peered from them. Fluss is only suggesting to do this more thoroughly as it will help avoid reaching premature and incorrect conclusions.

That takes unfortunately unholy amount of effort & time, which everyone lacks. So all my respect to people with patience to fiddle with this. More tests, less assured conclusions.
talcikdemovicova.com  Website and blog
be.net/jurajtalcik   Our studio Behance portfolio
Instagram   Our studio Instagram, managed by Veronika

2020-05-04, 16:01:56
Reply #158

ynotsop

  • Active Users
  • **
  • Posts: 39
    • View Profile
I completely agree with the fact that in the end things have to be investigated in a practical manner. It's cool to see these comparisons lolec and lupaz are doing. It's also saddening to see that this issue has been already brought up way back in 2017 (or maybe even before that?).
« Last Edit: 2020-05-04, 18:04:08 by ynotsop »

2020-05-04, 16:34:52
Reply #159

lupaz

  • Active Users
  • **
  • Posts: 763
    • View Profile
@Fluss
You said it yourself:
So photorealism really isn't that simple to define as there is a whole lot of intricacies. The key is to use references and observe the world around you. I often look at light interactions around me and a lot of these observations are making me think "If I'd see this in one of my renders, i'd probably think this is a bug".

It would be really interesting to make a thread with great images and review them to highlight why it works well.

The reason for showing renderings and also photos is to observe.
You can do all the tests you want being methodical and practical, but in the end there's also a perception component that we can learn from looking at different images.
So I disagree. It's not garbage. It's learning. Isn't AI based on this?

If I may add my two cents, this discussion can be boiled down to braking bad's quote: "after all, how pure can pure be?"
For most people 96% purity is more than enough and for some people difference between 96 and 99 is not just noticeable it is like tremendous gulf.
and finally get conclusion "fstorm is 18% more photorealistic" who cares?

The whole discussion, I think, it's because we strive to make more compelling images. Images that sell.
Clients will "buy" faster (18%? :)) an image if they FEEL they like it.

Some Fstorm images have a feeling that is very attractive.
We're trying to pin-point it, right?


Original contrast destroys blacks, oversaturates colours to a point that you are no longer able to see the original texture used which is really annoying when working with product design or anything that requires more than just a punchy-candy picture.

How do you guys know that this is the cause? A bad tone mapping...
Could it be that the shadows are wrongly represented by the engine in the first place, making them more bland with a gradient that makes contrast to flatten them?

That's what I'm trying to do with the black and white noisy tests:
I want to discard tone mapping, sharpening, LUT, bloom, DOF, color, textures, contrast; and I want to compare the core of the engine: How it resolves light and shadows, antialiasing, BRDF and sampling of small objects and details.



2020-05-04, 17:11:38
Reply #160

lupaz

  • Active Users
  • **
  • Posts: 763
    • View Profile
I notice you didn't enable LUT or modify Hightlight compression. Any reason why? Fstorm applies a LUT by default, so this is not a oranges to oranges compression, just wondering if you have a reason?

What I tried to do was to remove from the Fstorm frame buffer all post production. For this test I didn't care about Fstorm defaults. If we remove post production stuff and end up with the same rendering as corona, that would mean that you can get an Fstorm render with little effort, with Corona's controls or other software.

What I want to do next is add DOF to the test and see what happens.

2020-05-04, 17:23:48
Reply #161

Fluss

  • Active Users
  • **
  • Posts: 553
    • View Profile
LMAO, you're quoting me totally out of context. If you still did not get my point then there is not so much I can do. Anyway, I'll leave it here as I don't want to pollute the thread even more. let's move on
« Last Edit: 2020-05-04, 17:31:45 by Fluss »

2020-05-04, 17:42:45
Reply #162

Designerman77

  • Active Users
  • **
  • Posts: 507
    • View Profile
I did not ask anyone to shut up.

I'm not neglecting anything.

We are trying to answer one question. Not "did you see it ?", not "Do you like it ?", but "Why?". That's the point.


Then I'm totally with you. :)

2020-05-04, 17:54:14
Reply #163

Designerman77

  • Active Users
  • **
  • Posts: 507
    • View Profile
I think we should wrap this up at this point to a conclusion that everyone is trying to contribute towards the same goal - improvements in the area that as we can all see is really really wanted by many.

Current tone mapping is not working as it should in many areas that were discussed back and forth many times.
Original contrast destroys blacks, oversaturates colours to a point that you are no longer able to see the original texture used which is really annoying when working with product design or anything that requires more than just a punchy-candy picture.
LUTs are also not the way to go, they should be used to get a certain degree of mood imho, not the base acceptable picture.

What concerns me the most is the fact that this in one of many topics abut all those issues and we still can't get noone from the Corona team to participate.
We still get all those "key new features" like faster color picker and the important stuff as we can see and extrapolate from the most wanted features poll gets neglected.

Those discussions basicaly IMHO end up in a vacuum and we get this BS policy of a roadmap that is all but "no promisses".
If the community wants certain features and they end up not being developed and we can't get any promisses then I vote for bringing back the box licence so I can upgrade when and only when the features that community wants get implemented.

My few cents.




With your observations about tone mapping misbalance, burning colors, exaggerated contrasts in blacks, etc. you 100% nailed it!
And that's also basically my point: there is no way to talk those facts away.

However, regarding your assumption that Corona devs "don't care"... hmmm... I am sure the do care and read this thread very carefully, since it
criticized the very core of Corona - in a pro active way. :)
About one year ago, some guys and me massively complained about the new denoiser messing up round edges & bump textures to a point that one could not work anymore.
Relatively quickly. the devs implemented the new AI denoiser that does a good job.

Wouldn't be surprised to see a new, advanced tone mapping in Cor 7.
Let´s hope.







2020-05-04, 18:05:08
Reply #164

Designerman77

  • Active Users
  • **
  • Posts: 507
    • View Profile
Regarding your recent comparison, yes there is a difference between the two (perhaps not 50%) but I can clearly see the black crush in the shadows issue I have mentioned before in trying to achieve a contrasty punchy image whilst retaining not overly clamped highlights (look at how much detail is lost in the lower right shadow on the floor in the Corona version, along with the shadow at the base of the exterior sphere, the soft shadow gradient on the left hand interior wall, and the tendancy for Corona to go to pure black in corner gaps - the dynamic range of the shadows have been butchered). Now scale this issue up to images that mostly deal with subtle tonal variations (i.e moody or brightly lit product shots).

As a side note outside of tonemapping, you can also see the filtering issue I mentioned previously in the Corona example. Look at how much detail has been smoothed out of the concrete walls and wood material in the Corona version.
[/quote]






Brilliant eye & observation, Sir!

Exactly those things ( black shadow areas, wrong dark-light transitions, etc.) make you literally feel like somehow not being abele to look at an image and get the information your eye looks for.
A bit like listening to a music recording that has distorted frequencies.