This was discussed in another thread relating to a specific scenery package:http://forums.avsim.net/dcboard.php?az=sho...id=264366&page=Given that in flight simulator we are viewing the world through a limiyed resolution screen I wondered what the effect of this is. I analyised this to determine what distance on the ground would be represented by 1 pixel on the screen:http://forums.avsim.net/user_files/122618.jpgThe values I assumed were:SD 600mm (approx 2 ft)SH 300mm (corresponding to a 21" monitor)SVR 1200 (corresponding to a 1600 * 1200 screen) resolution)The results were:H = 100m ( 328ft) R = 0.67mH = 200m ( 656ft) R = 1.34m H = 400m (1312ft) R = 2.67mH = 500m (1640ft) R = 3.34mH = 720m (2362ft) R = 4.80m*H = 1000m (3281ft) R = 6.68mH = 2000m (6562ft) R = 13.36m* Standard FS9 resolution.This suggests that improving the present resolution would only really be of benefit when flying below 2300ft. Above that height one pixel on the screen would be the combination of more than one pixel on the ground. I haven't considered the possibility that better resolution might give a better subjective visual appearance because I've no idea how this might be measured.Also increasing the resolution would increase the file sizes by the square of the ratio between them. Halving the resolution to 2.4m/pixel would increase ithe bitmap file size of 42.7Kb (UK VFR scenery) by a factor of 4 to 171Kb. This would reduce the minimum height to 1312ft. The best widely available resolution of photographic images on the UK is 0.250m/pixel. Using this would increase file sizes by a ratio of 368, with a file size of 16Mb - using the very best at 0.125m/pixel the file size would become about 62Mb. I think we can rule out the use of these. Incidentally, does anyone know the resolution of Googe images?Comments welcome, specially if I've got something wrong!

When flying above 2300ft the FS9 photographic scenery image still doesn't look nearly as good as an actual photo viewed with the same number of pixels on the same monitor. Why?Is it possible, in the not to distant future, for FS to have a 0.5 meter resolution so the ground would look much better when flying low? Maybe FS has to use autogen over the small area you're flying over to make this possible.Any guesses what the resolution would be for FS10? FS9 resolution is 4.76 meter.

>When flying above 2300ft the FS9 photographic scenery image>still doesn't look nearly as good as an actual photo viewed>with the same number of pixels on the same monitor. Why?>I admit I don't know. Could it be the difference between JPEGs and bitmaps? JPEGs create their effect by deceiving the eye/brain into seeing more than is actually there.>Is it possible, in the not to distant future, for FS to have a>0.5 meter resolution so the ground would look much better when>flying low? Maybe FS has to use autogen over the small area>you're flying over to make this possible.>0.5m resolution is about 10 time better than the FS standard of 4.8m so that the file sizes woyld go up by a factor of 100. A UK VFR bitmap of 42.7 Kb would become 43Mb. I don't think this could be loaded quickly enough.

>When flying above 2300ft the FS9 photographic scenery image>still doesn't look nearly as good as an actual photo viewed>with the same number of pixels on the same monitor. Why?>Further thoughts on the subject.If you're viewing an actual photo then you are viewing something of high original resolution so it should look good. If, on the other hand, you're using FS VFR scenery then you're viewing something that's already had its resolution reduced to 4.8m/pixel so I'd expect it to look worse.

>When flying above 2300ft the FS9 photographic scenery image>still doesn't look nearly as good as an actual photo viewed>with the same number of pixels on the same monitor. Why?Because (in my humble opinion) there is a bug in mgh's calculations or in his general approach. I don't know where it is (I wished I had time to look for it) but his approach seems unnecessarily complicated.If we want to calculate the altitude from which 1 pixel on the screen will equal roughly 5 meters in real life then why not forgo this complex trigonometry and perform this really simple arithmetic/experiment: if my monitor has 1200 pixels in horizontal line then if each pixel is 5 meters we must be able to fit almost 6 km of terrain in our view (5 x 1200 = 6000). So this is exactly what I did - I positioned myself above KSEA airport and looking straight down at the airport I was elevating myself above the airport until I could practically fit 2 runway lengths in the horizontal view (the longest runway in KSEA is about 11,000 ft = 3 km). I had to go well above 20,000 ft to get the desired effect. So I am sorry but this 2300 ft can't be correct or a zero was dropped somewhere by mistake.Michael J.WinXP-Home SP2,AMD64 3500+,Abit AV8,Radeon X800Pro,36GB Raptor,1GB PC3200,Audigy 2http://www.precisionmanuals.com/images/forum/747400.jpghttp://www.hifisim.com/images/asv_beta_member.jpg

I now know what's wrong with your approach. Your equations are correct and your numerical calculations are flawless but the issue is with your oblique angle. Your 2300 ft altitude is artificially low because you took under consideration natural compression of vertical dimension when viewed at an angle. The problem is however that in this case the worst case will be horizontal pixels that do not get compressed. In other words the vertical dimension is "sharpened" by the fact we look at the terrain at an angle but the horizontal dimension is not. And when carrying out calculations for the horizontal pixels using your numebrs one gets that altitude of roughly 15,000 ft (1/4 mmm pixel over 300 mm = 1/1200, 4.8 m x 1200 = 5000 m = 15,000 ft in approximation).Michael J.WinXP-Home SP2,AMD64 3500+,Abit AV8,Radeon X800Pro,36GB Raptor,1GB PC3200,Audigy 2http://www.precisionmanuals.com/images/forum/747400.jpghttp://www.hifisim.com/images/asv_beta_member.jpg

Interesting. This places a mathematical explanation on something that I have noticed in FS for the last two incarnations. Flat terrain never displays as well as hilly or mountainous terrain. Can't stand flying around Meigs because the terrain never works right to my eye, no matter the res, the mesh or the anisotropic filtering. But throw in a few hills and whew!Crossing back to the other thread about trees and forests, this poses a question as to whether the ground textures would actually be better if they were `

I don't agree with your definition of best/worst case.At 2362ft a pixel on the screen occupies 4.8m vertically and 1.16m horizontally on the ground. Increasing the aircaft height above this figure will increase the 4.8m value, so that area covered by the pixel on the screen will consist of parts of more than one pixel on the ground, immediately reducing the resolution below 4.8m/pixel.My calculations show that the 1.16m becomes equal to 4.8m at 9739ft. However at that height a pixel on the screen would cover 19.82m vertically, i.e. over 4 pixels so there would be a significant loss in resolution - 4 pixels on the ground would be combined into a single one on the screen.Finally, I dont't understand your calculation which seems to be the number of metres covered by 1200 pixels at 4.8m/pixel. I have two points about this: (1) 4.8 * 1200 = 5760m (approx 18900ft) and (2) how does this relate to height?

>I don't agree with your definition of best/worst case.This is not a 'definition'. You selectively decided to carry out computations for the vertical dimension only - try to carry them out for the horizontal dimension as well and then your result will be significantly different. If you managed to apply such fancy trigonometry to the vertical case you should have no problem with the horizontal case - which is much simpler. Yes, you have to pick the "worst" case result - well, if you don't understand why I am at loss. ABC of mathematical logic. :-hmmmMichael J.

I note that you did not address the points I made in my last post, so I'll reiterate them more simply.I did carry out the horizontal calculations and showed them in my last post. They showed that at a height of 2362ft, a pixel on the screen showed 1.61m of the ground horizontally, and 4.8m vertically The latter as before.4.8m is the FS resolution so that at any greater height the pixel on the screen would be built-up from more than one pixel on the ground vertically with a consequent loss of resolution. The loss in horizontal resolution would occur at a greater height.The height when the loss in resolution begins is the critical one - that is the worst case. At lower heights there would be no loss of resolution due to this effect.

>I note that you did not address the points I made in my last>post, so I'll reiterate them more simply.Mr. First Name (mgh),Because I have no intention to 'baby-sit' your calculations. I am at the end of I wanted to say anyway. If my numbers, methododolgy does not convince you then perhaps something inside you (intuition?) should alert you to the fact that your main result: "This suggests that improving the present resolution would only really be of benefit when flying below 2300ft." - has no connection to reality. The simple fact is that I can sim-fly at 8000 ft (high above your 'magic' number of 2300 ft) and still see 4.8 m resolution 'fuziness' without straining my eyes or using a microscope, I can even see individual 4.8 m 'tiles' at altitudes above 5000 ft. So I am afraid if FS had improved scenery resolution we would all see it and benefit from it at altitudes high above 2300 ft.Michael J.

For those who may still be interested, the resolution of what you can see in FS is determined by the resolution of the screen and the resolution of the underlying image shown on it.My calculations and assumptions show that looking down and forwards at the bottom of the screen then at 2360ft the vertical screen resolution is 4.8m/pixel. This is the same as the maximum underlying resolution of photoscenery so one pixel on the ground is represented by one pixel on the screen.At heights greater than 2360ft the screen resolution worsens so that at 4717ft the screen resolution becomes 9.6m/pixel. In this case two pixels on the ground are being compressed into one pixel on the screen with an inevitable loss in overall resolution. At 9739ft the vertical resolution is 19.8m and the horizontal resolution 4.8m. This results in four pixels on the ground being compressed into one on the screen with a further loss in resolution.Increasing the resolution of the ground image with no other changes would result in more pixels on the ground being compressed into a single pixel on the screen.