Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

What's the verdict with ICX-694 cameras?

This topic has been archived. This means that you cannot reply to this topic.
53 replies to this topic

#26 hytham

hytham

    Surveyor 1

  • *****
  • Posts: 1990
  • Joined: 25 Dec 2012

Posted 20 May 2014 - 07:35 PM

I would get the KAF 8300 just to see Pavel/Konihlav blow a fuse.

#27 blueman

blueman

    Photon Catcher

  • *****
  • Posts: 6925
  • Joined: 20 Jul 2007

Posted 20 May 2014 - 07:45 PM

Well, I for one am not unhappy with the lowly KAF8300 chip. Noisy, well I don't know, it is not bad.
Sony is fine, I like the chips, but are they really superior? I am not convinced.
Blueman

#28 blueman

blueman

    Photon Catcher

  • *****
  • Posts: 6925
  • Joined: 20 Jul 2007

Posted 20 May 2014 - 07:53 PM

blueman, JJK: I understand that you don't like to hear what I wrote (AFAIK I just wrote KAF-8300 is very noisy which is just simply true, not much to argue about) and you do not have to support your 8300 based camera by providing some of your images (there are billion of them on thee Internet). I really believe it works well and very well for you and maybe it's your only camera. After you use and test many more cameras you learn more. You both did not pay attention that I said, the case where the noise matters is narrow band imaging and lucky imaging only. In LRGB space, the 8300 outperforms (due to bigger FOV) ICX-694, especially under a heavy polluted skies. What I wanted to point out is how BIG the difference in total exposition time is in this narrow band imaging case comparing 5e- vs 10e- RN cameras. That is something 99.99% do not have clue about and do not realize.

just for your curiosity. There's one CCD chip that is worst ever on planet Earth, real mess. And I have, myself, intentionally purchased a camera equipped with such a mess. Why? chip size 36x24mm and price about 5000 USD (very cheap camera). Since I use it with F/3.7 system and ONLY for LRGB imaging, I really win every time I shoot with it under dark sky with LRGB filters. Narrow band with it makes no sense, it's totally blind.



Phil: Peter in Reno promised to send some biases and flats to me so as I can measure his new CCD and tell if the newest QSI cameras are truly perfect (so far I had one from them and it was very, very good, even a hair better than Atik).

If you want to make a big leap forward in equipment improvement and your image quality improvement (apart from post processing skills) I would recommend 36x24mm CCD camera (or even the 36x36, but that gets very expensive with filters). Bad fact is that such a leap forward calls for an expensive astrograph :-(

nowadays I have a system that is 10x more effective in LRGB imaging then my old little Borg with ICX-694. The answer for this was the F/3.7 10" aperture system with super noisy KAI-11002 that requires lot of cooling and precise, very precise callibration...


Maybe your idea of how much that noise effects an image is a bit overblown. Most of us cannot see small increases in readout noise or dark noise in an image.
But hey, if you really like the Sony chips, great! I think everyone should choose for themselves.
Blueman

#29 JJK

JJK

    Gemini

  • -----
  • Posts: 3465
  • Joined: 27 Apr 2008

Posted 20 May 2014 - 08:52 PM

blueman, JJK: I understand that you don't like to hear what I wrote (AFAIK I just wrote KAF-8300 is very noisy which is just simply true, not much to argue about) and you do not have to support your 8300 based camera by providing some of your images (there are billion of them on thee Internet). I really believe it works well and very well for you and maybe it's your only camera. After you use and test many more cameras you learn more. You both did not pay attention that I said, the case where the noise matters is narrow band imaging and lucky imaging only. In LRGB space, the 8300 outperforms (due to bigger FOV) ICX-694, especially under a heavy polluted skies. What I wanted to point out is how BIG the difference in total exposition time is in this narrow band imaging case comparing 5e- vs 10e- RN cameras. That is something 99.99% do not have clue about and do not realize.

just for your curiosity. There's one CCD chip that is worst ever on planet Earth, real mess. And I have, myself, intentionally purchased a camera equipped with such a mess. Why? chip size 36x24mm and price about 5000 USD (very cheap camera). Since I use it with F/3.7 system and ONLY for LRGB imaging, I really win every time I shoot with it under dark sky with LRGB filters. Narrow band with it makes no sense, it's totally blind.



Phil: Peter in Reno promised to send some biases and flats to me so as I can measure his new CCD and tell if the newest QSI cameras are truly perfect (so far I had one from them and it was very, very good, even a hair better than Atik).

If you want to make a big leap forward in equipment improvement and your image quality improvement (apart from post processing skills) I would recommend 36x24mm CCD camera (or even the 36x36, but that gets very expensive with filters). Bad fact is that such a leap forward calls for an expensive astrograph :-(

nowadays I have a system that is 10x more effective in LRGB imaging then my old little Borg with ICX-694. The answer for this was the F/3.7 10" aperture system with super noisy KAI-11002 that requires lot of cooling and precise, very precise callibration...


Pavel, I know my way around a CCD chip/camera spec sheet, and know how to do noise calculations. While the S/N ratio is paramount in measurement, there are other factors that can and do influence camera purchases.

Some folks just want to get their toes wet in the imaging game, and frankly, any decent used camera will give them a great learning experience (assuming they're willing to put focused study and time in).

Others (like me) have constraints on their time and would prefer to have their equipment spend its time gathering data with a reasonably wide FOV than spend their own time stitching porthole views together in a mosaic.

I have no skin in this game when it comes to a particular computer, telescope, camera, etc. At the moment, I prefer cameras that provide a wide FOV with the scopes that I currently have.

BTW, to change the subject, the stars in your Pac Man nebula image are not round.

#30 Konihlav

Konihlav

    Surveyor 1

  • -----
  • Posts: 1653
  • Joined: 05 Mar 2009

Posted 21 May 2014 - 02:37 AM

Phil: my advice on equipment would be to replace (if you do not already have) the tube of your newtonian for a thick (7mm) carbon tube (rigid - holds collimation, temperature stable - holds focus position all night long) then upgrade the focuser for something really great (I have ASA motorized 3" version) and then put an corrector/reducer of high quality based on image circle of your newly selected CCD camera. If it is tiny chip (KAF-8300, SONY ICX-xxx) you may be OK with saving much money on 2.5" focuser instead of 3" focuser needed for 3" Wynne Corrector that is suitable for 36x24mm camera (which I highly ! recommend). Also, I highly recommend to stay at about F/4 with your final system, faster becomes nightmare due to strict tolerances in squareness of your imaging train...

#31 Konihlav

Konihlav

    Surveyor 1

  • -----
  • Posts: 1653
  • Joined: 05 Mar 2009

Posted 21 May 2014 - 03:20 AM

Jon Rista: you got my respect from your last post. Nicely summed up. The only thing that I stand (still) behind is that FOR MY CASE (narrow band imaging - notice none of you guys care about what I wrote - for NARROW BAND ONLY not for general purposes) the 5e- (ICX 694 cameras from QSI 4e- QHY 4.5e- ATIK 5e- SX 7e-) camera versus 10e- (KAF 8300 cameras from SBIG, Moravian 7.5e- over Apogee, QSI 9e- to FLI 10e- to SX/QHY 12e-) the difference in total integration time is ULTRA BIG like 6x. If you make 30min subs with KAF instead of 10min subs (what I do with ICX) then the result would be like a factor of 5x instead of 6x (it obviously helps to make longer subs with higher RN camera). The other thing - the RN is propert of the chip and it's in about +-0.5e- range (from my findings) and ONLY the manufacturer's electronics can mess things up and make final product with higher noise. That's why I made that research years ago, to figure out a camera maker who makes nice and clean electronics. At the moment in EUropa it is Moravian Instruments and Atik and in USA it is (still was SBIG) and now, QSI seems to finally (after years of KAF 8300 experience) approach the best ones in the same ball park...

but as you know, all of the makers put some extra features to their cameras to make them attractive for the buyer. Nothing wrong with that :)

anyway, thank you for your nice post that summed up the reality.

BTW you also wrote:
I also believe that the problem can be mitigated with more total integration time.

and that is obviously the way to go. And as I pointed out, it can be mitigated by noise reduction that everyone uses (majority overuse it to aggressively smooth background).

#32 Konihlav

Konihlav

    Surveyor 1

  • -----
  • Posts: 1653
  • Joined: 05 Mar 2009

Posted 21 May 2014 - 03:21 AM

Knez (orion69) wrote:
From what I see maybe you saved me some big chunk of money.

- As you see I easily made only 10min subs, but you did 30min subs... wow, that's challenging.
- I have presented the image in 1:1 full scale, one pixel is that 4.54um but you have presented the image in 50% therefore one pixel is virtually 10.8um and that is the reason why at the same total integration time (14 hrs) you see more or less equally looking image

one key factor is the final image scale you show present your image at. I, with small pixels, did not have much choice to downsize (for me, every pixel counts and that's why I am "mad" or crazy on camera properties). When you downsize you get super pixel scale and boost your "gain" by say factor of 4x (4x more light = twice as better SNR twice less noisy image).

Least factor is the noise reduction, I have also applied some soft one :)

so the answer to the same looking image is, from my perspective, your downsized final image resolution.

#33 Konihlav

Konihlav

    Surveyor 1

  • -----
  • Posts: 1653
  • Joined: 05 Mar 2009

Posted 21 May 2014 - 03:21 AM

hytham:

your post is not worth a reply, but anyway. If I am given two small chip cameras, two "tiny" :-) chips (8300 and ICX694) then I would get the clean one, that's obvious, especially if I ever care on shooting Ha, SII and OIII channels where the clean one "rulez". But I can understand that 8300 is good compromise for all the rest. I also sometimes have to recommend 8300 cameras to beginners because it all depends on the context.

If you want to make a big leap forward then forget these small 8300 or tiny ICX694 chips and get a REAL camera, I mean a full frame, full size (size matters) 36x24 or 36x36mm, unfortunately Kodak one. I do have it, it's the worst chip, but since it is VERY BIG surface, I truly win (in LRGB imaging, narrow band is not worth a try ever except of Ha and very bright emission objects, otherwise it is blind).

May budget be a problem? like in my case, look how much cheaper is the **** noisy KAI-11002ME camera, mine with filter wheel (integrated 5 pos only since no need for SII and OIII) was about 5-6000 USD. This interline can still produce fine images in LRGB space due to 9um pixel size and 4032pixels wide. If you downsize to 50% you make a superpixel (virtually 18um) and reduce the noise significantly. That was my point.

With 36x36mm there is problem that it's twice as expensive then old STL-11K and you need twice more expensive filters (Astrodon's obviously) and I do not like the rectangular format, I prefer 3:2 aspect ratio. Also there's problem with OAG and quality image circle. You need top notch expensive astrograph to "feed" the camera with photons :)

#34 Konihlav

Konihlav

    Surveyor 1

  • -----
  • Posts: 1653
  • Joined: 05 Mar 2009

Posted 21 May 2014 - 03:31 AM

blueman: finally we agree :) it's everyone's choice, our needs differ, even with totally **** noisy camera you can make a superior image (like some of my images done with worst KAI-11000 chip ever - but done in LRGB space where it DOESNT matter at all, the importance of readout noise, but comes to play the pixel size, that parameter is most important then, pixel size).
You are also right that I may a bit "overblow" but really a little. In narrow band, with H-a filter, full Moon up in the sky, you still can make decent images. But when the Moon is out (new moon) then the boost in H-a images quality is also notable :) my friends laugh me that I shoot Ha also during new moon :) but I know why. On my calculations I "gain" by 50% :)

so it's everyone taste. We all must understand why we do this and with which equipment we gain in which situation (under which circumstances it matters and when it doesn't).

ad my comment on subexp duration:

it is obvious that readout noise affects the noise in image after every readout of the image. So if you make more shorter subs you have more final noise then if you make less but longer subs provided total integration time is equal. That's something everyone understands and believes to. Now, when we talk about the special case of narrow band imaging when the background is limiting to zero only the camera specification matters. And here is the "miracle" the fact that 5e- RN camera is 4 times better then 10e- RN camera and even if you would like to mitigate the high readout noise with that 10e- case by making much longer single subexposition (going up to 120 minutes subs, as I said) it won't be the way to go. Not that only you may have hard time to make your mount tracking for two hours a single sub, but as you yourself pointed out (good point) the full well capacity would not be big enough with 8300.

not much more to add to this I guess.

I may repeat that for general purposes, imaging at lousy skies, imaging in LRGB space, you may not ever care about the readout noise. That's the reason I purchased TWO cameras instead of one. - there's no telescope good for visual and imaging at the same time - you need to have two or more. The same with CCDs.

8300 is compromise. It's cheap. It's good for camera makers as the chip itself is very cheap (lot less then overpriced SONY). Sure it is PITY that SONY doesn't make real, big sensors :(



#35 Konihlav

Konihlav

    Surveyor 1

  • -----
  • Posts: 1653
  • Joined: 05 Mar 2009

Posted 21 May 2014 - 03:42 AM

JJK: finally I can again say I completely agree with what you wrote. Nothing wrong there. It is our hobby and we want to make some imaging. It's better to make some imaging then no imaging at all.

my personal approach is that for capturing large nebulas and portions of sky I prefer to have as big CCD chip as possible (with as big pixels as possible) in my camera and for capturing details on limiting image scale (what seeing permits) I prefer to have technologically superb SONY clean camera with reasonably small pixels (not too small) since for e.g. imaging of planetary nebulas or galaxies I do not need big FOV. Therefore there are these two basic "use-cases" every done better with proper equipment.

8300 is, from Kodak family, probably "best" in the compromise terms and reason why so many people own it. I am different in my "chase" for best equipment for the particular "use-case".

HOWGH :)

clear skies to all of you, need to leave NOW.

#36 vdb

vdb

    Apollo

  • -----
  • Posts: 1457
  • Joined: 08 Dec 2009

Posted 21 May 2014 - 03:50 AM

I didn't read all the posts in detail, my 2 cents are it all depends on how uniform your bias is, I have had/have 694, 814 and 8300 ...
SX's for the sony and a QSI and ATIK for the 8300, what it all boils down to is stability ... and for me that is, believe it or not the ATIK 383L+ ...
Easiest to callibrate ... (my QSI is not the best in class ... I do realise it should be the top performer)

I know SX don't offer the lowest read noise for the ICX chips, so maybe my experience is not representative of what the true potential is of the Sony chip ...

So in the end, a bit of luck with the sample you get is what counts ;-)

#37 orion69

orion69

    Viking 1

  • -----
  • Posts: 998
  • Joined: 09 May 2010

Posted 21 May 2014 - 05:58 AM

Knez (orion69) wrote:
From what I see maybe you saved me some big chunk of money.

- As you see I easily made only 10min subs, but you did 30min subs... wow, that's challenging.
- I have presented the image in 1:1 full scale, one pixel is that 4.54um but you have presented the image in 50% therefore one pixel is virtually 10.8um and that is the reason why at the same total integration time (14 hrs) you see more or less equally looking image

one key factor is the final image scale you show present your image at. I, with small pixels, did not have much choice to downsize (for me, every pixel counts and that's why I am "mad" or crazy on camera properties). When you downsize you get super pixel scale and boost your "gain" by say factor of 4x (4x more light = twice as better SNR twice less noisy image).

Least factor is the noise reduction, I have also applied some soft one :)

so the answer to the same looking image is, from my perspective, your downsized final image resolution.


That is exactly why I was looking at ICX814 chip. It's possible to resample from full resolution to 2560 pixels.
It's really frustrating not to be able to predict improvement to 8300 chip.

It's too much money to spend without knowing result.

#38 alpal

alpal

    Soyuz

  • -----
  • Posts: 3998
  • Joined: 15 Jun 2009

Posted 21 May 2014 - 06:44 AM

The problem is simple:

You need both cameras if you're serious -
an 8300 for wide fields & a 694 for faint small galaxies or other small targets
where large size doesn't matter.

#39 JJK

JJK

    Gemini

  • -----
  • Posts: 3465
  • Joined: 27 Apr 2008

Posted 21 May 2014 - 07:17 AM

Jon Rista: you got my respect from your last post. Nicely summed up. The only thing that I stand (still) behind is that FOR MY CASE (narrow band imaging - notice none of you guys care about what I wrote - for NARROW BAND ONLY not for general purposes) the 5e- (ICX 694 cameras from QSI 4e- QHY 4.5e- ATIK 5e- SX 7e-) camera versus 10e- (KAF 8300 cameras from SBIG, Moravian 7.5e- over Apogee, QSI 9e- to FLI 10e- to SX/QHY 12e-) the difference in total integration time is ULTRA BIG like 6x. If you make 30min subs with KAF instead of 10min subs (what I do with ICX) then the result would be like a factor of 5x instead of 6x (it obviously helps to make longer subs with higher RN camera). The other thing - the RN is propert of the chip and it's in about +-0.5e- range (from my findings) and ONLY the manufacturer's electronics can mess things up and make final product with higher noise. That's why I made that research years ago, to figure out a camera maker who makes nice and clean electronics. At the moment in EUropa it is Moravian Instruments and Atik and in USA it is (still was SBIG) and now, QSI seems to finally (after years of KAF 8300 experience) approach the best ones in the same ball park...

but as you know, all of the makers put some extra features to their cameras to make them attractive for the buyer. Nothing wrong with that :)

anyway, thank you for your nice post that summed up the reality.

BTW you also wrote:
I also believe that the problem can be mitigated with more total integration time.

and that is obviously the way to go. And as I pointed out, it can be mitigated by noise reduction that everyone uses (majority overuse it to aggressively smooth background).


No one here would argue that a higher read noise camera is better than a lower RN camera when the signal is weak (e.g., narrow band). However, much of this difference can be made up with longer integrations.

If one already has a telescope, then the choice of a particular camera is not just based on noise. The FOV required to frame the subject can be just as, or perhaps even more important to some, particularly if their mount and support equipment allow long sub exposure times (as mine does).

If I were completely new to imaging and had no equipment, I'd look differently at the scopes and cameras I'd purchase. However, a lot of us cross over to the dark side (imaging) after being visual observers for a long time and have several (or more) scopes that might not be ideal for imaging. If you aren't sure that imaging isn't just a passing fad, it makes no sense to sell all your equipment and buy new stuff. You work with what you have, in the way you see fit, and make the best of it.

I currently use an AP Traveler for imaging as my short focal length scope. Perhaps the Officina Stellare Veloce RH200 200 mm f/3 Riccardi-Honders is a better choice (same focal length, twice the aperture), with the understanding that at f/3, you really have to pay attention to getting the camera dead-on square with the optical path. However, I bought the Traveler to travel with (fancy that). It went to a lot of places, was a fun visual instrument, and has done well in its conversion to a very capable astrograph. Will I get an RH200? It's likely. However, that scope is capable of illuminating a large image circle. It seems like a waste to not use it by adding a small porthole FOV camera.

#40 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23551
  • Joined: 10 Jan 2014

Posted 21 May 2014 - 12:32 PM

Jon Rista: you got my respect from your last post. Nicely summed up. The only thing that I stand (still) behind is that FOR MY CASE (narrow band imaging - notice none of you guys care about what I wrote - for NARROW BAND ONLY not for general purposes)


Sure, I think it's great that you've found a way to minimize your total exposure and integration time. Looking at your work, you have an incredibly expensive collection of equipment, including some ultra-fast 10" ASA telescopes and a wide range of high end, large format, large pixel cameras on top of the smaller format Sony cameras. Not everyone has the option of *absolutely zero* compromise...you have all the best equipment, for every possible situation, so you have the option of selecting the ideal bits and pieces to perfectly suit whatever it is you point your scope and camera at. That's an exceptionally rare position to be in.

I do believe everyone read that you stated low read noise is important for narrow band. I think it's just that they don't think it is quite as important as you are pushing. ;)

the 5e- (ICX 694 cameras from QSI 4e- QHY 4.5e- ATIK 5e- SX 7e-) camera versus 10e- (KAF 8300 cameras from SBIG, Moravian 7.5e- over Apogee, QSI 9e- to FLI 10e- to SX/QHY 12e-) the difference in total integration time is ULTRA BIG like 6x. If you make 30min subs with KAF instead of 10min subs (what I do with ICX) then the result would be like a factor of 5x instead of 6x (it obviously helps to make longer subs with higher RN camera). The other thing - the RN is propert of the chip and it's in about +-0.5e- range (from my findings) and ONLY the manufacturer's electronics can mess things up and make final product with higher noise. That's why I made that research years ago, to figure out a camera maker who makes nice and clean electronics. At the moment in EUropa it is Moravian Instruments and Atik and in USA it is (still was SBIG) and now, QSI seems to finally (after years of KAF 8300 experience) approach the best ones in the same ball park...


There are ways of mitigating noise that don't require excessive integration (which, BTW, does not blur the image...it's a constructive form of averaging pixels vertically rather than a destructive form of averaging spatially...if you average enough subs, you complete your images, that does not cause blurring.) Advanced multiscale wavelet and median noise reduction tools give you maximum control over where denoising algorithms are applied, which makes it a lot easier to reduce noise without unduly affecting important detail.

I'd also offer that the difference between 10m exposures and 30m exposures is 3x, not 5x. If you take the same number of subs, and expose for 30m, then you have increased your total integration time by 3x. If you increase your exposure time AND take more subs, then you would have more than 3x total integration time. Again, if the background sky is truly zero, then the only thing that more exposure time per sub is going to do is change the nature of everything except the background sky. In other words, since the ICX has ~20ke FWC and the KAF has ~25ke FWC, a 30m exposure will result in greater contrast and lower noise overall in every other respect except the background than the 10m exposures with the ICX, since otherwise both cameras are similar in terms of well capacity and overall sensitivity.

Remember, while the ICX has 37.5% more Q.E., it's pixels are still 41.5% smaller, which still leaves the KAF with a 4% overall sensitivity edge. The only key differences for NB imaging would be dark current and read noise...since RN is only ~7-8e- in the KAF, if you expose for too much longer than with the ICX, dark current will rapidly become your most significant source of sensor noise. That means you really can't expose for significantly longer, 30m is probably pushing it, 120m would certainly result in unmanageable dark current noise (at 0.02e-, after 7200 seconds, were talking about 144e- worth of dark current noise...that swamps read noise completely!)...you end up with a significantly different image...more deeply exposed in the nebula parts, possibly even clipped in significant regions of non-background sky for a 120m exposure, but increasingly noisy in the background sky as dark current builds up. That leaves shorter exposures but more of them for more total integration the only real option for reducing background sky noise when read noise is higher.

I still believe the important factor here is S/N, not absolute read noise levels. At ~4.5e- RMS RN and say a 50% 10ke- exposure out of 20ke- FWC, your S/N is 20log(10000/4.5), or 66.9dB. At ~7.5e- RMS RN and a 50% 12.5ke- exposure out of 25.5ke-, your S/N is 20log(12500/7.5), or 64.4dB. In terms of stops, that is only ONE stop. That would mean that to get the same overall exposure/DR with the KAF, you only need to expose for 1200s (20m). Even at this level of exposure, your RN is -7-8e-, but your dark current is 24e-. The best solution for noise reduction, again, is going to be integrating more subs...so instead of 10x10m, you might take 20-30x20m. And, ironically, the best trait of the Sony ICX sensor is not the RN levels...it's the dark current levels. If you needed a 30m exposure, the ICX sensor is really the best option for that, since with it's 0.003e-/s/px dark current, you only end up with 5.4e- dark current noise after 1800s.

#41 blueman

blueman

    Photon Catcher

  • *****
  • Posts: 6925
  • Joined: 20 Jul 2007

Posted 21 May 2014 - 01:25 PM

I was looking at your gallery Konihla and all of the photos are quite small so it is impossible to see any details in the photos. When you click an individual photo you just go to another set of photos. Do you have any at full resolution so they can be analyzed?
The one photo you linked I could not help but notice how the stars were not round at all. Are you unable to track properly for these long subs?
Blueman

#42 hytham

hytham

    Surveyor 1

  • *****
  • Posts: 1990
  • Joined: 25 Dec 2012

Posted 21 May 2014 - 01:44 PM

hytham:

your post is not worth a reply, but anyway.


My post was tongue in cheek with regards to your steadfastness at very specific objective characteristics of a camera's properties, but completely neglecting the success in which it has been deployed to produce very fine images (post noise reduction aside) by the folks here on this board and beyond.

If Sony had developed a 36x36 chip with similar properties to their ICX 694, I would have not even entertained the KAF-16803 I have today ... provided it did not come at a significant price increase, of course. :)

#43 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23551
  • Joined: 10 Jan 2014

Posted 21 May 2014 - 02:08 PM

If Sony had developed a 36x36 chip with similar properties to their ICX 694, I would have not even entertained the KAF-16803 I have today ... provided it did not come at a significant price increase, of course. :)


Agreed, a large format ccd by Sony would be pretty awesome. I suspect, though, that it probably would come with a hefty price premium for a while at least.

#44 Robert York

Robert York

    Messenger

  • -----
  • Posts: 462
  • Joined: 05 May 2013

Posted 21 May 2014 - 07:38 PM

Just for a data point, here's 5 frames of 300sec, with only flat subtraction, no other processing done. This was taken with the ICX-694C (one shot color), so not as good resolution as the mono should have.

http://www.astrobin.com/37539/

Again, no noise reduction, no dark removals, I didn't even do hot pixel removal. Pretty impressive for just 5 frames averaged, right out of the camera. No idea what a KAF-8300 looks like in similar fashion, but it would be an interesting comparison.

Edit: Also, this shows the slight blooming I had, which has been fixed with a little adjustment to the camera. Also, this was taken from a red zone, with just an IR/UV block filter.

#45 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23551
  • Joined: 10 Jan 2014

Posted 21 May 2014 - 10:57 PM

@Robert: That's a pretty bright target. Might not be the best, given it's dynamic range (the core is pretty heavily blown out in your shot.) I think the concept is sound though...maybe another target? Rosette? Monkey head?

#46 JJK

JJK

    Gemini

  • -----
  • Posts: 3465
  • Joined: 27 Apr 2008

Posted 21 May 2014 - 11:26 PM

Just for a data point, here's 5 frames of 300sec, with only flat subtraction, no other processing done. This was taken with the ICX-694C (one shot color), so not as good resolution as the mono should have.

http://www.astrobin.com/37539/

Again, no noise reduction, no dark removals, I didn't even do hot pixel removal. Pretty impressive for just 5 frames averaged, right out of the camera. No idea what a KAF-8300 looks like in similar fashion, but it would be an interesting comparison.

Edit: Also, this shows the slight blooming I had, which has been fixed with a little adjustment to the camera. Also, this was taken from a red zone, with just an IR/UV block filter.


Robert, what specifically do you find impressive about that image?

#47 Robert York

Robert York

    Messenger

  • -----
  • Posts: 462
  • Joined: 05 May 2013

Posted 22 May 2014 - 12:25 AM

Maybe I'm just a bit new to imaging, but I was surprised how little noise and how much dim gas structure came out. I figured it'd take dozens of subframes to see anything pretty. It was the first shot I took with my ICX-694, and I was impressed at what I could do and the dim structures I could see with so little processing.

I've done some more since then, but I've been struggling with guiding issues, and seeing/weather/light pollution so haven't put much else together. I can dig through my archives and see if I can find anything else.

But again, maybe my excitement is just my own lack of experience. And yes, not the best target, but for me, I'd shot it before with other telescopes and other cameras. I wanted to see what this could show me that I hadn't been able to capture before. This is what I was used to getting with my SLR and newtonian:

http://i.imgur.com/x3mDn.jpg

And pardon the colors, I was having fun with it. So to me, it was a whole new world of imaging. Maybe it's not the fact that it was a 694, so much as just having proper CCD. I'm not sure. I could probably go back and do something similar with an SLR at this point, if Orion were still up.

#48 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 23551
  • Joined: 10 Jan 2014

Posted 22 May 2014 - 12:57 AM

@Robert: I think it's an interesting test, but the Orion nebula is so incredibly bright. These are my very first two astrophotography images that I ever took (with an equatorial mount):

Posted Image

Posted Image

These were taken with a Canon 7D. I was pretty amazed with the amount of nebulosity I was able to capture in IC434, and was also amazed that I was able to bring out some of the dark dust lane detail between Running Man and Orion nebulas. And these are with an APS-C DSLR sensor...and the 7D no less, which is well known as being a particularly noisy DSLR with some pretty nasty fixed patterns and such (and certainly a MAJOR chore to process out...I have to spend an exorbitant amount of time working noise to bring it down to a reasonable level, and that usually kills a lot of the fine structure detail in the process.)

Personally, I'd expect just about any cooled CCD to do an order of magnitude better than my 7D, and be able to lift out a lot more of the dark/Ha nebulosity surrounding Orion nebula. I'm not saying the 694 isn't amazing, but the Orion nebula is an easier target.

#49 Konihlav

Konihlav

    Surveyor 1

  • -----
  • Posts: 1653
  • Joined: 05 Mar 2009

Posted 22 May 2014 - 02:56 AM

Blueman: please, what browser and OS do you use? I guess some new Internet Explorer under some Windows? I know my gallery isn't perfect as I do not fully support MSIE due to various reasons (all "products" from Microsoft have different set of bugs and are inconsistent between versions of the same software and for me it's a nightmare to cover all exceptions because MS doesn't care about any standard). Every other browser should work well as every other cares about internet standards... maybe you have just by default disabled popup windows.

anyway, many of my images still do have some kind of problems that is visible in 1:1 (100%) scale. Sometimes it's collimation (I am new to newtonians) or focus or kinda tracking - I have to setup every day from scratch, I am mobile photographer, no observatory, no fixed place, I put all gear to my car and go to dark place...

here are some direct links to images and full scale:
http://www.astro.cz/...rsehead.jpg....
http://www.astro.cz/...0minHaD-W6fi...
http://www.astro.cz/...OR.jpg.html?...
http://astrofotky.cz.../1379338239.jpg
http://www.astro.cz/...g.html?g2_im...
http://www.astro.cz/...gether-W9fin...

and here are some in 50% only:
http://astrofotky.cz.../1396881171.jpg
http://astrofotky.cz.../1393598615.jpg
http://astrofotky.cz.../1381166843.jpg

and one of my best, 5.5hours total (2h Lum + rest for color) in full scale:
http://www.astro.cz/...g.html?g2_im...

this image above demonstrates the 10x improvement in LRGB imaging when using big aperture, fast focal ratio and 36x24mm chip though totally lousy chip.

here I used small 77mm aperture Borg, but with superb clean high QE SONY chip, but as you can see, the chip surface is 8x lower and low readout noise is of no importance for LRGB (if you shoot LRGB 2min subs or 4min subs is not a big deal to mitigate RN) AND the aperture 250mm vs 77mm has another total impact on final SNR so after 5.5 hours with tiny refractor and small surface chip I got only this noisy image:
http://astrofotky.cz.../1351541816.jpg

#50 Konihlav

Konihlav

    Surveyor 1

  • -----
  • Posts: 1653
  • Joined: 05 Mar 2009

Posted 22 May 2014 - 03:10 AM

CONCLUSION:
if I am a beginner on tight budget I would get one of the two introductory camera options with small/tiny chips (KAF 8300 if I care about FOV and general use or ICX 694 if I care about detailed, hi-res work and imaging small targets like galaxies, PNs, Arps etc. OR am about to do best narrow band images. I would personally prefer ICX-694 over KAF-8300 because it requires just 1.25" filters and even without calibration you get nice results.

if I am not beginner any more I would go straight with some big chip surface camera (36x24mm), either the **** noisy KAI-11000 or KAI-16000 or with the 36x36 expensive beast KAF-16803 (not much else to choose from). The big chip camera offers option to downsize to 50% and still end up with decent resolution image to make some printouts. It also very much helps hide the noise even after noise reduction. Big pixels are important as well as big chip surface for high SNR.

it's again a compromise between detail (if seeing permits) captured in low image scale (hi res) images where decent SNR takes longer total integration time and between large nebulas imaging with high image scale (low res) setups where decent SNR is easier to reach, but you pay by the missing detail :)

TIPS:
I no longer use LPS/LPR filters, even the IDAS LPS-P2 (I call it "soft LPS") is now unused in my recent-most setup. Rather a travel to dark sky location helps very much to get better signal...



CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics