Bayer CCD vs. LRGB imaging

What do you think about the color Bayer mask CCD versus the
monochromo+filters LRBG as far as results is concerned?

Antonio Zanardo

On Thu, 19 Oct 2006 22:15:01 GMT, "Antonio Zanardo"
wrote:

What do you think about the color Bayer mask CCD versus the
monochromo+filters LRBG as far as results is concerned?

The best images are RGB via separate filters. Next is LRGB. Worst is
one-shot color (Bayer RGGB, CMY, etc). The problems with the one-shot
are partly because of artifacts from the color components not lining up
perfectly, but mostly from the fact that the dye filters over the pixels
are a bit leaky, they don't have the correct bandpasses for astronomical
emission sources, and in many cases the passbands aren't even monotonic
(so there are ambiguous color combinations).

That's not to say that you can't get fair results with one-shot color
cameras, especially with targets that lean heavily towards continuous
emissions, but the results won't usually compare with what you can get
by combining individual RGB or LRGB frames.

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com

On 2006-10-20 12:47 +0900, Chris L Peterson wrote:

The best images are RGB via separate filters. Next is LRGB. Worst is

Hi, Chris. Why would RGB be better than LRGB? Is it something to do with
binning on LRGB exposures?

I apologize if it's a stupid question. I haven't done any imaging yet
apart from some afocal lunar shots with a cheap hand-held.

Thanks,

trane
--
/////////////////////////////////////////////////////////
// Trane Francks Tokyo, Japan
// Practice random kindness and senseless acts of beauty.

Antonio Zanardo wrote:
What do you think about the color Bayer mask CCD versus the
monochromo+filters LRBG as far as results is concerned?

Antonio Zanardo

Hi:

There's no doubt that a filtered monochrome CCD (whether LRGB or some
other technique) produces superior results. HOWSOMEEVER...there's no
easier way to get color deep sky shots than with a one-shot color
camera (and there are some very capable models out now, like the SBIG
one-shot color St2000). For a lot of folks "easy and good enough"
trumps "difficult and more better." ;-)

Peace,
Rod Mollise
Author of:
Choosing and Using a Schmidt Cassegrain Telescope
and
The Urban Astronomer's Guide
http://skywatch.brainiac.com/astroland

On Fri, 20 Oct 2006 15:23:47 +0900, Trane Francks wrote:

Hi, Chris. Why would RGB be better than LRGB? Is it something to do with
binning on LRGB exposures?

RGB provides the purest chrominance information. With LRGB, your
luminance is normally out of balance with your chrominance (in many
cases, the L even has an IR component included). In other words, the
luminance signal is weighted by the spectral responsivity of the
detector, which is different than the spectral responsivity of the
convolved RGB components. Furthermore, the main reason that people use
LRGB is to reduce the total exposure time by collecting high resolution
luminance and low resolution color. So LRGB images contain less
information, even though this is usually not readily visible to the eye.

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com

On 20 Oct 2006 05:56:42 -0700, "RMOLLISE" wrote:

There's no doubt that a filtered monochrome CCD (whether LRGB or some
other technique) produces superior results. HOWSOMEEVER...there's no
easier way to get color deep sky shots than with a one-shot color
camera (and there are some very capable models out now, like the SBIG
one-shot color St2000). For a lot of folks "easy and good enough"
trumps "difficult and more better." ;-)

I disagree. I don't find getting data with a one-shot camera any easier
at all- with a dedicated astrocamera, either way you push a button and
let a data collection program run. The only complexity that using
filters adds is the need to take additional darks and flats, but those
normally come from a library, so the added work is needed only
occasionally, not routinely. However, when you use a one-shot color
camera (unless you simply don't care about reasonable color at all) you
hugely increase the time it takes to process images. A separately
filtered image might take 30 minutes to completely process; a one-shot
image can take hours.

IMO, it is false economy to view one-shot color cameras as simpler or
easier to use.

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com

Chris L Peterson wrote:

I disagree. I don't find getting data with a one-shot camera any easier
at all- with a dedicated astrocamera, either way you push a button and
let a data collection program run. The only complexity that using
filters adds is the need to take additional darks and flats, but those
normally come from a library, so the added work is needed only
occasionally, not routinely. However, when you use a one-shot color
camera (unless you simply don't care about reasonable color at all) you
hugely increase the time it takes to process images. A separately
filtered image might take 30 minutes to completely process; a one-shot
image can take hours.



HI Chris:

You are allowed to disagree! Well...this one time, anyway. :-)

However, most of the rest of us--especially those who, like your Old
Uncle, could be best desribed as "imaging dabblers"--are gonna disagree
with _you_, since to do tricolor imaging:

You need three (or four) good exposures of the target, not just one.
Many of us have a hard time getting that one.

The exposure has to be right for each image.

Exposure time goes up for filtered shots, too, giving us more time to
screw up.

The resulting shots must be calibrated and have to be properly exposed
and registered and balanced. I've seen some "wizards" struggle with
this on occasion.

I've never seen that a one-shot image takes "hours." Quite the reverse.
OTOH, I don't aspire to appear in the back pages of S&T or challenge
Jack Newton. For my purproses, monochrome or one shot color does just
fine. I like black and white a lot, but when I have to have color, one
shot does fine for me.

'Course I'm in awe of what you and those other folks who actually know
what they are doing have accomplished. ;-)

Peace,
Rod Mollise
Author of:
Choosing and Using a Schmidt Cassegrain Telescope
and
The Urban Astronomer's Guide
http://skywatch.brainiac.com/astroland

On 20 Oct 2006 07:26:30 -0700, "RMOLLISE" wrote:

However, most of the rest of us--especially those who, like your Old
Uncle, could be best desribed as "imaging dabblers"--are gonna disagree
with _you_, since to do tricolor imaging:

You need three (or four) good exposures of the target, not just one.
Many of us have a hard time getting that one.

But you need the same total exposure time for the same S/N. Why is
taking three or four images more work than one? Either way, you push a
button and come back when it's done.


The exposure has to be right for each image.

But it's always the same ratio, based on your filters. How hard is that?


Exposure time goes up for filtered shots, too, giving us more time to
screw up.

No, it doesn't. For a given S/N, you need a given amount of photons. If
you are shooting shorter one-shot exposures, it just means you are
settling for lower S/N. You can shoot the tri-color shorter too, and get
the same results.


The resulting shots must be calibrated and have to be properly exposed
and registered and balanced. I've seen some "wizards" struggle with
this on occasion.

Exactly. This is by far the hardest part, and it is very much more
difficult with one-shot images. With good tri-color images, you can
normally calibrate on standard star colors and you're 90% of the way
there. Once you've worked out the calibration factors (once!), color
processing is almost automated. In fact, you can get quite acceptable
results with no user intervention at all. Not so with one-shot color.
Because of the leaky filters and poor bandwidths, you can't calibrate on
star colors and you need to apply manual color tweaks to selective parts
of the image (and selective colors) to get something reasonable unless
you have nothing but continuous sources in the image.


I've never seen that a one-shot image takes "hours." Quite the reverse.
OTOH, I don't aspire to appear in the back pages of S&T or challenge
Jack Newton. For my purproses, monochrome or one shot color does just
fine. I like black and white a lot, but when I have to have color, one
shot does fine for me.

I prefer monochrome astroimages myself. But when discussing color, I
think you fall into a logical fallacy comparing tri-color and one-shot.
The reason you find one-shot simpler is because you are willing to cut a
lot of corners (shorter exposure time, simplified processing). You could
cut the same corners with tri-color and you'd still end up with better
results.

The only time I sometimes find one-shot simpler is when using a DSLR. In
a field setting, the ability to stick a camera on the back of a scope
without worrying about setting up a laptop and control software can be
very liberating. But I still end up with images that take far more
processing than would be required for tricolor. But the acquisition
component is somewhat simpler.

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com

On 2006-10-20 23:01 +0900, Chris L Peterson wrote:

Hi, Chris.

On Fri, 20 Oct 2006 15:23:47 +0900, Trane Francks wrote:

Hi, Chris. Why would RGB be better than LRGB? Is it something to do with
binning on LRGB exposures?

RGB provides the purest chrominance information. With LRGB, your
luminance is normally out of balance with your chrominance (in many
cases, the L even has an IR component included). In other words, the

[ snip ]

Thanks for an easy-to-understand explanation. It is very much appreciated.

trane
--
/////////////////////////////////////////////////////////
// Trane Francks Tokyo, Japan
// Practice random kindness and senseless acts of beauty.

Along the same topic, but slightly different...

How do you feel about the oversaturation applied to the color in many
astro shots? I think it is taken to the extreme at times and can give
the uninformed unrealistic expectations for observing.

My personal opinion is that the new mural at the Griffith Observatory
is a little "over the top" if the color saturation shown in the APOD
pic of the same accurately reflects the public display:

http://antwrp.gsfc.nasa.gov/apod/ap061011.html

---
Michael McCulloch