CATE accuracy - reference points - moved result

F4Browse, CATE, Tacedit, Terrainview & TheaterMaker etc utility related

Moderators: Lone Wolf, Snake Man

Post Reply
User avatar
Luk
Brig. General
Posts: 455
Joined: 2007-04-23 09:59:36

CATE accuracy - reference points - moved result

Post by Luk » 2009-10-19 13:02:37

Hi,
I wanted to update my terrain mesh for Eu-mod-testing yesterday (some road fixed,new objectives,leveling... etc).
So I wanted to update X1 colors in this pass as well.
I have also some BMPs for finall X1 pass in progress, but they are not finished yet.
It is because my resource map is WGS84 I think. (F4 -Eu uses sinusoidal projection on WGS 84 spheroid, if I can remember correctly).
Unfortunally I dont have acess to ArcGiS software at the moment and dont want to look for some black software just for 1 transformation.
I have ArcGIS project with referenced resources and good projection from past. If somebody has ArcGis sofware, help with 1 picture conversion is very wellcome. (I am not able to choose good projection in free ESRI viewer).
256color cropped samples are here: (projection just about stretched in bimap editor)
foto colours - more noisy
img]http://www.maxipicture.cz/verejne/09/43 ... ur_256.gif[/img]
solid coloured
img]http://www.maxipicture.cz/verejne/09/43 ... ur_256.gif[/img]
I will experiment with them. It can be blend of both pictures in final....

So I had to make some temporary X1 bitmap. I have used *.ter image as basemap and edited with some few clicks.
Basemap created with "THR_Creator_64_128".
img]http://www.maxipicture.cz/verejne/09/43 ... om_thr.gif[/img]

The result is still a little bit agressive. But I have used quite agressive colours in my experimental bitmap.
BTW - I did not have to use lxnormal fix (I was suprised). Perhaps I have some good SPTinstall version......?
Image
--------------------------------------------------------------------------------------------------------------------
And after all this TOPIC:
Tiles are coloured. But moved about 3 tiles to west!
You can see eastern border of the theater here:
Image
view from NW
img]http://www.maxipicture.cz/verejne/09/43 ... _10_41.jpg[/img]
My source X1 map for CATE was of course 2048 x 2048 pixels (like terrain mesh).
Image
Why is it moved?
Begining of *.conf file:

Code: Select all

BMPFileName=d:\Cate\X1.bmp

FullRegionTest=No

#UpdateOceanTiles=OK

[Rules: 0]
Color=0,0,0
ForceFog=0 * *
[EndRules]
...etc....
I think my conf file has standard begining....
Where is the problem please?

Thanks
Luk
P.S. the X1 update process is very fast now. Thanks to MAPEditor and 4 byte CATE. It takes only a few minutes while processing CATE.
Last edited by Snake Man on 2009-10-19 13:13:07, edited 3 times in total.
Reason: please dont hotlink very large (file size) images. 100kb per image is ok, 300kb is not.

User avatar
Luk
Brig. General
Posts: 455
Joined: 2007-04-23 09:59:36

Re: CATE accuracy - strongly moved result

Post by Luk » 2010-01-08 13:33:20

Hi all.
As I have written in another post, the previous problem was solved.
In conclusion, I was able to update whole theater X1 value by 4byte CATE correctly.
BTW - 2byte CATE result was totaly messed up noise. 4byte CATE problematic and about few tiles moved result was
probably caused by my source map. I have used keyboard arrow keys to move some layers by one pixel. It worked well,
no matter how much zoomed my image was (still 1 press - 1 pixel). I dont know what happened, but once I moved map about 3 pixels/1press. Just for explanation....

Current PROBLEM
I am retiling whole theater now. I have painted continental Eu terrain by hand (cost, cities, rocks, snow and
some forrests, slopes and lakes). It is painted acording to features (road/rivers) position, so I am preventing tiling errors.
The source bitmap is not finished yet, but in advanced stage - so I want to tune my confings now (then finish it).
This is zoomed random mountain range part of my source map.
Image

This is my updated 1st. passed terrain in TV. Mountains tiles are wrongly placed in coastline area, but the segment itself is correcty tiled(highs are of course bad).
Image

It is strange. I have extracted terrain bitmap from "thr" file. It shows the problem:
Image
WTF is this? There is no need to set 64/128 theater for CATE I think...
BTW - I have used 4 byte CATE in pictures. But I think 2byte CATE makes the same.
Thanks
Luk

P.S. of course my source bitmap is placed accurately now - It can not be the reason...
Image
Last edited by Luk on 2010-01-09 11:24:19, edited 3 times in total.

Snake Man
Commander-In-Chief
Posts: 9338
Joined: 2000-07-31 22:01:01
Gaming Interests: ArmA, ArmA 2, Falcon 4.0 and OFP.
Editing Interests: All, I (try) to edit everything.
Location: PMC
Contact:

Re: CATE accuracy - strongly moved result

Post by Snake Man » 2010-01-08 15:39:46

Luk wrote:It is strange. I have extracted terrain bitmap from "thr" file. It shows the problem:
http://www.maxipicture.cz/verejne/10/01 ... 10-thr.jpg
WTF is this? There is no need to set 64/128 theater for CATE I think...
That's the THR view from Tacedit?

I recall seeing some problematic Tacedit THR views during development, but I cannot remember what was the cause.
PMC Tactical Forum New User Registration please read new info here.

PMC since 1984

Editing knowledge, visit PMC Editing Wiki
The leading, most detailed and comprehensive modification made for the Vietnam War - Vietnam: The Experience homepage
View our videos in PMC Youtube channel

PMC Tactical forum Advanced Search is power.

"ALPHA BLACK TO PAPA BEAR. ALL RUSSIANS ARE TOAST. OVER."

User avatar
Luk
Brig. General
Posts: 455
Joined: 2007-04-23 09:59:36

Re: CATE accuracy - reference points - moved result

Post by Luk » 2010-01-08 16:01:40

That's the THR view from Tacedit?

I recall seeing some problematic Tacedit THR views during development, but I cannot remember what was the cause.
Yes, THr view is from tacedit. But THR is produced by Phoenix's THR creator. It is created well.
The problem is tiled terrain - the source for THR creator. You can see the moved result also in second picture (terrainview).

CATE have tiled it moved about a half of theater.
Well, I am starting to feel like an artillery member. 1st shot 10 miles behind, second 10 miles in front...the 3rd bingo!
I can modify my source bmp and "divide" it to match that crapy offset. But I would be really happy when find some clean and
"straight" solution.

To be more clear- the question is:
Why the CATE tiles the terrain so moved in case of BMP rule?
(feature pass is positioned well - datas come from "TDF" in that case)

Did you have similar experience?

Luk

Snake Man
Commander-In-Chief
Posts: 9338
Joined: 2000-07-31 22:01:01
Gaming Interests: ArmA, ArmA 2, Falcon 4.0 and OFP.
Editing Interests: All, I (try) to edit everything.
Location: PMC
Contact:

Re: CATE accuracy - reference points - moved result

Post by Snake Man » 2010-01-09 09:11:20

Sorry don't have any idea why it might do that.
PMC Tactical Forum New User Registration please read new info here.

PMC since 1984

Editing knowledge, visit PMC Editing Wiki
The leading, most detailed and comprehensive modification made for the Vietnam War - Vietnam: The Experience homepage
View our videos in PMC Youtube channel

PMC Tactical forum Advanced Search is power.

"ALPHA BLACK TO PAPA BEAR. ALL RUSSIANS ARE TOAST. OVER."

User avatar
Luk
Brig. General
Posts: 455
Joined: 2007-04-23 09:59:36

Re: CATE accuracy - reference points - moved result

Post by Luk » 2010-01-20 17:30:26

Some additional info to the topic of CATE accuracy.
I did new configs for eu theater in last weeks. The configs are for terrain types, transitions and features tiling.
I can say, I was quite suprised, it is not so complex science(just boring/time consuming experience).
I have tested my configs and they do, what I want. (We have snow, rocks, slopes, lakes and even the rivers in the cities ...of course cityrivers looks fine in large cities, in small towns they need some fixes).

But CATE "bitmap-handeling/BMP rule" is still big pain for me.
I moved my source map to match the "inacuracy" and it worked well for the first time(source bitmap moved about half of theater + one tile north).
Then I was working with result and I tested other rules with it.
But I have also tuned my source map during the proces.
Then I wanted to update my first (BMP) pass. CATE was unable to load my source map again.
I have discovered some process how to load bitmap(pointing to 1st BMP-error, change it to another identical + loading) but the result was still bad.
It looked it is still operating that 2weeks old source. I have painted some obscene symbols in source map but CATE was unable to update terrain.
I have tried a lot (25x). I was frustrated and tried to run exactly same files in another PC with win2000system.
It worked fine. But the movement of the bitmap should be a little bit different - about 4 tiles more to west.
Ok, I have updated source bitmap - but I was unable to load it again!
Of course I restarted CATE everytime. I started to restart the system every time. But it did not helped.

I have installed virtual PC machine with W2k on top of my XP laptop. I did a snapshot of the system and I lunched CATE.
It loaded my last bitmap without problem(but it takes 20-30minutes in virtual PC). The applying of rules took 20-30seconds.
I am saveing the result now (about 45minutes). I am very curious, wheather the movement of the source map will be the same for both W2k systems.
......I write qite slow in english.......
Ok after about 45 minutes or an hour virtual machined CATE saved the result. The result is fine and matches the same offset/movement/deformation needed for previous W2K system(which was unable to load the updated *.bmp for second time ). It looks the problem is system dependent and CATE uses some CACHE/temp data which makes a problem. I am unable to locate these temporary datas/entries for deleting.

So my question is - is it possible for somebody competent (Phoenix?) to roughly check the code and find the "temporary datas" location please?
OK I have some solution now - to restore my W2k virtual PC state again and again...but it is really slow..

Luk

User avatar
Luk
Brig. General
Posts: 455
Joined: 2007-04-23 09:59:36

Re: CATE accuracy - reference points - moved result

Post by Luk » 2010-03-12 13:04:27

I am sorry for double post, but I dont want to misguide somebody.

I think I have solved my problems:

1st problem - loading a bitmap, loading changed bitmap etc.

Do not use relative paths in your configs. Relative means - "../datas/example.bmp" for example.
Use complete path like "d:/CATE/datas/example.bmp".

With complete paths the CATE works great!


2nd problem - moved result when bmp rules used
Your 8 bit BMP file must have full 256 colour palette. In case you have for example 20 colours in BMP file, you should fill your palette with some "balast" colours to have all colour slots full. Rather do not use "logical" colours ("black, white" named in save dialogue). You can use them as generic colours "rgb 0,0,0" etc.
2048x2048 (128 theater) bmp is fine when 4 195 384 bites in size. Other sizes causes moved result. If you miss few colours in your palette, your result will be moved just few tiles. If your palette is more space-saving, your result can be moved about half of theater...

So mentioned problems can be easily avoided. It is not 2byte/4byte CATE version dependent, neither system dependent.

Luk

Post Reply

Who is online

Users browsing this forum: No registered users and 18 guests