View Full Version : May/June 2011 Lite Challenge Entry - Compound Interest

Master TMO
06-02-2011, 11:53 AM
Finally got the file upload issue resolved - IE9 doesn't like the script for some reason. Changing to FF allowed me to upload.

As I mentioned in the main thread, I had an interesting idea while looking at all the great maps for this challenge. I don't know if I am capable of pulling it off, but it can't hurt to try. I'm not out anything but some time if I fail, and I will probably learn something along the way. Already have, in fact - it took me about a day to figure out how to do the eyescreen properly so I can repeat it as necessary.

What would a planetary survey look like as seen through compound eyes?

### Latest WIP ###

06-02-2011, 03:16 PM
sounds interesting - but there is something that bugs me *lol*

06-02-2011, 04:02 PM
Ya know, I ask myself that same question at least once a day. :) Kidding. Great idea, man.

Master TMO
06-02-2011, 04:24 PM
Well, if I can get it working right the way I'm picturing in my head, it will be genius. If I fail and don't change it, it will be brain-bendy and eye-splody. Odds are I'll have to change it slightly, but I'm aiming for Genius first. :D

Master TMO
06-02-2011, 06:25 PM
The concept I am aiming for is to have 4 different maps all being shown in overlapping facets, each with a different primary color. For us 2-eyed (or maybe 4-eyed) folks, focusing on a single color might allow us to view the map by ignoring the facets with different colors in between. I honestly don't know if this will work or not.

I haven't picked all 4 layers yet. I know I'm going to use a false-color infrared map for the reds, and probably an elevation map for the white. That leaves green and blue, and I've got temperature, rainfall, climate, and visible-light (really just a pretty version of the climate) layers immediately available to me. I could also theoretically change the primary color for each, maybe using CMYK scales vs RGB.

I'm attaching the template I've created for laying out the facets, but I'm not going to inline attach it due to exceedingly strong levels of eye-splodyness. Although if my preview of this post is anything to go by, it still shows. If so, my apologies in advance for any injuries.

EDIT: Yup, it shows a thumbnail. Just count yourself fortunate that you're not working with a 4096x4096 version of it like I am. ;)

Master TMO
06-05-2011, 01:51 AM
Foof. Real life really cuts into the time I can spend mapping. ;)

Since there's no real way of automatically converting a planetary image into the compound vision short of manually copying and pasting everything one hex at a time, it's not something I want to do more than once, so I have to make sure I have all my planetary views completed first. For the moment I think what I've got is good enough, although I'll think about it some more. Past midnight like this I might not be using my best judgement.

I still need to build the background starfield, which I also have to have finished before I convert.

I have 5 planet views - visible light, elevation (white), temperature (red), rainfall/humidity (blue), and vegetation (green). The concept is that the Giant Voracious Space Dragonfly (From Space!) is looking for planets with plenty of plant biomass to consume. It has eye cells that sees 3 different wavelengths (white, red, blue), which it combines to make the green, which points out where the most plantmass to consume is. And yes, the green layer is calculated from the other 3, although I probably could have used the Image Climate function from FT to do the same thing.

Oh, and it is intentional that none of the views but Visible has the shadow, clouds or sunlight. Those views don't use the visible light spectrum and therefore can't see any of that.

There are going to be a few eye cells that see the visible light spectrum though, which is why I went ahead and made one.


06-05-2011, 02:02 AM
still cool - but you sure got your work cut out for you :)

Master TMO
06-05-2011, 02:07 AM
I'll be honest, I'm not even really sure if a Giant Voracious Space Dragonfly (From Space!) looking at a planet searching for food really qualifies as a 'planetary survey', but we'll just say I'm doing this one for my Muse and some experience, and folks are welcome to not vote for it if they think it doesn't meet the criteria of the challenge. ;) It's definitely got me thinking outside the box, and I've learned quite a few new tricks along the way. And I haven't thrown my laptop across the room yet...

(Although it did just occur to me that I don't have to have the starfield done to work on the 4 non-visible views - starlight won't be visible on them any more than sunlight will)

06-05-2011, 02:18 AM
if the Giant Voracious Space Dragonfly (From Space!) is surveying - then its a survey :)

Master TMO
06-05-2011, 09:59 AM
Man, starfields just are NOT my forte. I've never tried one before, concentrating solely on planetary mapping to this point.

Master TMO
06-05-2011, 11:26 AM
New Visible-wavelength image with cheesy starfield. I did remember to block out some of the stars on the lit side of the planet, although I'm not sure if I did enough. And the shadowed portion of the planet may not be dark enough.


06-05-2011, 12:59 PM
Looking good! There are some great starfield tutorials (http://www.webdesign.org/photoshop/articles/make-a-realistic-star-field.3811.html)on the web, I think a couple may also be posted in the tutorial section but not sure.

Master TMO
06-05-2011, 01:53 PM
That's the tutorial I followed, although I modified a few steps. I replaced the manual random erase step with blocking sections out with layered Cloud and Diff Clouds. And the manual cloning of stars was replaced by creating 4 more layers of small and large starfields. While it seems to do a decent job at small clusters of stars, it doesn't do a large Milky Way-style band of brighter stars, because in a 4096x4096 image, the clouds are proportionally small compared to the canvas. It's the same issue for the cloud cover on the planet itself. Neither stars nor clouds is something I've attemped before, so I haven't figured out what works for me yet.

For the planetary clouds I followed a tutorial on this site (for doing whole planets), but again, the canvas size in the tutorial is much smaller, so the resulting clouds are smaller. I'm not sure if just transforming the clouds and making them larger would work or if they'd pixellate obviously.

06-05-2011, 06:43 PM
All right, here's a quick and dirty rundown of my starfield technique (no pictures). I'm assuming this is in GIMP.

1. Start with an all black image.
2. Rename your background layer "small stars"
3. Go to Filters>Noise>HSV Noise. Make a noisefield with the following values: Holdness 4, Hue 0, Saturation 0, Value 200.
4. Go to Colors>Brightness-Contrast. Increase the contrast by 50. Leave the brightness alone.
5. Make a new all black layer named "bright stars".
6. Make another HSV Noise field with the same values as the first.
7. Go to Colors>Brightness-Contrast. Increase the contrast by 100 this time.
8. Scale up the bright stars layer to double its original size.
9. Go back to the small stars layer and select everything by using the magic wand with a threshold of 255.
10. Go back to the big stars layer.
11. Go to Layers>Crop to selection. Now the excess big stars layer is deleted.
12. Now we're going to give the brighter stars a little color. Make two new layers. Name one "reds" and one "blues".
13. Paint both all black and make HSV noisefields in them as well. Same ol' values.
14. Gaussian Blur both layers with a radius of 5 pixels.
15. Brightness-Contrast both layers. Set Brightness and Contrast both to 100.
16. Add layer masks to both layers. Initialize to "Grayscale copy of layer"
18. Get out the bucket fill with threshold set to 255. Paint the "Reds" layer all red (RGB 255, 0, 0 or HSV 0, 100, 100). Paint the "Blues" layer all light cyan (RGB 0, 255, 255 or HSV 180, 100, 100).
19. You should now see your bright stars on a mottled red and cyan background. This is not what we want. So, set the modes of both the color layers to "Color".
20. Now some of your bright stars are blue and some are red. This is looking good, so merge both color layers down into the "bright stars" layer.
21. Set the "bright stars" layer to "screen" mode so that the smaller stars are visible.
22. Hit "flatten image".
23. There you go, a decent-looking starfield!
24. With practice, you may be able to modify this technique for even greater realism, but always remember the following:

HSV noise is your friend.
The Brightness-Contrast tool will make your noise look more like stars and less like TV static.
To color your stars, make separate red and blue layers. This will color some stars, but where the layers interfere with each other you'll still get plenty of white stars.

Also note: the starfield that other tutorial teaches you to make may look cool, you would never see something like that outside a dense star cluster. A realistic starfield should look like the night sky, but with no light pollution or twinkling. However, you could consider adding a "milky way" band, although I'm not sure the best technique.

Master TMO
06-05-2011, 11:16 PM
Thanks for the post! I'll give it a try later tonight. I use PS though, so I'll translate as best I can. Have a rep pt as thanks.

I have started copying cells over.

06-06-2011, 10:40 AM
This is such a great idea and so much work. Good for you for going for it. :)

06-06-2011, 12:53 PM
Yeah, thanks for taking the time to share your technique!


Master TMO
06-06-2011, 02:53 PM
Figured out a way of copying cells over a bit faster, but still not quite just copy/paste/move on. I've built 4 layers of overlapping Hexagonal selections. I copy the entire layer under one of those selections, then paste it into the target layer. I move the whole layer so that the center of the copied hexes is on the center of the right eye hexes for that color. Then I use Marquee selections to slide out all the surrounding hexes until they're on their correct view hex cells. Then once everything is in place, create a copy for the left eye as well, then Spherize each of them into an eyeview. Yes, the view across the eyes is duplicated. Each eye is centered on the same point.

I've finished the Elevation (white) layer. Just Temp, Rain, Vegetation like this to go, then I will go through and add some Visible Light cells that replace the others.


Hmm. Now that I see what the view range is, I may have to recenter the view so you see something other than sea. :)

06-06-2011, 03:05 PM
This is a winner for sure. A great idea and great tips on how it was(is being) done. :)

Master TMO
06-06-2011, 06:23 PM
Thanks for the compliments folks, they mean a lot.

I'm taking a step backwards, to recreate the planet views with a new center for better scenery. I'm also experimenting with creating better cloud cover for the visible light view. The clouds I've generated actually look pretty good to my uneducated eye.

The steps I took, in PS:
- Create a new cloud layer, and Render/Clouds to fill it. Set the layer to Screen.
- Transform the layer, and set the Width to 250%.
- Add a layer mask, fill it with Render/Clouds, and transform the mask the same way.

- Create two more Cloud layers, press Ctrl-G to have them affect only the main Cloud layer.
- Fill both layers with Render/Clouds, and fill their layer masks with clouds also.
- Transform one of the Cloud layers and set the width to 150%. Do the same to its layer mask.

- Apply the Layer Masks to all 3 cloud layers.

You now have 3 cloud layers, 1 main one and 2 sub-layers. Change the order of the sub-layers and play with the layer settings to see what looks best. Maybe the 150% layer should Multiply, while the 100% layer Screens, or vice-versa. Color Burn and Hard Light also have strong effects.

I haven't gotten far enough yet to post any screens of it, but I wanted to write down the process before I forgot it. ;)

Master TMO
06-06-2011, 08:46 PM
Here's the sample output from the clouds. I used Linear Light and Color Burn on my sublayers. Once I got them arranged, a few twirls where appropriate to give the clouds a bit of motion.


Master TMO
06-07-2011, 11:34 AM
Built new PS files to do the compoundization, as I figured out that part of the problem with the view size on the original was that I hadn't accounted for the way the hexes spread out due to the interleaving of 4 different maps. So only the central chunk of the view was available, with everything else getting clipped by the edges of the canvas. The new Hex file has the dimensions of 8192x8192, and I have successfully imported in the Elevation view and plotted it. However, the file is now too big for the Save for Web. *rofl* It throws an error when I try to save a JPEG, even a smaller one. So no more thumbnails until I have everything imported and arranged, and I can flatten and shrink the file back down.

Master TMO
06-07-2011, 11:08 PM
Now I'm zoomed out too far, but that's easier to fix than too close. ;) And now I can see that I need to tweak the colors, as the white elevation cells are much more visually dominant than the green, which is the supposed target.

### Latest WIP ###

Master TMO
06-09-2011, 04:17 PM
Walked away from this for a day or two just to clear my head a bit before tackling the cleanup.

Just a quick upload of the color corrections, without the compounding. I've got less than a week to finish and polish this. I've got what I want the result to look like in my head, so it should be doable. I think the green is prominent enough in this version now, although not as much in the thumbnail, I notice.


06-09-2011, 11:50 PM
Amazing...the concept of this I'm speechless.. :)

Master TMO
06-11-2011, 12:23 AM
The Exoanisoptera, or, as popularized in recent pulp horrorholo "Invasion of the Giant Voracious Space Dragonflies from Space!!!", the GVD, or Geeveed, is a unique lifeform like nothing we have ever encountered before. In this article, I will attempt to consolidate what is known about the creature, what is supposition, and what is fantasy.

Description: The only adult specimen that has ever been observed was approximately 10 meters long including the tail, and 2 meters wide at the widest point in the body. The body was approximately 1.65 meters in height, and stood 2 meters above the ground when it landed. It was completely covered with what appeared to be a thick exoskeleton. 7 pairs of wings extended along the body and down the tail, with the longest of them reaching 3 meters beyond the body. It had two large compound eyes, each measuring .85 meters across.

History: The first known encounter with a 'Geeveed' was when one appeared in low orbit around the planet Chongatta. 'Appeared' is the best term, as there was no sign of it approaching the planet from a distance. It suddenly appeared on the scanning boards of 5 different satellites. Unfortunately, none of the satellites were equipped with visual surveillance cameras angled in the proper direction to capture its arrival, leaving us with collision avoidance sensor logs as the only records. Chronological analysis of the logs place it's arrival time to a window of no more than 1.5 minutes. The Geeveed remained motionless in orbit for 4.95 minutes, then began to descend over the continent of Rutland, using its wings to control its speed and direction of descent. It moved at an angle across the planet's surface, directly from where it had appeared to it's planetfall location in the Beanaugh river basin.

We do not have any direct holo or video of its landfall, only weather satellite imagery. Months of skilled enhancement of the images has given us a general idea of its immediate actions. It landed, or hovered close to the surface, and remained motionless, or mostly so. After 37.45 minutes it moved again and flew 20.73 km in a northeasterly direction, repeating its initial actions by landing and remaining motionless. After 27.6 minutes it took off again and flew 35.3 km to the north, then landed again.

It is at this time that an aircraft from the city of Risengi (located on the southern coast of Rutland, at the Beanaugh river mouth) arrived, having located it using surface radar, and providing us with our first visuals of the creature, as well as direct observation of its activities. According to the crew, the Geeveed (called 'the bug' in the audio logs) had apparently killed and was eating a Kinys antelope (an indigenous herbivore). The aircraft was able to stay in position for over 2 hours before needing to return for refueling. During that time, the Geeveed killed and ate 5 more antelope. It killed them by flying above them, then landing on their back and grasping it with its legs and biting at the back of the neck until it brought the animal down. It then scooped out the abdominal portions whole with its jaws before moving on to another prey animal and repeating the procedure. With night approaching, another aircraft was not dispatched to replace the first one when it left the scene.

The next morning the aircraft was dispatched, arriving just before local dawn, but was unable to locate it using radar or optical searches, and had to return to base after a few hours of fruitless searching. They did observe several more dead animals, including a large Polman croc. (Later analysis of the satellite imagery indicated that the Geeveed had moved north and west from its last spotted location, and had not resumed movement until later in the morning, after the search craft had passed it by.) Another search was made that afternoon, and the Geeveed was found and followed again, with the same behaviour being observed. This general pattern was repeated for 4 days, with the Geeveed occasionally moving large distances as soon as hunting began to be difficult in the current location.

On the fourth day, the first reports of small animal corpses with abdominal wounds began to come in from scattered remote areas. This prompted a lot of confusion, as the bodies were in areas not visited by the creature. Due to a lack of resources, there was very little response to these reports at the time, considering them either local animal activity, crank calls, or hysterical reports based on news reports of the creature. Reported sightings of bugs escalated and spread until it was obvious that there was something to them and search teams were dispatched. They quickly confirmed the existence of small bugs. The original Geeveed was nicknamed "Echidna", after the mother of all monsters in Greek legend. It was not until our analysis of the satellite imagery that the theory that she laid young upon first arriving was confirmed. Although the young could not, of course, be seen on the imagery, the locations Echidna stayed motionless in coincide with the centers of reported geeveed activity later.

4 days later Echidna vanished, disappearing from view while being filmed. She rose up to 50 meters above the ground, accelerated the speed of her wings, and then disappeared.

2 days later the first human casualty was reported, Haley Brigitte Wood, an infant sleeping in her crib. The next day the first confirmed geeveed kill was reported, by a local farmer with a hunting rifle. For the next week, there were sporadic reports of geeveed and human deaths. Examinations of these early bodies proved that they were vaguely similar to terrestrial insects, with a strong exoskeleton, segmented body, and breathing spirules. The connective tissue holding the internal organs in place was thicker and stronger than insects, and body motion appeared to provide more circulation and respiration than is possible for terrestrial insects, which explains how they were able to support a much larger body than is normally seen. Any damage capable of penetrating the exoskeleton tended to be strong enough to practically liquify the interal organs, so the results of these early examinations were not very conclusive or informative.

For the next 4 weeks, the conflict between the species escalated. Due to their speed and maneuverablity, shooting them was very difficult. As the geeveeds matured, their exoskeletons got thicker and more resistant to damage, even as they got larger and slower. After 4 weeks, their exoskeletons seemed to become completely invulnerable to hand arms fire, and human casualties skyrocketed.

Currently, the entire planet of Chongatta has been placed in a state of Planetary Emergency. Military forces have been dispatched for protection and peacekeeping, and most of the populace has been evacuated to less habitable portions of the planet or onto space- and seacraft. The population is having to be supplied by food shipments from neighboring planets. The death toll has risen to in excess of 250,000 out of the original 4 million on the planet, or 1/16th of the populace. There are an estimated 10,000 remaining Geeveed left on Chongatta, and they have done an estimated 7.9 trillion pounds worth of damage.

* Biology: The Geeveed exoskeleton is an unknown organic compound, exhibiting a high carbon crystal content. It seems to only get thicker and stronger with age and maturity, with no currently known upper limit, although we are assuming at this point that Echidna was a fully mature specimen. We do not know how strong or thick Echidna's exoskeleton was, as she vanished before the situation escalated. The breathing spirules appear to have an exoskeleton plug. We assume that this is what allows the Geeveed to venture into space, although we have not seen any of the immature Geeveed do so. It appears to operate as a pressurized lock - if the atmospheric pressure drops below a certain amount, pressure from the internal organs forces the plug closed, trapping the air and gasses inside and preventing a loss of pressurization. The Geeveed eyes appear to have specialized eye facets capable of detecting ambient humidity, temperature, even elevation. They appear to combine into a single visual feed that flags whether or not an area is a good place to find food.
* Intelligence: The Geeveed have not displayed any recognizable signs of higher intelligence yet, seemingly reacting in a purely animalistic and intuitive manner.
* Interplanetary Travel: The biggest mystery is how Echidna managed to appear and vanish from Chongatta. Many people from many disciplines are either in complete denial or undergoing nervous breakdowns trying to explain this. None of the immature Geeveed have been witnessed disappearing, but given how many of them there are, the amount of ground they are spread out over, and the fact that there are few active watchers, it is likely that even if some had gained the ability we would not have been able to see it. Nothing in any of the bodies we have examined have had any parts or organs that might even begin to explain this ability.

* The Geeveed swarm appears to be tearing apart the Chongatta ecosystem. No large animal is ignored by them, and because there is almost always more than one Geeveed around, even if one fails in an attack, another one seems to swoop in and take advantage to finish the prey off.
* Sidearm munitions seem to no longer be effective against the Geeveed once they have passed a certain maturity level. Strong enough explosives should still kill them, but successfully targeting them is difficult given their maneuverability and speed. An area blast is still likely to succeed than aimed fire. As a last resort, nuclear devices should be guaranteed to kill all Geeveed in the blast radius, at the usual cost to the environment and bystanders.
* Any planet without satellite monitoring devices should have some installed immediately. Any Geeveed detected should be bombed immediately, before it has a chance to deposit any offspring. Young Geeveed should be killed before they grow large enough to become invulnerable.
* If a Geeveed could be captured, we could study it while alive to determine any weaknesses or vulnerabilities. Immediately obvious applications we might learn is the secret of their exoskeleton, with possible uses in armor plating and personal protection. And their "teleportation" ability would be invaluable to us all, perhaps even allowing instantaneous travel beteween star systems.

Artist's Rendition of Echidna's first view of Chongatta:
### Latest WIP ###

Master TMO
06-11-2011, 12:40 AM
GM Notes:
* Teleportation: The Geeveed have somehow gained or evolved access to an 'Inverse Dimension', where object's dimensions are essentially 2-dimensional, and proportional to it's mass in the normal universe. So stars and gas giants are truly massive bodies, while interstellar space is practically nonexistant. Meaning that it is fairly easy and quick to navigate across interstellar distances, and choosing a planet of the desired mass is merely a matter of measuring the curve of the edge you can see, since all mass is equally dense in this dimension. Geeveed do not develop this ability until they grow to full maturity.

* Also note that travelling by Inverse Dimension means that the easiest travel is to the closest neighboring systems. Meaning that unless a solution is found for killing off the 10,000 remaining Geeveed before they mature, the Humanity Combine is most likely going to have some very rough years ahead of it as neighboring systems get invaded en masse. Fortunately for them, Echidna was one of the farthest-travelling Geeveed of her swarm, but it's entirely likely that even if this swarm is taken care of, eventually others will find a Combine planet and begin the whole thing all over again.

* Planetary Survey notes: There are 5 layers of planet views overlayed in the image. The first one si the regular visible light spectrum. The next three are rainfall (blue), temperature (red), and elevation (white). Those 3 layers combine together to make the 'vegetation' layer (green). The bright green hexes are those that should support the most life, and are therefore the best places to hunt and gain a lot of food quickly.

Steel General
06-11-2011, 08:01 AM
This is really quite cool, and a bit disorienting :)

Master TMO
06-11-2011, 01:16 PM
This is really quite cool, and a bit disorienting :)

Alien viewpoints will do that to you. :D

06-11-2011, 02:21 PM
Excuse me waiter? There seems to be a fly in my map... :) Just awesome. I think you've got a winner here.