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

IMX 455 vs Kaf-16803-bases cameras

  • Please log in to reply
17 replies to this topic

#1 Rasfahan

Rasfahan

    Sputnik

  • -----
  • topic starter
  • Posts: 34
  • Joined: 12 May 2020
  • Loc: Hessen, Germany

Posted 11 August 2020 - 07:20 AM

Hello,

 

I am in the market for a new scope/camera combination, and both of the above seem to be on the short list (the IMX 455 would need to be binned). Does anyone have experience and can compare the two? I currently image with an ASI 1600MM Pro, and find noise reduction rather tedious. Also, the very low FWC means I have to process 400-800 subs per (broadband) target (Pixinsight needs about 2-3 days to process with WBPP), so a larger FWC would come in handy.

 

Thanks for your insights,

Torben



#2 pyrasanth

pyrasanth

    Gemini

  • *****
  • Posts: 3,117
  • Joined: 08 Jan 2016

Posted 11 August 2020 - 07:45 AM

I have the ASI6200 camera which uses the IMX455 sensor. It is a fabulous camera and using demure denoise in PI the noise is almost none existent on any stack with average SNR. I would not go back to CCD after using this camera. I did have a 16803 sensor and the 11002 & I can tell you that the comparison between the two sensor types is unfair- the CMOS really does wipe the floor with both on noise quality.

 

The 16803 is a great sensor however it has met its match when compared with the IMX455. The only advantage I see with the 16803 is the big pixels and field width at 1x1 bin but that is no show stopper. The IMX455 binned 3x3  (which I use a lot) gives fabulous results with a resolution of just over 11 um per pixel at 3x3 bin. The well depth on the IMX455 is tremendous & it only needs cooling to between 0 & -5 to give great low noise results.

 

I know that everybody does not have access to the latest computers but on a 3x3 bin my system will process including cosmetic correction, 600 subs in an hour using WBPP and that includes creating 4 integrated flats at 100 subs each. It might be worth considering an upgrade to your existing hardware.


Edited by pyrasanth, 11 August 2020 - 07:50 AM.

  • psandelle, leviathan, Jon Rista and 1 other like this

#3 Rasfahan

Rasfahan

    Sputnik

  • -----
  • topic starter
  • Posts: 34
  • Joined: 12 May 2020
  • Loc: Hessen, Germany

Posted 11 August 2020 - 09:58 AM

Thanks for sharing your experience, it is good to know you prefer the 6200MM to your former KAF-based camera. 
 

At the moment I use the ASI1600 unbinned, producing 16 MP images (compared to 9 MP for a 3x binned 6200). They are processed on a server with 4x10 cores and 256GB RAM (for this VM) and a PCIe SSD. The server is older, but in the Pixinsight benchmark it performs very well and is only trumped by the latest AMD Ryzen-based machines. The number of subs above was meant per channel (so up to over 3000 subs). Still, there remains a discrepancy. May I ask what system you are using?

 

- Torben


  • segoviamarket likes this

#4 gregbradley

gregbradley

    Vostok 1

  • -----
  • Posts: 140
  • Joined: 23 Jan 2008

Posted 11 August 2020 - 04:37 PM

I am planning to get an ASI6200mm but have been using a FLI Proline 16803 as my main imaging camera now for over a decade.

 

I usually judge new gear by the images I see posted as often people are enthusiastic about the latest gear. To be honest I have not seen any ASI6200 images that are outstanding yet. A couple that were impressive. I saw a good Saggitarius Trio image from a QHY600 that was very short exposure yet it looked like a long exposure image. That showed off how well matched this camera is to an FSQ type setup.

 

FWC of the 16803 is 105,000 electrons with 1x1 binning. It handles very bright stars very well.

 

So keep that in mind. 1x1 binned images have 121mb file sizes. 16803 are 32.8mb. 3.76nm pixels ideally only suit shorter focal lengths, 9 micron pixels do the full gamut from FSQ106 to CDK17.

 

The 16803 is all about its huge field of view. Its 50% bigger than the 6200. That's massive and the huge FOV of the 16803 creates a unique image that seems to really please viewers. 

 

The 16803 does form vertical lines in the sensor over time. My 16803 is about 13 years old possibly a tad older. It had no vertical lines in it for a long time but now it has one brightish line and sometimes a few smaller not bright vertical lines. These sometimes calibrate out and other times they don't. You can set up a defect map or they are easy enough to remove using Photoshop healing tool in the later versions of Photoshop that tool will remove a line by clicking the start and then holding down the shift key and clicking the end of the line it will remove a straight line. PI also has a cosmetic correction App which I believe will remove them if set up correctly.

 

Binning on CMOS is only software binning. You still get an increase in SNR but its not as effective as binning on a CCD. So the QE advantage of the 6200 would be weakened compared to 2x2 binning on the 16803. The smaller pixels would mean oversampling on many scopes also weakening the QE advantage. I see some QE advantage of my 84% QE  ASI183 but not as much as you might expect compared to a KAF16200 or 16803. Its there, its not massive though. 

 

Smaller pixels do have some advantages though. I find I get rounder stars as each star is made up of more pixels and gives a rounder appearance.

 

But they also make it harder to focus. On my 16803 focus was pretty easy and easy to see when you nailed it. With the 183mm its much harder but not super hard but I wanted to make the point its harder. Also much more prone to the seeing.

 

As to noise my 16803 is very clean at -30 to -35C. Sure there is some but noise is not an issue as what little there is calibrates out easily.

 

CMOS sensors do seem prone to varying fixed pattern noise. This looks like little horizontal lines of darkness which vary from sub to sub. It won't bias out because its shifting. Luckily it seems to disappear in processing once stacked. I would use dithering though with these cameras.

 

So bias on these cameras is an odd ball. They don't seem to work.

 

Cooling on the ASI6200 seems quite weak. 0 to -10C seems practical in the sense of not much noise but dark current isn't the only gain from strong cooling. Maybe less important with a CMOS but there are still gains. I would have to do a side by side comparison but with the extensive imaging I have done so far with my ASI183mm, -25C cooling seemed nicer images than -10C. Not 100% sure of that though as other conditions may have been more relevant. The Proline 16803 does -40C often and -30C all year round, perhaps even -35C.

 

Having to shift gain and offset is a bit of a pain. CCDs are simpler there and less complication with having to have a large suite of darks/flats.

Everything has to match. This is worse with QHY as they have currently 3 modes on top of gain and offset plus they are planning to offer 8 modes. Sounds good in one way and a total pain when it comes to processing your files trying to get everything matched up exactly.

 

There are some reports of banding where some say its not a problem and some say it can be.

 

Owners of the 6200 though seem to be very happy with the camera and would recommend it. The 6200 is about 1/3rd the price of 16803.

 

My `6803 is a FLI Proline and its very large and heavy. Probably weighs something like 2-3kgs. It puts stress on the focuser. The 6200 is light in comparison. The 16803 requires 50mm square filters which are expensive the 6200 50mm round which are still expensive but less so.

 

But new 6200 type cameras will continue to come out as Sony makes new sensors for its mirrorless cameras (the 6200 sensor I believe is the same as what is in the Sony A7riv camera). Whereas the 16803 has ceased production so if you want one then now is the time whilst there is some stock left.

 

Having said all the above I am still personally excited about getting a 6200 camera but I would never sell my 16803 for it. 

 

I don't quite get why people are doing super short exposures and stacking hundreds and taking days to process the calibration files. I use 300 and 600 second subs on my ASI183mm at -25C and it works great.

 

By the way lowered dark current is not the only benefit from deeper cooling. 

 

In summary I think it would depend more on your planned setup. If you have a long focal length scope like a CDK12-20 inch or similar type then the 16803 shines in this sort of application. If you are planning to use an FSQ or similar shorter refractor setup then the 6200 should shine there as well but the 16803 still works just fine and will give a 50% larger field of view (that also means an expensive scope to be able to illuminate such a large field - 54mm versus 44mm for the 6200)

 

My 2c worth.

 

Greg.


Edited by gregbradley, 11 August 2020 - 04:41 PM.

  • WebFoot and leviathan like this

#5 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 781
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 11 August 2020 - 04:54 PM

Both Cameras have their purpose.

 

One is about 4k and the other is still 13k.

 

6200 has very low read noise which = short subs

6200 has very low/no amp glow = darks aren't as critical

 

6200 doesn't need darks. You can use Dark Flats, Flats and Lights

 

You can probably cool your 6200 to just "0c" and have nothing discernable to cleanup.   I have the 533 and 2600 varients.

 

Integration time is king.  With very low read noise, you don't take excessively long subs just to overcome your sensor so you hit diminishing returns quick.  You're better off taking more smaller exposures and dithering X number of frames.

 

It's pretty trivial to do a mosaic to double your field of view and with short subs and quality of photos out of the 6200, you could probably do a 2 frame mosaic in a single night and the resolution/fidelity of that mosaic would be mind boggling.

 

I guess if i spent 12k+ on a camera i'd want a decade plus.... so in reality the value of the 6200 is simply phenominal.

 

these new sensors should stand on their own feet btw... comparing modern CMOS with CCD is beginning to be apples/oranges.  You bin on CCD, you don't bin on CMOS, you take long exposures on CCD, you take short exposures on modern CMOS... they're becoming unique products in their own way.



#6 Rasfahan

Rasfahan

    Sputnik

  • -----
  • topic starter
  • Posts: 34
  • Joined: 12 May 2020
  • Loc: Hessen, Germany

Posted 11 August 2020 - 05:10 PM

Thank you for the advice and experience.

 

Concerning taking hundreds of short subs: The ASI1600 has a really small FWC, even in comparison to the ASI183. I blow out lots of stars in 60s Lum exposures with the 1600, so usually use 30s. With my 183 I can do 240-600s depending on subject.



#7 WebFoot

WebFoot

    Surveyor 1

  • *****
  • Posts: 1,812
  • Joined: 02 Jun 2005
  • Loc: Redmond, WA, USA

Posted 11 August 2020 - 07:27 PM

If I had to buy a new camera for my FSQ106 (to replace the STL11000M), I'd seriously look at an IMX-455-based camera, for the small pixels and high sensitivity (possibly doing away with guiding altogether).  But I haven't looked seriously at one.

I absolutely love the FOV of the 16803, and the 9 micron pixels match well with my long focal length scope.

Acknowledging that CCD is going away, I still don't believe that the best CMOS cameras are a match for the best CCD cameras, especially if, like me, you like self-guiding cameras, and love adaptive optics.  There are lots of reasons other than inertia why the most eminent imagers all seem still to use CCD.  If you're very price-sensitive, you may have a different conclusion.

I would not even consider any CMOS camera for my big scope as long as the STX-16803 still is available; but the small pixels of the best CMOS cameras can make an interesting case for themselves, when mated with a very good, short focal length scope.  Especially if I could be convinced that they don't have calibration issues.

 

Mark



#8 gregbradley

gregbradley

    Vostok 1

  • -----
  • Posts: 140
  • Joined: 23 Jan 2008

Posted 11 August 2020 - 10:28 PM

Both Cameras have their purpose.

 

One is about 4k and the other is still 13k.

 

6200 has very low read noise which = short subs

6200 has very low/no amp glow = darks aren't as critical

 

6200 doesn't need darks. You can use Dark Flats, Flats and Lights

 

You can probably cool your 6200 to just "0c" and have nothing discernable to cleanup.   I have the 533 and 2600 varients.

 

Integration time is king.  With very low read noise, you don't take excessively long subs just to overcome your sensor so you hit diminishing returns quick.  You're better off taking more smaller exposures and dithering X number of frames.

 

It's pretty trivial to do a mosaic to double your field of view and with short subs and quality of photos out of the 6200, you could probably do a 2 frame mosaic in a single night and the resolution/fidelity of that mosaic would be mind boggling.

 

I guess if i spent 12k+ on a camera i'd want a decade plus.... so in reality the value of the 6200 is simply phenominal.

 

these new sensors should stand on their own feet btw... comparing modern CMOS with CCD is beginning to be apples/oranges.  You bin on CCD, you don't bin on CMOS, you take long exposures on CCD, you take short exposures on modern CMOS... they're becoming unique products in their own way.

 

Doing mosaics is not a trivial exercise. I have done many. Shifts in the weather being the main issue. So you only get one panel or part of 2 panels done, differing times again determined by the weather mean different gradients in the image. But I see your point about the 6200 potentially being faster to get the panels. Faster on a widefield scope. Not necessarily on a longer focal length scope. As I mentioned the 183mm doesn't seem to have required shorter total exposure times to get an excellent result.

Stacking short subs sounds like more work than doing longer ones that are deeper so long as you don't blow out highlights. 

 

I have seen excellent results from large numbers of short exposures, surprisingly so but I think you would get the same result from longer subs dithered so long as bright spots are not overexposed. The 6200 should be fine with longer subs with the large full well depth especially 2x2 binned. The main advantage of the large numbers of subs is it takes pressure off the mount so a poorer mount can still be used. There is no mathematical advantage to smaller subs = better image. The total exposure time is still the king. I was referring to the earlier posts by guys saying their computer took days to process a big stack. That's silly and would be a deal breaker. My 16803 images are slow enough to process but measured in minutes not days.

 

As an aside I was told by FLI they bought a lot of 16803 sensors and they have nearly sold them all so it may keep pressure on prices for 2nd hand 16803 cameras to maintain higher prices.

 

Greg.



#9 pyrasanth

pyrasanth

    Gemini

  • *****
  • Posts: 3,117
  • Joined: 08 Jan 2016

Posted 12 August 2020 - 06:22 AM

Thanks for sharing your experience, it is good to know you prefer the 6200MM to your former KAF-based camera. 
 

At the moment I use the ASI1600 unbinned, producing 16 MP images (compared to 9 MP for a 3x binned 6200). They are processed on a server with 4x10 cores and 256GB RAM (for this VM) and a PCIe SSD. The server is older, but in the Pixinsight benchmark it performs very well and is only trumped by the latest AMD Ryzen-based machines. The number of subs above was meant per channel (so up to over 3000 subs). Still, there remains a discrepancy. May I ask what system you are using?

 

- Torben

Hi- thanks for your post.

 

I'm using a Ryzen Gen9 3950X over clocked to 4.2 GHZ with 128GB of RAM and Firecuda NVME PCI4 drives in a stripped array which gives me a theoretical burst rate of 8.4GB per second. The speed comes from the fact that I've previously used 3x3 bin files which at under 13 mb each- the system can process nearly 600 per hour.

 

It would slow down on bigger files but I suspect 3000 subs would take a matter of a few hours even at 1x1 bin instead of days. I had a previous intel Xeon E52967 v2 dual CPU rig and that gave 5700 on Cinebench R20- this computer is nearly 9700 so very much faster. It might be useful to test your computer using Cinebench R20 so you can get a meaningful comparison.


Edited by pyrasanth, 12 August 2020 - 06:23 AM.

  • psandelle likes this

#10 ManuelJ

ManuelJ

    Surveyor 1

  • -----
  • Posts: 1,577
  • Joined: 19 Dec 2005
  • Loc: Madrid, Spain

Posted 12 August 2020 - 01:35 PM

I don't know what WBPP does, but if it takes 2-3 days to process a 400 subs stack, it is doing something very wrong.

If you are going to use a 6200 binned (rescaled down, you mean), I see no point because you can buy a 11002 (and non-China brand) with filters second hand for a lower price.
  • rockstarbill likes this

#11 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 781
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 12 August 2020 - 01:50 PM

The 6200 and the KAF 16803 solve imaging goals in different ways.

 

I wouldn't bin a 6200 down and any notion to conceive of doing such is applying CCD knowledge that isn't useful on CMOS which seems to be the entire premise of the KAF vs 6200 here...

 

If you buy a 6200 you don't bin, you take more subs, you hit diminishing returns quicker, you spend more time on integrating more subs but that is a trade off that improves that glorious SNR.

 

As for Mosaics, they're trivial in a single night. If you're not too far in northern hemisphere or too far south and you get astrodark most of the year, getting a 2panel mosaic done with modern CMOS in a single night is doable especially with a 6200 as its a very nice, fast, clean camera. You're spending more time imaging than you are spending time overcoming read noise. Trade time for saving 9 grand if you need to push the subs over 2 nights (or more..)

 

If you want the big KAF, get a KAF :)



#12 rockstarbill

rockstarbill

    Cosmos

  • *****
  • Posts: 7,524
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 12 August 2020 - 02:25 PM

The 6200 and the KAF 16803 solve imaging goals in different ways.

 

I wouldn't bin a 6200 down and any notion to conceive of doing such is applying CCD knowledge that isn't useful on CMOS which seems to be the entire premise of the KAF vs 6200 here...

 

If you buy a 6200 you don't bin, you take more subs, you hit diminishing returns quicker, you spend more time on integrating more subs but that is a trade off that improves that glorious SNR.

 

As for Mosaics, they're trivial in a single night. If you're not too far in northern hemisphere or too far south and you get astrodark most of the year, getting a 2panel mosaic done with modern CMOS in a single night is doable especially with a 6200 as its a very nice, fast, clean camera. You're spending more time imaging than you are spending time overcoming read noise. Trade time for saving 9 grand if you need to push the subs over 2 nights (or more..)

 

If you want the big KAF, get a KAF smile.gif

You do get an advantage of binning the CMOS chip, but the same advantage can be had in post since its just software binning anyhow. The advantage isn't identical to binning a CCD, but you do get a SNR boost from doing so. If you have every intention of just binning in post, you can just do it at the time you are imaging and save yourself some time and hard drive space. ;)

 

I have a 6200 (and an 11002 for that matter) and I would bin the 6200 with specific scopes I own, and in some of those cases it is useful to do so. 

 

I would love for you to share why it isnt. 


  • psandelle likes this

#13 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 781
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 12 August 2020 - 02:57 PM

You do get an advantage of binning the CMOS chip, but the same advantage can be had in post since its just software binning anyhow. The advantage isn't identical to binning a CCD, but you do get a SNR boost from doing so. If you have every intention of just binning in post, you can just do it at the time you are imaging and save yourself some time and hard drive space. wink.gif

 

I have a 6200 (and an 11002 for that matter) and I would bin the 6200 with specific scopes I own, and in some of those cases it is useful to do so. 

 

I would love for you to share why it isnt. 

 

I should be clear. I'm not saying there is 0 reason to bin :) 

 

The mere act of owning a 6200 means I wouldn't bin for shorter sub durations, diminishing returns is already astonishingly fast on the 6200/2600/533 (similar sensors).

 

Binning for image scale, I have no image scale that would be better served with a binned 6200.. but... there is probably someone who does. I guess a huge TEC160 or TEC200 in OK seeing is better served binned 2x2 on paper

 

My statements are less about there is 0 reason to bin on a 6200, mostly about how the KAF is framed as an alternative as if the way the KAF works by design is how the 6200 should work by design. They're different. 

I normally match the native sensor size to scope vs binning it.. but that's just me.



#14 rockstarbill

rockstarbill

    Cosmos

  • *****
  • Posts: 7,524
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 12 August 2020 - 03:24 PM

I think the key to the IMX455 chip lies in its versatility. Examples:

  • 130mm APO, 873mm FL - Bin 1x1 - 0.89"/px
  • 254mm iDK,1720mm FL - Bin 2x2 - 0.9"/px
  • 254mm RC, 2300mm FL - Bin 3x3, 1.01"/px

One camera, 3 different scopes, all relatively close in terms of image scale to one another. If your goal was to get around 0.9-1"/px scale for your skies (like me) the one camera gives you more scope options to go with. Those options obviously have different FOV, but the scale is relatively similar. Examples:

 

FOV_Example.JPG

 

 

 

 

 

 


  • psandelle likes this

#15 sn2006gy

sn2006gy

    Viking 1

  • ****-
  • Posts: 781
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 12 August 2020 - 03:41 PM

So binning in that regard would work.  Those are big spender decisions. Maybe when my kids are out of school ;)

 

If i binned and had those scales - i'd still use the modern techniques of shorter exposures or doing a mosaic to get larger FOV using a 6200 vs choosing a KAF which wouldn't necessarily be binned for the same reasons. 



#16 rockstarbill

rockstarbill

    Cosmos

  • *****
  • Posts: 7,524
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 12 August 2020 - 04:18 PM

So binning in that regard would work. Those are big spender decisions. Maybe when my kids are out of school ;)

If i binned and had those scales - i'd still use the modern techniques of shorter exposures or doing a mosaic to get larger FOV using a 6200 vs choosing a KAF which wouldn't necessarily be binned for the same reasons.


Not sure why shorter exposures is referred to as modern. At any rate, I would expose any camera as long as needed to get good quality frames. At gain 0 and gain 100 with the IMX455 you have great options for exposure lengths and great dynamic range. In my case I don't want a ton of files. So I would shoot my broadband at gain 0 and my narrowband at gain 100. Exposure times in this case are roughly the same as my KAF16200 for those scenarios.

So one can't just rest the debate on CMOS and exposure time entirely. Not for good quality discourse anyhow. That lots of short images in huge stacks paradigm came from the Panasonic sensor in cameras like the 1600. The two cameras in this discussion are leagues beyond that and require more nuance in the discussion.
  • WebFoot, psandelle and xthestreams like this

#17 Rasfahan

Rasfahan

    Sputnik

  • -----
  • topic starter
  • Posts: 34
  • Joined: 12 May 2020
  • Loc: Hessen, Germany

Posted 14 August 2020 - 02:21 PM

Hi- thanks for your post.

 

I'm using a Ryzen Gen9 3950X over clocked to 4.2 GHZ with 128GB of RAM and Firecuda NVME PCI4 drives in a stripped array which gives me a theoretical burst rate of 8.4GB per second. The speed comes from the fact that I've previously used 3x3 bin files which at under 13 mb each- the system can process nearly 600 per hour.

 

It would slow down on bigger files but I suspect 3000 subs would take a matter of a few hours even at 1x1 bin instead of days. I had a previous intel Xeon E52967 v2 dual CPU rig and that gave 5700 on Cinebench R20- this computer is nearly 9700 so very much faster. It might be useful to test your computer using Cinebench R20 so you can get a meaningful comparison.

Sorry for tangenting in my own thread. I cannot run cinebench, because it is a linux machine, but in geekbench, the 8-year-old server scores 75% in multithread of a rig similar to yours. With an optane 905P i/o should be comparable. Single thread performance is only 25%, though (2.2 Ghz, Turbo 2.8). Still, today I opened up the server, and cleaned out the dust. It seems, there is some driver problem with the heat sensor so it did not report overheating. Processing time for the 3000 image stack plus calibration frames went down to a more reasonable 8h. Thanks a lot, that comparison helped to pinpoint a problem. The 6200 still looks quite nice, though, especially with the noise profile it has.


Edited by Rasfahan, 15 August 2020 - 01:32 AM.


#18 xthestreams

xthestreams

    Sputnik

  • -----
  • Posts: 42
  • Joined: 18 Feb 2020
  • Loc: Melbourne, Australia

Posted 21 September 2020 - 01:32 AM

Agree with the sentiment that you need to recalibrate your thinking between CMOS and CCD.

 

The other benefit of binning with CMOS (at least for me) is even shorter exposure times, helpful if conditions demand it (for example if a few of the world's billionaires decide to launch 1 orbital satellite for every man, woman and child and there's not an arc second of sky left that doesn't have some bloody piece of junk flying through it - gives new meaning to the term lucky imaging ;-)

 

Back to the OP, great question. I'm (kinda) torn myself.

 

Owning both a 1600MM  and a 183MC, I am committing to a 16803 for a new site with all the the ridiculous expense in filters, wheels, OAGs and I am sure something else I've forgotten that this entails, but I have the scope, mount and the skies to support that decision. The mission brief for that platform also has other users who have the same sensor in other locations with existing workflows and in some cases data they wish to combine that means this is a good choice (albeit a potentially foolish one given the half-life of a $12000 CCD sensor seems to be measured in femtoseconds and the death of KAF sensors means it's the end of the road).

 

Prior to making the decision on 16803, the CMOS IMX455 platform was and is on my radar, but as I've been reading in another CN thread, it's not without its teething problems too - but it's fair to say that the market will find its way with CMOS (just perhaps don't run out an buy an FLI Kepler just yet - read the other CN thread for that little disaster). CMOS is the way forward, it's just a question of "at what cost?" - Richard Wright's talk on TAIC lays out most of the pros and cons well, if you are planning on doing anything science oriented where issues like RBI, linearity and other matters might be an issue, then you need to ensure the sensor platform support your decision.

 

Point being that as it looks like everyone above has tried to communicate, that comes down to your goals - thanks to the miracles of computer post-processing most sensors either will do what we need them to or can be made to for a pretty wide set of criteria - the trick is knowing what really matters to you and your budget.

 

 

Not sure why shorter exposures is referred to as modern. At any rate, I would expose any camera as long as needed to get good quality frames. At gain 0 and gain 100 with the IMX455 you have great options for exposure lengths and great dynamic range. In my case I don't want a ton of files. So I would shoot my broadband at gain 0 and my narrowband at gain 100. Exposure times in this case are roughly the same as my KAF16200 for those scenarios.

So one can't just rest the debate on CMOS and exposure time entirely. Not for good quality discourse anyhow. That lots of short images in huge stacks paradigm came from the Panasonic sensor in cameras like the 1600. The two cameras in this discussion are leagues beyond that and require more nuance in the discussion.




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