Re: [thredds] Re-enabling a auto-scaling defaultColorScaleRange in ncWMS

Hi Heiko,

The only difference between the behaviour of ncWMS and THREDDS in this
respect is that ncWMS will automatically scan each dataset and deduce a
"sensible" colour scale range for each variable, then save this in the
config file ($HOME/.ncWMS/config.xml).  THREDDS has no such background
process for scanning datasets (at the moment), so you have to enter this
information manually in wmsConfig.xml.  (You can use THREDDS' built-in
Godiva2 client to help you work out sensible defaults if you like.)

Both of these mechanisms set the *default* colour scale range for each
variable.  This is used if the COLORSCALERANGE parameter is omitted, and
hence allows backward-compatibility with standard WMS clients.  An
"auto" colour scale range is quite different, and sets the scale range
on a per-image basis.  As you say, this isn't particularly useful except
for quick-looks and debugging.

See for full
details of the COLORSCALERANGE parameter in ncWMS.  THREDDS behaves
identically, except from the point of view of configuration, as
described above.

I know this is a little confusing - sorry about this.  If you configure
your colour scale ranges manually (sorry) in wmsConfig.xml your external
WMS users should be able to access your images without a problem.

(One day we might be able to port the ncWMS dataset-scanning mechanism
to THREDDS, but we're concerned about performance and scalability.
THREDDS servers can have a *lot* of datasets.)

Hope this helps,

-----Original Message-----
From: Heiko Klein [mailto:Heiko.Klein@xxxxxx] 
Sent: 21 September 2010 13:21
To: Jon Blower
Cc: thredds@xxxxxxxxxxxxxxxx
Subject: Re: Re-enabling a auto-scaling defaultColorScaleRange in ncWMS

The method you implemented before you added the defaultColorScaleRange 
parameter in ncWMS was different from what you describe here. I'm quite 
sure that all images had the same scale and the values where adapted to 
the data-values within the file, i.e. for sea-ice concentration data, I 
got a scale between -5% - 105%. This was much better than what I get 
currently with the default -50 - 50.

Seen from a performance point of you, it makes sense that an 
auto-colorrange tries to adjust the scale per image-data - but in the 
tiled WMS world, this doesn't make any sense, and if you plot a scale, 
you will need to decide what data to scale to, since no region is 
selected at all.

And since the COLORSCALERANGE is a ncWMS extension, it is difficult for 
a general purpose wms-client to adapt to it. With the new ncWMS in 4.2, 
we will need to adapt our client to make use of these extensions to get 
useful pictures. Those producing the data are partly external users, and

have only very limited rights, and definitely no administrative rights 
to configure the wmsConfig.

Best regards,


On 2010-09-21 12:49, Jon Blower wrote:
> Hi Heiko,
> The problem with colorscalerange=auto is that the scale range is
> automatically chosen *for each image*.  The purpose of setting values
> the wmsConfig.xml is to ensure that each image from the same variable
> gets the same colour scale for consistency with the legend and between
> tiles in tiling clients.
> You can still use the URL parameter&COLORSCALERANGE=auto if you want
> your image to be automatically contrast-stretched, but beware that the
> image won't match the legend.
> Does this make sense?
> Cheers, Jon
> -----Original Message-----
> From: Heiko Klein [mailto:Heiko.Klein@xxxxxx]
> Sent: 21 September 2010 09:04
> To: thredds@xxxxxxxxxxxxxxxx
> Cc: Jonathan Blower
> Subject: Re-enabling a auto-scaling defaultColorScaleRange in ncWMS
> I upgraded to a snapshot of thredds 4.2 and I have seen the new
> configuration options to ncWMS. In particular, the
> defaultColorScaleRange seems to be new.
> I liked the old auto-colorScaleRange which determined the max and min
> value from the datasets. Is this option still available within the
> wmsConfig.xml file? I tried
> <defaultColorScaleRange>auto</defaultColorScaleRange>  and it didn't
> work.
> Best regards,
> Heiko

Dr. Heiko Klein                              Tel. + 47 22 96 32 58
Development Section / IT Department          Fax. + 47 22 69 63 55
Norwegian Meteorological Institute 
P.O. Box 43 Blindern  0313 Oslo NORWAY

  • 2010 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the thredds archives: