Beta Test / Technical

[yachtvalhalla]

On both 1.3.4 and beta build 1122 I cannot find a way to display textual data on the cm93 charts. I'm referring to island and town names, etc. I've tried different settings for vector charts without any help. Is it me?

[Ptizef]

Yachtvalhalla :
With "mariner standard" and "select al"l it should work
clear the box "show important text only" and don't forget the A+ button on the main window
JP

[philocat]

Ptizef:

Thanks for this tip.
I didn't realise this option existed (yeah not too familiar with program yet - I'm not mobile and running in armchair mode as my boat and laptop battery not working )

If the first time default option on starting OpenCPN is "show important text only" then perhaps it should be changed to show all unless user then elects to filter displayed info. I may have inadvertently altered this option so disregard this and apologies if not the case already.

I also have suggestion which I shall post in technical thread regarding separating the user entered waypoints info from the config file.

Phil

[Ptizef]

Otya
Very happy if I'd helped you

JP

[manimaul]

I have a few color palettes to experiment with vector charts. Is anyone aware where in the source the vector color palette is defined?

Thanks,

Will

[bdbcat]

Will....

Vector chart color palettes are in the file s57data\S52RAZDS.RLE

At the top of this file are lines like

CCIE   31LANDA0.36000.400049.00brown

Note the "escape" characters (0x1b)

The colors are encoded in a hard-to-interpret CCIE color map.
The names are somewhat descriptive (e.g. LANDA for land area, etc.)
There is code to convert from ccie to RGB near s52plib.cpp(387). No inverse transform (RGB to ccie), though. Google is your friend.....

If you browse this file you will see that the colors are referenced by rules which render the various features, e.g.

0001    500090
LUPT   40LU00090NILLNDAREA00001SPLAIN_BOUNDARIES
ATTC    1
INST   10AC(LANDA)
DISC   12DISPLAYBASE
LUCM    612010
****    0

Decode: LUPT......LNDARE.... (a LNDARE feature)
INST......AC(LANDA) (AreaColor LANDA)
etc.....

Anyway, you can fiddle with the values in this file, and no recompile is necessary. S52RAZDS.RLE is read as a data file. You can specify its location in the config file like this:

PresentationLibraryData=/home/dsr/Projects/opencpn_sf/opencpn/data/s57data/S52RAZDS.RLE

It has been on my mind to fool around with some alternate palettes that might be more pleasing to some folks used to traditional cm93 presentation in other ECS's.

Good Luck
Dave

[manimaul]

Great, I'll see If I can play with this a bit. I really like the Brazil and Canadian raster chart palettes. So, I'd like to duplicate them for our vectors.

OpenEv has a side project called libs52. There is some great information there on the standard color palette. S-57 data - GRASS-Wiki

libs52 palette:

[PeteCross]

Originally Posted by bdbcat

Will....

Vector chart color palettes are in the file s57data\S52RAZDS.RLE

Good Luck
Dave

If anyone is interested here is my first hack at Vector chart colors. Just some minor adjustments that seem better on my display...

The attached file after download needs to be renamed s52razds.zip and then unzipped. The resulting s52razds.doc file then needs to be renamed to s52razds.rle

/Pete

[idpnd]

A few options for palettes could be published online.. I like the opencpn default, although I agree that a lot of the prevalent paper/raster charts have a more classic/naval look

[Robert_1]

Hello,

My first message on that forum ... I apologize for my poor English writing

CM93 charts: There are some problems with the background colours of the chart and the font colours and size. Hardly readable in some cases, not enough contrast between text and background, and text overlapping.

Best regards, Robert.

Figure attached : comparison of 3 identical charts shown by OpenCPN, CMapWorld, and raster from the same area near Antibes in France. And some suggestions.

Windows XP SP3, screen 1400x900
OpenCPN :1.3.6
Depth shallow 2m, Safety 3m, Deep 15m
Display cat : mariner standard
4 colours
Zoom details : 0

[manimaul]

Nice work guys, I'm still confounded by what the heck "ccie" is. Is this CIE XYZ 1931 color space?

CIE XYZ looks promising, but the RGB conversion is just not working out. Example:

CCIE 31LANDA 0.3600 0.4000 49.00 brown
If .36 .40 49 in XYZ is supposed to be brown, how come it is blue?


HTTP = #002BBE
Web safe = #0033CC
RGB 0-255 = 0.00 43.19 190.40
RGB 0-FF = 00 2B BE
RGB 0-0.1 = 0.00000 0.16937 0.74665
CMY 0-0.1 = 1.00000 0.83063 0.25335
CMYK % = 100.000 77.316 0.000 25.335
XYZ = 0.360 0.400 49.000
Yxy = 0.400 0.00723 0.00804
CIE-L*ab = 3.613 -0.827 -119.449
CIE-L*CH = 3.613 119.452 269.603
CIE-L*uv = 3.613 -8.852 -20.896
HunterLab = 6.325 -0.908 -454.927

Any more useful pointers?

[bdbcat]

Will...

I went round and round on this some years ago....
I could never find a colorspace expert that I could communicate with.

Finally concluded that these coefficients are xyL space.

To convert xyL to XYZ

 
                        X = ( x * L ) / y;
                        Y = L;
                        Z = ( ( ( 1.0 - x ) - y ) * L ) / y;

I don't pretend to understand colorspace conversion, but this is (part of) what opencpn does to get RGB from the CCIE numbers, empirically developed.

Also, I would suggest that folks working on new PLIB data files use alternate names, instead of S52RAZDS.RLE. Might save some confusion later. You can specifiy the PLIB file location in the opencpn config file, and it can be named anything.

Good Luck
Dave

[Robert_1]

Originally Posted by bdbcat

Will...

Finally concluded that these coefficients are xyL space.

To convert xyL to XYZ

 
                        X = ( x * L ) / y;
                        Y = L;
                        Z = ( ( ( 1.0 - x ) - y ) * L ) / y;

I don't pretend to understand colorspace conversion, but this is (part of) what opencpn does to get RGB from the CCIE numbers, empirically developed.

Could this reference help you ? May be you now this software already ...

http://white.stanford.edu/~brian/scielab/scielab1-1-1/lab2xyz.m

function xyz = lab2xyz(lab,whitepoint,exp)
%lab2xyz(lab,whitepoint,exp)
%
% Converts colors in CIEL*a*b* coordinates into XYZ coordinates.
%
% lab should be in the form [L-plane a-plane b-plane]
% whitepoint is a 3-vector of the xyz values of the white point.
%   If not given, use [95.05 100 108.88] as default (not recommended).
% exp is the exponent used in the CIELAB formula. Default is cube
%   root as used in standard CIELAB. If specified, use the number
%   as exponent. (note this exponent here should be the same
%   as the exponent used in xyz2lab.m)
%
% xyz is returned as [x-plane y-plane z-plane].
%
% Created by Brian Wandell
% Last modified 4/15/98 by Xuemei Zhang

if (nargin<3)
  exp = 1/3;
end
exp = 1/exp;

if (nargin<2)
  disp('Using XYZ values of normalized D65 (95.05, 100, 108.88) as white point.');
  disp('You should really provide your own white point data to ensure reasonable results.');
  whitepoint = [95.05 100 108.88];
end

if length(whitepoint) ~= 3,
  error('White point is not a three-vector');
end

m = size(lab);
lab = reshape(lab, prod(m)/3, 3);

Xn = whitepoint(1);
Yn = whitepoint(2);
Zn = whitepoint(3);

% Usual formula for Lstar.   (y = Y/Yn)
fy = (lab(:,1) + 16)/116;
y = fy .^ exp;

% Find out cases where (Y/Yn) is too small and use other formula
% Y/Yn = 0.008856 correspond to L=7.9996
yy = find(lab(:,1)<=7.9996);
y(yy) = lab(yy,1)/903.3;
fy(yy) = 7.787 * y(yy) + 16/116;

% find out fx, fz
fx = lab(:,2)/500 + fy;
fz = fy - lab(:,3)/200;

% find out x=X/Xn, z=Z/Zn
% when (X/Xn)<0.008856, fx<0.206893
% when (Z/Zn)<0.008856, fz<0.206893
xx = find(fx<.206893);
zz = find(fz<.206893);
x = fx.^exp;
z = fz.^exp;
x(xx) = (fx(xx)-16/116)/7.787;
z(zz) = (fz(zz)-16/116)/7.787;

xyz = [x*Xn y*Yn z*Zn];

% return xyz in appropriate shape
xyz = reshape(xyz, m);

[PeteCross]

Originally Posted by Robert_1

Could this reference help you ? May be you now this software already ...

http://white.stanford.edu/~brian/scielab/scielab1-1-1/lab2xyz.m

[/CODE]

Hi,

Personally speaking, what I am searching for is a rgb2cie.c function to produce a little utility to help make a good color palette. There are endless utilities to produce RGB values. This rgb2cie.c could then give the input for the Presentation Library.

Is there such a function out there?
Or am I barking up the wrong tree?

/Pete

[PeteCross]

Originally Posted by PeteCross

Hi,

Personally speaking, what I am searching for is a rgb2cie.c function to produce a little utility to help make a good color palette. There are endless utilities to produce RGB values. This rgb2cie.c could then give the input for the Presentation Library.

Is there such a function out there?
Or am I barking up the wrong tree?

/Pete

Googled this:....

void RGB2CIE(float R,float G,float B,float &x,float &y,float &z) {
float x1,z1,sum;
x1 = 0.478*R + 0.299*G + 0.175*B;
z = 0.263*R + 0.655*G + 0.081*B;
z1 = 0.020*R + 0.160*G + 0.908*B;
sum = x1 + z + z1;
x = x1/sum;
y = z/sum;
}

So I'll see if I can do something with it....
Any observations anyone??

/Pete