Monochrome vs Color CCD

[cloud_cover]

It will be closed though if I use the focal reducer or use a filter with 2" tube...

[orly_andico]

i realized you must keep the reducer insulated from the camera nosepiece because if it gets cold, the fog will form on it..

[cloud_cover]

found this graph from our friend Christian Buil:



Look at that line up there (the Kodak KAF-3200 used in some SBIG cameras)

The QE of a Bayer sensor is 1/3rd of a mono sensor, because you've got 4 pixels (2x2 matrix) 2 green, 1 red, 1 blue, then you interpolate that into 4 pixels each with RGB values. The intrinsic QE of the DSLR sensor is about the same as the mono sensor, but it gets flattened by the Bayer filter.

Does that graph mean an astro CCD is about 7x more sensitive than a DSLR sensor?

[orly_andico]

It's 0.8 vs 0.16

That's 5X.

But as mentioned... the QE of the DSLR sensor is really about 0.6 to 0.7, but the Bayer sensor destroys this.

Consider: on a mono sensor with 1.0 QE (ideal) every photon that hits the pixel produces 1 photo-electron. So 100 photons = 100 electrons.

Now take that sensor and put a Bayer filter on top of it. The Bayer filter has the following characteristics:

Red = 0.299 Luminance
Blue = 0.114 Luminance
Green = 0.587 Luminance

So your 100 photons coming in, translate into 30 photons of red, 11 photons of blue, and 59 photons of green.

So in a 2x2 pixel matrix, you have two green pixels (each gets 59 photons), one blue pixel (11 photons) and 1 red pixel (30 photons). A total of 159 pixels.

If there was no Bayer filter, you'd get 400 pixels over that same 2x2 matrix.

So there's about a 60% (3/5ths) light loss because of the filter. Now the DSLR sensor isn't built to the same standards as the KAF-3200ME, which accounts for the further loss of sensitivity.

However if you're looking at a Starshoot Pro or QHY8 (both of which use DSLR-type sensors) then their only advantage is cooling.

Makes that SBIG 8300M look even more drool-inducing don't it :-)

[cloud_cover]

So generally if I'm looking at maximising the efficiency (i.e. reducing my exposures), then a monochrome CCD is virtually inescapable? (yikes...... $$$$$$$$$)
The 8300 is drool worthy but its USD2K, new, not counting filters and filter wheel (add another 2K for whole set)....
So that's USD 4K for something I can achieve with a modded DSLR but at a higher exposure level.......
So spend that 4K on a better mount/AG or even AO and a modded camera or 8300C?

[timatworksg]

You can still get great shots with an unmodded DSLR. Of course with a modded one things are better as far as nebulas go. For Star Clusters an unmodded DSLR is good, since you don't want any filters when shooting clusters. Unless you intend to stack filtered and unfiltered images in PS.

To start with, the good ole DSLR unmodded will do. You may not get as much detail as those with purpose built/modded cams, but you will still get a great shot! A monochrome CCD is more light sensitive compared to a OSC. But you can get great images with OSC. All a matter of trying and seeing what works, where and when is the best time to shoot.
For me, I shoot DSLR when targets are at least 30degrees to avoid LP.

If using a H-a filter, you can use it with a DSLR or OSC but you need longer exposures so it will pick up more light. But with a DSLR longer exposures bring noise and LP.

Dont forget to grab a programmable Bulb Remote for your cam so you can set Bulb exposure timings and how many to take, unless you can hook it up to software.

[cloud_cover]

Thanks Tim!
For me I try using my UHC filter to reduce LP; I seem to have it bad even at Zenith and I'm in the far north! (facing P. Ubin and the dark side of johor)
I'm just curious how a HA filter will work; I haven't tried my OIII yet (the UHC is simply a broader OIII+HA)
Practically the max I can push my cam to at 30deg C is 10mins at ISO1600 or 5mins at 3200 before amp glow starts to take over. Seeing I don't have a guider yet though, that 5 min limit may not be reachable due to mount factors anyway. 1600 is virtually noiseless unstretched and even 6400 is very acceptable if not stretched
I'm really quite keen to see how much HA my low-pass filter will let through, combined with the fact I can do ISO 12500 with some noise (although that's probably distinctly above unity gain for a 14bit ADC).
I do have a nice programmable bulb remote. Delay before shooting, exposure, interval between shots and number of shots
Of course now what I need is a GEM.......

[rcj]

The decision-making process of going via the DSLR or astro CCD route can be both a mind-boggling and daunting activity, though sometimes if you just want excellent quality pictures (irregardless of molar), the way to go is still getting a dedicated monochrome CCD astro-camera. However, once we factor into the convenience of image yield, sky opportunities, money, accessories, option for terrestrial photography, the decision invariably becomes not as clear cut, and going either route is a plethora of pros and cons.

Before we can all decide (and note that every single budding astrophotographer views the pros and cons in his own perspective and situation, and where he/she is coming from), here are some insights:

1) Image dynamic range.
For most DSLRs, we are talking about 12-bit to 14-bits. Basically, this means in the image histogram, any image pixel can hold an intensity value from 0 to 4096 (for 12-bit cameras) or to 16384. For CCD cameras, most are now 16-bit (0 to 65536). What this means when we look at a picture is that "higher bit" cameras are more refined in the allocation of intensity levels, and may appear to be intensity-smooth to the person viewing it. But note that some software (like Nebulosity, etc) automatically pads or scale lower-bit camera production to 16-bit, so a distinction has to be made that even if we see a 16-bit image from a Canon 350XT, which is actually a 12-bit camera. Now, depending on what you want to take, sometimes the higher the number of bits does not mean the better. When it comes to imaging Jupiter, for example, the dynamic range is quite limited as compared to capturing Orion Nebula with all its faint tendrils to the bright core highlights.

2) ISO and the equivalent gain in CCD
Kevin was asking about the equivalence of ISO in the CCD astro camera world. Basically, this translate to the gain in the CCD. Low ISO is akin to using low gain in the CCD camera, and vice versa. Low ISO (low gain) has an effect of widening the dynamic range since one exposes longer to reach a certain saturation level (say white point). This in effect also gives lower noise and errors resulting from quantization when converting to intensity level of an image. High ISO (high gain) has greater detectivity, but higher noise (higher quantization errors). Nevertheless, regardless whether one chooses DSLR or CCD, it is good to be flexible in ISO/gain settings for various imaging targets.

3) Option for REGULATED Cooling
Now, we have been talking about the obvious choice going to getting a photon detector with cooling facilities, but the fact is, even if it is a cooled camera, you will not get consistent image quality if the cooling is not REGULATED. Why is this the case? Any detector that we use, requires us to obtain a set of dark frames for dark frame subtraction. But a dark frame is only useful if it is taken at a predefined sensor temperature (in this case, at the same temperature as when a true image is taken). Note that not all CCD cameras offer regulated cooling.

4) Noise attributed to detector circuitry
Different DSLRs, and different CCD makes, even if they are having the same sensor chip, can exhibit different image profile character. Read noise for one comes to mind. Basically, when the number of photons is extracted from a photo-site, it is not a perfectly clean process, and some errors in photon count is expected. This manifests as read noise (the processing from reading out from these photo-sites (or wells). For example: The ST8300M and QSI583 may be using the same chip, but the latter has a better circuitry (with cleaner or lesser noise profile). This needs to be kept in mind especially for those going for very short exposures since these will have more read-noise dominant.

5) Monochrome and OSC
Thought I will include this point to clear up any remaining misconception about monos vs OSCs and how pixel information is extracted from the detector. Take for example a monochrome CCD camera vs an OSC version - ST8300M vs ST8300C. If you look at SBIG's website for these cameras' specifications, you may be surprised to see that either camera has the same resolution dimension of 3326x2504. Actually both are using the same sensor, but the only difference is the OSC version has an additional filter array placed on top of the sensor. As Orlando has pointed out and explained well, the Bayer Matrix facilitates the derivation of colour information from any image target. Whether mono or OSC chip, the same number of photons are received per site. The only difference is that in order to generate a colour image, the OSC version will have to do an additional job to "guess" what the intermediate colour information (or photon intensity level) per site, since photons are collected in 2x2 sites (2 green, 1 red, 1 blue). This method of guessing is what we come to know as "debayering". In the early days of OSC cameras, the guesswork is usually just a simple average interpolated value between two pixels. This was extremely inefficient since it did not take into account the fact that image data can have "edges" or places where there are strong intensity gradients (in our layman case - star edges, etc). Nowadays, there are better debayering algorithms out there (like VNG, etc), so much so that the effect of an image taken by a monochrome (through RGB process) and an OSC is getting competently close. Nevertheless, a monochrome gives one the added flexibility in deriving how the eventual colour profile an image can have. If one is shooting Pleiades for example, having a monochrome camera allows you to focus on taking more exposures via the blue filter, but this attribute is absent in an OSC. However, if you stay in an area where clear sky time is limited (sounds familiar), an OSC is more productive - you get colour images faster than a monochrome (three times longer in terms of RGB or 4 times for LRGB, etc).

Now, having said all the above, hopefully it gives everyone a clearer understanding of what is out there in our market, but more importantly, before anyone decides what kind of detector to get, one needs to know some of the points discussed above. Of course, budget is also a primary element of consideration. But do keep in mind, that even dedicated monochrome CCD cameras like the ST8300M is considered "cheap" in terms of astro CCD standards. Imagine how much the ST8 used to cost years back! Also, you may not really need to get a filter wheel (though it is convenient), cause in terms of practicality (for our kind of skies), how often can one image the full suite of filters in one night, knowing that one needs to take a couple of exposures per filter! And with the current low USD, you can get the ST8300M for about SGD$2.6K shipped. Filters like Baader would be useful, probably start of with LRGB or LHaOIIISII, not necessarily getting the full suite.

Hope this helps!

[cataclysm]

How do you keep the chilled sensor from fogging though?

Fogging: you have to keep everything sealed up so that the sensor doesn't fog.

I overcome my fogging issue on the chilled camers by recirculating dry air at the front of the sensor. I fashioned it from a portable aquarium pump, only needs 1.5V to power it, drawing very little power. It will work even on an open system, the gist is to maintain a curtain of dry air infront of the sensor. Yes this entails you need to run one more device on the imaging setup, but it has proven to work well even under the most testing condition that I had once experienced at Punggai.
Here is the pic:

[cloud_cover]

Thanks guys for all the advice! Its really helped me understand more about CCDs and the whys or why not to adopt one.
I guess for me until I actually get a permanent setup and can afford to "power up and image", I don't think Monochromes are for me. With 2 kids, imaging time will be severely limited!
I think parts for this thread should be put into the wiki as info on CCDs and DSLRs! Thanks to everyone again!