Gigapixel project...

[SamuraiCatJB]

Now why didn't I buy one of these.... :eek: :eek:
http://www.gigapxl.org/
;);)

[jbbooks]

I want one!

[the_iceman]

ok... I read through that website (after looking at the awesome pics)... and my head hurts..... There was so much technical information ...... WOW. I can't even pretend to state that I understand it all... LOL

[SamuraiCatJB]

now one of us needs to build one... anyone won the lottery recently? ;) :p

of course you did notice it is basically a film camera where lens/film/scanner are chosen to optimize resolution to the grain of the film/optics to precisely match each other.

[phsv]

Maybe if we all pitch in, we can do our own. Anybody in?

[Tam Hanna]

Looks lovely.
But I don't know how big it will be in the end! I don't think one will be able to carry it in a jacketpocket, like one can with the SX1.
BTW, did you cehck out the images of the golden gate bridge already? In ther really high zoom modes there is a lot of flicker/artifacts to be seen....

[the_iceman]

I haven't gone back to the site.... and think the Bridge was using a 'standard' 480mm zoom. Still... for the distance you are talking about (showing the ship way off in the distance) .... I wouldn't know how to calculate the actual distance between the camera and the ship behind the pilons ..... it is still incredible you can see the ship so well by cropping the picture.

The website does state that atmospheric conditions play a much bigger role with the camera than with most. Because of the highest scan matching it does against the film... etc. (again... there was a lot of the technical specs I didn't fully understand/absorb). I remember it stating ... the eye 'processes' the atmosphere much differently than a camera does. Our eyes learn to 'ignore' a lot of atmospheric conditions whereas a camera doesn't. A tiny little amount of 'mist' coming down 1 mile away ... we won't notice it... but a camera such as the giga certainly will when it goes to process the picture.

[SamuraiCatJB]

I haven't gone back to the site.... and think the Bridge was using a 'standard' 480mm zoom. Still... for the distance you are talking about (showing the ship way off in the distance) .... I wouldn't know how to calculate the actual distance between the camera and the ship behind the pilons ..... it is still incredible you can see the ship so well by cropping the picture.

The website does state that atmospheric conditions play a much bigger role with the camera than with most. Because of the highest scan matching it does against the film... etc. (again... there was a lot of the technical specs I didn't fully understand/absorb). I remember it stating ... the eye 'processes' the atmosphere much differently than a camera does. Our eyes learn to 'ignore' a lot of atmospheric conditions whereas a camera doesn't. A tiny little amount of 'mist' coming down 1 mile away ... we won't notice it... but a camera such as the giga certainly will when it goes to process the picture.

it isn't that we learn to ignore it... my boss can explain it much better, but it deals with pupil size vs lens. When you look at a scene with your pupil you are getting near a straight line view through turbulence, it minimizes turbulence because it is looking through the fewest possible "pockets" of turbulent air. When you look through a camera or telescope you draw a larger light area and focus it to increase light gathering power, but now you are looking through 10 times the number of pockets of air in a camera and 100 times in a reflecting telescope. You gather more light, but at the cost of increasing the effect of atmospheric distortion.

[Gizmo]

Maybe if we all pitch in, we can do our own. Anybody in?
I'll supply the duct tape! :D

[SamuraiCatJB]

I'll supply the duct tape! :D

I got a screwdriver. :)

[Gizmo]

OK, Sam and I did our jobs. Anyone want to donate 20 grand? Do I hear 10?

[SUCHY]

gigapixel, I wonder what would be the size of 1 photo, ! gig probably.

[SamuraiCatJB]

gigapixel, I wonder what would be the size of 1 photo, ! gig probably.

3 gig, 3color. :)

[Tam Hanna]

Store that......
3Gb is impossible, I do not know of any codec/machine that would be capable to handle such a huge amount of image data. But, there is 64bit coming-so who knows...

[SamuraiCatJB]

if you read the detail on the process... they are combining film and scanner such that the scanner resolution matches the grain resolution of the film. The film being 9 inches wide has a lot more grain resolution than 35mm film.

Since you are not talking about a CCD or CMOS sensor, and scanners work line by line, Image size is indeed possible for 3gig. I deal with much larger images in satelite imagery. In fact they even talk about the fact that the first gigapixel imaging system was satelite based sensor. Although the satelite sensor works very differently, they wanted to achieve the same result. Now it is possible they are talking gigapixel equivalent to a CMOS or CCD based camera, in which case it would be a 1gig image. With the exception of only a couple of cameras, there is not a sensor for each pixel of rgb, there is one or another and the filter process recombines them into full rgb for every pixel based on nearby rgb sensors. Wavelet compression works well into the terrabyte range in still image capability, there are several used for satelite imagery ECW, MrSID, and others.

[SamuraiCatJB]

okay, my mistake, the first 1 gig pixel camera was a night sky camera, astrocam. I read too much into it. It was 1000 megapixel. The gigapixel land camera claims to be up to 1250 megapixel

100 pixel/mm scanning, military 9"×18" format film (104,500 mm2). 10000 pixels ber mm square would be 1,045,000,000 RGB pixels by my count. Each image is stored on its own DVD. I assume they used simply used a low grain film to achieve the 1250 megapixel....

they hope to reach 4000 megapixel in months.

[SamuraiCatJB]

Store that......
3Gb is impossible, I do not know of any codec/machine that would be capable to handle such a huge amount of image data. But, there is 64bit coming-so who knows...

"The ECW JPEG 2000 Compressor has the ability to compress images of up to 500MB in uncompressed size. To compress imagery of practically unlimited size (up to and including the multi-TB range) to ECW and JPEG 2000 you can use ER Mapper’s flagship product, ER Mapper."

We just aren't used to seeing the codecs because they are a specialized business, but they have been around for about 6+ years now. :) it's how I justify my 2.5 TB disk array. :D

[SamuraiCatJB]

just for giggles and grins was thinking how I could do this by image stitching... its still a massive prospect, but had to see....

Their current film process is 1045159200=22860x45720 pixels....

thus from a 6 megapixel camera, an equivalent image would be 15.24x 4/3 (for image overlap) 20.32 images wide and 15.24 images tall. More work than I am willing to do... and more than I have memory anyhow... :) But... something to think of... if I had an electronically operated tripod, like I have had with telescopes... you could command a sequencial perfect sweep. It would be very easy in a computer controlled tripod. Even easier if I were take one of our 1024x768 motion cameras at 120fps and simply used a movie sweep to cover the area.... hmmmm... :)

Which also shows why their image has to fit on a dvd. it occupies 2.92gig of diskspace.

[the_iceman]

and now they are 'done' with the proof of concept for the gigapixel camera ..... and I believe they are working on the 4gigapixel camera.... that will require a few more computer resources.... lololol :)

[SamuraiCatJB]

and now they are 'done' with the proof of concept for the gigapixel camera ..... and I believe they are working on the 4gigapixel camera.... that will require a few more computer resources.... lololol :)

I may just have to do the movie thing for work, since I will be working with those cameras soon... :D :p

[the_iceman]

Sweeeeeeeeeeeeeet Would be very Cool to see.

With digital cameras like the rebel and d70..... what I did for 'overlap' was I panned from left to right. After my first pic. I looked where my far-right focus sensor was and then moved the middle focus sensor to that area to get the 2nd pic. I know I ended up with a bit more overlap this way (about a 50% overlap). However it seems the more you overlap the better the picture comes out. I did try a test with about a 30% overlap and the picture was a little more 'skewed' looking. It has been fun testing doing panos.... that is for sure :)

I look forward to your first gigapixel picture without a gigapixel camera :)

[SamuraiCatJB]

That will be a while. :) we have a 2 year project for those cameras. :) the purpose is actually to do something else... but I may have to go through the same process as a middle step.