Oolite Bulletins

Well, its not related to FOV. It does seem to be related to the lines scale with X resolution somehow.
Maybe they should relate to Y resolution instead?

top 1680x1050, bottom 5040x1050)


The lines are much thicker (borders, radar, space compas, crosshair)

@another_commander: I created a new pull request implementing fov manipulation from javascript
https://github.com/OoliteProject/oolite/pull/130

@Vincentz: I think I've seen in the code that the hud lines width are proportionnal to the width of screen. It would totally explain your problem. Would it be useful to report a bug? A quick solution might be to make the lines proportionnal to the smallest of screen height/width.

Day wrote:

@another_commander: I created a new pull request implementing fov manipulation from javascript
https://github.com/OoliteProject/oolite/pull/130

@Vincentz: I think I've seen in the code that the hud lines width are proportionnal to the width of screen. It would totally explain your problem. Would it be useful to report a bug? A quick solution might be to make the lines proportionnal to the smallest of screen height/width.

If its fixable, I think it should always stick with height. Take 2 resolutions 1024x768 and 1280x768 the lines would be thicker on the 1280, which they shouldn't. It just shows a bit more in the sides.
So yeah, please do report it as a bug

If its fixable, I think it should always stick with height. Take 2 resolutions 1024x768 and 1280x768 the lines would be thicker on the 1280, which they shouldn't. It just shows a bit more in the sides.
So yeah, please do report it as a bug

I'll let you do it
https://github.com/OoliteProject/oolite/issues

Vincentz wrote:

So yeah, please do report it as a bug

It looks like scaling the HUD line width with increasing screen width is intentional. What was probably not considered is the existence of ultra-wide setups which result in overscaling. An attempt to fix that was done in commit ba9fb32 by clamping the maximum line width to a value which I consider reasonable (but have no means of testing). It will be in for testing in tonight's build. Vincentz, can you try that tomorrow and let us know if it makes things better for you?

another_commander wrote:

Vincentz wrote:

So yeah, please do report it as a bug

It looks like scaling the HUD line width with increasing screen width is intentional. What was probably not considered is the existence of ultra-wide setups which result in overscaling. An attempt to fix that was done in commit ba9fb32 by clamping the maximum line width to a value which I consider reasonable (but have no means of testing). It will be in for testing in tonight's build. Vincentz, can you try that tomorrow and let us know if it makes things better for you?

Great. Thanks While it did look a bit strange with thick lines, it didnt render it unplayable though.
In fact most modern games doesn't do as well as Oolite does when it comes to triple display. (especially now with the FOV changable ingame).

Looks much better, though its still (a little) thicker, and space compass flickers a bit.
Still thinks it would be better to scale with vertical, which is afaik a standard for doing that (examples 720p 1080p)

Though with regards to viewing angle and FOV, I think (if its possible) it should automatically scale for horizontal (which afaik is a standard for FOV).

So if the "intended" FOV for Oolite is 57.5 for a 640x480 (4:3) resolution, then when changing resolution, the FOV should follow it.
Now, before I go all mathematically here, keep in mind I have a 6 month daughter and a 3½ year old, very energetic son running around, so my calculations could be way off

But if...
4:3 (640x480 (1024x768) - 1:33) = 57.5 FOV (43.125)
Then...
5:4 (600x480 (1280x1024)- 1.25) = 53.9 FOV
16:10 (770x480 (1680x1050)- 1.6) = 69 FOV
16:9 (853x480 (1920x1080) - 1.77) = 76.7 FOV
and if I'm not wrong... for triple monitor setup (48:10)
48:10 (2310x480 (3x1680x1050)- 4.8 ) = 207 FOV

So, to get the same zoom level as a 4:3, the 1920x1080 would need a FOV of 76.7 (the zoom is same, just more to see on the sides)
and a triple monitor setup with 3 1680x1050 would need a 207 FOV.
Which, if implemented and still wanting a FOV manual setting in Settings, it should imho be a % of the resolutions FOV.

Examples to illustrate different aspect ratioes with 57.5. The greater the widescreen, the less is actually shown (vertically).
Even with FOV 80, a triple monitor setup will be zoomed way too much, and be difficult to control.

Thanks for testing. Much appreciated.

Vincentz wrote:

Still thinks it would be better to scale with vertical, which is afaik a standard for doing that (examples 720p 1080p)

I am not fond of fixing this by scaling on the vertical, because this way you just move the problem to those who might enjoy playing on ultra-high setups. If you think this is too rare, we had a thread in this very forum about people who were playing in portrait screen setups not too long ago. For this reason, I am fine with just a little bit thicker lines on very wide resolutions, as long as they don't distract too much due to their thickness itself. If you think that we can clamp the line thickness to somewhat lower and it will look better, we can certainly try to.

Though with regards to viewing angle and FOV, I think (if its possible) it should automatically scale for horizontal (which afaik is a standard for FOV).

This is something we can certainly look at, although I won't consider it a show stopper if things remain as they are now.

4:3 (640x480 -1:33) = 57.5 FOV (43.125)
Then...
5:4 (600x480 - 1.25) = 53.9 FOV
16:10 (770x480 - 1.6) = 69 FOV
16:9 (853x480) - 1.77) = 76.7 FOV
and if I'm not wrong... for triple monitor setup (48:10)
48:10 (3072x480 - 4.8 ) = 207 FOV

So, to get the same zoom level as a 4:3, the 1920x1080 would need a FOV of 76.7 (the zoom is same, just more to see on the sides)
and a triple monitor setup with 3 1680x1050 would need a 207 FOV.

There are technical problems which prohibit going over the top with FOV. Even a 180 degrees FOV is not an option; things go haywire with the display. Also, the interior of ship models starts to become visible in high FOVs so that's another thing to take into account when deciding limits. That's why the current limit is set to 80 degrees - and this takes into consideration the possibility of scripts applying a multiplier to FOV in the future. Most games that I've seen set FOV limits to about what we have now (what is Elite:Dangerous doing by the way with this? What is its FOV limit?)
So I think that the 80 degrees we currently have is a reasonable compromise between what is technically achievable and what is good for overall gameplay.

But if...
4:3 (640x480 (1024x768) - 1:33) = 57.5 FOV (43.125)
Then...
5:4 (600x480 (1280x1024)- 1.25) = 53.9 FOV

I'm sorry, you lost me.
What's the calculation behind that?

another_commander wrote:

Also, the interior of ship models starts to become visible in high FOVs so that's another thing to take into account when deciding limits.

Without testing it, I can't say for sure, but I don't think it will show IF x-resolution is high enough.
I can see the tip of the cannon if I use 80 FOV on a normal 16:10, however it wont show due to the increased zoom if I use triple monitor.
This can be tested on a normal monitor too, if game-window is resized to same width but 1/3 the height of the monitor.

Though, my 207 calculation looks to be somewhat wrong, there is a great tool here :
http://www.wsgf.org/fovcalc.php
It says for a 4:3 game with 57 FOV, that a 3 x 16:10 would be around 125 FOV (which sounds more reasonably)

Day wrote:

But if...
4:3 (640x480 (1024x768) - 1:33) = 57.5 FOV (43.125)
Then...
5:4 (600x480 (1280x1024)- 1.25) = 53.9 FOV

I'm sorry, you lost me.
What's the calculation behind that?

If the 4:3 is 1.33 in aspect ration, then the FOV aspect ration is 57.5/1.3 = 43.125
Then the 5:4 aspect ratio would be 43.125 * (5/4) = 53.9 FOV

(but as mentioned, I'm not in a position to do math
I think the FOV calculator above is better)

Vincentz wrote:

another_commander wrote:

Also, the interior of ship models starts to become visible in high FOVs so that's another thing to take into account when deciding limits.

Without testing it, I can't say for sure, but I don't think it will show IF x-resolution is high enough.
I can see the tip of the cannon if I use 80 FOV on a normal 16:10, however it wont show due to the increased zoom if I use triple monitor.
This can be tested on a normal monitor too, if game-window is resized to same width but 1/3 the height of the monitor.

I am not talking about seeing the laser when changing FOV - that is fine, we are still looking at exterior parts of the model. I am talking about seeing the inside walls of the model at high FOVs, which is a no-no, unless all models are redesigned and we don't want to go there.

another_commander wrote:

... what is Elite:Dangerous doing by the way with this? What is its FOV limit?

I think the default is 60, though some have tinkered with the files and upped it - they seem to prefer 75.

Not to beat a dead horse here but are the parts still showing if you use 125 FOV and the game-window height was reduced to a third of the screen (to simulate eyefinity)?

Vincentz wrote:

Not to beat a dead horse here but are the parts still showing if you use 125 FOV and the game-window height was reduced to a third of the screen (to simulate eyefinity)?

That depends on the model itself and I have not tested each and every one of them. For the Griff Cobra (early version of the model) the answer is yes for FOVs around 130-144 degrees. The higher the FOV, the higher the chance that this will happen with some model, so keeping the FOV limit to a conserative 80 degrees (OK, maybe it could be increased a little bit still) will hopefully avoid such problems.

If the 4:3 is 1.33 in aspect ration, then the FOV aspect ration is 57.5/1.3 = 43.125
Then the 5:4 aspect ratio would be 43.125 * (5/4) = 53.9 FOV

(but as mentioned, I'm not in a position to do math
I think the FOV calculator above is better)

No problem, and thank you for your answer

Ok, I get it. Yes the calculator is better

Original fov, that is to say horizontal angle of view is 57.5°.
57.5° / 1.33 does not give the vertical angle of view.
It would be 2 * atan(tan(57.5 / 2) / 1.33).

This does not invalid any part of the discussion, so no worries.