GUIDE: how to improve performance in IL-2 Korea (VR, Quest but not only)

[Individ]

Also make attention on this, if u have 3-4 joysticks.
https://gist.github.com/Kost9lbrik/c697a5b045186a707ad2a529bd5d403d

[YoYo]

The Anti-Aliasing section has been updated to include FSR AA, one of the more interesting anti-aliasing methods for IL-2 Korea as well. It has a negligible impact on FPS, yet resembles the resource-intensive DLAA.

[Husar]

9 hours ago, YoYo said:

The Anti-Aliasing section has been updated to include FSR AA, one of the more interesting anti-aliasing methods for IL-2 Korea as well. It has a negligible impact on FPS, yet resembles the resource-intensive DLAA.

There are also less resource-intensive presets for DLAA, such as K and F, but they produce more noticeable smearing than the in-game FSR AA.

 

[Antonovomatic]

"FOV Tangent Multiplier" isn't doing what you suggest here - it narrows the actual FOV, as if you are applying a slight zoom.

On Quest Pro and 3 there isn't even any significant "unused" display area that could be discarded, because of the much improved optic stack over the Quest 1/2.  That's why the jump in clarity and fidelity was so high for even the Pro despite barely any display pixel increase over the Q2 and more than solely res increase would lead one to expect for the Q3.

Edited September 27 by Antonovomatic

[YoYo]

1 hour ago, Antonovomatic said:

"FOV Tangent Multiplier" isn't doing what you suggest here - it narrows the actual FOV, as if you are applying a slight zoom.

On Quest Pro and 3 there isn't even any significant "unused" display area that could be discarded, because of the much improved optic stack over the Quest 1/2.  That's why the jump in clarity and fidelity was so high for even the Pro despite barely any display pixel increase over the Q2 and more than solely res increase would lead one to expect for the Q3.

Sorry, but no and I cant agree. It works exactly as described and has a noticeable impact on FPS. Source (if you want to call into question the results of the work done by Meta's developers):

https://developers.meta.com/horizon/documentation/native/pc/dg-debug-tool/

The FOV Tangent Multiplier in the Oculus Debug Tool (ODT) is a setting that allows you to limit the rendered field of view (FOV) in VR headsets. It acts as a multiplier for the default field of view on both the horizontal and vertical planes.

How does this work in practice? The default value in ODT is 0;0 (which corresponds to a default setting of 1.0;1.0, or 100% of the full field of view). The value format is: Width;Height (e.g., 0.85;0.85). Lowering the value (e.g., to 0.85;0.85) causes the graphics card to render an area that is 15% smaller both horizontally and vertically. Instead of a full image, black borders - often referred to as the "ski-goggle effect" or "vignetting" - will appear at the outer edges of the field of view. Why does this yield a significant performance boost? Due to the lenses in VR headsets and the anatomy of the human face, a substantial portion of the rendered image at the edges of the display never actually reaches your eye. By lowering the FOV Multiplier, you stop wasting GPU power on rendering pixels that are not physically visible.

You can clearly see the narrowing effect, and you can narrow the field of view to the point where a black border appears halfway across the image. For me, that’s an extra 5–15 FPS (depending on the scene), and there’s no arguing with that evidence. The thing to note is that everyone can adjust this to their own preference; for me, 0.75 works well and doesn't obstruct my field of view. Below that value, I start to see the frame. Since the FOV reduces the rendering area, this is also reflected in the resolution; normally I would have around 3200 pixels with upscalling (PPD), but with the FOV setting, I have about 2700 and this does not come at the expense of sharpness. That difference of 500 pixels represents a huge saving. I’ll also add that there are quite a few videos on YouTube about FOV in ODT for example, in the context of MSFS2020/24, featuring experts who are even better than I am. Cheers.

Edited September 27 by YoYo

[Aapje]

Keep in mind that how much of the screen you see depends on your face shape and the size of the spacer you use (third party facial interfaces can allow very good fine tuning to minimize the distance). For me the limit is quite high: 0.9; 0.8 

[YoYo]

That's right; it also depends on the gasket being used and how close your eyes are to the lenses, so everyone needs to check for themselves how it works in their case. The aforementioned 0.75 can serve as a starting point.

[VISHNU]

Thanks for this.    It was very helpful, and it did improve my FPS.  THANK YOU!

[Antonovomatic]

On 9/27/2026 at 4:19 PM, YoYo said:

Sorry, but no and I cant agree. It works exactly as described and has a noticeable impact on FPS. Source (if you want to call into question the results of the work done by Meta's developers):

https://developers.meta.com/horizon/documentation/native/pc/dg-debug-tool/

The FOV Tangent Multiplier in the Oculus Debug Tool (ODT) is a setting that allows you to limit the rendered field of view (FOV) in VR headsets. It acts as a multiplier for the default field of view on both the horizontal and vertical planes.

How does this work in practice? The default value in ODT is 0;0 (which corresponds to a default setting of 1.0;1.0, or 100% of the full field of view). The value format is: Width;Height (e.g., 0.85;0.85). Lowering the value (e.g., to 0.85;0.85) causes the graphics card to render an area that is 15% smaller both horizontally and vertically. Instead of a full image, black borders - often referred to as the "ski-goggle effect" or "vignetting" - will appear at the outer edges of the field of view. Why does this yield a significant performance boost? Due to the lenses in VR headsets and the anatomy of the human face, a substantial portion of the rendered image at the edges of the display never actually reaches your eye. By lowering the FOV Multiplier, you stop wasting GPU power on rendering pixels that are not physically visible.

You can clearly see the narrowing effect, and you can narrow the field of view to the point where a black border appears halfway across the image. For me, that’s an extra 5–15 FPS (depending on the scene), and there’s no arguing with that evidence. The thing to note is that everyone can adjust this to their own preference; for me, 0.75 works well and doesn't obstruct my field of view. Below that value, I start to see the frame. Since the FOV reduces the rendering area, this is also reflected in the resolution; normally I would have around 3200 pixels with upscalling (PPD), but with the FOV setting, I have about 2700 and this does not come at the expense of sharpness. That difference of 500 pixels represents a huge saving. I’ll also add that there are quite a few videos on YouTube about FOV in ODT for example, in the context of MSFS2020/24, featuring experts who are even better than I am. Cheers.

That's an old article written before the optics stack was completely redesigned, let alone the single display used in the Quest Pro vs split-dual on other models.
Having experienced the described "black border" effect in the Q2 when changing the setting, I can firmly say it does not behave at all similarly on the QP.  It narrows/zooms the entire perspective.  There is no "gasket"/face interface acting as a standoff in the QP to prevent me from seeing the very edges of the display area in the lenses.  The "wasted" display area is already miniscule compared to the fresnel lenses of the Q2, and you *really* have to get your eyeball up to the lens surface to even find it.

Edited October 1 by Antonovomatic

[YoYo]

I disagree completely about what you wrote; it’s not just about theory and facts, I’m speaking from my own experience with how FOV works. It not only affects the visibility of the area, but the impact is clearly evident in the increase in power and reducing rendering times (not only in IL-2 Korea but also in MSFS, DCS and IL-2 GB) and the numbers don't lie, so there is really nothing to debate here for me, but fine, let’s agree to disagree.

Edited October 1 by YoYo