Celestial Navigation

[YellowFin]

I want to set up SH IV to allow for celestial navigation and have started with preliminary research. I understand the basics of celestial navigation.

From what I've read the main idea is to obtain the angles between the horizon and celestial bodies (stars, moon, sun). These angles define a circle on the earth's surface. When the angles of at least two celestial bodies are obtained, these circles intersect on the earth's surface and give two possible positions. Reasoning from whether these positions are on a land mass or at sea and proximity to previous positions allow to eliminate the incorrect position solutions.

I have several questions WRT how SH IV models the various involved real world elements:

• Is the night sky a cylindrical dome as opposed to a spherical one?
• How does that fact affect the calculations? I suppose there is a fix conversion that would still allow to translate in game angles to real world angles?
• How would I measure angles in game?
• How does the game model time and clocks? Is DST modeled in-game?
• I suppose the in-game clock is accurate. Can anyone confirm or deny?
• Is there an easy way to get rid of the own boat marker on the map?
• And to switch it back on for testing purposes?
• If I understood another poster correctly a frame rate of > 30 causes time to drift in game. How do I limit the game's frame rate to 30?

I will check on the existing celestial navigation tools / mods. Any other hints / answers will be greatly appreciated.

[YellowFin]

probeanie's initial comment on my inquiry as to the possibility and reliability of CN:

I forgot about your navigation question... sorry 'bout that. But yes, it is possible, though there are a LOT of things that can get "off", or out of sync, in the game. The "world" is a cylinder, while the sky appears to be a squished 1/2 dome, and the sun and moon are separate from that... What we found last year was that the Time Compression used can influence the "sync" of these parts of the game. Front Runner was the one who discovered this. He would play the game a 1x (real) time as he worked around the house, and would notice a time-drift over the course of time, where sunset and sunrise times were "drifting". He started a thread called "Base Time 2018", which has a few more links to other threads and resources for "discoveries" he found. The big one was limiting the frame rate of the game to 30fps, which is possible to do rather easily with an nVidia card, but not so with an onboard Intel, and we haven't found anything for an AMD card yet either. Once you limit the framerate, then the time sequence in the game doesn't seem to drift out of sync with the various "celestial bodies" like it did previously, plus with the work that CapnScurvy did on the moon (it seemed to be "backwards" before), you might now have relatively stable base to calculate from (though we have not had anyone try and comment on succes or failure yet). However, the only way you find out that the game is out of sync with itself, is when you run aground in the middle of the night, though by your calculations, you're 30 nautical miles from the nearest land mass...

[YellowFin]

Front Runner's invaluable investigation on de-syncing of the celestial sphere as well as save game loading related problems with time and celestial sphere:

https://www.subsim.com/radioroom/sho...d.php?t=239188

[YellowFin]

This website seems to do what I need.

http://www.jb.man.ac.uk/almanac/

For instance I can input time and location and obtain the position of certain celestial bodies

I can also perform the reverse and input the positions of celestial bodies as I see them and my time and solve for a position.

I do struggle with the correct input for "West Longitude", see below link for their specification.

Apparently instead of a longitude, such as

122°E 03' 00''

they require their eastern longitutdes to be negative

-122° 03' 00''

and in addition to that they should be in hours (divide degrees by 15). Obviously this can return fractional values, as it does indeed in this case:

122°E 03' 00''

as far as I understand their specifications it should be

-8.1333... 03' 00''

Their online application doesn't accept non-integer values for the hours however.

Does anyone understand what exactly they want? Should I convert the fraction back into minutes:

-8 11' 00''

like so?

http://www.jb.man.ac.uk/almanac/help.html#longitude

[YellowFin]

I just did above procedure converting decimal fractions to minutes and degrees for the latitude. It's right before sunrise. My position is roughly 122°E 03' 0'', 15°N 01' 00''. My ship clock shows 0641 as the time and it is December 17, 1941.

For the sun the almanac returns an elevation of -72°, where it should be around 0°.

I got a difference in azimuth for the sun of only 8 degrees: The almanac returns the sun's azimuth as 238.7°, which is equivalent to 360°-238.7° = 101.3° true bearing. The game displays the sun at 110°, so 8° degrees too far south.

I don't know if

• My game suffers from a desynced celestial sphere
• A time / clock problem
• I wrongly use the almanac

I strongly suspect that something is wrong with the input values I used for the almanac, as at around 7 in the morning the sun should be much closer to the horizon than -72° in the Philippines. I have no cause to assume that the game clock displays an incorrect time.

Any help, maybe from someone familiar with real celestial navigation would be greatly appreciated.

[Sean C]

I was going to suggest the USNO's celestial navigation data page, but I tried it just now and couldn't even get it to load. Anyway, when it does work, it's very easy and useful.


I'll email them and let them know that several of their pages seem to be down. This happened once before and was promptly fixed.



BTW, ↓

[YellowFin]

Originally Posted by Nathaniel B.

I was going to suggest the USNO's celestial navigation data page, but I tried it just now and couldn't even get it to load. Anyway, when it does work, it's very easy and useful.


I'll email them and let them know that several of their pages seem to be down. This happened once before and was promptly fixed.



BTW, ↓

Thank you. I also found the USNO sites to be down, which is why I ended up using jb.man.ac.uk.

There is something I really don't understand on their site, which is the transformation of longitude into what they call "west longitude" in hours, minutes and seconds, instead of the more familiar degrees, minutes and seconds.

They require the longitude in DMS to be transformed into hh mm ss, which is to say degrees into hours. I believe the correct way to do that is to convert DMS into DD (decimal degrees) and to convert that into hours, by dividing DD by 15. The decimal fraction of this last division is reconverted to minutes and seconds by multiplying the decimal fraction of the hours by 60 and the remaining decimal fraction of that multiplication by 60 again, to obtain seconds.

I tested their system: If for example you don't input the position yourself, but use one of their preset locations, say, Tokyo, Japan, you will get an output page that shows a very odd position for this city:

Latitude: 35 42 00 Longitude: 14 40 55

While the latitude is fine, the longitude I can't make sense of. It's not DMS, as the DMS longitude for Tokyo, Japan is 139°41′32″E according to wikipedia. Converting that longitude into DD gives me 139.6922° and converting that into hh mm ss gives me 9 57 32. So their outputted longitude is neither DMS, nor DD nor hh mm ss.

What gives?


With regard to the USNO website: How exactly would I use that one, as opposed to jb.man.ac.uk? Does it have an input form or do I look up values and make my own calculations? I don't have any formulas as of yet. I simply have a rough understanding of which values to input, namely time, date, position, and which values to compare with the in-game sky, namely azimuth and elevation.

[Sean C]

Congratulations! You've found the most confusing and convoluted web page claiming to offer almanac data on the interwebs!


Seriously though, I can only assume that this page is asking for longitude in the format of hours minutes and seconds. When I tried using Tokyo, Japan as an input, I also got a longitude of 14h40m55s or 14.6819444...h. If we multiply this by 15, we get 220.229166...°. Converting that to DMS (by taking the decimal part and multiplying by 60 and doing the same again) we get 220°13'45". If we subtract that figure from 360°, we get 139°46'15". This is approximately the East longitude of Tokyo.


BUT WHY DO IT THIS WAY?! Why not just use DMS for longitude like every other [useful] application I've ever seen?


I could go on about the many other ways this site is poorly designed, but I can't be bothered. It's simply terrible and way more complicated than it needs to be. Until the USNO site is sorted, I suggest you download and use the excellent software "Teacup Navigation" by Rodger Farley - available here. (If you need help determining "Delta-T" for a specific time, I can help you do that.)


There are also other web sites which can give you almanac data much more easily. Here is one example. Here is another. There are many others. Again, I must congratulate you on finding the absolute worst one. (Please don't take this wrong, I'm just dumbfounded at how bad that site is. It's no reflection on you, personally.)


The USNO site does have an input form. You enter the date, time and your location and it gives you data for all visible bodies - [IIRC] GHA, Dec., Hc, Zn and all of the corrections necessary to reduce a sight of each body. IOW, everything you need for celestial navigation. Very simple.

[YellowFin]

Originally Posted by Nathaniel B.

Congratulations! You've found the most confusing and convoluted web page claiming to offer almanac data on the interwebs!

See? I'm a talented man

Thanks for the links and explanation though. For some reason I could only find the jb.man.ac.uk and the USNO (out of service). Or maybe it was Google who was being greedy.

[Sean C]

Originally Posted by YellowFin

Thank you. I also found the USNO sites to be down, which is why I ended up using jb.man.ac.uk.

FYI: The USNO got back to me and their site appears to be fully functional now.