FPS - is it everything?
-
And for a rig with ample perf (over say 90fps) and a 60hz fixed-refresh monitor, I recommend you try Triple Buffering – with a frame-rate cap in the NVidia or AMD control panel, set to about 30% over your refresh rate. (So, for a 60hz monitor, cap at about 80 or 85 fps.)
Thanks for this recommendation. I thought I’d give it a try (under Windows as life is just easier that way). Ultra Low Latency on. Triple Buffering on, frame rate cap at 85fps, Vsync On. BMS set to no Vsync, no triple, quality 1 or whatever the setting is called.
I didn’t notice any real difference EXCEPT I would occasionally get a little stutter - most noticeable when looking sideways (e.g. looking at the runway over my left shoulder in overhead pattern).
Is Vsync On part of your recommendation, or without?
To be honest this is just tinkering ‘cos, well, it’s fun - I’ve been a strictly Mac and Linux man for 20+ years (and the one regret was leaving FlightSims behind, Falcon 4 in particular) , so I’ve never had to worry about GPU settings, gaming etc until I hit BMS in November.
-
Is Vsync On part of your recommendation, or without?
V-sync OFF. Part of the premise of triple-buffering is that there will be no tearing. (There are 2 back-buffers and 1 front buffer… so it is always safe for the gpu to render, unconstrained, into one of the back-buffers – while the other one stands ready to flip to front when needed.)
If you have NVidia graphics, setting v-sync to ‘Fast’ in the nvidia control panel, is a way to enable triple-buffering for all games. (Then leave tb=off and vsync=off, within BMS.)
I wrote a much longer bit on triple-buffering, here, a few weeks ago … because I was learning about all this myself.
https://www.benchmarksims.org/forum/showthread.php?40961-Triple-Buffering/ -
Why?
Hard to explain in words and I haven’t found a good article with diagrams, or youtube video… I’d maybe make one if I had a little bit more knowledge and expertise in this field. (I may sound like I know what I’m talking about but I do not. I’m a veteran software engineer, but almost no experience when it comes to graphics, DirectX and game development… I just like learning how things work.)
Here’s my earlier thread researching this:
https://www.benchmarksims.org/forum/showthread.php?40961-Triple-Buffering/Depending your rig you may not need a fps cap … but a modern, mid/high-end gaming rig that could conceivably produce 200+ fps on cpu and gpu, to feed a 60hz monitor, is a huge waste of energy. Unless it’s winter in your hemisphere and you need to heat your room. lol
So a fps cap is probably a good idea … then the question is, how high to set the cap?
The key metric I try to focus on, to optimize both latency and smoothness, is the average “age” or “staleness” of a frame … ie. how many milliseconds from the time your joystick was sampled, to the time the result appears on-screen. I want that average-frame-age to be (a) consistently low, but also (b) actually somewhat randomized (within a low range) so that occasional frametime hiccups don’t become obvious, noticeable glitches in smoothness.
(And for whatever reasons, BMS regularly produces a lot of ~5-10ms frametime hiccups.)
So, for those reasons, I recommend avoid capping at any even multiple of the refresh rate… go with an odd number like 79 or 83 (for a 60hz monitor) so the ages of adjacent frames end up highly randomized. You could instead opt to go a little higher than 1.5x… say 97 or 103… it’s a tradeoff of dropped frames (energy wasted) vs driving down the average-frame-age by a couple milliseconds.
Just my 2c, every system is so different … and honestly I haven’t tested with U1 much yet, so caveat emptor!
-
Anybody that would want to set the correct FOV, which obviously depends on the screen size and set resolution, should take the formation checklist and open “CANOPY ANGLES CUES”. Then set the FOV in Falcon BMS accordingly to match the angles given in the checklist with measured between the line of sight to the gun cross and the given cues on the screen.
Thanks airtex2019 for the hints on the graphics. I am waiting for mine card to come (position 260 in the queue for it in mine store). I will use your experience as it comes.
-
Anybody that would want to set the correct FOV, which obviously depends on the screen size and set resolution, should take the formation checklist and open “CANOPY ANGLES CUES”. Then set the FOV in Falcon BMS accordingly to match the angles given in the checklist with measured between the line of sight to the gun cross and the given cues on the screen.
Thanks airtex2019 for the hints on the graphics. I am waiting for mine card to come (position 260 in the queue for it in mine store). I will use your experience as it comes.
Do you mean that the angular FOV that your monitor takes up in your vision should be matched to the angular FOV in game?
Nikolai
-
Do you mean that the angular FOV that your monitor takes up in your vision should be matched to the angular FOV in game?
Nikolai
I think it’s a measured angle (on your screen) from gun sight cross to canopy cue as marked.
-
On almost every first or second flight lesson I specifically will fly close and parallel to an U.S. Interstate at about 1000’ AGL. I have the new student look down and point out
that even in our piston single plane we are passing (hopefully) all the traffic. This usually the first time it hits home while we are going slow on airplane scheme of things, it is still normally at least 110 mph, for most it is faster than they have ever controlled a vehicle. But there is near no sensation of speed, it feels slower than a car. The scale speed is different than you normally experience due there being no close objects to judge the rate of closure. Objects in your peripheral vision have more of an effect on speed sensation since they always move through your field of view faster. But it doesn’t take a very high altitude at all for your peripheral vision to no longer have almost anything in it besides for sky.On the flip side, the fastest I have ever felt in a plane is a mock crop dusting run when I was doing my aerobatics class with a former crop duster. Again, it was slow on the overall airplane range of 100-120 knots (I don’t think I ever looked down at the airspeeds for the whole two passes of the 3/4 mile field), but we where at 3 feet of altitude and had a lot of things whipping past like tree lines to the sides, cut corn stalks, irrigation crawlers, and also a very disconcerting set of of power lines at the field edge in front, at least until I realized we could fly under them without a problem (but at just a few seconds after thinking that and when I felt entirely too close, the instructor pulled up and we still missed the lines). There’s a reason why flying through canyons feels faster than even flat ground though because your FOV constantly has things moving through it.
Yes, for flight sims FOV and objects movement are key to speed sensation but you should only ever feel it down really low, and also down there in RL the stereoscopic effect of your vision also finally comes back into play with things somewhat (which BMS doesn’t do yet, but it really isn’t that much of the equation still). I remember playing Aliens vs. Predator when it first came out and the fisheye FOV effect of the Alien’s view made it feel so fast, as will fisheye lens camera in Youtube videos will do.
So yes, cranking your FOV up will make you feel faster, but doing that alone on a narrow monitor with distort many other important things. I agree with Dee-Jay, use a FOV calculator and make your default BMS FOV match fairly closely to the real one. I love to fly in sims tree top level on a triple monitor setup, close to a large 4k TV, or in other sims, my Pimax 8k X. The extra FOV does help feel faster at appropriate altitudes and helps a few other things like flaring to land and formation flying, but most of the time it shouldn’t seem any faster. I feel like the speed sensation due object movement is sufficient in BMS. That portion of the speed sensation doesn’t take much, a fairly featureless desert will apparently still feel somewhat fast for the same given speed and altitude due to the still noticable terrain variations and sagebrush etc.
As for FPS, I grew up playing F-19 Stealth Fighter, Aces Over the Pacific/Europe, and then Falcon 3.0 so I feel like my brain is wired to deal with a little choppiness with flight sims I can still feel pretty fast at 30-40 fps in VR even down low without big stutters.
-
Do you mean that the angular FOV that your monitor takes up in your vision should be matched to the angular FOV in game?
Nikolai
Yes, if you want the angles to be the same in BMS comparing to RL. Then the immersion of speed should be similar, especially if you cover more of the area around you and not only infront of you (due to what others pointed above). This means bigger screens or your nose more near to the screen and FOV matched with it.
Wysłane z mojego SM-A520F przy użyciu Tapatalka
-
So yes, cranking your FOV up will make you feel faster, but doing that alone on a narrow monitor with distort many other important things. I agree with Dee-Jay, use a FOV calculator and make your default BMS FOV match fairly closely to the real one.
At the risk of turning this into yet another “what’s the correct FOV?” thread… the various FOV calculators reckon 59 degrees for my screen and measurements (distance from screen). Which is bonkers. I can only see the HUD at that FOV.
-
At the risk of turning this into yet another “what’s the correct FOV?” thread… the various FOV calculators reckon 59 degrees for my screen and measurements (distance from screen). Which is bonkers. I can only see the HUD at that FOV.
These calculations are nonsensical. We are viewing a projection of a faux 3d space onto a flat (or nearly flat) screen, which is typically fairly narrow view of the world, unless you have one of those ultrawide monitors or a 3 monitor setup. It’s not going to be completely 1:1 with real world. You need to set it up to what feels right, which is just a matter of some experimentation. Change FOV by increments of 10 degrees until you find the spot that seems the best, then change 5 degrees both ways around that. That should get you there.
-
At the risk of turning this into yet another “what’s the correct FOV?” thread… the various FOV calculators reckon 59 degrees for my screen and measurements (distance from screen). Which is bonkers. I can only see the HUD at that FOV.
Thats probably pretty accurate, too.
Im with zerg - you should use a FOV which looks good to you, and which you are comfortable with. For me, that used to be one click out from the default. When last I was flying regularly, I was just using the default and that felt fine.
Using the correct FOV will likely leave you feeling like you are looking through binoculars.
I suggest that BMS has a pretty damn good sense of speed already. More details on the ground would help with that, so I am looking forward to new terrain. At 30,000, it has the correct sense of speed. Even at lower levels like 5,000ft, it has an accurate sense of speed I think.
Sure makes the GA aircraft I fly feel slow in comparison…
-
I suggest that BMS has a pretty damn good sense of speed already. More details on the ground would help with that, so I am looking forward to new terrain. At 30,000, it has the correct sense of speed. Even at lower levels like 5,000ft, it has an accurate sense of speed I think.
I don’t disagree! I started the thread only because my “speed” differed or appeared to differ from those of, say, Lorik on YouTube. Not because I thought there was a problem. I’m very happy with it!
-
the various FOV calculators reckon 59 degrees for my screen and measurements (distance from screen). Which is bonkers. I can only see the HUD at that FOV.
I think there are 2 wholly different approaches … pitbuilders trying to create a physically immersive and dimensionally accurate space. And those of plebes with an ordinary desktop PC and a single monitor… maybe ultrawide if we’re lucky.
I want what’s on my screen (no matter how big or small it is) to represent what I’d be able to see, irl.
(What is a typical peripheral vision fov with a fighter helmet on? I’m not sure.)
I just do the thing where I hold up my hands next to my ears, and slowly move them inward, to determine my effective peripheral vision – and that seems about 120. So, I set BMS to that, and it feels right… yes especially eg. flaring to land. And at 3440x1440p, I can still (barely) read the MFDs and DED etc.
-
At the risk of turning this into yet another “what’s the correct FOV?” thread… the various FOV calculators reckon 59 degrees for my screen and measurements (distance from screen). Which is bonkers. I can only see the HUD at that FOV.
is simple, uman eyes fov is 130-135º vertically and 200-220º horizontally. To know if what we have in bms is correct, we must take a photo inside the cockpit that allows us to have a vision of 135º v and 220º h, then compare that photo with the cockpit of bms and you will know if the fov it’s correct.
-
is simple, uman eyes fov is 130-135º vertically and 200-220º horizontally. To know if what they have in bms is correct, we must take a photo inside the cockpit that allows us to have a vision of 135º v and 220º h, then compare that photo with the cockpit of bms and you will know if the fov it’s correct.
This is depending on your screen. If you mimic human FOV with “classic” screen, you will have deformation. To me, it is not the way to go.
The FOV is dictated by the screen (size and resolution) you are using … unless you won’t be able to read MFDs and will have a kind of distorted view. -
This is depending on your screen. If you mimic human FOV with “classic” screen, you will have deformation. To me, it is not the way to go.
The FOV is dictated by the screen (size and resolution) you are using … unless you won’t be able to read MFDs and will have a kind of distorted view.in practice it would be a deep image of the cockpit, with the instruments too far away to be read.
-
And that’s another thing I’ve been trying to solve - what is the “correct” FOV
From my seated position to the edges of the 34” monitor I measured it to be approx 80 degrees, so I set that in the sim. Seems t work great. Of course, your mileage may vary.
-
I think there are 2 wholly different approaches … pitbuilders trying to create a physically immersive and dimensionally accurate space. And those of plebes with an ordinary desktop PC and a single monitor… maybe ultrawide if we’re lucky.
I want what’s on my screen (no matter how big or small it is) to represent what I’d be able to see, irl.
Cant do it without huge distortion. Basically set your FOV in BMS to about 160 or so and you will be in the right ballpark.
Strictly speaking, the same distortion is already present at 80 degrees, but it is far more obvious the higher your display FOV is compared to the apparent width of your physical display.
-
This is probably a daft question. If so, my apologies.
I Run 4.35 on Linux and on Windows. On both, I can get 90-100 FPS @ 3440x1440@60Hz, but tend to use Vsync to minimise tearing, which limits to 60fps.
I don’t observe any performance difference between Vsync or not Vsync, unless I put the FPS counter on.
But what I do observe is that my speed perception/feeling of speed is significantly lower than that that I see on many people’s published videos (regardless of whether I’m at 60 or 100FPS). And that’s the bit I don’t understand. I had assumed that FPS was everything, i.e. the slower the FPS, the lower that feeling of speed. But that doesn’t seem to be the case. Unless I’m just making a fair comparison between my actual output and the YouTube version of someone’s flight?
I can’t make out FPS above 60. 30 is bare minimum IMO.
-
@Jammer:
I can’t make out FPS above 60. 30 is bare minimum IMO.
What hz is your monitor?