Veegie

I appreciate your patience!

HDR is a tricky space to execute. Even when it is "correct," it can still subjectively feel washed out, compared to an SDR image with an out-of-range gamut expansion that has become increasingly common as manufacturers try to get their image to "pop."

There was a lot of foundational visual technology we built as part of this effort that should, over the long term, benefit both SDR and HDR visuals. But for this initial release, we wanted to focus on providing a friendlier, more flexible HDR experience. Using DIM isn't very fun when HDR required exclusive fullscreen.

Yeah that is generally the reason.

Lights that are part of the player sandbox can appear anywhere in our world, so we can't carefully cater the scenes GPU cost towards them, so they have to render quite cheap, achieved through both limiting their size or the complexity of the lighting effects — thankfully these lights tend to be temporally quite rapid so you can get away with it. Since we know that combat won't be present around flashlight areas we can use more expensive lights that are actually casting dynamic shadows on their surroundings.

Generally speaking the cost of a light is proportionate to the amount of pixels that it hits on screen, rather than the shape of the light (omnidirectional, spotlight, line, etc.). Things like shadow casting are so costly that we enforce a total number of shadow casting lights that can be on screen at a single time, this is why you won't ever find the effect on any lights that are part of the player sandbox.

Hi there! I'm the person you can blame for letting these issues leave our walls and get into the wild.

The current items I'm aware of that I would like to address:

• SDR brightness setting affects the HDR image, requiring you to sanitize it to neutral before using HDR
• The HDR black point is setting is functionally unusable and should be left to its default setting, ignoring the calibration
• Certain combinations of white/black point values can cause subtle colour inversions in the very dark range of values

Some of the above problems tend to be triangular dependencies, making them less than trivial to address.
I am still hopeful we can find a way to responsibly allocate the technical resources towards these efforts, it's just a matter of finding the right timing so that we don't starve other critical efforts of the same required resources.

In the meantime we have begun to standardize our minimum level of bright...

I love this. Thank you.

I think I changed it a few years ago when we were implementing PBR between D1 and D2 because it had some nice curvy surfaces to showcase reflections and improved normal quantization.

You know what's up.

They definitely take a while, especially when trying to find best representative comparisons that are fully deterministic between captures and settings changes, but also don't show any spoiler content.

I think I spent a week compiling all those shots (and technical info with approachable phrasing), ensuring that they were as close to pixel perfect as possible.

It was borne of a frustration I had reading various guides where it would be totally different lighting conditions between a comparison of something like antialiasing options.

Glad you enjoy! They're a labor of love.

Neat! What would your expectations be for colorblind support on this differentiating aspect?

One wrong parenthesis caused quite a bit of damage. Thanks for letting me know!

That would be cool! Currently that isn't on the trivial side of things to implement due to aspects¹ of the way the shader system works. I think right now if we were to cut that deep in the code we would prioritize some of the bigger fish to fry But I honestly do appreciate hearing your feedback ^{(I think I saw someone else post this the other day, too}) about things you'd like t...

As someone that has worked on the gear rendering system for many years, I appreciate the feedback on the subject and have personally seen this given many times — I browse this subreddit daily like many of my coworkers.

I can't say that it is always, but very often these areas are the result of current technical limitations, artist time to handle more time-involved workarounds of said limitations, and concepted designs that continue to push into new territory to give you all exciting new offerings.

There are some fairly challenging technical hurdles to overcome the combinatorial complexity that Gear tends to carry that has made it harder than it may appear to do this. We're tackling *a lot* of things this year as you're all now starting to see, this means that it can be harder to dedicate resources towards things like this until we're safely on the other side of some of these investments.

I don't intend that to come off as an excuse, but rather a transpar...

Thanks for the report! I'll investigate this.

We talked about some of our development strategies here if you are interested.

EDIT: Looks like GDC website is having some problems. You can view a Google cached version ...