Amiga.org
Coffee House => Coffee House Boards => CH / Science and Technology => Topic started by: blobrana on December 10, 2004, 11:56:18 PM
-
Hum,
"the construction of what will be the world's tallest building is set to begin in Dubai, United Arab Emirates. The building contract was awarded to a consortium led by the South Korean Samsung Corporation on Thursday.
The Burj Dubai tower will stand 800 metres tall - just 5 metres shy of half a mile - once completed in 2008. That will be a full 350 metres taller that the tallest floored in the world today, the Petronas Towers in Kuala Lumpur."
link (http://www.newscientist.com/news/news.jsp?id=ns99996781)
-
:-o
-
That could be the start of the "half-mile high club"
-
:lol:
-
X-ray wrote:
That could be the start of the "half-mile high club"
If you do it twice, does that count as joing the mile-high club? ;-)
-
Hum,
well i`m a `base-jumping` virgin.
----------------------
[if in doubt - google]
-
I wonder how long it would take a chunder from the top to reach the ground? And what kind of area it would cover?
-
Hum,
The object just falls at a fixed speed set by its mass and air resistance (eg, ~200km/h for a falling human = 14.4 seconds)
As for the `splatter` (the technical term used to describe how many aliens can be spread on walls) best to download this splatting algorithm (http://mpire.sdsc.edu/Splatter/Splatter.html)
(Doh, needs Cray super computer)
-
What is a chunder (besides an AO member?)
-
@ Blobzie
That Splatter stuff is pretty cool. Those features are now standard in all CT scanners. I was doing 3D reconstructions of the spines of traffic accident victims like that (lower quality because I seed the voxels manually) back in 1995 on a lowly Philips Scanner attached to a SUN Sparc station. These days the quality of the reconstruction is dependant on the patient's ability to keep still during the scan, and the z-value for the stack (in other words, how thin the 'slices' are). If you compare the female head in your link to the dry male pelvis, you will see that the head has 'stepping' in its surface contour. This is probably because they used slices of about 3mm to 5mm thickness, not overlapping. The pelvis was probably done on 1mm or 2mm thickness, possibly overlapping. Of course, when you have a dry pelvis you don't have to worry about irradiating the 'subject' as much as when you have a living person. I was once brought an archaeological relic, hundreds of years old, and asked to scan its insides and I got much better pictures on that than a moving patient.
As a matter of interest, the same 3D technology is now being used to do virtual colonoscopies by means of CT. Instead of physically driving an optical device through someone's lower alimentary tract, the person is CT scanned and then the slices are reconstructed based on tissue density levels chosen by the radiographer. You then have a virtual colonoscopy and you can 'fly through' the person's bowel, looking for polyps and other pathologies, just like a game of Wipeout.
Edit: You shouldn't need a supercomputer for this, after all they are doing this sort of thing on today's CT scanners. I've seen much smoother reconstructions on our current hospital hardware. All they need to do is write some software that takes a series of preformatted slices (I could get these quite easily from the CT scanner at work) and then have the software 'stack' the slices in a column, using a corner annotation as a registration mark, and Robert will be your father's brother.
In fact, I could probably retrospectively reconstruct these gunshot spine injuries, just by scanning the films on my Epson 4870, adjusting the levels in Photoshop so that we have an edge pixel value of 255,255,255 and then slap those onto some slabs in Cinema 4D. The only thing I would have to figure out is how to 'trim' the excess black from the rendered stack.
-
Karlos wrote:
I wonder how long it would take a chunder from the top to reach the ground? And what kind of area it would cover?
Chunder :- to Barf, Blow Chunks, Vomit etc.........
-
Ooooooh, to call for Ralph. Now I get it.
-
X-ray wrote:
In fact, I could probably retrospectively reconstruct these gunshot spine injuries, just by scanning the films on my Epson 4870, adjusting the levels in Photoshop so that we have an edge pixel value of 255,255,255 and then slap those onto some slabs in Cinema 4D. The only thing I would have to figure out is how to 'trim' the excess black from the rendered stack.
Well, wouldn't you just make the black transparent?
-
@ Karlos
I would love to make it transparent, but I don't know how, because it would be like mapping a fried egg onto a small paperback book: Cinema will see that as one object. If I etch the outline of the white objects out from their black backgrounds in Photoshop, I then have a problem of using that bitmap to delineate a spline boundary from which I can extrude an individual slice in Cinema.
Maybe I'm not thinking on the right track. One thing is for certain: any attempt to manually delineate the borders of a shape in Cinema will result in a very jagged reconstruction.
-
Presented with a problem like this, I'd get annoyed and then hit a C compiler ;-)
Assuming the scans are false colour, I'd assign a transparent colour.
I'd load each image in turn, pre-rotate it into a desired perspective (either using it as texture for an OpenGL quad, or perhaps some custom rendering routine) and stack them depth sorted (farthest away first) according to the current viewing direction.
It wouldn't give you a realtime view, of course (unless your machine has some real grunt) but it would give you a 3D reconstruction from the slices.
-
I'm starting to hate you, Karlos :-P
-
@X-Ray
TBH, it wouldn't be ideal. It's conceptually not any different from printing your slices (with the black set transparent) on acetate and stacking them.
-
@ Karlos
I have found a solution (but it only works for an outside skin):
You start in Photoshop, import all the images/CT bitmaps/whatever, make sure they are all the same size and orientation, then batch 'adjust levels' so that all the backgrounds are 0,0,0 (I found it doesn't matter if the edges aren't 255,255,255 as long as they are whitish). Then save all the images in sequential order filename (I used TIFF, uncompressed).
Then in Cinema: Objects/Spline Primitive/Vectorizer...
You then specify the texture for each spline in sequence as the image files you saved earlier. Then for each 'vectorized' spline you choose Objects/NURBS/Extrude NURBS and set the extrude amount. If it is a series of CAT scan images being stacked then the extrude amount can be the slice interval in millimetres (you can read that from the scan), or at least make the extrude amount proportional to the Photoshop bitmap/ actual bitmap (from the CT relative to the actual slice interval as seen on the scans).
It sounds like mumbo-jumbo, but I can't explain it better and it is actually very simple (well it must be if I can do it)
-
@X-Ray
Alas 3D modelling and animation is to me what C++ is to you perhaps :-)
However I'm glad you found a way :-D
-
Update: I have the proper solution, but you may not like it:
I found that you need a hell of a large bitmap to make the vectoriser accurate. The reason why I thought you could only have the outside skin was because my bitmap was too small (it was 512 x 512).
So I made a 6000 x 6000 bitmap with a white solid circle and inside that one I drew a black filled-in ellipse. Saved as TIFF, uncompressed.
Used the Vectoriser technique in Cinema, system almost puked with the strain, but after a while I had the spline, with the elliptical hole cut in it, fairly smooth. I must say I'm very impressed with this. I increased all the subdivisions upon extrude, and it is really nice.
Bad news: it is hogging 885mb of RAM during rendering :-o
And I only have 2gb, so at this rate I'll be down to half my RAM very soon :-D
Edit:
"..Alas 3D modelling and animation is to me what C++ is to you perhaps.."
Trust me, the programming is more difficult, has to be. At least with raytracing you always get a result and you can tinker
-
Good to see you back Blob. :pint:
800 meters, that's roughly 2,400 - 2,500 feet (beer addled estimate).
-
aw man, I knew about this, yes, but I didn't knew that it's already in construction
half a mile - one should have volunteers to work on one of the top floors of that building, and I will NEVER be one of them (well, that is, I once stated the same about the tallest tower in paris -not the Eiffel tower, btw- but ten minutes later I was at there) :nervous:
how the heck can they make it hurricane/earthquake/excessive-temperature-changing proof?
-
Speelgoedmannetje wrote:
how the heck can they make it hurricane/earthquake/excessive-temperature-changing proof?
With a disclaimer. :-P
-
:roflmao:
-
Imagine having a top floor suite there: the ultimate in penthouse views.
Imagine also getting out of the lift at the ground floor, feeling in your pockets for the car keys, and realising you left them upstairs :lol:
Even better: "I'm sorry, sir the lifts are out of order, you'll need to use the stairs"
-
@Eyso:
how the heck can they make it hurricane/earthquake/excessive-temperature-changing proof?
There is a huge steel ball, at the base of the tower, that somehow counters the effects of an earthquake.
I'm sure someone else here will give an explanation of how it works.
I think it's quite a cool looking building.
-
Its probably a pendulum effect of some kind that absorbs the translational energy and keeps the building vertical when all the rest are shearing side to side and bouncing up and down...
Or something...
-
There was a program about this on TV earlier this year, and they had a huge ball mounted on hydraulic limbs just like Wilse says, and sensors all around the surface of the building would feed into a computer that would drive the hydraulics so that the building's center of gravity could be changed slightly to directly counter the effect of wind.
Edit: they are called Tuned Mass Dampers:
http://www.columbia.edu/cu/gsapp/BT/BSI/TMD/tmd.html
http://nisee.berkeley.edu/prosys/tuned.html
-
Hum,
do you think they`ll use microsoft to provide the software?
--------------------------------------
A tree that does not bend in the wind,
is in a green house.
-
"..Hum,
do you think they`ll use microsoft to provide the software?.."
---------------------------------------------------
They had better not, or there may be a crash
;-)