Normal mapped Moon

[starbase1]

Well, I was pleased to find a normal mapped moon in the sample scenes. Less pleased to find it bright orange!

Anyway, I turned off the glow, and replaced the orange fractal noise texture with an image map. I then turned down the strength of the image map to about 40%. It should really be even less for realism, but then it would not show off the normal mapping so well.

Here are my results, 2 of the front of the Moon, 2 from the back.

[Bliz]

That's really nice! Looks pretty realistic to me.

[starbase1]

If anyone would like the relevant files or sufaces, let me know.

But the difference between a bump map and a normal map really shows with those craters! I wonder how theybuilt the normal maps for the planet in the first place???

Nick

[UnCommonGrafx]

That's pretty.

You've seen the threads about something called "Melody" from nvidia that will change bumpmaps to normal maps; displacement maps, as well.

Yes, I would like a zip of your version. I bet it's impressive in layout if the mesh subd number is high. Is there any aps in action, here?
Looks really nice.

[starbase1]

No, not seen this 'Melody' thing, time to go to google I think...

Give me a day or so, and I'll bundle up all the maps used, and release the object. Nag me if I take too long!

:D

[harlan]

Nag nag nag!!!!! ;)

[oDDity]

I've worked a lot with generating normal maps for a doom mod, and I can tell you that the 2d normal map generators are rubbish compared to the 3d ones. THey're only a small notch up from a photoshop emboss.
THe moon looks great though, I'd guess they modeled it in zbrush using projections and also generated the normal maps there.

[starbase1]

That would be one **** of a task - I've had a look around the web, and the supplied normal map appears to be based on stuff from the Celestia Motherlode web site - they have some VERY large versions!

There is mention of some radar data from the Clementine probe...

There is a LOT of real information in there, things like valleys linking craters that only show up at shallow angles of illumination. If I turned down the strength of the normal map a tad, I'd call those images photo real.

Data from Mars is MUCH more complete, and detailed. (I mean the scientific data). I'm really looking forward to playing with THAT!

Nick

[starbase1]

OK, it's not exactly elegant, but here are all the bits you need.

I have taken the modified moon object, the image map i used, and a sample scene. You will find that the normal map came with the Lightwave 9 content.

For some reason the conference does not like my 2 mb ZIP, so here's a direct link to that file on my web page...

http://www.starbase1.co.uk/Moon.zip

I'll be VERY interested to see any tweaks...
:goodluck:
(Not that anyone ever bothers to post them...)
:cry:

Nick

[DonJMyers]

Maybe newtek can add your upgraded version to their library of samples.

The normal map of the moon may have been made from 3D landscape data (USGS DEM file) taken by a lunar orbiting probe.

[starbase1]

OK, all done now. I have made cleaned up copies of my nodal / normal mapped Moon and Mars objects available on my downloads page. A sample tidy scene is now included.

Fairly obviously you will need Lightwave 9 to use them!

I'd love to see what you come up with, (but no one ever bothers, so what the heck...)

http://www.starbase1.co.uk/downloads.htm

Cheers,
Nick

[zapper1998]

ok how do you create the Normal maps ??????????


Michael

[starbase1]

I'm doing some reseach on that... There are programs to create normal maps from bump maps, and for planets, my suspicion is that this will work extremely well.

Most stuff is about taking high poly count models and turning them into low poly count models, with normal maps, but that does not apply here.

n terms of downloadable normal maps for planets, the Celestia Motherlode seems the best place - but many of there large ones are a strange format, or broken up into tiles, (or both). If anyone can work out how to reassemblye these into one bug image again, I'd really like to hear about it...

Nick

[Bliz]

I'd love to see what you come up with, (but no one ever bothers, so what the heck...)

http://www.starbase1.co.uk/downloads.htm

Cheers,
Nick

I will 'try' and have a play with it inbetween working overtime and pacifying the wife because I'm working overtime. :)

[Bliz]

Ok, downloaded these last night and did some quick test renders. The mars object is superb (wish we had this when we were doing Captain Scarlet :cry: ) Big thanks for supplying that Nick.

The moon object is also very useful but I think the diffuse texture needs a bit of work to eliminate the strong lighting and shadows at the poles. I'll attempt this tonight or the next couple of days.

[MarkG]

I was watching 'Space 1999' on DVD last week and thinking how fake the moon model looked: they could really have done with Lightwave and these files back then :).

[Matt]

Cool thanks Starbase! :)

[avkills]

Those look really good starbase1, but there is something nagging me when I look at them, the diffuse falloff seems a bit unrealistic -- it is much too sudden and you would have a different falloff depending on the height/bump.

Just an observation; btw I will be working more tonight on my updated hi-res earth object...right now I have a surfacing problem with the atmosphere. The APS on the clouds is only so so also.

Hmmm, looking at them again it seems the falloff is varied, but it seems much to sharp for my tastes. hmmm.

-mark

[starbase1]

AVKills, and others - PLEASE have a go and see if you can improve it, all I ask is that you share the results if you think you have something interesting!

I think the terminator should be sharp - the lighting is not at all like the earth in this regard. Here's a photo to show what I mean:
http://www.nwgis.com/greg/images/c-cresnt.jpg

This is controlled by the diffuse sharpness setting, on the advanced tab.

I'd be particularly interested in the views of those who have seen the moon through a decent sized telescope.

The other aspect that is not built into the model is that the moon has a strong tendency to throw light back strongly at the source - this is why the brightness kicks up a LOT when it is very close to full moon. It's caused by a large number of square crystals on the surface material, which act as corner reflectors.

I guess you could do this with a gradient on the diffuse channel.

Nick

[avkills]

Nick, does the normal require the use of APS? I am downloading now on my laptop just to take a quick peak; I will do the same at home tonight on my G5 and see what I come up with, although I was going to do some work on my earth atmosphere RIM.

** edit ** Another quick question, can you turn the normal map into grayscale and have it still work?

-mark

PS, I dig the Nebula and the sun solar flare... very cool looking.

[starbase1]

Those look really good starbase1, but there is something nagging me when I look at them, the diffuse falloff seems a bit unrealistic -- it is much too sudden and you would have a different falloff depending on the height/bump.

Just an observation; btw I will be working more tonight on my updated hi-res earth object...right now I have a surfacing problem with the atmosphere. The APS on the clouds is only so so also.

Hmmm, looking at them again it seems the falloff is varied, but it seems much to sharp for my tastes. hmmm.

-mark

I'll be VERY interested to hear how you get on with your high res Earth, as I am still working on mine... Are you doing anything special to get hi res maps in? best I have managed is 9000x9000 using 8 but PNG, Which is a shame as I have just got my hands on 42000x21000 normal map for the whole Earth!

The clouds are still the bit I am least happy with...

Nick

[starbase1]

Nick, does the normal require the use of APS? I am downloading now on my laptop just to take a quick peak; I will do the same at home tonight on my G5 and see what I come up with, although I was going to do some work on my earth atmosphere RIM.

-mark

PS, I dig the Nebula and the sun solar flare... very cool looking.

APS? Don't understand...

Normals require LW9 or I beleive there is a free plugin out there...

Nick

[pixelinfected]

no, aps mean that you build real poly displacement, Normal mapping is a fake, is like a superbump, that not only change light and shade, but bend light to simulate a rounded shape also where polys is not enought to round a shape.

[avkills]

Where did you get that normal map?

Although I have a bump map from Nasa that is 21600x10800. My earth color map is actually only 8192x4096 from Blue Marble. The bigger color maps do not include ice at the poles and the water looks sorta darker than it should.

APS (adaptive pixel subdivision) lets you add real geometry to a subdivided object. So I can do like a 8 way tessellated sphere for the earth and turn it into over 1 million polygons that can be deformed with a displacement bump map. I'll post a exaggerated grayscale so you can see it. It is cool.

-mark

[avkills]

Here is a APS subdivided earth with just over 1 million polygons. It looks as though there are a few tweaks I need to do to the displacement map, but that should be easy with photoshop... just darken the offending areas.

** note that this scale is not actual and exaggerated to get the effect to be more noticeable.

-mark

[starbase1]

I'm getting help from one of the Celestia guys with the normal maps, (though I have just discovered he has sent me a bump, not a normal!) There are huge normals for Celestia, but they are broken up, and often distributed as DDS files. The DDS photoshop filter from NVidia does not cope with large images.

I should have picked up on APS, all is clear now! Nope, normal maps don't require APS.

I'm not that bothered about poles, I am mainly looking for low earth orbit views, (and manned polar orbits don't happen!).

The thing with subdivision is that the level of detail is still relatively small - I would have thought not good enough for low earth orbit views at good size images. For example, and asuming a cubic earth for simplicity (!), a million polygons is about 1000x1000 if mapped to an image, for the whole globe - thats 500 points pole to equator.

Mount Everest could fall between two point at that scale...

The image maps I am using are 9000 pixels pole to equator, (over in another thread I started). Now if I can use normal maps at that scale, (instead of bump), I think the results would be seriously impressive. And simply nt possible with geometry, (unless I am missing something about APS?)

Nick

[Anttij77]

AVKills, and others - PLEASE have a go and see if you can improve it, all I ask is that you share the results if you think you have something interesting!

I think the terminator should be sharp - the lighting is not at all like the earth in this regard. Here's a photo to show what I mean:
http://www.nwgis.com/greg/images/c-cresnt.jpg

This is controlled by the diffuse sharpness setting, on the advanced tab.

I'd be particularly interested in the views of those who have seen the moon through a decent sized telescope.

The other aspect that is not built into the model is that the moon has a strong tendency to throw light back strongly at the source - this is why the brightness kicks up a LOT when it is very close to full moon. It's caused by a large number of square crystals on the surface material, which act as corner reflectors.

I guess you could do this with a gradient on the diffuse channel.

Nick

Or you could try the Minnaert shading model, which was originally developed to render the moon "correctly".

[avkills]

I'm getting help from one of the Celestia guys with the normal maps, (though I have just discovered he has sent me a bump, not a normal!) There are huge normals for Celestia, but they are broken up, and often distributed as DDS files. The DDS photoshop filter from NVidia does not cope with large images.

I should have picked up on APS, all is clear now! Nope, normal maps don't require APS.

I'm not that bothered about poles, I am mainly looking for low earth orbit views, (and manned polar orbits don't happen!).

The thing with subdivision is that the level of detail is still relatively small - I would have thought not good enough for low earth orbit views at good size images. For example, and asuming a cubic earth for simplicity (!), a million polygons is about 1000x1000 if mapped to an image, for the whole globe - thats 500 points pole to equator.

Mount Everest could fall between two point at that scale...

The image maps I am using are 9000 pixels pole to equator, (over in another thread I started). Now if I can use normal maps at that scale, (instead of bump), I think the results would be seriously impressive. And simply nt possible with geometry, (unless I am missing something about APS?)

Nick

Well you are right about the scale. At the orbit views you are talking about, I wonder if one would even see any displacement in real life? Anyway, my main scene so far is displacing about 2 million polys for the earth, and another 2 million for the clouds. The clouds look ok, just enough displacement to get some poof in them. I am actually using 2 cloud layers, one that is displaced and one that isn't.

I will try and post some renders tonight. I am having some issues with the atmosphere RIM as usual. ;)

-mark

[starbase1]

Or you could try the Minnaert shading model, which was originally developed to render the moon "correctly".

I'd like to know more about that - any references?

[starbase1]

OK, I'm getting the hang of making normals from bumps... Attached are 4 images of Earth, 2 with coloured surface, and 2 with plain gray. I chose one image over the Himalayas, and one over the Americas.

The map I am using for normals is a relatively small 5400x2700 pixels, so it can be pushed a good bit further...

Nick

[nthused]

OK, I'm getting the hang of making normals from bumps... Attached are 4 images of Earth, 2 with coloured surface, and 2 with plain gray. I chose one image over the Himalayas, and one over the Americas.

The map I am using for normals is a relatively small 5400x2700 pixels, so it can be pushed a good bit further...

Nick

These look great! Can't wait to get some time to learn how to use these features.

[avkills]

OK, I'm getting the hang of making normals from bumps... Attached are 4 images of Earth, 2 with coloured surface, and 2 with plain gray. I chose one image over the Himalayas, and one over the Americas.

The map I am using for normals is a relatively small 5400x2700 pixels, so it can be pushed a good bit further...

Nick

Hmmm, maybe the normal maps make more sense for speed and ease of use. They do look quite good. Although I notice some strange diffuse stuff going on towards the poles. The normal map definitely has more detail to the bumps. I wonder if a normal bump map will work for a normal map?

-mark

[avkills]

Good call on that Nick! I just messed around with the bump map I have from NASA and it works real well. Here is a shot above North America pretty darn close. You can see the color map of the earth is bit fuzzy, but the bump map is sharp as can be!

I think I will switch to the normal mapped method... probably do the same for the clouds.

I will have a render of everything here in a couple of hours.

-mark

[starbase1]

Hmmm, maybe the normal maps make more sense for speed and ease of use. They do look quite good. Although I notice some strange diffuse stuff going on towards the poles. The normal map definitely has more detail to the bumps. I wonder if a normal bump map will work for a normal map?

-mark

Nope, not a hope. A bump map is gray scale, a normal map is a colour image, where two channels are mixed to descibe the suface vector at that point. And only a very limited range of cololurs is valid.

[avkills]

Nope, not a hope. A bump map is gray scale, a normal map is a colour image, where two channels are mixed to descibe the suface vector at that point. And only a very limited range of cololurs is valid.

Interesting...the grayscale image I have seems to do an somewhat ok job, maybe not? Anyway, so you have been getting the normal map from the Celestial site then?

-mark

[starbase1]

Nope, anything of decent size for Celestia is broken up into chunks that are a nightmare to reassemble. I've had a lot of help and pointers from Jestr who provides a lot of material for that site. I'm working on getting it into a more lightwave friendly format. (You try feeding Lightwave an 80 megabyte Jpeg pf Martian altiude data! - it's about as stable as a one legged drunk opn a unicycle!)

Also when you start getting to very high resolutions, you are likely to run into issues about keeping the normal and image maps in perfect registration.Or approximate registration come to that...

I managed to make those normal maps from bumps, but I think I can make them a LOT better - a straight "load in the bump and hit the button" will not deliver best quality.

But I have a real mix of tools, methods, and samples in a big confusing pile at the moment. I plan on putting together a tutorial on normal mapping planets as soon as I understand what I am doing!

(I also need to dig out my registration info or I won't be able to make normal mapped images at all in a couple of days!)

Nick

[avkills]

I just downloaded a plug-in for photoshop for windows that is suppose to let me convert a bump map to a normal map... stay tuned.... I may have a 21600x10800 normal here real soon. :D

I am on a Mac, luckily I am at work with a ton a PCs around so I should get this to work I hope.

LW has been rock stable on my G5 once I disabled the Hub with very large images.

-mark

[avkills]

Well things are looking down. I going to try and do this by tiling, the PC just can't handle that image; I think it is more a GPU issue since I am pretty sure that PS plugin is using it and RAM.

-mark

[starbase1]

I know the NVidia one has a limit of about 2 million pixels. Stay tuned, I may have a Mac and PC solution shortly...

[avkills]

Well I have it done and it is enormous. Heh heh, anyone willing to give it a shot, I can try and save PNG and see how big it is. Had to do it by tiling by hand into 8 tiles, then re-stitching it together.

-mark

[Tesselator]

I'd like to know more about that - any references?


It's the subject of my next video. ;) The video is about 60% done and sitting
on hold while I create some more LW9 Content. It shouldn't be too long now.

BTW, Glad you liked the original L9 moon scene enough to invest a little time
to make it better! Thanks!



Edit - Oh, and that shouldn't have been rendering orange! Was that in the render
or just the OpenGL? The OpenGL is indeed Orange. ;)

[starbase1]

Nope, the actual render came out orange on my machine! And strongly glowing! (That's why I am claiming some credit for the moon images - the stuff that came out of the sample scene looked more like Titan on an off day!)

[starbase1]

OK, here's how to make the normal maps from the bumps, from images too big for the NVidia free software to handle. (And thats something like over 2 megapixel images).

Stage 1, download and install the open source graphics software, Gimp. It's completely free, and you can get it from many places, including here:

http://www.gimp.org/downloads/

Gimpl is available for Windows, Linux, Unix, and Mac OS-X. I wasn't expecting a lot when I tried it, but I was pleasantly surprised.

You can then get a plugin for the Gimp that does norml map conversion here:
http://nifelheim.dyndns.org/~cocidius/normalmap/

The bad news is that there is no Mac version here - but there is the source code. I suspect a Mac version may be out there somehere, but I can't find one - maybe you Mac users know better places to look, (or a friendly coder with a compiler!)

As far as I can tell, the filter works up to a size of about 20000x20000 pixels, (400 megapixels), though I have not tried it yet. That would handle the highest resolution Blue Marble Next Gen in 2 tiles.

The filter also has the capability to normalise normal maps - which is important if you do ANY processing that changes the colours of ANY pixels.

If anyone does find a Mac solution, let me know and I'll add it to the tutorial I am working on.

Cheers,
Nick

[avkills]

Nick,

To be honest, I found a combination of using APS with real displacement and just using a regular bump map (NASA grayscale one) going to the bump parameter in the Node editor good for my tastes.

I was probably doing something wrong, but the normal maps do not seem to cast shadows correctly. You get shadows even if you have a light pointing straight at the object directly behind the camera exactly perpendicular to the sphere, if that makes sense.

Let me get a screen shot of what I have so far...sans clouds. I sort of started over doing everything nodal.

-mark

[avkills]

The first image is something I found from NASA as a reference, second image is a render with the same general area of interest. Objects include earth and lower atmosphere layer.

Everything is node based now. Atmosphere reddening may need some tweaking, but it seems cool to me.

-mark

[starbase1]

Nick,

To be honest, I found a combination of using APS with real displacement and just using a regular bump map (NASA grayscale one) going to the bump parameter in the Node editor good for my tastes.

I was probably doing something wrong, but the normal maps do not seem to cast shadows correctly. You get shadows even if you have a light pointing straight at the object directly behind the camera exactly perpendicular to the sphere, if that makes sense.

Let me get a screen shot of what I have so far...sans clouds. I sort of started over doing everything nodal.

-mark

I'd guess that you might have had a combination of normal map too strong, and needed a bit more diffuse sharpness, that might do it...

Anyay, if we go different ways it will be more interestng!:thumbsup:

[avkills]

Nick,

Upon further experimentation and reading, It looks like to get the Normal maps from the Celestial Motherlode site to work correctly, I needed to Invert the X parameter in the node, plus also use a standard grayscale bump map applied to the bump parameter.

I have some new screen shots, and will post them to the WIP area shortly.

FYI I patched together the 32k tiled normal map from that site from the JPG files. I had to scale them each a little, but the resulting map is 25600x12800, way to big for lightwave to handle with all the other large images I am using. So I scaled it down to the same size as my colormap (8192x4096) and I am still using the 21600x10800 grayscale bump map.

If you have an FTP site I can FTP the map as a JPEG to you if you would like it.

I am still using all images as Targas; I discovered that even though the PNG take less room on the HD, LW expands them out to the same memory size as a Targa, so I am just using the Targa images. If you are having problems with LW crashing, I might suggest disabling the Hub -- it caused nothing but problems for me on my Mac.

-mark

[starbase1]

Thanks for the info, much appreciated. I have been working on other stuff this weekend, so have not been playing with normal maps much.

I'd love to see those maps - would it be possible to get the file via "www.yousendit.com" - I find them a rather good way of moving large stuff around for free...

Nick

[avkills]

Map should be on its way. The rest of them can be had at our favorite earth texture repository Blue Marble (Thanks NASA). :)

-mark

[starbase1]

Yep, picked up the file, thanks.

[oDDity]

Is there any way to get fprime to render normal maps?

[avkills]

If you have fprime ODDity I can send a small res version of the normal map if you want to experiment.

-mark

[Jeffers]

Just found this site that may be of interest:

http://amber.rc.arizona.edu/lw/normalmaps.html

I couldn't see if it had been mentioned in this thread already!!!

[Shulmanator]

Well, I was pleased to find a normal mapped moon in the sample scenes. Less pleased to find it bright orange!

Anyway, I turned off the glow, and replaced the orange fractal noise texture with an image map. I then turned down the strength of the image map to about 40%. It should really be even less for realism, but then it would not show off the normal mapping so well.

Here are my results, 2 of the front of the Moon, 2 from the back.


What's with the perfectly straight line between night and day? Polygon errors?

[connerh]

There's no reflected light coming from anything, so the diffuse line should be very very sharp.

[starbase1]

What's with the perfectly straight line between night and day? Polygon errors?

See reference image links above... The moon does not light like a simple sphere - it sends a lot of light straight back at the source due to the cubic crystal nature of much of the surface dust. It also does not fade as much near the limb, and requires strong diffuse sharpness. No atmosphere means no scattered light.

The moon is also deceptively flat, even the big craters. Expectations were shaped by stuff like Chesley Bonestell wonderful illustrations, and early SF films. For most of the big craters if you stood in the centre, the walls would be below the horizon.

[toby]

I think he's talking about where it gets even sharper, like polygon sharp, at the lower quarter of the image. I've had that too, where the surface is smoothed so you can't see polygon edges, but shadow rays can. The polygon edge is right at the center of the diffuse shaded area and blackens everything beyond it. That's what it looks like it might be -

Starbase, THANK YOU for the downloads! I've never been to your webpage, lots of great stuff!

Hey has anyone tried Infinimap Pro for the normal map images? I think he supports them. Infinimap ( by our own resident Lightwolf ) lets you use maps of ANY size, by just loading peices of the map that are getting rendered. Even Terrabyte size maps!

[starbase1]

Thanks Toby!

And as for the moon issues, the object is there for download so if anyone wants to try and fix it...

:D

[Intuition]

Have you guys used Infinimap for any of this?

I also wonder if Infinimap can utilize Normal maps?

I think it can. What would be good is to have the largest Normal map made converted into ECW/Jpg2000. Then you don't have to worry about maxxing out your ram from loading the images.

[starbase1]

By converting to 256 colour PNG there are not really any memory issues anyway...

Nick

[toby]

As a matter of fact one of the reasons I wasn't able to get Infinimap into DD is because he doesn't support PNG yet. We want to keep the smaller files of 8-bit PNGs even if we can use gigabytes of files in LW, because we don't want to clog up our network. With all the HD stuff we do now it would sure come in handy.

Back OnT, how does one generate Normal maps? There's a Modeler plugin, but it doesn't seem to work -

[starbase1]

What kind of normal maps are you trying to make?

There's a pluging that will map polys as normals onto a lower poly version, not tried it.

If you see page 3 of this thread, you will find how to turn bump maps into normal maps. This works VERY well with heightfields, you may have problems with more complex normals.

Nick

[Lightwolf]

How could I have missed this thread ... :foreheads

Appologies for jumping in, I hope this isn't too much OT...

As a matter of fact one of the reasons I wasn't able to get Infinimap into DD is because he doesn't support PNG yet. We want to keep the smaller files of 8-bit PNGs even if we can use gigabytes of files in LW, because we don't want to clog up our network. With all the HD stuff we do now it would sure come in handy.

Have you tried it using the unlicensed version? (Hey, it allows for free planar mapping on Y using the shader...).
I actually think even uncompressed JPEG2000 files are compareable to 8bit PNGs in size (8 bit palette, right? infiniMap Pro does support 8bit greyscale - as well as higher bith depths). Also, infiniMap Pro only loads the parts of the compressed image it needs across the network, and caches as much previously used image data as possible when you render an animation (it makes sense to farm out an animation in sequences of frames per CPU because of that).
If you want to, I can get you in touch with some other users that are using infiniMap Pro on render farms.

I also wonder if Infinimap can utilize Normal maps?

I think it can. What would be good is to have the largest Normal map made converted into ECW/Jpg2000. Then you don't have to worry about maxxing out your ram from loading the images.

Hm, good question. If you use lossless compression and JPEG 2000 it might work (you wouldn't want lossy compression artifacts in a normal map), however, infiniMap Pro uses triliniear interpolation/filtering on the image data, so it might destroy the normals. Off the top of my head it shouldn't though.

Cheers,
Mike

[toby]

Appologies for jumping in, I hope this isn't too much OT...

Have you tried it using the unlicensed version? (Hey, it allows for free planar mapping on Y using the shader...).
I actually think even uncompressed JPEG2000 files are compareable to 8bit PNGs in size (8 bit palette, right? infiniMap Pro does support 8bit greyscale - as well as higher bith depths). Also, infiniMap Pro only loads the parts of the compressed image it needs across the network, and caches as much previously used image data as possible when you render an animation (it makes sense to farm out an animation in sequences of frames per CPU because of that).
If you want to, I can get you in touch with some other users that are using infiniMap Pro on render farms.
There's Mike! Hi Mike!
One of the other problems is that we have nothing to create or edit JPG2000 files with. Getting a plugin installed for photoshop or another program installed will require as much nagging and justifying as it would to get Infinimap. Getting them to support JP2 in the other apps we use, like Nuke, would be yet another battle. But I'm still working on it, I'll have to remember to make a JP2 map at home and take it in -

Do you have plans to support the usual formats, png, tiff, tga, or psd? I think that would encourage more buyers!

[Lightwolf]

There's Mike! Hi Mike!

;D Hi Toby!

One of the other problems is that we have nothing to create or edit JPG2000 files with.
...
Do you have plans to support the usual formats, png, tiff, tga, or psd? I think that would encourage more buyers!
Well, you can use infiniMap to convert, but only for images of up to 500MB uncompressed.

As for other image file formats: All of these would more or less need to be translated to another, internal image file format. Using something like TIF or PNG as is would be _****_ on the network. Even rendering a low-res image, would require infiniMap to completly read in the image file. The advantage with J2K/ECW it that they can be read in parts... both local parts (i.e. only upper left corner) or resolution parts (i.e. whole image, but 10th of the resolution) - without having to read in the complete file.
I am working on OpenEXR support though, which is likely to be the native format for infiniMap, and will include a converter for that. _Maybe_ even an automatic conversion process it you load in another image file format. Still thinking of how to approach this.

Cheers,
Mike

P.S. Photoshop CS2 writes JPEG2000, not brillantly, but it saves them. I believe the plugin is not installed by default.

[toby]

Well, you can use infiniMap to convert, but only for images of up to 500MB uncompressed.
I did not know that!

As for other image file formats: All of these would more or less need to be translated to another, internal image file format. Using something like TIF or PNG as is would be _****_ on the network. Even rendering a low-res image, would require infiniMap to completly read in the image file. The advantage with J2K/ECW it that they can be read in parts... both local parts (i.e. only upper left corner) or resolution parts (i.e. whole image, but 10th of the resolution) - without having to read in the complete file.
I did not know that!

I am working on OpenEXR support though, which is likely to be the native format for infiniMap, and will include a converter for that.
That would help a lot, that's the other format we use. EXR is awesome, it should really be the standard right now. For anyone that doesn't have it yet: http://www.flay.com/plugs/downloads/openexr.zip

P.S. Photoshop CS2 writes JPEG2000, not brillantly, but it saves them. I believe the plugin is not installed by default.
Yes - this one even gives you the option of storing an alpha:
http://www.fnordware.com/j2k/
But so far I haven't been able to use it at work.

[Lightwolf]

That would help a lot, that's the other format we use. EXR is awesome, it should really be the standard right now. For anyone that doesn't have it yet: http://www.flay.com/plugs/downloads/openexr.zip

Actually, for anyone that doesn't have it yet: Wait another couple of hours and check our website later on today, we're releasing exrTrader today. Supporting all platforms, and it includes a free OpenEXR loader and saver :D

I'll be posting an announcement though.

Cheers,
Mike

P.S. It seems like we need to re-work our product information pages ;)

[toby]

Just tried the EXR plugin, thanks Mike! Especially for making a Mac version! But it looks like it's been rabbit-punched by the same bug in [9] that broke HDR output :censored: No problems in [8].
I guess when the bug gets squashed it will kick in - I hope.

I'll get the full version as soon as I get out of this hole! I shouldn't have bought this Mini Cooper!

[3dworks]

...

If anyone does find a Mac solution, let me know and I'll add it to the tutorial I am working on.

Cheers,
Nick

some time ago i did some experiments with normal maps and found this application for OSX here:

http://homepage.mac.com/nilomarabese/Menu13.html

maybe this is useful for anyone...

cheers

markus

[starbase1]

Thanks! I'll pass it on!
Nick

[toby]

some time ago i did some experiments with normal maps and found this application for OSX here:

http://homepage.mac.com/nilomarabese/Menu13.html

maybe this is useful for anyone...

cheers

markus
Great find! It works pretty well.

I'm liking Normal maps better than bump maps, even when they use the same map source. But I'd really like to learn how to make them from geometry. If/when I find anything I'll try to make a tutorial - if it doesn't already exist -

[Chrusion]

The only problem with all of these normal maps is that they're derived from shaded relief images and not actual high-rez topographic data. The Clementine radar topo data is so freakin low rez as to be totally useless. But looking at it one can clearly see the difference. Crater rims are similar to light value fuzzy edged rings on a nominal gray BG with depressions gradiating to black and peaks gradiating to white, etc. Topographic images created from highlights and shadows doesn't result in elevation data... just another shaded image.