IP and NASA images
49 messages in this thread
Ben:
I am having some problems with your NASA PI module. I am using it with IP
6.16. The browse images work just fine but I have not been able to get any
of the compressed images to display. Mostly I get GURU's. Occasionally it
will load the image but the display is blank.
As it works with uncompressed images I figure I have installed it
correctly. Any ideas?
I am very impressed with the job it does on the browse images!
Thanks,
-=Martin=-
– via Whap!
Probably you don't have enough memory, that's my guess. It does work. Those
images are 800×800, and that means they take up 4,480,000 bytes for ONE
image if the undo buffer is on.
You need a meg for the image processor, too, room for the screen, the
operating system, and the PI Module. Scary, huh? <grin>
I would say that you do have it installed correctly.
One thing you can try is turn off the undo buffer. That may help.
Also, select the lowest resolution display screen you can, using the least
resources – that would be either AVG or LUMA modes. You don't need color
anyway, as they have none.
Don't load any other images first.
If you're close on memory, maybe these hints will bring you over the edge
so it will work for you. Good luck!
–Ben
Are most of the NASA images at 800×800? If so, given the memory
requirements, maybe you could alter the FITS PI module to allow reading the
800×800 FITS image into IP in a couple different ways:
1) into an arbitrary sized buffer in IP. As you read the image in, you
scale it to fit the buffer size specified. (ie. a 800x800fit image may be
imported as a 320×400 buffer) I know that this would cut down resolution,
but it would allow users who didn't have extra-large amounts of ram to make
use of the images.
2) allow importing only a subset of the full image. For example, bring in
only the top-right 320×400 of an 800x800fit image.
Of course you could always combine the two, where the user would be able to
bring in an subset of the image into an arbitrary sized buffer. 🙂
Just a thought….
–jp–seattle–
The new FITS module allows importing as 400×400, which will take 1/4th the
memory. Should help quite a bit in marginal cases.
Yes – most images are 800×800, at least on those CD's.
–Ben
Martin, I just spoke with Pete (he wrote it, not me) and we'll add a "vut
in half" option. That should help a lot – it'll load the non-browse files
as 400×400, which will take 1/4th the memory. Give us a few days to do it.
–Ben
Ben:
I am not sure memory is the problem. I have 8M fast and 2M chip on my 3000.
IP reports over 7M free. I tried the memory conservation measures you
suggested and I still GURU. The GURU seems to happen immediately after I
select "don't save the info file", i.e. it doesn't get that far into the
file.
Do you have the name of one of the NASA images that you know works? That
way I could try that specific file on my system.
I will look for the new PI module, that should determine whether or not
it's a memory problem.
Thanks for your help,
-=Martin=-
– via Whap!
Martin, the only images I have are the ones that John Foust sent; they
don't have names, just numbers (got'v work, so what else is new?)
Tell me about the guru. a 3? a 5? What cpu are you running?
–Ben
If you get a CD-ROM drive, I'd be glad to loan you a couple discs. I'm sure
one would be sufficient. 🙂
Might just take you up on that one, John. OTOH, if I got a CDROM drive, I'd
probably just buy the disks… certainly they're inexpensive enough.
What I need is more slots. :^( Or to change disk controllers; the 2090a
isn't smart enough to drive the SCSI CD-ROM units.
–Ben
I didn't just change controllers, I used the CD-ROM drive as an excuse to
get the latest in GVP 030/SCSI combos. 🙂 I talked to someone who had
talked to someone in the USGS, they're working on new disc sets of Mars
imagery, including a color version of the (B/W) discs they're publishing
now, with an improved index. With Magellan out of commision, their plans
for the full 300 disc set of radar maps has been cut short, too.
Know what I'd like to see? A complete set of Terra maps… they certainly
have the data.
For that, I'd instantly buy a CD-ROm drive, no questions asked.
–Ben
I'm sure there's a "reason" why they don't – can you say ICBM or
terrain-following bombers? It's a shame, it would be fun. There's the DEM
files, but they haven't put those on CD yet. They have the line-oriented
DLGs on CD, though.
BTW, thanks for the mention in AmigaWorld. Guess we finished that loader
just in time, eh? <grin>
–Ben
John,
Great article on CD-Roms in Aw arrived today! Now I have some sources
for CD-Rom catalogs. Thanks.
My CD-Rom drive (Toshiba 3301B) arrived on Thurs. along with an outdated
version (1.2 May 1991) of the HMC FF disc. Works great with the CDrom-FS
handler. NASA is behind in shipping but I was promised my discs would be
here by Monday.
….Mark Sterling 74335,1677
John:
I want to second that… great info on CD-Rom & the Amiga. A terrific
reference for those of us who've been wondering about the what-when-how &
where's of CD-Roms.
regards,
Michael
I thought it was kind of funny they talked about the NASA roms but didn't
mention HOW you were supposed to view them… at the time they wrote that,
I guess there was no way.
Now you can use Imagemaster; and since John kindly gave the code and sample
images to Perry over at ASDG, I presume you'll be able to use AdPro to do
the same at some time in the future. We've already written Amiga World to
let them know that we support the format.
–Ben
Ben:
Hmm… I thought the article mentioned IM could handle the NASA images… I
must have gleaned that from here. Didn't John send ASDG the format at their
request?
regards,
Michael
I don't know why he did it – you'd have to ask him.
–Ben
>>terrific reference … on CD-ROMs …<<
Whoops! I missed this — please repeat?
— Eric A.
Eric– March '92 AmigaWorld… Fouts has a nice piece on CD-Rom
Michael
Test Piloting 0.02A
>>nice piece on CD-Rom<<
Got it! Just arrived today — thanks!
— Eric A.
I haven't seen the article yet! What's on the cover?
MULTIMEDIA SOLUTIONS is the cover headline
Michael
Test Piloting 0.02A
John,
The AW cover story is "Multimedia Solutions". There is a stylized sorta
touchscreen deal with Apollo 17 images done by Hoel Hagen and Karen Watson.
Your article is hyped on the front. Your article has three images. 1. A
collage of space images 2. A desert Storm pic 3. The red spot on Jupiter.
Tomorrow I get to fool with the real images!
….Mark Sterling 74335,1677
Ben:
I get:
NJPLRD
Program failed (error # 8000000B)
Suspend or Reboot
Then I get the GURU, same error code, when I select reboot.
I am running a 3000-25, 68030 CPU. KS: 37.165, WB: 37.52
I'd be happy to upload some of the images if you need them.
Thanks,
-=Martin=-
– via Whap!
Hum. are you running the image processor from a shell or the WB? If a
shell, how big is the stack? I'm using 40k here, for reference, and have
seen no failures. If none of that helps, then EMAIL me an image that fails,
and I'll try it here. If it crashes, we'll fix it (of course).
–Ben
Ben:
I normally run ip from the workbench. I just checked the stack size and it
was 4096. So I changed it to 40,000. The GURU changed to 80000004. Went up
to 100,000. Same result.
Then figured I would try running from the shell, WShell actually. Stack
size at 40,000. It ran fine! Still can't get it to run from the workbench.
I am going to check out my system see if I can find a problem. If not I'll
upload the image. Not sure why it would work from the shell and not from
the WB. Any ideas?
Thanks,
-=Martin=-
– via Whap!
Probably it is a stack problem, one way or another. It's obviously somthing
to do with the way that the system invokes WorkBench processes as opposed
to shell processes; we run it from the Amiga shell (never got our upgrade
notice from Wishful Thinking, and the new Amiga shell really is much better
than the old Amiga shell) and have not seen problems.
For the time being, run it from the shell and you should be fine. I'll
check with Pete about it's stack usage… part of the problem may be that a
large amount of the code in that loader came from NASA, and we're not
intimately familiar with it all. But we'll see what we can do.
Thanks for chasing it down.
–Ben
Hmmm. A slightly old version of kickstart and workbench might be a
problem. My 3000 has KS 37.175 WB 37.67. I am still waiting for my NASA
discs to arrive. I will be able to try them in a couple of days. Nothing to
feed the FITS IP module yet.
Which NASA discs do you have? I will have the 12 Voyager discs and 4
Mars Viking discs.
….Mark Sterling 74335,1677
Martin, you might need to get the latest version of 2.04 before your NASA
loader will work properly. The version I have for the 3000 is Kickstart
37.175 and Workbench 37.67.
John – K7KB (via Whap!)
John and Mark:
Thanks for the info, I thought I had the latest. I will make a trip to my
dealer and see what he's got.
-=Martin=-
– via Whap!
KS 37.165 ?? 37.175 is the 'release' version. Unless you're a developer
you shouldn't even have .165 and if you are, why haven't you upgraded
175 has been out for quite a while.
I just tried the FITS loader with my 12mb 3000/25, KS 37.175 WB 37.67
and it DOES crap out if launched from WB. Running it from CLI with a
healthy stack eliminates the problem for me.
BTW. Wow, it works great with the NASA CD-roms. I just got 16 discs of
Voyager and Magellan images. Loads are quick and I've got great image
processing at my fingertips. Next I need to: 1.Get the aspect ratio right
2. Learn how to erase those black dots en masse 3. Assemble pics shot with
filters into full color images.
Many thanks to BBS for coming through. Haven't found any images that
won't load yet but I have barely scratched the surface (and I mean JUST
barely).
….Mark Sterling 74335,1677
The FITS pixels are square, aren't they? Yes, the black dots are a pain…
I bet it wouldn't be hard to write a little IM program to massage them away
automatically – just average the pixels from the right, left, top and
bottom around them. What are the Magellan images like? As for filtering,
I'd love to hear your ideas.
You don't have to "average" the pixels. IM provides a process called
"Remove Feature" which is intended for exactly this type of artifact.
It allows radial or linear re-building of a region from the region's edge
data. Certainly it would be far less distracting to look at than the NASA
injected correlation spots.
–Ben
Is "remove feature" simply given a pixel X,Y location, and it does its
stuff? then it would only take an ARexx script to do this, right, if
NASA's spot are always in the same place.
Remove feature removes an area – any area that is marked in the mask.
The area can be any size from a pixel to an odd-shaped large region.
–Ben
John,
Yep the FITS pixels are square. Images look squashed on my 1080 monitor
with the FC24.
Thats a good idea about averaging the pixels from each side. They don't
look real clean when I zoom in on them. I tried using Light24 and cutting a
little brush out right next to each dot and then stamping it down to erase
the dot. Works OK but is mighty slow.
Oops, did I say Magellan? Sheesh, I should have said Viking orbiter. I
have four discs from Viking/Mars and not the Magellan stuff which I think
is not available right now. The Viking discs do not work for me. As soon as
I load a Viking disc I get a crash 8000000003. All four do the same thing.
Voyager discs are OK. I use CDRom-FS ver. .702. Time to call Allen Purtle.
Booting up with a Viking disc in there gives the same result.
I have made databases of all the image data (index) with Superbase so I
can search on various criteria. I hope to look for consecutive images taken
with only a filter change. I should be able to assemble three 8bit images
into one 24 bit one. Might take some digging. Have you tried something like
this. I also hope to look for images that can be made into an animation.
….Mark Sterling 74335,1677
Maybe you should send me one of those Mars discs, and I'll tell you if it
works on my PC and Mac. 🙂
As for calling Allan Purtle, he called me yesterday, cursing the AmigaWorld
article only from the standpoint that his phone has been ringing off the
hook – not with sales, necessarily, but strange questioners with stranger
questions.
Aww c'mon. Those discs are so cheap there is no excuse for not getting
them. You probably wouldn't want them if they didn't work.
I did call Allan Purtle today (I guess I'm one of those strangers) and
he plans to get those discs to see what the problem is. It may be a while
before anything happens. He does go on.
….Mark Sterling 74335,1677
Glad you got it working; have fun!
–Ben
Ben,
Works fine. Full images load in 8-10 secs and browse images load in a
flash. I like the option to save the Image Data header as a separate file.
Ordered IM/Fc 7xx yesterday when I finally got through. I think the
operater was dazed from all the "Morph" phone calls.
….Mark Sterling 74335,1677
Yeah, it's been an interesting week.
Something else that is in the 7.02 version of the software is _vastly_
improved rendering for Amiga mode images. 7.02 adds a number of dither
selections to choose between, and we've got some more capabilities cooking
in that regard also.
Something we heard quite a bit is that the users really wanted to create
Amiga-mode output images, and we'd never really put as much work into the
quality there as we should have.
We're quite happy with the output now… we've got a bunch of other
products that do high-color to Amiga rendering, and none of them do as well
(now).
–Ben
Ben,
I do plan to render many images in regular Amiga modes. Its good that
you haven't forgotten to do a good job with them too.
OK, how about a Toaster Framestore loader saver? Anything cookin' on
this? I hear all these Toastoids whining about getting images in and out of
their boxes.
….Mark Sterling 74335,1677
Ben:
I see I am not the only one who has been having fun with the new NASA
loader. The Amiga is now the easily the best platform for viewing and
manipulating NASA images. The IBM software that came with the CDROM's is a
joke. The ability to load FITS format images, coupled with the tremendous
power of ImagePro, ARexx and other tools makes the Amiga the platform for
working with NASA CDROM's
Just wanted to thank you and everyone at BBS for adding the format to your
products. Great job!
-=Martin=-
– via Whap!
Be a pal and upload some of those NASA images in either HAME or JPEG
format!
Nick
I plan to UL some good NASA images when I get them processed right. If they
are B&W 8bit (256 grey scale) then GIF might be most easily read. I can do
HAME or JPEG with no problem.
….Mark Sterling 74335,1677
Vic:
I was a developer but when it came time to renew I decided it wasn't worth
it. I thought I had the release version but obviously not. I'll try and get
it from my dealer this weekend. Nothing unofficial going on, I just wasn't
paying enough attention.
-=Martin=-
– via Whap!