1 hour ago, mSparks said:

I just quickly recorded a "through the lens" VR view.

The transition is only visible in 2D

he very black edges you get in 2D.

I'd say the root cause is probably not having a proper FOV setup, putting the eye adaptation out of whack.

I didn't really notice this before, because 2D sucks in so many more ways than that I can't really bare to look at it.

EDIT: yeah, if I set a "proper" field of view the transition all but vanishes, but then of course in an airliner even on a 19 inch monitor the cockpit isn't exactly usable either

https://sampsoid.com/fov-calc/

Monitor width including bezel:

19"

Visible width excluding bezel:

17.4"

Viewing distance:

24"

Field of View when driving:

42°

thanks for sharing that ... think goes to show a bit of compromise here helps with different setups

3 hours ago, mSparks said:

I'd say the root cause is probably not having a proper FOV setup, putting the eye adaptation out of whack.

Most of the images/videos posted concerning dark cockpits have FOV way off.  The logic used is this - I can see a wide FOV in the real aircraft, so I want the same FOV on my monitor.  Makes no sense to me when the simulated world is being viewed through a frame the size of the monitor.  A FOV way too wide for the monitor size distorts distance and speed, and also makes dark cockpits an issue.

My video above shows what a correct FOV looks like in the default cub. So many screenshots I see look like the virtual pilot is sitting in the back seat. There is no dark cockpit in my cub video - that fact is just ignored, judging by the zero responses saying there isn't a dark cockpit in the video.

 

16 hours ago, efis007 said:

Do you think it is important to talk about the programming techniques used?
The customer doesn't care at all.

When you go to a restaurant, do you judge a dish by the list of ingredients, or by the taste?
You can use the best ingredients/tools in the world, but if the dish then turns out to be indigestible because you cooked it wrong, you failed the dish. 

Same thing for XP12.
You can use all existing XYZ programming languages, even "alien techniques" if you like, but if your product then generates dark panel bugs, you've failed the dish.
It's not hard to understand.

Actually Yes, it is important to understand how an effect is reached. 
Because the solution for your problem would be very different.

And you would understand that although the result is the same, the underlying technology is vastly different.

to stay in your analogy you can add too much salt or swap sugar by salt. In both cases the meal is ruined, but the cause is different because the recipes are different ( one recipe needs a little salt, the other needs sugar).

This not hard to understand neither…

5 hours ago, GoranM said:

I will however agree, that I also laughed, but at how ridiculous this entire thread is

In that case i wont bother posting in it any more. 

And I want to add that to a certain degree I agree that the cockpit could be too dark at times, it needs more tweaking. 
At the other hand the idea how to implement this is imo an improvement over previous models.

Edited March 13, 20233 yr by soaring_penguin

13 minutes ago, MrBitstFlyer said:

In that case i wont bother posting in it any more. 

Don’t get me wrong. It’s not the people refuting efis’ posts. I’m beyond stunned that he would go to such lengths, typing such enormous posts, for shadows. 
I do admire your patience, though. 

Edited March 13, 20233 yr by GoranM

Oh wasn't this topic locked yet? 🙂

 

Here's a video from Michael Brown, owner of Xforce PC and good friend of Austin. 
This is what he had to say about 12.01 and dark cockpits. Not sure if 12.04 is any better

 

9 minutes ago, peroni said:

Oh wasn't this topic locked yet? 

I'm kind of imagining the mods throwing their hands up in the air, and saying, "We give up!  You guys sort it out yourselves!"

😄

1 hour ago, MrBitstFlyer said:

Most of the images/videos posted concerning dark cockpits have FOV way off.  The logic used is this - I can see a wide FOV in the real aircraft, so I want the same FOV on my monitor.  Makes no sense to me when the simulated world is being viewed through a frame the size of the monitor.  A FOV way too wide for the monitor size distorts distance and speed, and also makes dark cockpits an issue.

My video above shows what a correct FOV looks like in the default cub. So many screenshots I see look like the virtual pilot is sitting in the back seat. There is no dark cockpit in my cub video - that fact is just ignored, judging by the zero responses saying there isn't a dark cockpit in the video.

 

so the choice is to not be able to see anything (narrow field of view) ... or not be able to see anything (dark cockpit) ? 😅

Is there a better example of the clown show on display in this thread than below? This is apparently an example of a beautifully bright and contrasty image, both inside the cockpit and out 🤣

https://imgur.com/6QBioS4

 

The 

Edited March 13, 20233 yr by 2reds2whites

16 minutes ago, 2reds2whites said:

Is there a better example of the clown show on display in this thread than below? This is apparently an example of a beautifully bright and contrasty image, both inside the cockpit and out 🤣

https://imgur.com/6QBioS4

 

The 

Keep those insults flying.  You're doing great.

32 minutes ago, level7 said:

so the choice is to not be able to see anything (narrow field of view) ... or not be able to see anything (dark cockpit) ? 😅

The cheapest choice is

https://github.com/AIRLegend/aitrack

The best choice is VR

The most expensive choice is

If you are viewing 80 degrees FOV (new XP default iirc) on a single small display, a dim cockpit when you aren't looking at it is the least of your "it doesn't look realistic" issues.

Edited March 13, 20233 yr by mSparks

2 hours ago, peroni said:

This is what he had to say about 12.01 and dark cockpits. Not sure if 12.04 is any better

Well, what I think I would do, faced with an issue like that, would probably be something like

 

4 hours ago, peroni said:

Oh wasn't this topic locked yet? 🙂

 

Here's a video from Michael Brown, owner of Xforce PC and good friend of Austin. 
This is what he had to say about 12.01 and dark cockpits. Not sure if 12.04 is any better

I follow him and saw this clip also. As we can see, it is a valid issue being brought up by many in the community and by some of the biggest fans of xplane. Why should the topic be locked if we are discussing a valid issue?

Somehow, I started coming across as a flat earther to others that did not like the conversation/my opinion. This is typically where things go off the rails. I've seen it many times. Otherwise, a civil debate back and forth is no problem to me, 1 page or 30 pages.

1 hour ago, mSparks said:

Well, what I think I would do, faced with an issue like that, would probably be something like

That shows that it can be improved by default.

Hi @GoranM,

I believe you have a few misconceptions regarding this issue. I'd be more than happy to discuss the details privately as it includes things I can't post here, but here are a few points that I'd like to address publicly. If you don't want to, that's fine too.

14 hours ago, GoranM said:

A statement like that requires numbers.  I can say the majority of people say that Boeing is a better company than Airbus.  Without figures and percentages, that statement is nothing but an opinion.  So yes.  You need numbers to back up your statement.

While I can't bring you numbers without doing a comprehensive survey, I think there are very good reasons to do believe that majority of the users are unhappy with the current implementation. Firstly, I think it's safe to assume that LR has already made such a survey, in the form of bug reports. Therefore, if LR thinks there is an issue affecting the majority of the users, it's safe to assume they would publicly acknowledge it. And indeed, the dark cockpits issue was mentioned in multiple videos from LR, including this video which is only one month old: 

Watch the chapter titled "Dark Cockpit" which starts at 0:30 - quoting Michael Brown himself, "But the bottom line is, most people feel this is too dark and it is being addressed", which is a good evidence that majority of the users are indeed unhappy.

14 hours ago, GoranM said:

However, if we want to actually prove the shadows are too dark, we need measurements from a light meter inside a real cockpit, at a certain time of day, and then compare that to measurements from within the sim.  THEN, we can have a truly productive discussion.  And if someone disagrees, the person with the concrete figures can say, "Don't have to believe me.  Believe the numbers."

This is why I said that I believe you have a few misconceptions regarding this issue. As I have mentioned multiple times in this thread, I am not claiming that there is an issue with the rendering side of things. LR already did measurements in real cockpits and found that their rendering engine is producing accurate results. So the rendered HDR scene is accurate and I never claimed otherwise. However this is only half of the story, and the issue is about how this HDR scene is converted into an SDR representation to be displayed on a monitor in a process called tonemapping. In other words, I and many others believe that the tonemapper used in X-Plane 12 is perceptually incorrect, because if the rendering side of things is accurate but the results are not reflecting human experience good enough, then there is only tonemapping left to blame.

The keyword here is perceptual, which is inherently subjective. As there is no way for me to show my experience of being in a cockpit to anyone (which would make such debates much easier), there is no way for me to bring you any numbers. However it shouldn't be dismissed just because it's subjective, due to two reasons. Firstly, while human experience is subjective, given we all share the same basic biology, it is likely that what people experience is more or less consistent between each other. Secondly, a significant part of rendering science already relies on human experience. In fact, a lot of tonemappers are evaluated by asking subjects to rank different tonemappers by how perceptually correct they are. In short, such decisions are usually made based on what the majority thinks is perceptually correct. And to determine what the majority thinks, there isn't any other way than trying to describe our subjective experience using analogies.

14 hours ago, GoranM said:

Some people are happy with the shadows.  I am.  Others are.  So what?  Why do we have to keep reading the same stuff from people who think they're opinion matters more than others.  If I'm happy how shadows look, why does someone else have to tell me, "How can you be happy??  No, you should not be happy.  They're too dark."

You have all the rights to hold the opinion that the cockpits look right. However, I need to point out that this issue disproportionately affects large dark-colored cockpits. While the CL650 with its light-colored cockpit looks perfectly fine, the default A330 and 737, Zibo 737, Toliss Airbuses, Flight Factor A320 and Rotate MD-11 look too dark. Therefore if your opinion is only based on the CL650 and/or aircraft with small or light-colored cockpits, you likely didn't even experience the issue I'm mentioning.

14 hours ago, GoranM said:

Just think, if, hypothetically, everything posted in here about the shadows being too dark is actually correct, think what LR could have done with this information!  Instead it's being posted here.

While I don't know about the other people in this thread, but me, Jean-Luc from RXP and several other people already presented this information to LR multiple times, with their response to us being that it is indeed an issue and it will be addressed. Just check the reply thread I made on Slack which has 149 posts. While me and Jean-Luc have disagreements on what the issue is caused by (tonemapper vs rendering), we all agree that the cockpits look too dark. In that post, I mentioned an approach which can help - measuring the dynamic range of the rendered HDR scene and blending between different tonemappers based on the dynamic range of the scene to take advantage of different tonemappers which work best for different dynamic ranges. I even showed them a test implementation which blends between X-Plane 12's Narkowicz ACES tonemapper and a gamma-like tonemapper depending on the dynamic range of the scene. While I chose the second tonemapper to reduce the contrast when the dynamic range of the scene is too high for the scene to be accurately represented by Narkowicz ACES (which is only the case for people who don't use VR or eye/head tracking) as I was too lazy to implement a local tonemapper instead of a global tonemapper, it could easily be replaced with a local tonemapper which would preserve local contrast. Such a local tonemapper would roughly look like @Murmur's composite A330 cockpit screenshot in page 10. By blending between Narkowicz ACES and such a local tonemapper, the current cockpit lighting would be preserved when the dynamic range of the scene is not too high and the scene would still be able look perceptually correct when the dynamic range of the scene is too high. Therefore VR and eye/head tracking users would not be affected as the dynamic range of the scene is usually not too high for their case.

Technicalities aside, you can see why I'm really frustrated now. Not only me and other developers spent several days discussing the issue and potential solutions with LR, but LR themselves acknowledged the issue multiple times and said that they will find a solution. Meanwhile we have people in this thread saying that I'm making a Mount Everest out of a mole hill, it's not a bug and I'm a chameleon, dismissing the points I make without even trying to understand them and even conveniently cutting out parts that change the context when quoting me. Just to prevent a possible misunderstanding, I'm not saying that you were doing these, I'm saying that they were done by some people in this thread. This short interjection aside, I don't think they constitute to a healthy discussion and I wish they drawed your attention as much as @efis007's posts did.

Edited March 13, 20233 yr by Biology