Bone-fancying boffins in the States have issued an impassioned call for the world of paleontology to standardise on a digital file format for 3D images of fossils before it's too late. Ellis W. Shuler's photo of Glen Rose track Until recently, fossils were recorded by investigating boffins using moulds or casts of plaster or …
no universal standard accepted by software manufactures
Oh how that will echo through the history of slothful business.
If any manufacturers refuses to make its software compatible with some open standard you should never, ever buy their software again.
If they cant do it they are either completely crap at software or just completely crap at providing a service. Either way not to be touched.
This is not a title
But what if there isn't actually an open standard?
Has anybody actually gotten around to ratifying a standard for 3D images which are at reasonable high res? Was it ratified by something with a bit of clout and certified (as far as reasonably possible) to be without IP issues?
I'm speaking from ignorance but the press release sounded part a plea for somebody, anybody, to come forward with a reasonable standard that was open for use and part advert for the sponsors who wouldn't mind becoming a de-facto standard.
Astronomers use FITS for multidimensional data
Flexible Image Transport System http://fits.gsfc.nasa.gov/documents.html
but it is old (and consequently simple)
Newer astronomy systems are moving to HDF5
I agree completely.
Working in the DTP field, it's so annoying to have all the Quark and InDesign hassles.
ie Old versions of Quark only being readable by a slightly newer version... which then needs a slightly newer version to convert to the latest version.
InDesign is slightly better - but not much, and that's probably down to it being not quite so old.
What we need is some kind of universal postscript/pdf/whatever. So that you could open it in either program - or new yet to me written software. Grrr.
...before MPEG_LA comes along and says "we are sueing you for using a single standard that benefits all mankind"?
OK, I'll get off me soapbox.
Would be a great article....
....but now using Flickr to display the pic's.
Which of course many of us work people can't see.
Videos using YouTube I understand. but pictures, c'mon.
And howlong before M$ adds an extension to the file format that makes all files incompatible with old versions of M$ XP DinoScan forcing you to upgrade while at the same time sueing anyone who produces DinoScan compatible software for other platforms
As there's a free version of Sketchup available they should investigate exporting to that.
It would also map the data to Google Earth so you'd see exactly where the footprints are situated. Sketchup runs on Mac Windows and Linux under Wine.
Sketchup uses a proprietary file format, and Google keeps changing it to keep the "don't wanna pay for premium" users frustrated. I've been told it's not a very nice format.
there already is a standard... kinda
wavefront .obj as mentioned. everything 3d will open it, 3ds max, lightwave, maya, hell even poser :D
most 3d kit wil read it, almost as ubiquitous as autodesks dxf
as above for .obj but moreso
either format will do as they are basically a plaintext file of xyz values. I use proper big lidar datasets all the time, they tend to be in ascii x,y,x 10 chars per field (which i tihnk is the bare bones of what you get in an .obj). Makes for mahoosive files sometimes, but what can you do? yuo want big things modeled in minute detail, dataset will be large.
Wavefront .obj format.
Yes, it is plain ASCII.
It's not perfect. For one thing, there doesn't seem to be any way of recording the scale (except as a comment in the file). 1 unit in Poser is around 8 feet (it's changed between versions for no apparent reason), while in other systems it might be a metre or a foot or a half-fathom.
There's usually three sub-sections to the file. First, the set of vertices. These can be referenced by the modified line-number. Then the UV coordinates of each point (Mapping the 3D point onto a 2D texture--limited use here). Finally the polygons, made up of edges defined by pairs of vertex numbers.
Poser uses 6 digits after the decimal place, which for a scale of 100 inches per unit means you can define a shape to 1/10000 of an inch. Is that excessive detail?
There are elements of the format, such as defining the material of a polygon, which have little obvious use. In the example of the dinosaur footprint one might want to distiguish the footprint from the surrounding surface. And the UV map data might be useful for rendering software, allowing 2D images to be mapped onto a 3D model.
It's maybe easier to translate to some future format than some file formats are. Does it record enough the right data for this use? I've noticed quite a few resolutely proprietary formats in the world of 3D models. They can be a pain in the butt.
My thoughts too...
I thought DXF straight away, then got on to thinking about other widely available 3D formats...
If only there was some precedent for dinosaurs being moddled in 3d, maybe it's been done in film, or on the telly?
of course UV's
forgot about them!
the terrain mapping lidar i use has the capacity for uv mapping digital imagery captured as part of the survey, then the files become unbelieveably massive (considering the point data alone can come in at about 8Gb per km surveyed) - takes forever, but looks incredible.
as for the detail, im not so sure about the reverse engineering type of kit these guys are using, presumably a couple of orders of magnature better than survey lidar, which is typically +- a couple of cm's.
of course the real bugbear with this is any kind of processing takes forever, which is probably the problem, click to convert and wait 3 days...
propriatary formats arew always a PITA, but as i said earlier OBJ is the way to go - especially with the UV coords I had forgotten about. and there is already tons and tons of software out there to crunch them numbers.
UV mapping shouldn't be an issue with high resolution models; if anything UVs aren't particularly reliable, they aren't a great way of storing accurate (uniform) colour information, and require an external texture map.
At the kind of resolutions they're working at (or should be working at, 1m polys is woeful for archival data) they could be using a UV-less system such as PTex:
Any zBrush or 3DCoat artists here familiar with polypainting on multi-million poly models will know how good life is without being locked into bitmap textures with their working models.
OBJ is fine for moving data between applications, but it's not a good archive format.
Why don't they just write a format a be done with it.
Not very impressed by their scanning resolution either...people never think of the future. Just too lazy to haul rocks and bones around, i dunno...
RE:Not very impressed by their scanning resolution either.
1.2mm is pretty good accuracy if we are talking absolutely.
I, for one...
...Welcome our laser-scanner-totting, three-dimentional dinosaur overlords.
"A statement from Souther Methodist University"
Are they even allowed to believe in dinosaurs?