Page 2 of 2

Re: VIIRS swath to grid - missing display types

Posted: Wed Mar 04, 2020 12:26 pm
by bobc
Hi HP,

Are you trying to use a derived field in the Field Selector after adding a VIIRS data source? If this is the case, this part is not yet working. I apologize for not mentioning that earlier. The plan is to get these formulas working as derived fields, but right now they do not.

As of right now, you should be able to use all of the "end-user" VIIRS formulas. These are in the same place that the swathToGrid and VIIRS_M_RGB formulas were located from the plugin. You should see them if you go to the Field Selector tab of the Data Explorer, select Formulas in the Upper Left, and expand Imagery. To evaluate these formulas you'll have to manually select each band and make sure that you subset in one of the Region tabs in the Field Selector window that appears when you hit Create Display.

If you are using the end-user version of these formulas and seeing the error, please let me know what platform you are running on.

Thanks!
Bob

Re: VIIRS swath to grid - missing display types

Posted: Sat Mar 07, 2020 4:56 pm
by hproe
Hi Bob -

I am not using derived fields (cannot see derived fields anyway) as you describe in your second para. And then thing unfold as I described in my former email. McV is running under Windows 10 64-bit.

HP

Re: VIIRS swath to grid - missing display types

Posted: Mon Mar 09, 2020 1:56 pm
by bobc
Hi HP,

As far as the derived fields, these should list as long as you have one data source that contains all of the bands necessary to create the product. For example, on my Windows 10 machine running today's nightly, if I load a data source containing SVI bands 1, 2, and 3, I see:
Field Selector tab of the Data Explorer
Field Selector tab of the Data Explorer

Under Image you find the normal bands, calibrations, and QF fields. Under Imagery, I see the derived fields listed that should be able to be produced by the data source. Again, these are just a placeholder for now since the derived fields don't currently work.

With your NDVI image below, I've seen results similar to that (squares of data, not a seamless image) if I didn't do any geographical subsetting in the Region tab of the Field Selector window when selecting my bands for the display. When selecting your fields in the Field Selector window, are you doing Shift+Left-Click+Drag in one of the Region tabs to draw a green bounding box? If you don't do this, you end up with the small squares of data in the display. This is written up as a bug/limitation inquiry that this subsetting is required right now when displaying JPSS data through a formula.

I'm struggling to replicate the problems you are seeing unless I actually select the derived field when manually selecting the bands/calibrations for the RGB. For example, when going through the True Color RGB formula, if I choose "VIIRS True Color RGB" in the Field Selector window (instead of M5 Reflectance as the label says), I see the text string you mentioned ("image values must be mapped to Red, Green or Blue only"). If instead in this window, I expand IMAGE and select the M5 Reflectance field, then things work. Can you send me an image of what your Field Selector window looks like when you are selecting the bands/calibrations for the RGB?
Field Selector window
Field Selector window

Another thing that could be worth verifying is that no local code/settings might be interfering with things. Could you try starting McV using a fresh McIDAS-V directory to see how things work? An easy way of doing this is by going to a Command Prompt window, cd to your McIDAS-V-System directory, and start up McV with "runMcV.bat -tempuserpath". This will start McV as if it were the first time ever run on your machine. Do the formulas work better now?

Also, could you please let me know the names of the JPSS files you are using? There are several ways these files can be ordered from NOAA CLASS, and perhaps we are working with different types of files. If you didn't get your data from CLASS, please let me know that as well.

Are you able to get any of the VIIRS formulas working correctly?

Thanks,
Bob

Re: VIIRS swath to grid - missing display types

Posted: Tue Mar 10, 2020 9:26 am
by hproe
Hi Bob -

Sorry I have been misled by your 'derived fields'. I have been thinking of the explicitly named 'Derived' fields, that go eg with the grid formulas. Yes, I tired this formula. I ow understand the issue and everything works fine up to one point, we have discussed earlier on (with no clear answer if I remember correctly), namely:
When, after a 'VIIRS Natural Color SVI RGB' I do another one, no image is displayed and McV hangs (the slider in the bottom right corner continues to rock). In detail I see the following after having given the OK to the band selection:
1. the label 'RGB Composite' appears in the Legend
2. the layer label from the former image hops up 1 line
3. alder continues rocking
4. 'Stop current loads' allows to continue the session
Interestingly, doing a 'VIIRS Dust RGB' instead of a second 'VIIRS Natural Color SVI RGB' work just fine. There are certain combinations that just do not work.

cheers, HP

Re: VIIRS swath to grid - missing display types

Posted: Tue Mar 10, 2020 6:54 pm
by bobc
Hi HP,

Are you adding multiple displays to the same display panel? If so, I have the feeling you're running into the bug written up in Inquiry 2296 where only one RGB Composite display can be added to a display panel. When creating these RGB functions I compared the output from the "3 Color (RGB) Image" and "RGB Composite" display types for each formula. Many of them I found looked best with RGB Composite, which is why several of the formulas only list RGB Composite as a valid display type. Others, such as Dust and Nighttime Microphysics, seemed to give a good display with both display types. I'll check with the programmers to see if we want to re-evaluate inquiry 2296 now that we're adding more RGB functions to McV.

As for the derived fields showing up under "Imagery", this is just the default, where they list under the same category that the formulas are under in the Field Selector when Formulas is selected. I like the idea of listing the derived fields under "Derived" for ABI/AHI/MODIS/MSG, and "Derived (Under Development)" for JPSS. This, as you mentioned, would be consistent with gridded derived fields and would make it easier for users to what they are. I'll look into making this change later on this week.

Thanks,
Bob

Re: VIIRS swath to grid - missing display types

Posted: Wed Mar 11, 2020 8:30 am
by hproe
Hi Bob -
I am fine with your comments and pleased that you revived Inquiry 2296. Anyway, beyond what I have said there about my preference for display type 'RGB Composite', I cannot see any difference to the other RGB display type.
HP