Mc-V hanging with AVHRR GAC

[kbedka1]

I created a local AVHRR Level 1B server and I am trying to look at Global Area Coverage data. I am able to bring the data in, see the various image times in my folder, but when I select an image and try to choose a parameter (i.e. raw, temp, brightness) for a specific channel, McIDAS-V hangs for a LONG time (i.e several minutes). Eventually I am able to subset and/or display, but many users will likely give up and force kill the process.

[Rick]

Hi Kris,

Are you making a local ADDE dataset? I just downloaded a GAC pass from the CLASS system and it worked just fine (loaded within a minute and display in about the same amount of time).

Here is the file I used

ftp ftp.ssec.wisc.edu
user anonymous
password e-mail
bin
get
NSS.GHRR.NN.D10305.S2332.E0114.B2808788.GC

I did have to move the image to a vertical level of 0.

[kbedka1]

Yes I am creating local ADDE, I tried your file and am experiencing the same behavior. I'd say the lag was about 90 secs, if I click temperature, and then I say, "oops, I want radiance", I'd have to wait another 90 secs with a blue spinning ball on my screen (Mac OS-X) Based on Mc-V behavior with other local or remote) ADDE data, this is quite unusual behavior. Usually the available display types, times, region, etc.. pop right up. If anything were to take a "long time", it would be the display of the user selected data, especially if over a large domain.

There seems to be some unique behavior with this AVHRR data. Even though the AVHRR GAC swaths are ~12000x400, other data types, i.e. full disk SEVIRI, have more total pixels but behave more efficiently in V. Is this caused by the fact that the entire image navigation has to be built from anchor points stored in the GAC files? If so, this issue will not go away in the near future...maybe the navigation manipulation should be done before the list of fields pops up in the data chooser? A long initial data load would seem more reasonable to a user than a delay in providing display types and region subset info. That's my two cents at least...

[joleenf]

Kris,

For what it is worth, I tried loading some HRPT data that I had used in the past. It is also slower than I remember with about a 26 second lag time when choosing the data that I wish to load. However, if I choose something like 3.74 micron Temperature, then change the selection to a different choice (still see lag time), but then go back to the original choice of 3.74 micron Temperature, there is no lag. I don't know if that makes any sense or any difference.

Joleen

[kbedka1]

I still experience the lag, even if I go back to a parameter that I had previously selected.

[Rick]

I guess our definition of LONG time is a bit different I am seeing a similar lag. The majority of the problem is due LALO navigation and the new displaying the coverage the region tab. Reading in all those points takes a long time. We are going to be working on performance issues with the image chooser, hopefully some of that work will help.

[kbedka1]

It would be OK if this only happened once up front and did not happen again every time a user chooses a new field. Maybe a message could be displayed alerting a user that Mc-V is going to take a while to bring up the region tab, otherwise, one might think that it has hung.

[joleenf]

I agree with Kris. Previously, this may not have been noticed much because in the old chooser, the user had to select the individual band, region, etc. and then load into the field selector. Therefore, a time lag is expected when new fields are selected because the user would manually go through the entire process again. The new format of the chooser nice all the data available is listed. However, the user might expect more speed because it appears that all the data has been queued and is ready to load into the display. A warning message would help relieve some fear that the session has hung permanently.