Viewspeed v2

[seventhfrost]

20 minutes ago, Drimzi said:

A lot of my desktop to 3D 'muh feels' testing as been from playing a song in Shoot the Beat, and then going back to the desktop when the songs finished.

I've been considering trying out comparing shoot the beat to osu today. maybe you should find someone who can do osu maps blindfolded and just see how it goes lol.

Anyway after just a bit of using 325 DPI, it really is nicer for me, both ingame and on aimbooster. just barely overflicking, so it's a shame i can't just set 320. might consider a mouse where i can. but since I'm just returning to it, I can't tell anything about matching feels, really. it looks similar, but that's not good for much. I am curious as to why you wanted to know about people trying this, though. 

 

Edit: I should note that 325 with this most recent formula put me between 350 distances of two of the other formulas I liked, so that explains some of the bias i felt.

Edited December 15, 2017 by seventhfrost

[Drimzi]

Edited January 28, 2018 by Drimzi

[seventhfrost]

2 minutes ago, Drimzi said:

I don't think it will work like that, they won't be able to do it blindfolded. Your brain should interpret the rotational speed vs the cursor speed automatically and adjust the mouse distances as necessary.

Yeah that part was more a joke than not, but this is pretty much how I was understanding viewspeed to work, so I'm glad I was at least understanding the core concept. 

Though if you're right about the part that matters to how things 'feel' being the middle that's less distorted, is there a way to scale with that instead of scaling with the fov overall? I believe those change at two different rates, right? Or do you have to keep tweaking the current formula with this method of scaling until it just matches closely to how much of the screen isn't distorted?

[Drimzi]

Edited January 28, 2018 by Drimzi

[Cocyx Skeleton]

Hmm, I don't mind it feeling too slow personally -- I can just adjust over time. I just want the most possible accurate conversion from desktop --> in-game mouse movement.

[Drimzi]

Edited January 28, 2018 by Drimzi

[Drimzi]

Edited January 28, 2018 by Drimzi

[potato psoas]

8 hours ago, Drimzi said:

One thing I have noticed is that for sensitivities to feel the same for me across FOVs, the feeling needs to remain the same within the non distorted area of the screen.

If you go up to 170 FOV for example, aiming straight ahead at the tiny little targets feels like it's suppose to, the rest of the screen, the highly stretched and distorted mess, is vastly ignored. Zoom in to a very low FOV, and you now have the whole entire screen useable, and that same feeling is expressed to the full screen. As the FOV increases, that window where your sensitivity is matched just moves further and further away, shrinking into the distance, while the rest of the screen, the distortion, is ignored.

This is what viewspeed does. It is just 1/sin(x/2) after all. If you maintain a constant circle size, then the distance the chord is from the diameter is the scaler for the sensitivity. This accounts for the distortion because it essentially ignores it in practice. I found that viewspeed v2 did not work so well within the extreme ranges though. With the addition of the diagonal multiplier, which will be 1.414 at 0, and 1 at 180 FOV since the horizontal, vertical, and diagonal all converge at 180 FOV, the extremes are fixed and no FOV feels odd. I'm just plugging in random things so I don't know why this works.

If you do a static monitor distance match, so that the 2D distance is the same regardless of FOV, then it is not accounting for the distortion. The distortion is included within the measurement. Don't judge whether these formula are too fast or too slow by basing the 2D distances across FOVs, that will only match with a static monitor match %.

With this formula, aiming within this red box feels the same as aiming within the whole screen at a low FOV where there is minimal distortion. If you concentrate properly at what your looking at, you pretty much ignore everything outside that red box. This is why so many CS:GO pros use 4:3 black bars, as it just crops off the distortion.

 

For testers, I encourage you to make use of the script above to generate all of the CS:GO sensitivity values, and to go into a workshop map and try a wide range of FOV.

This is honestly using the exact same principle that I figured out as well, except it gives you more customisation over what you think the useable window is. Mine just uses the unit circle to adjust the size of the window based on FOV:

18 hours ago, potato psoas said:

Just imagine the monitor match value is where the virtual window's edge is (as a percentage of the entire screen).

7 hours ago, Drimzi said:

This is why matching 2D and 3D is so difficult and there is no known factual method yet.

There won't ever be unless the game itself accounts for distortion and in that case 100% MM would be the perfect method.

Edited December 16, 2017 by potato psoas

[potato psoas]

I'm seriously beginning to think there's no point in matching 2D to 3D. I'd rather just stick to 100% monitor match. And convert my 2D to 90 FOV or something.

[Bryjoe]

30 minutes ago, potato psoas said:

I'm seriously beginning to think there's no point in matching 2D to 3D. I'd rather just stick to 100% monitor match. And convert my 2D to 90 FOV or something.

But 100% monitor match simply is as far as possible from accuracy in the center of the screen. Are you referring to a different 100% monitor match than exist in the live calculator?

In the calculator, it actually was originally 50, as being in the middle is good for flicks and good for aiming. It was changed to 75% because many games already use that method and most people seem to like it. I have heard some people like 0%, but almost no one prefers the 100% monitor match, it feels bad unless you are playing a game where pinpoint accuracy in the center of the screen doesn't matter. It is hard for me to keep up with your implementation versus Drimzi's, so I apologize if you're referring to 100% monitor match implemented in a different way.

If you truly think the current calculator implementation of 100% monitor match is the best please explain why.

Edit: Drimzi's seems to be close to the original Viewspeed, so that can't be the same as 100% monitor match, can it?

Edited December 16, 2017 by Bryjoe

[potato psoas]

4 hours ago, Bryjoe said:

But 100% monitor match simply is as far as possible from accuracy in the center of the screen. Are you referring to a different 100% monitor match than exist in the live calculator?

In the calculator, it actually was originally 50, as being in the middle is good for flicks and good for aiming. It was changed to 75% because many games already use that method and most people seem to like it. I have heard some people like 0%, but almost no one prefers the 100% monitor match, it feels bad unless you are playing a game where pinpoint accuracy in the center of the screen doesn't matter. It is hard for me to keep up with your implementation versus Drimzi's, so I apologize if you're referring to 100% monitor match implemented in a different way.

If you truly think the current calculator implementation of 100% monitor match is the best please explain why.

Edit: Drimzi's seems to be close to the original Viewspeed, so that can't be the same as 100% monitor match, can it?

Well I did explain. The distribution of circumferential rotation as it is projected onto the 2D image becomes more equal as you approach 0FOV. And when the distribution is equal the best monitor match to use is 100% as 100% flattens the arc so that it is the same length as the chord (your monitor). And therefore 100% would be perfectly in sync with the 2D sensitivity. But because there is distortion the circumferential intervals are squished at the center and stretched at the edges so the higher the FOV the slower it feels at the center, at least compared to 2D. But if you just match your 2D to your most common FOV (90) and convert from it to higher FOVs using 100% then you reduce a lot of the slowdown at higher FOVs, while still allowing a lot of the screen to be useable.

The 50% approach also makes a lot of sense but you still want your monitor match to scale to 100% as you approach 0FOV. So you could try that too, scaling from 50% at 180FOV to 100% at 0FOV.

FYI the FOV with middle ground of 50% is 120:

This is probably where I would have the cut off for my FOV range.

Edited December 16, 2017 by potato psoas

[Drimzi]

Edited January 28, 2018 by Drimzi

[potato psoas]

9 minutes ago, Drimzi said:

Monitor match in the calculator is based on desktop horizontal and actual horizontal fov. All my methods use 1:1 aspect ratio to avoid any issues with matching to a point that's not in the center of the screen. I already made a vertical monitor match formula and it felt okay, but the new one is much better. My latest formula is diagonal monitor match, using the 1:1 diagonal fov and the diagonal desktop length. Since it is 1:1, all of the matching points are going to be fairly close to the center of the screen. The way it works is it will take your half of your 1:1 fov, since you only concern yourself with half of the screen, find the diagonal, and match to that point on the monitor. So you match at the max frame size possible for every 1:1 fov, ignoring all the distortion gained from the extra fov granted by wider aspect ratios. As you approach 0 FOV, the vert/hor/diag all converge, it should be the same as 100% monitor match in the calculator, regardless of the 1:1 aspect ratio. Try setting 1 FOV in the calculator and in my formula, match the bases, and they will have the same result.

Yeah we definitely need to scale using an aspect ratio independent formula. I'm just trying to figure out what is the best way to scale.

Edited December 16, 2017 by potato psoas

[potato psoas]

2 hours ago, potato psoas said:

I'm seriously beginning to think there's no point in matching 2D to 3D. I'd rather just stick to 100% monitor match. And convert my 2D to 90 FOV or something.

This method actually gives very similar results to Viewspeed v2 in the calculator.

Edited December 16, 2017 by potato psoas

[Drimzi]

Edited January 28, 2018 by Drimzi

[potato psoas]

5 minutes ago, Drimzi said:

Honestly I don't think you want 100% monitor match at 0 FOV. Low FOVs will be impossible to use. You wouldn't even be able to rotate without doing 1000 swipes of the mouse.

I just tested my formula, it is 100% monitor match towards 180 FOV. It physically took roughly the same distance to move my mouse diagonally to the bottom corner of the screen ingame and with the cursor at 170 FOV.

The 0 FOV is the exact same as 100% in the calculator, same 360 distances, but in reality it's not matched to the desktop. If it was, it would actually be incredibly slow. Try it yourself.

But you can't try 0FOV for yourself because 0FOV = 2D. In a real game situation you would never use a FOV lower than like 5. Battlefield 4's 40x scope is like 3 FOV. And low FOVs are meant to be slow. cos(3/2) = 0.999657324975557. So the monitor match at really low FOVs is very very close to 100%.

[potato psoas]

So I was thinking about how when we use higher FOVs we aren't using higher FOV for precision we are using it for speed, and speed doesn't require that much accuracy. So distortion wouldn't affect aim too bad at high FOVs if you're just trying to get a fairly similar feel. But the lower the FOV the more precision we require, which is why we ADS - for precision. And funny enough that is when the distortion affects it the least and it maintains more consistency with the 2D sensitivity. So I think everything ends up working out in the end.

This makes me think about different games. Games with different zoom levels you don't need to worry about but games like CS:GO with only one high FOV you may need to have a hipfire sensitivity and an ADS sensitivity. Even though you're not actually ADSing what you do is bind a DPI adjustment to your right click to help with precision. But the DPI adjustment has a lower monitor match so that it is not as slow.

Edited December 16, 2017 by potato psoas

[iBerggman]

Do any of you know if there's a way to make a script either in the csgo console or maybe logitech lua scripting that would randomly cycle between different fovs and sensitivities on it's own? I like testing sensitivities on the aim training maps from the workshop but I find manually switching the fovs distracts me too much from the aiming to really pay attention to how it feels. I guess the next best thing would be to have a button that toggles through a list of different fovs/sensitivities and set that on a macro so I'll start with that.

Edited December 16, 2017 by iBerggman

[Drimzi]

Edited January 28, 2018 by Drimzi

[Drimzi]

Edited January 28, 2018 by Drimzi

[Cocyx Skeleton]

1 hour ago, Drimzi said:

For 1920x1080, 6/11, this formula would equal 2.5 sensitivity in CSGO.

Woah, exactly 2.5? 

[Drimzi]

Edited January 28, 2018 by Drimzi

[Drimzi]

Edited January 28, 2018 by Drimzi

[seventhfrost]

I'd definitely be willing to try a slower formula. I'm not sure how I feel about this fixed version of the most recent formula. It feels like it's scaling better between 90 and 81 fov with this fixed version, though before it was the lower fov (ADS) that felt like it was too slow, while the higher one (hipfire) felt good. Now they both feel more similar to eachother.

I'm getting the feeling that it's all a bit fast in comparison to 2D, currently, but I don't really trust that since I've only been at this DPI for about a day now.

[Drimzi]

Edited January 28, 2018 by Drimzi