G_fSmartScalingThreshold
-
Depends on the algorithm. but I think the offset should be a great idea.
Looking forward to viewing it.
-
Do you know where to find the paper?
Off the top of my head, no, but it was handed to me by someone that found it in the public domain. The basic premise of the paper is that the “right” answer is to use a display of sufficient resolution…the problem is that displays aren’t providing close to 20/20 visual acuity yet. At the time of writing the state-of-the-art was 20/40 equivalent; I’m sure it’s better now but even 4k doesn’t get you where you need to be (I recall doing the math at some point to convince myself it was worth writing the code but that was a while ago). Primary reason the config button for the scaling option is there at all comes down to the idea that one day display rez will in fact catch up…then we don’t need the scaling turned on at all…mangled (a bit or a lot) or not.
-
-
Ah, there it is – thanks, Robert!
-
stuff like gun funnel slightly out of proportions,
This has always been my question, especially with modeling the longer range of the PGU-28 round’s longer range of 1.3 NM. BUt never had a good evidence either way. Tried to have a discussion on it in a thread here but I think I only got one taker.
Off the top of my head, no, but it was handed to me by someone that found it in the public domain. The basic premise of the paper is that the “right” answer is to use a display of sufficient resolution…the problem is that displays aren’t providing close to 20/20 visual acuity yet. At the time of writing the state-of-the-art was 20/40 equivalent; I’m sure it’s better now but even 4k doesn’t get you where you need to be (I recall doing the math at some point to convince myself it was worth writing the code but that was a while ago). Primary reason the config button for the scaling option is there at all comes down to the idea that one day display rez will in fact catch up…then we don’t need the scaling turned on at all…mangled (a bit or a lot) or not.
I’ve done the math, by pure visual PPI at normal montior distance around a 37" or less 4K monitor is the right PPI for 20/20 or better. But I haven’t digested that paper yet for the other factors.
-
I’m not an optics guy per se but my understanding is that size of display is only marginally relevant. Discrimination of vehicle detail is all about resolution – think about it this way: a distant a/c reduced to a single pixel shown as a physical millimeter or two across give you no more orientation cues than the same pixel on a jumbo-tron.
-
I’m not an optics guy per se but my understanding is that size of display is only marginally relevant. Discrimination of vehicle detail is all about resolution – think about it this way: a distant a/c reduced to a single pixel shown as a physical millimeter or two across give you no more orientation cues than the same pixel on a jumbo-tron.
Yep, at about 30" distance, if you have only 20/20 vision your eye cannot resolve a single pixel smaller than about 130 PPI (can’t find my original sheet of math right now). The size of the display matters because the larger it is the less PPI you have for the same given resolution.
http://jaredjared.com/2012/10/visual-acuity-dpi/
Edit to add using the Jumbotron analogy: a 32" 4k monitor has much smaller pixels than a 32 foot 4k Jumbotron. You have to be further away from the Jumbotron to not notice individual pixels (screen door effect).
Do you guys know the magnification factor with standard (set 0) Smart Scaling of an aircraft at 1.3 NM?
-
0 is " use original smart scaling" or 0 is 0nm?
-
-
-
Yep, at about 30" distance, if you have only 20/20 vision your eye cannot resolve a single pixel smaller than about 130 PPI (can’t find my original sheet of math right now). The size of the display matters because the larger it is the less PPI you have for the same given resolution.
http://jaredjared.com/2012/10/visual-acuity-dpi/
Edit to add using the Jumbotron analogy: a 32" 4k monitor has much smaller pixels than a 32 foot 4k Jumbotron. You have to be further away from the Jumbotron to not notice individual pixels (screen door effect).
Do you guys know the magnification factor with standard (set 0) Smart Scaling of an aircraft at 1.3 NM?
We’re at crossed purposes. The size of the screen is essentially irrelevant to the scaling problem Serfoss’ paper reads on - he was studying what pilots can reasonably discern at a given range about aircraft orientation which requires greater density of pixels to show aspects of vehicle shape. The range at which your nose is placed from the screen such that you can make out a single pixel or not is a different problem and one that the code can’t and won’t address ever…if you aren’t close enough to see – move or buy a bigger display What the scaling code tries to do is, presuming you are at a suitable nose-to-screen distance from your display, give you enough pixels on the screen arranged to deliver the same equivalent ability to discern orientation as you’d have sitting in a Viper cockpit up in the wild blue yonder.
-
We’re at crossed purposes. The size of the screen is essentially irrelevant to the scaling problem Serfoss’ paper reads on - he was studying what pilots can reasonably discern at a given range about aircraft orientation which requires greater density of pixels to show aspects of vehicle shape. The range at which your nose is placed from the screen such that you can make out a single pixel or not is a different problem and one that the code can’t and won’t address ever…if you aren’t close enough to see – move or buy a bigger display What the scaling code tries to do is, presuming you are at a suitable nose-to-screen distance from your display, give you enough pixels on the screen arranged to deliver the same equivalent ability to discern orientation as you’d have sitting in a Viper cockpit up in the wild blue yonder.
It is what I am talking about and you Your visual cone for IDing aircraft (foveal vision) is only 1 degree, so really screen size makes no difference, it is the number of pixels in your 1 degree cone of “IDing vision.” It’s not screen size that is the magic number here, it is PPI. I am saying how many pixels you have to draw the object for all those details or as you say “greater density of pixels to show aspects of vehicle shape.”
But it just so happens that if you don’t change the resolution and you up the screen size PPI drops. Smaller screen size for same resolution = more pixels to draw stuff. We have industry standard resolutions but we don’t really pair them with screen size. It is the same reason a 27" 1920x1080 looks worse than a 27" 2560x1440.
I redid the math, for 30" (typical monitor viewing distance) is 115 PPI= 20/20 visual acuity. My monitors are actually 92 PPI, so from the get go I have less than 20/20 on the monitor on both my side 20" 1600x900 and my center 32" 2560x1440 monitors (which is why I use Smart Scaling )
So for 30"/76 cm viewing distance and common monitor resolutions here are the largest sizes you can have to have a 20/20 PPI (115 PPI). Smaller screen sizes than listed will actually give you better than 20/20 acuity:
3840x2160=38"
2560x1440=25"
1920x1080=19"
1600x900=16"https://www.sven.de/dpi/
It is buying a monitor to get a PPI and since we sell them by screen size and not PPI you have factor that in.Obviously for the lower resolutions you also tend to not want them because they are smaller FOVs and have less immersion. In general 115 PPI+ is still hard to make in large displays.
-
With my eyes with glasses on at 30" away or arms length, 2560 1600 30inch monitor is great, 40" 4K also great. Smaller monitors are also good with those resolutions, but anything larger and SDE becomes more and more apparent as the size increases and they need to be pushed back further and further.
-
My new 32" 2560x1440 is much better for my failing eyes.
-
PPI isnt really relevant here- supposing you can discern every pixel in the screen, the correct metric would be to compare number of pixel in width vs FOV (default 60°). Distance to screen is not really irrelevant.
In full HD, that would be 60° (default FOV)/1920 pixel = 1/32 ° per pixel ~=2 minute of arc (MOA) per pixel. 20/20 vision is supposedly being able to discern 1 MOA.
-
errr didn’t get the math actually but seeing ppi and monitor sizes and such, so if I have a 440ppi 4,7" monitor (a smartphone lets say) I will have better view of falcon than in my 24" monitor?
Also is known that many large hd tv’s 32" 40" are crap as of ppi or dpi and the comparison is a bit harsh…So I believe the doc and the scaling algorithm doesn’t take in account your actual screen resolution or ppi or dpi but tries to make things better for u. Yes size does matter as for monitor size but detail is also crucial as ppi.
and what we are talking here is few pixels differences… like a needle in the barn.
-
PPI isnt really relevant here- supposing you can discern every pixel in the screen, the correct metric would be to compare number of pixel in width vs FOV (default 60°). Distance to screen is not really irrelevant.
In full HD, that would be 60° (default FOV)/1920 pixel = 1/32 ° per pixel ~=2 minute of arc (MOA) per pixel. 20/20 vision is supposedly being able to discern 1 MOA.
Isn’t an inch (as in PPI ) a FOV amount when a given viewing distance is factored in? IIRC from my FOV math at 30" viewing distance 1" is approximately 2 degrees. I keep pointing out it is at normal viewing distance. If you have a different viewing distance (i.e. as I’m about to talk about with Arty’s post) your PPI requirements for 20/20 change.
How I got it from linked previously http://jaredjared.com/2012/10/visual-acuity-dpi/
Visual Acuity
A circle is made of of 360 degrees. A single degree is made of of 60 arc minutes. So 1 arc minute is 1/21,600 of a full circle.
When we say a person has 20/20 vision we mean they have the acuity to discern a detail of 1 arc minute.
Some people have worse vision such as 20/40 and they can only discern at 2 arc minutes – some have better vision such as 20/10 and they can discern 0.5 arc minutes.
There is a physical limit to human visual acuity as there are only so many cones (color sensors) in the eye and while 20/8 is possible it is rare – the kind of thing you see in fighter pilots.
Size, Distance, and Density
So when we look at picture, poster, or monitor we are really talking about 3 different elements.
The size of the thing – and based on the size of the thing we choose a distance to be from it when we view the thing, and then from there we can determine the density of dots that our visual acuity can actually resolve.
For my first example – let me start with my TV. I have a 55″ 1080p HD LCD TV.
TVs are measured diagonally and in reality the TV is approximately 50″x30″.
So based on the size – how far away do I want to sit? Well that depends on how much of my Field of View (FOV) I want to fill.
While the human Field of View is very wide – a comfortable (practical) Field of View is 40 degrees. Pick too wide a FOV and yours eyes dart around – too narrow and the image just isn’t large enough.
With these 2 variables – Width of target and Field of View – I can calculate an ideal viewing distance.
Distance = (Width /2) / Tan(Viewing Angle/2)
Distance = (50/2) / Tan (40/2) = 68.7″ or 5.7′
So now I know that the ideal distance from my couch to my TV is 68.7″ or about 5 1/2′.
My TV is 1080p which means it has a resolution of 1920×1080 pixels.
1920 pixels spread over 50″ results in 38.4 DPI – this seems really low at first – but how does this line up with my visual acuity.
To answer this – I have to ask: If I am sitting 68.7″ away from something – what is the smallest sized dot that my visual acuity can resolve?
For this I need just 2 variables – the Distance from the target – and my own Visual Acuity which I will give as 20/20 (I wish).
From my Visual Acuity I can calculate my Visual Resolution – the smallest number of visual degrees I can detect.
Visual Resolution = (1 / Visual Acuity) * (1 / 60)
Visual Resolution = (1 / (20 / 20)) * (1 / 60) = 0.0166667 degrees
(1 / Visual Acuity) tells me how many arc minutes I can detect so (1 / (20/20)) = 1, (1 / (20/10)) = 0.5, and (1 / (20/8)) = 0.4 (which I said before is as good as it gets).
Multiplying the Number of Arc Minutes * (1 / 60) converts the number from arc minutes into degrees.
Smallest Dot Size = 2 * Distance * Tan (Visual Resolution/2)
Smallest Dot Size = 2 * 68.7 * Tan (0.0166667/2) = 0.019984 inches
Dots Per Inch = 1 / Smallest Dot Size
Dots Per Inch = 1 / 0.019984 = 50 DPI
So with my 20/20 vision I cannot resolve more than 50DPI on something that is about 5 1/2′ away. So my TV’s 38.4 DPI isn’t too bad after all.
And let us not forget the reality:
-
I am not looking for individual dots on the screen
-
I have worse that 20/20 vision
-
I sit more than 5 1/2′ from my TV
errr didn’t get the math actually but seeing ppi and monitor sizes and such, so if I have a 440ppi 4,7" monitor (a smartphone lets say) I will have better view of falcon than in my 24" monitor?
Also is known that many large hd tv’s 32" 40" are crap as of ppi or dpi and the comparison is a bit harsh…So I believe the doc and the scaling algorithm doesn’t take in account your actual screen resolution or ppi or dpi but tries to make things better for u. Yes size does matter as for monitor size but detail is also crucial as ppi.
and what we are talking here is few pixels differences… like a needle in the burn.
You will have better screen visual acuity, but putting a 4.7" screen at 30" is not very immersive This is also why smart phones and VR headsets have crazy resolution and high PPI for the size, you are viewing them much more closely. If you are using a smart phone at 12" you need 286 PPI by my math and sources to get 20/20.
EDIT to add:
Hehe and yes this is all academic unless you have one of those monitors. Anybody actually know the amount Smart Scaling size up an aircraft at cannon range? If I have that I can figure out if I want to use this new config line anyway
-
-
Isn’t an inch (as in PPI ) a FOV amount when a given viewing distance is factored in? IIRC from my FOV math at 30" viewing distance 1" is approximately 2 degrees. I keep pointing out it is at normal viewing distance. If you have a different viewing distance (i.e. as I’m about to talk about with Arty’s post) your PPI requirements for 20/20 change.
How I got it from linked previously http://jaredjared.com/2012/10/visual-acuity-dpi/
This is why I was saying : “supposing you can discern every pixel in your screen”.
-
the cannon starts to become effective at 2 nm (correct me if i’m wrong ofcourse as i’m not entirely sure here) so that’s where it stops the smart scaling for me right now.
i see people talking about 20/20 vision. but what if people like me who have significantly reduced vision in 1 eye compared to the other with bad eyes all around for me really factor into that equation… wouldnt that make the entire argument invalid for me as i am far below snuff for “pilot level eyes” and as such wouldnt that make the setting variable for different people? to simulate 20/20 vision as if you were a pilot?
-
the cannon starts to become effective at 2 nm (correct me if i’m wrong ofcourse as i’m not entirely sure here) so that’s where it stops the smart scaling for me right now.
i see people talking about 20/20 vision. but what if people like me who have significantly reduced vision in 1 eye compared to the other with bad eyes all around for me really factor into that equation… wouldnt that make the entire argument invalid for me as i am far below snuff for “pilot level eyes” and as such wouldnt that make the setting variable for different people? to simulate 20/20 vision as if you were a pilot?
The 4.33 manual has pictures and as been discussed here, it scales all the way to 0 range, not 2 nm (1.3ish NM is guns range according to your HUD and the -34)
Yes you are at a disadvantage to what your eyes can get anyway, but Smart Scaling won’t make it any worse. It set to 0 or just slightly higher like .2 is probably your best option.