[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[IDV #RGL-474313]: IDV v2.9u4: Reading Satellite data from NOAA's CLASS archive



> 
> On Apr 6, 2011, at 9:30 AM, Unidata IDV Support wrote:
> 
> > Hi Dave,
> >      I tried all three files you uploaded, and all worked fine in the IDV.
> >      There are a few points you need to pay attention. The first, all three 
> > files are pretty big in dimensions. With the screen resolution available in 
> > your computer, you need to subset them before creating the display. Second, 
> > the IDV need 2G memory allocated to perform better in these three plots.
> 
> Yuan,
> 
> OK, I've increased the RAM allocated to the IDV from 1.5 GB to 2 GB. (Too bad 
> I can't do this on our older MacBook Pro laptops, which have only 2GB of RAM.)
> 
> In the Dashboard, under Data Choosers > Sat & Radar > Images, I've accessed a 
> GINI 1km  visible West CONUS image.
> 
> I think I understand now that specifying the image center location and the 
> image size together, controls subsetting, right?
> 
> Magnification appears to control pixel sampling and hence image display 
> resolution.
> 

Yes.
> Now, the raw image dimensions are 5120x4400 pixels, and my screen dimensions 
> are 1920x1080. Although I can't plot the image with its raw dimensions, nor 
> can I plot the image at 95% of its raw dimensions (4864x4180), I can plot 
> images with full 1 km resolution (magnification of 1) by specifying image 
> dimensions of 4608x3960 (90% of the raw image dimensions), which exceeds my 
> screen dimensions by a wide margin.
> 
> So, there's something about this that I don't understand--why can I view 
> images with dimensions much larger than my screen dimensions and up to at 
> least 90% of the raw image dimensions, but not 95% or larger? (The raw file 
> size is on the order of 5120x4400 ~= 20MB, which seems as if the available 
> RAM should handle easily. Am I missing something there?)

If your screen resolution is 2000x1000, and likely your IDV is a small portion 
of the whole screen, this is why I feel like 1000x500 should be enough. 
However, you can certainly select the dataset resolution as 5120x4400, it is 
just a waste and not getting any benefit in the display.
> >
> > goes11.2008.... and 1102041400_sir.mca are AREA files, you can specify the 
> > data type as "McIDAS AREA files" and no need to subset, the IDV will do it 
> > for you.
> 
> I am now able to display the "goes11.2008..." file (downloaded from NOAA's 
> CLASS image archive), even though there appears to be no way to subset the 
> region or increase the stride. SInce the only change I think I've made is to 
> increase the memory allocated to the IDV from 1.5GB to 2GB, I assume that 
> that must have done the trick. (The image file size is about 70MB.)
> 

If the image is coming through the ADDE server, you can do the subset or 
strike, otherwise, user has no control.

> Interestingly, although that image supposedly is a 1 km visible image, 
> zooming in and inspecting the pixels seems to show that the resolution is 
> actually about twice that (24 pixels span a distance of 12 km). I don't 
> understand that!

this image line and element resolution is 
line_resolution = 4; // int
element_resolution = 8; // int

I am not sure the unit, but I don't think the IDV changes the resolution.
> 
> I can also display the 1102041400_sir.mca AREA file, which must also be due 
> to the increase in memory allocation. This is great--I've been wanting to do 
> this for years now!
> >
> >
> > 110306200_vis1km.nport is grid file, you can specify the data type as "Grid 
> > files", and the dimesnions are 4400 X 5120, you do need to set the strike 
> > as every 3 or 4 point, unless you have very big memory allocated to the IDV.
> 
> I specified a data type as "Grid files" and tried a stride of 4 points. After 
> an interminable period waiting for the image to load, I gave up.  I also 
> tried spatially subsetting the area to plot. Same result. The files are 
> around 16MB.

It took about 55 seconds in  my machine. One trick is that you don't want to 
create the contour plan view, but choose the color shaded plan view. To create 
the contour is very expensive.
> 
> NOAA's Weather and Climate Toolkit (WCT) seems to get around the file size 
> problem by using Google's progressive refinement approach, which means that 
> users don't have to set appropriate image size and magnification before 
> trying to view the image. That's a big advantage, and it's the kind of thing 
> that attracts users such as my colleague, John Monteverdi, to more 
> special-purpose applications like the WCT instead of the IDV. Any chance that 
> the IDV could incorporate progressive refinement for viewing large image 
> files?

I strongly agree that progressive refinement is great enhancement feature, I 
will think about to implement it in the future.\\\


Yuan
> 
> -- Dave
> 
> 
> ************************************************************
> *   Dr. Dave Dempsey                        |       ^   ___       \|/     *
> *   Dept. of Geosciences                    |  ) ^     /||_||\    --0--  *
> *   San Francisco State University   | )  )    /  ||_||  \    /|\     *
> *   1600 Holloway Ave.                      |  )  )  /   ||_||    \          *
> *   San Francisco, CA   94132          |  )  ) /    ||_||      \   ^   *
> *                                                             | )  )  )    
> ||_||        \      *
> *   Phone:  (415) 338-7716               |  )  )  )~||~||~~~~ \~~*
> *   FAX:      (415) 338-7705               | )  )  )  ) ~  ~  ~ ~ ~ ~*
> *   Email:   address@hidden        |  )  )   )  ) ) ~  ~   ~ ~ *
> ************************************************************
> 
> 
> 
> 
> 


Ticket Details
===================
Ticket ID: RGL-474313
Department: Support IDV
Priority: Normal
Status: Open