[ncl-talk] RasterFill plot using RGBA colour map

[Rick Brownrigg]

Hi Kay,

I've looked carefully at the computed colors in your colormaps, and I agree
its perplexing as to the resultant colors in the maps and the labelbar in
the unpaneled case.  certainly perplexed as to the resultant colors. We
have a known issue with RasterFill and RGBa colors -- it manifests rather
differently, but I'm wondering if your case is related.

How much trouble would it be for you to provide your scripts and data for
us to take a deeper look at the issue?  If the data are to large to email,
they can be posted to an FTP site:


On Tue, Mar 10, 2015 at 5:19 PM, Kay Shelton <kay.shelton at gmail.com> wrote:

> Hi Rick,
>
> Thanks for the response. I had followed that example to make one of my
> plots, but for the others I wanted to avoid needing to know the contouring
> values and number of required colours.
>
> For the first plot (following example 18), I have:
>
>   opacity_val = 0.5
>
>   colour_map0 = "BlueYellowRed"
>   cmap0 = read_colormap_file(colour_map0)
>   colours0 = (array_append_record(2,floattoint(fspan(255.,129.,4)),0))
>   colours1 = (array_append_record(colours0,floattoint(fspan(128.,2.,6)),0))
>   colours = (array_append_record(colours1,255,0))
>   colour_levels = cmap0(colours-2,:)
>   colour_levels(:,3) = opacity_val
>
>   resFa0 at cnFillOn = True
>   resFa0 at cnFillMode = "RasterFill"
>   resFa0 at cnRasterSmoothingOn = True
>   resFa0 at cnLevelSelectionMode = "ExplicitLevels"
>   resFa0 at cnFillColors = colour_levels
>   resFa0 at cnLevels = (fspan(0.,100.,11))
>
> For the second plot I have:
>
>   colour_map1 = "WhiteBlueGreenYellowRed"
>   cmap1 = read_colormap_file(colour_map1)
>   cmap1(:,3) = opacity_val
>
>   resFa1 at cnFillOn = True
>   resFa1 at cnFillMode = "RasterFill"
>   resFa1 at cnRasterSmoothingOn = True
>   resFa1 at cnLevelSelectionMode = "ManualLevels"
>   resFa1 at cnMinLevelValF = 0.5
>   resFa1 at cnMaxLevelValF = 10.
>   resFa1 at cnLevelSpacingF = 0.5
>   resFa1 at cnFillPalette = cmap1
>   resFa1 at cnSpanFillPalette = True
>
> Both methods seem to have the same results with the hideous colours (note
> the actual colours that result vary depending on the opacity_val I set),
> when panelled the label bar colours are correct.
> (Also note, cnRasterSmoothingOn = True seems to be entirely ignored,
> turning it on and off make no difference to the plot. Separate issue, but
> odd.)
>
> A pdf showing the resulting plots can be found here:
> https://drive.google.com/file/d/0B6ntGQfT_YcxZGdFMEhGUEx6aTQ/view?usp=sharing
>
> If I set cnFillMode to "AreaFill", I get exactly what I want, except it
> takes a fair while to plot. For the moment, I can apply smoothing to the
> fields to make them plot quicker, but for publication, I would prefer not
> to have to do that, so I'd rather use a raster plot.
>
> I am a little stumped by this.
>
> Any help would be greatly appreciated.
>
> Kay
>
>
>
>
> On 10 March 2015 at 03:28, Rick Brownrigg <brownrig at ucar.edu> wrote:
>
>> Hi Kay,
>>
>> I'm not totally sure I understand what you are after, but does RGBa
>> example #18 address your issue?
>>
>> http://ncl.ucar.edu/Applications/rgbacolor.shtml
>>
>> It tested it with RasterFill and it behaves as expected with regard to
>> the opacities of the labelbar.
>>
>> Hope that helps..
>> Rick
>>
>>
>> On Sun, Mar 8, 2015 at 8:24 PM, Kay Shelton <kay.shelton at gmail.com>
>> wrote:
>>
>>> Hello NCLers,
>>>
>>> (I am running 6.2.0 and 6.2.1)
>>>
>>> Should I be able to make a contour plot using RasterFill and setting
>>> cnFillPalette to a colour map with alpha values <1?
>>>
>>> I am trying to do this because I need to use RasterFill as I have high
>>> resolution global data that takes an age to plot as an AreaFill contour
>>> plot, but I want the opacity level reflected in the colour bar. The
>>> alternative is to set cnFillOpacityF, however then the plot colours do not
>>> match the label bar very well.
>>>
>>> My current result when using RasterFill and the colour map with edited
>>> alpha values is hideously coloured plots, but with required opacity level.
>>> For plots that are not panelled the label bar matches the plot, but is in
>>> hideous colours, for panelled plots the plots are hideous but the label bar
>>> is correct.
>>>
>>> Is what I am trying to achieve even possible?
>>>
>>> Thanks,
>>> Kay
>>>
>>> _______________________________________________
>>> ncl-talk mailing list
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20150311/502f2b81/attachment.html