very wide field / high rez view of northern milky way

[ptc]

this is about 60 x 45 degrees roughly

was shot using the 39Mpixel Proline and a 45mm f/4 pentax 6x7 lens used at f/5.6

it could definitely benefit from more exposure time but that's all I got in 8.5 hours

http://www.narrowbandimaging.com/milky_way_p45_pl39km_HaGB_page.htm

[renormalised]

That's the only thing I don't like about the big Proline....you have to push out those exposure times to get good light. It's really more of a "daytime" camera, unfortunately. Needs lots of photons to make up for the low QE.

Kodak needs to get its act together and make a 39000 chip that's good for astropiccies. Or even something larger:)

[ptc]

If they used the same 6.8 micron microlensed NABG transparent gate pixel as in the 3200ME it would be the astronomer's wet dream....

But it is still a lot of fun to use. Raw files are 78MB each.

The one shot color, when deinterlaced as 32bit Floating point color are 450MB each roughly. stacking 50 to 100 of those can take some time!

[renormalised]

Yeah, you'd think Kodak would've thought of producing a chip for their astro market as well, in this size or larger. Imagine having a 6-8 micron/pixel, 60 or 80 megapixel chip with 60% and above, QE. Then you'd need some really serious processing power to stack and process those piccies!!!. Guys would start building cluster farms of linux units for their home computers!!!:):P

You'd probably need something like that to handle the pics!!.

[ptc]

I am confident they are aware of the market but I suspect their assessment of the volume opportunity fell short of what they need to justify a new product development. The sensor in my camera was developed for Hasselblad for their 39MP "medium format" camera.

It cost a lot of $$ to do the design and make a set of masks / fab a wafer lot so they probably have to see an annual volume way higher than they estimate with this market and no one willing to fund the development as Hassy is rumored to have done.

The sensor was optimized for the DSLR application too: the very fast lenses of down to f/1.4 would have such a sharp chief ray angle that they had to not use microlenses due to the poor admittance angle properties of their microlens technology

also since the sensor needed to be one-shot-color, there's a metal frame that surrounds each pixel to reduce color cross-talk. The net net is that the fill factor is only 53% and the QE about 20%

Despite the low QE it actually makes sense to use for RGB widefield imaging if you follow this logic

http://www.narrowbandimaging.com/incoming/PL39000C_value_proposition.pdf

btw in case you are interested here's a nice technical paper on the sensor's development:

http://www.narrowbandimaging.com/incoming/31Mp_and_39Mp_Full-Frame_CCD_paper.pdf