For the best result I also recommend to disable post processing filters in the User.cfg, otherwise the sharpening filter from MSFS overlays the one from GFE and it actually deteriorates the image.
Imo in general those filters dont add anything and can be disabled. But Enabled 1 must remain otherwise postprocessing is disabled completely and menus and flight planner will look wrong (ie there are other filters that are not reachable through the cfg file)
I had postprocessing disabled since MSFS came out but the thought just crossed my mind that only copying our settings for GeForceExperience in combination with default post processing might actually yield very bad results. We want the image to get perfect, not worse!
So if you follow the steps and your image becomes very grainy with a lot of extra black/white noise (like an old newspaper print) your sharpening is too high. It should look much improved, not worse, with sharpness vastly more realistic. I might make some comparison shots later on when I show my final settings.
All of this adjustment and tweaking. When I look at the comparison pictures above, I see a scene on one day and then the same scene on another day. Atmospheric conditions? Obsessive compulsive? Sheesh. I spend my time flying.
Not true. If you look at the comparisons on a decent IPS monitor, the original look like it’s watched through a pane covered with 10 years of nicotin residue and 5 dioptrines short sightedness on top.
But to each his own I guess.
No, reshade has even better performances than the Nvidia filter in most cases (but it is very similar).
Nvidia Freestyle filters and Reshade share the same shader engine, and filters used in Freestyle are the same as filters used in Reshade.
The difference is that Reshade is compatible with more apps, and advanced shaders (those that rely on the depth buffer) work better with Reshade and cover a wider range of games (this depth buffer access only work for offline apps, so it cannot work on MSFS).
Let’s say that for now, Freestyle is more user friendly (even if i prefer the Reshade UI personnally).
The fps impact depends mostly on those different factors:
the number of shaders you use
the complexity of those shaders
the resolution of the back buffer (the real resolution used by the game, not the one provided by TAA implementation, or DLSS). So if you choose a resolution in the graphic settings, and use TAA or DLSS, the backbuffer resolution can be different (mostly lower). This can also vary if you use downscale or upgrade techniques.
of course, the CPU and the GPU performances
We achieved great effort (especially Crosire) to reduce at max the impact of Reshade and its shader engine (and of course on Freestyle - for this particular program, Nvidia engineers also worked on it -). So if you use standard filters (adaptative or luma sharpen filters, color filters…), with a decent resolution or a native resolution, Reshade and/or Freestyle should not have a big FPS impact.
In Reshade, you can monitor this impact in the statistics tab.
Well, if you’re talking about my OP, the scene were shot no more than 1 minute apart, with the only difference being settings off vs. settings off. And the tweaking took me, eh, maybe all of 3 minutes? Well worth it to me to have a much more accurate white balance, with a good level of sharpening dialed in, too.
Not what I meant. The pictures look as if you took pictures of the same scene but on a different days where atmospheric conditions were slightly different.
My point is that you are applying exposure (photography) concepts to a weather simulation. A matter of personal taste, not a sim issue.
Also can you point to which specific images he posted that show different weather conditions? Even the clouds resemble each other like 1 egg to another.
OK, you guys are playing semantic games now. If you aren’t, please reread, slowly, what I posted.
Just in case, I’ll try one more time: the difference in GFX settings (clarity, sharpness, lighting) look as if you took the same picture on different days, i.e. first day the air was laden with moisture so everything is blurred a bit, but the next day the air was dry and crisp, so everything looks clear and detailed.
In other words, everything is relative. If you like your visuals overly sharp focused, have at it.
In fact I went to great lengths to set the same weather and time as in real life, fly from my home area and compare it with what I see, including the color and brightness of the sky etc.
Default in general is several hours (in terms of evenings) too dark and has too much green which results in a constant yellowish tint. It also has a couple of dioptrine short sightedness built in (that’s why you can’t read instruments before the right sharpening filter is applied and afterwards you can)
Of course, normally computer generated images should not be sharpened. But there is some sort of softening filter applied by default that cannot be switched off and it is currently the only way to correct it.
I also always found it weird that another filter in the postprocessing suite is called “Sharpening”. That is not the way to do it, first to make an images blurry to reduce artifacts and then sharpen it again.
I would say MSFS without those filters would look amazing but it would bring out other problems with the graphics. That’s why games are so relentless at making you use depth of field, as well as other techniques that reduce the image to a blurry background.
I spent considerable time to optimize my settings (for my own taste of course)
I must say that there were two quite different versions which achieved the same. So a “perfect” setting don’t exist. I think it would however help to start with settings that isn’t too extreme and work from there by making small changes.
Brightness/Contrast
Exposure: 0-10% (0% is actually more realistic but I like 10% more as it is brighter when I look at the sun)
Contrast: 15%
Highlights: -60%
Shadows: -20%
Gamma: 0%
in addition to that I use Sharpen
Sharpen 60%
Film Grain Reduction 100%
but the sharpen filter is even more subjective and should be adjusted separately, ideally after you are finished with the coloration and lighting of the scene
It is what I use for bright sunlight. I noted that another setting is superior during sunset/sunrise, so I think I will use a different one for minimum light conditions.
Always look at the full size originals to compare, otherwise a lot of the original detail is lost.
Your penultimate post above says you weren’t applying photographic concepts to sim images but then your last post says your settings were set by “personal taste”. Uh…
Can you understand that, relatively speaking, image quality could be explained as well by weather simulation rather than image quality settings? It could also be the monitor.
I have none of the issues you describe. So what’s the difference between our systems - or our expectations?
Tbh I see no point in further discussing in the direction you are constantly trying to take me. The whole “applying photographic concepts to sim images” doesn’t make a lick of sense to me. We are adjusting the images so that colors eg white match what you see irl, and we exchange the settings we think are best, and if the images you see are already better for you then why do you have to write it 100 times.
I have a degree in computer science and pretty long experience in image processing, if you are trying ot insinuate that I do this for nothing or that TN and IPS monitors are equivalent, then forget it. An image professional first calibrates his monitor and then, if an application does not produces proper whites, blues, reds etc, he matches the application.
Also as you can see in the images I showed (or should be able to see) the white in the second image is actually a dirty yellow and if you dont see it, then I would like to know the model and panel type of you monitor.
As to the “personal taste” I never claimed my settings or results were perfect, only that they are a vast improvement over the default imagery.
I like your idea of calibrating the monitor. Professional calibration tools are, I suspect, prohibitively expensive. I’d be grateful if you’re able to recommend your choice of a calibration procedure available to the home user?
or it will shut down the game)