Back in the day before digital imaging, much astronomy was done by comparing old photographic plates with newer ones, so it was more important that the emulsions were identical and stable (and optimized for each application) than that they be as sensitive as possible. It must have been very expensive for them to do this kind of quality control for a product with so little demand and sales, but to their credit, Kodak put out a quality product.

As a student, I worked with Kodak glass plates in stellar positional mapping, and we could be assured that plates exposed decades earlier had identical properties to the ones used today, and the images on each directly comparable. These old plates could not be replaced by newer technologies, they were a historical, nay, archaeological record. When a stellar image measured off one plate was noted to have moved relative to its position on another, or its brightness or color compared to a modern image, we knew any changes were due to the star itself, not the detector. Similar emulsion and bandpass filters were developed by Kodak for color photometry, spectroscopy, and other applications. Astronomers all over the world benefited from these products.

Years later, I worked with Kodak glass plates in photogrammetry, doing highly precise stereo mapping with images I knew were consistent and stable over time. Much of what geographers now know about the earth we can thank Eastman Kodak for.

But we know the sensor is overall 1600×1200, which is a 4:3 aspect ratio.

short side in degrees = 18.3 * 3 / 4 = 13.75

short_side / 4 = 13.75 / 4 = 3.43125

And the raw images in each band will be 1600×300.

So therefore and I rest my case.

So the images will in fact be Cinemascope® shape. The question for me is still whether they’ll do macro-level wizarding to compile those funny-shaped mediocre-def images into something beautiful we’ll all gasp over.

Trust the boffins.

Junocam looks to me like an innovative instrument designed, at least in part, to have minimal impact on what I’m sure was an already over-stretched budget. But I’m also sure it’ll treat us to some really cool hi-def space pics.

“Imaging boffins” indeed…

“HD” once had a technical meaning, but it’s a marketing term now. Generally we’re talking displays in multiples of 720×480 pixels. To me, “real HD” is 1920×1080. “4K: is the next step up @ 3840×2160, which produces amazing “photorealistic” pictures based on store demos I’ve seen.

But darn, 1600×1200, that’s disappointing…on the surface. But don’t overlook the ability of JPL’s imaging boffins. Consider this: The spacecraft is spin-stabilized at 3 RPM, plus there’s its orbital motion around Jupiter. Any time a camera pans across the scene, you have the opportunity to synthesize higher-resolution images out of successive images a fraction of a pixel shifted. Plus, image size in the direction of spin is effective unlimited, depending on how long you keep the shutter open and how efficiently you move partial images out of the buffer. It’s possible we’ll see long skinny images like Mars Global Surveyor, with better resolution on one axis than the other.

Not sure of the FOV discrepancy, but that sure doesn’t sound close to the 4:3 ratio of the imager. Something’s off. In any case, resolution of 15km/pixel, about 9 times better than Hubble, will be nice. I’m sure of it.

Maybe my excuse for upgrading displays is a mirage, but I still look forward to those images of Jupiter. Thanks for researching the camera.

That’s quite a bit more sophisticated than the “pushbroom” scanners I’m familiar with from the ’70s and ’80s, where a rotating mirror whipped the image past a set of CCD sensors, and where the effect of spacecraft motion was subtracted later during data reduction and processing. Back then, only a slight twist had to be removed from the image (by resampling the data) to compensate for the rotation of the earth under the platform.

I presume the resulting imagery will be fully rectified for public consumption (these purely geometrical effects removed and the resulting map flattened out and rubber-sheeted) to give an image that resembles more or less what we think of as a map projection.

Image processing and technology has come a long way since I meddled in it…

I’m appalled at how much I’ve missed.

They’ve bonded the color filters directly to the CCD. I haven’t tried the math, but I wonder if this is what gives rise to the “18.3 x 3.4″ degree view. Will that be the effective coverage of a single, “one filter” exposure?