CONDOR

Important: The process described in this post is incomplete and wrong. My post http://forum.condorsoaring.com/viewtopic.php?t=8141&postdays=0&postorder=asc&start=15 is correct
subject to one restriction. Your terrain must not span a UTM zone boundar or cross the equator.
See http://en.wikipedia.org/wiki/Universal_Transverse_Mercator_coordinate_system for details on zones.
=================

OK, I admit, I struggled with terrain calibration. So, I'm one of the dummies.

I know there are tutorials on this, but I can't find them.

With the help of some resources I found on the net I think I found a trivial way to do it. You experts out there please comment.

firstly, I found a spreadsheet which can convert from UTM to latitude/longitude. The spreadsheet says it good within 10 meters.

http://www.uwgb.edu/dutchs/UsefulData/UTMFormulas.HTM.

I didn't read all the math, I just used it.

I made one small change to the spreadsheet, I increased the number of decimals on the output of lat/lon.

Here is my .hdr file


All the information for the spreadsheet is here: map zone, and UTM coordinates.

Here is an example screendump of the spreadsheet (in openoffice).

So, all I do is enter the UTM coordinates of the four corners of my scenery from the hdr file. Now, the height data pixels are 90 metres
and I want the calibration points in the middle of the pixels in the 4 corners. So, for the lower right hand corner I use 45,45 (instead of 0,0)
and for the upper left I use 207315,115155 (instead of 207360, 115200).

So, my Calibration.txt should look like:



I created it in Notepad. Then I start the Landscape editor, load my
scenery and calibrate.

The only part I'm not sure about with this procedure is the offset of 45
metres. I seem to get a better result than with no offset.
In the middle the calibration is good, no error I can see with
the landscape editor. Close to the corners it looks like I am off by
almost 90 metres.

So experts, what do you think?

I am not a expert, but your method looks as a genial solution - if it is working. Please try set some tasks - triangles with different lenght at your scenery a compare task lenhts from condor with task lenght when you open igc file at SeeYou. You do not need fly task, only start it and save igc before take off. If you will have differences less than 1km, better about 500m to task with 300km, your calibration is excellent and it looks very usefull. Thank you for tip. Calibration is one of the most disgusting part from scenery develepment

_________________

I used similar method. I took corner points and converted them with similiar calculator. and ofcourse i used few more points located on scenery converted in same way. i still get around 2-3km error regarding to seeyou on longer tasks like 1000km triangles, 1000km out-and-return etc.

and who said seeyou is exact anyway? i have alot of weird data from seeyou. many airports are located on really great slopes, more or less all aiport heights are off for few hundred meters regarding to seeyou map height as well as comparing to hgt height maps. not to mention airport lenghts etc. with such input data its impossible to create a scenery for competition purposes. everything would be off and for example, SBC competition only uses sceneries that are seeyou compatible. doesnt matter if seeyou has errors or not.

any hint or ideas?

The method that is used in the Condor_3dem_calibration video, calibration points are taken from the very edge of a scenery development project and also the maximum height point in the terrain, the max height information can be taken by opening the hdr file and look at the maximum elevation point line.

When using Condor_3dem select from the menu Geo Coordinates then select decimal degree.

When using CST Landscape Editor, move to the very edge of each corner and enter the lat/lon coordinates that were taken from when using 3dem, find the maximum height point and enter the lat/lon coordinates then calibrate the scenery.

This can then be tested by creating a task in Condor then check this using SeeYou, when you enter the Condor GUI, you don't have to fly the task just exit and then save and export IGC file which can then be viewed in SeeYou.

This method will almost produce 100% accuracy in a scenery every time.

My scenery is quite small, 9x5 tiles or 207.3km by 115.2km.
I setup four airports in the scenery and defined a task.

Condor says the distance is 343.0 km, Google Earth says 342.726 km
(I don't have SeeYou).

The coordinates from Condor Task planner are:

45 24.660N, 76 21.930W
44 56.142N, 74 50.676W
44 56.736N, 75 56.418W
45 37.014N, 74 38.844W

I also tried a single leg, diagonal, Condor says 175.7 km, Google Earth
175.534. For this task the coordinates are:

45 24.660N, 76 21.930W
45 04.800N, 74 10.600W

I've seen some posts about 0 error in terrain calibration.
This process is not that good. Both my sample tasks work out to about 0.1% error.

To be honest, I don't have any idea how to improve the accuracy.
I've tried adding more calibration points.
I am not very good at looking at the elevation data and clicking in the right place.
I've added a total of 5 more points and the distances vary by
less than 100m.

I've tried moving the points around a little.
Maybe I'm dense, but I can't make that work either.

Download the Trial version of SeeYou, this is a very good way of checking how good your calibration of your scenery is.




This process is very good as it has alraedy produced a number of sceneries that are in the design process and have 100% accuracy and calibration.




If you place a lat/lon coordinate at the very edge of your terrain, these coordinates should be taken when using Condor_3dem, you will also need to take the maximum lat/lon calibration point, how do you know what the max point is this information can be taken by looking at your hdr file.

When using CST Landscape Editor this may take a short while finding the max height point, but helps if you have a rough guide where this is by looking at the flightplanner map.




You will only need to move the max height calibration.


Hope this helps.

I downloaded SeeYou and did the second task from my earlier post. Condor says 175.7 km, Google Earth 175.534km, and SeeYou 175.3km.

When I measured in Google Earth I was very careful to place the ruler endpoints in exactly the right place (fully zoomed in). I'm confident the Google Earth distance is as good as Google Earth can get.



I know there is a video tutorial of this process. I can't find it.

The only documented process I have is in the PDF file that came with the scenery toolkit.



Not so much.

My terrain is largely flat with some water features (lakes and large rivers).
I've tried using islands which I can pick out and in some cases stone quarries (much lower than the surrounding area).

Looking at the height data in the Landscape Editor and deciding where to drop the calibration point is highly subjective
(especially because the height data only has 90m resolution).

Based on the SeeYou data, it would seem I need to make changes. How?
As I mentioned, adding more calibration points does not seem to help.
And I don't understand how to move the existing calibration points to fix it.

would it be possible to edit hgt height files with custom points? for example if we could add fake vertical area at exact geographical location wich could be later used for calibration. this could be then flatened in CST. flat sceneries are really hard to calibrate if there are no hills etc.

one thing i noticed is that i dont have any hill as height as elev_m_maximum parameter in hdr file. not even seeyou has areas that high. is it possible that 3dem gets this values before scenery is converted to UTM and croped in 3dem ( F8 ) to final size?

Just a quick post before I leave for work. I will post more details this evening (N. America time).

I beleive I am very close to getting this method to work!!!



I am going to challenge common wisdom. I beleive that SeeYou distances are not completely correct. As I posted earlier, I see small differences
between distances reported by Google Earth and SeeYou.

I just tried the spreadsheet I pointed to in the first post of this thread
http://www.uwgb.edu/dutchs/UsefulData/UTMFormulas.HTM.
In addition to converting UTM to lat/lon it can do the reverse.
I converted the lat/lon for a straight-out task to UTM in meters.
I then used pythagorus theory to calculate the straight line distance.

I know pythagorus only works on a flat plane (not on a sphere), but for small distances at an angle near 0 or 90 degrees the error will be small.
For my straightout the heading is 104 degrees. The results I get are very close to Google Earth.

My straight-out is 150.0 km in Condor, 149.7km in SeeYou and 149.94km calculated based on UTM and pythagorus.

I will define a couple of straight tasks much closer to 0 degrees and 90 degrees and compare Condor, Google Earth, SeeYou and my calculated distance using UTM and pythagorus.

Ahhh yes. SeeYou is using the FAI and IGC published formula for calculating distances on a sphere. ( see the sporting code section 3 )

_________________
[url=http://condor.ephemeride.com/graduate/0/?id=540]
[/url]

Maybe I'm just a little tired tonite but I'm unable to create a calibration file as outlined here - also not getting anywhere with ASH33's suggestion above either.

First, CarlPotte's way:

This is my data:
data_format = int16
map_projection = UTM Zone 32N
ellipsoid = WGS84
left_map_x = 196380
lower_map_y = 6324480
right_map_x = 742860
upper_map_y = 6772950

The xl sheet say a couple of my numbers are invalid, anyone can help?
Also I dont get where CP get his quote "(instead of 207360, 115200). " numbers. I realise these numbers come from somewhere in hos scenery but where can I find mine?

TIA,
Haldor

_________________
Haldor
LS6a / LN-GHY
Norway

~~ if this is torture chain me to the wall ~~

My method is a work in progress.
My original post was intended to solicit comments.
I think I have all of the kinks worked out the process in more detail tonight (N. America eastern time).

Thanks for the pointer. The FAI sporting code Annex C http://www.fai.org/gliding/system/files/sc3c.pdf
contains a pointer to the following web page on page :
http://williams.best.vwh.net/gccalc.htm[url] which has a calculator for distances between two points.

It allows you to choose from a number of different earth models including
WGS84 and the FAI Sphere (referred to by the FAI sporting code.

From reading the Condor Scenery development PDF I beleive Condor uses WGS84. SeeYou uses the FAI Sphere.

I plugged in the coordinates:

45 24.660N, 76 21.930W
45 04.800N, 74 10.600W

For WGS84 it gives a distance of 175.73km and for the FAI Sphere 175.26 KM.

Clearly they are different. I am at work right now. I will investigate further tonight.
[/url]

AFAIK you can set SeeYou to WGS84 too

_________________

Correct! I found that just before I read your post.
The default setting is FAI Sphere.

To change select "Tools->Options" from the menu bar.
There is a pull-down menu at the bottom of the window.

Once you change it you need to reload your IGC file for it to take
affect.

With SeeYou set to WGS84 the difference between what Condor reports
and SeeYou is now 0.1 km (for my current calibration).

I am almost done with my work on this process. I have a couple of
little things I need to try and then I will post the process in a clear
step-by-step fashion.