CCD v CMOS

[centroid]

The first astro image in the 8 years since I gave up the hobby, with a somewhat disappointing result

 Just a test run under a Moonlit sky, when I would not have normally bothered with imaging deep sky objects, but after weeks of all sorts of problems, both hardware, and software, I had to give it a go.

 Just 15 x 90 second exposures, combined with 15 x 15 second exposures, and processed in Astroart 8, and Photoshop.

 Way below the quality that I would normally accept, but a new type of camera technology for me. Whereas I used to use CCD based astro cameras, I now have a 294c CMOS based camera.  More sensitive, and relatively cheaper than CCD, but they have some odd 'quirks', that take bit of getting used to dealing with.

 Yes, it could have done with a lot more subs, but even so a pretty lack lustre result, and one that I had to 'push' hard in processing the get that.

 I have included another image, of the same subject, that I took with an SX ccd camera, back in 2012.

 Early days yet, but the temptation to go back to ccd, while they are still available, is growing 😉

Edited January 15, 2022 by centroid

[Budgie1]

Well, to bring a bit of enthusiasm back for you, here's my effort from last year with ASI294MC Pro.

 

[centroid]

That is a pretty good result Martin, especially with just 7 x 90s subs, considering the disappointing result that I got with double that number of 90s subs.

In fairness, it was a bright moonlit sky, but I still would have expected better. M42 is hardly a feint object.

I will give it another go when  the moon is out of the way, and see what difference it makes.

I have an adapter arriving from Germany on Tuesday, that will allow me to attach the WO variable Flatttener/0.8x reducer to the 115 f/7 Apo, which will speed things up a bit from f/7 to f/5.6.

That said, call me a sceptic, but I still think that the CCD was much more 'user friendly. No severe amp glow/starburst, no black level, or gain setting to get right, no green colour cast, and no other strange colour renditions.

I do see some very odd colour images posted, having been taken with CMOS cameras.

Yes, I know they are cheaper, and more sensitive, but imo, compared to CCD, that is as far as it goes.

However, it is early days for me with a CMOS astro camera, that use re-purposed senors, not designed with astro imaging in mind.

I will see how it goes, and I can always go back to a loved, and trusted, Starlight Xpress camera. I note that FLO still has a good stock.

Terry Platt of Starlight Xpress, always provided excellent customer support, with their manufacturing based in the UK, and not China.

Edited January 15, 2022 by centroid
typo

[Whirlwind]

17 hours ago, centroid said:

That said, call me a sceptic, but I still think that the CCD was much more 'user friendly. No severe amp glow/starburst, no black level, or gain setting to get right, no green colour cast, and no other strange colour renditions.

I do see some very odd colour images posted, having been taken with CMOS cameras.

Yes, I know they are cheaper, and more sensitive, but imo, compared to CCD, that is as far as it goes.

However, it is early days for me with a CMOS astro camera, that use re-purposed senors, not designed with astro imaging in mind.

I will see how it goes, and I can always go back to a loved, and trusted, Starlight Xpress camera. I note that FLO still has a good stock.

Terry Platt of Starlight Xpress, always provided excellent customer support, with their manufacturing based in the UK, and not China.

There may be other factors as well.  From my own experience light pollution has got worse (not just brighter but now with greater use of bright white LEDs etc it is less easy to control.  The weather may have also played a part as last week was very misty/foggy which would both scatter more moonlight/light pollution more as well as reducing photons from space getting to us.

I don't think you can see CMOS as an improvement but just a change compared to CCD.  I too find that calibrating CMOS images is more of a chore compared to CCD.  I think this is down to the greater stability in CCDs compared to CMOS.  I've never had a problem calibrating CCD images but do so regularly with CMOS (a 183MM currently but am going to test a QHY268M shortly).  For example take a wide field image I was playing about with last week (1.57 hours total 94 x 60 secs on a stock canon 18-55mm lens using a 6nm halpha filter - please excuse the trees on the left!):-

Despite being calibrated there is still obvious banding remaining (the vertical element abut a third way from the left being particularly obvious, although there is less obvious horizontal banding).  I didn't get flats with this set up (I was really just testing a wide angle arrangement) but am not sure yet whether this effect will be corrected in this way.  The bias/darks have removed a lot (see below of those master frames)

In particular these have removed the amp glow but trying different combinations (with/without bias etc) leaves some banding so is either a flats issue or something that won't calibrate out (meaning either dithering or potentially longer exposures maybe).  Otherwise I have to use processing to remove it which ultimately does impact on the final image quality.

I hence do recognise the frustration compared to CCDs, however amateur CCDs won't be around for much longer in all likelihood so starting to learn the oddities of CMOS is likely going to be needed (though like you I am tempted to pick up one last CCD; although they are likely to left behind in the long term as CMOS should be able to be developed to allow part of an image to be used for guiding in the future, as an example).

Nevertheless the ability to take shorter exposures is a benefit placing less constraints on the mount you can use and can work in conjunction with short focal length instruments well that are 'weighty' which in the past might have needed more expensive mounts.

[centroid]

An interesting input "whirlwind" thank you.

It is very much a case of "swings & roundabouts", CCD has some very distinct advantages over CMOS in astro cameras.  Much cleaner images, no quirky green colour cast, and no gain or black point to set correctly,  but CCD is yesterday's technology, and more expensive. Plus the fact I seem to remember reading somewhere, that the failure rate in manufacturing the CCD 'wafers' was quite high.

CMOS on the other hand, is by comparison, much cheaper to produce, and has huge market in the digital camera world, including in mobile phones. The amateur astro camera market is miniscule in comparison.

As far as I can see, the main benefits of CMOS astro cameras, are lower cost, and greater sensitivity,. But, there is no such things as a free lunch", and the downside is, severe amp glow/starburst,and other quirks . As I understand it the amp circuitry is actually on the CMOS 'chip', whereas with the CCD camera, it is separate from, and behind the sensor.

The 'bottom line' is that sadly CCD for this application is now a thing of the past, and like it or not CMOS is here to stay, so we have to work with it, until something else comes along.

Fortunately for me, astro imaging is now very much a secondary interest to my photography.

I have seen several very nice astro images produced with CMOS astro cams, but I think that is more down to the processing skills of the imager,. Conversely I have seen quite a few very garish one.

Just for the fun of it, this afternoon, I had a 'play' with my image in the "digital darkroom".  Heavily processed in both Photoshop, and Lightroom to produce this psychedelic result 😀. Albeit it did extract some more detail, but quite over processed, especially the stars. No darks or flats.

In the mean time, I will give the 294c a fair 'crack of the whip' before making any decisions, and maybe, I might get to like it. 😉

Edited January 17, 2022 by centroid
Typos

[Lee_P]

I think that a big benefit of modern CMOS sensors is their ability to function well with short subframe lengths. This has a few benefits, such as putting less pressure on your mount's performance. Also note that the newer generations of CMOS sensors, such as those used in the camera I use (ASI 2600MC PRO) have no amp glow.

36 minutes ago, centroid said:

I have seen several very nice astro images produced with CMOS astro cams, but I think that is more down the processing skills of the imager,. Conversely I have seen quite a few very garish one.

I agree, processing skills make or break an image, regardless of sensor type used.

I'm sure that before long you'll be producing great pics with your new camera. But if not, you could always go back to CCD; just because it's old technology doesn't mean it's bad. All these cameras are tools for us to use, choose whatever suits you best and you enjoy using 

 

[DaveS]

What was the SX CCD you were using? the image is square, so a Kodak 4022 sensor?

[ollypenrice]

Are you comparing OSC with mono here? OSC will be very hard-hit by the moon. The fairly recent dual- or tri-band filters would open up the sky to your OSC CMOS.

I've just dipped my toes in the waters of CMOS and think it's fine, though there are other big variables when I compare results from both. Most obviously the CMOS is in an F2 system as opposed to F5 and F7 with my CCDs.

Give it a go without the moon, I'd say.

Olly

[CCD-Freak]

I was a skeptic at first too but after getting an ASI-533MCP and recently an ASI2600MCP I have been "assimilated".  8^P   

The newest CMOS cameras have little or no amp glow and I have never had the "green cast" using Astro Art 8.  Yes you do calibration a bit differently than CCD but I don't find it any more difficult.  The main thing I find is that I use shorter subs to keep from saturating the sensor and shoot way many more to get a long total integration.   Processing 300 or more subs can be a bit daunting if you computer is not a "power house". 

Sony no longer produces CCD sensors and there are only a few suppliers of scientific grade CCDs but the prices are way above my "pay grade".

Work with the new camera and figure out the things that are different and enjoy the CMOS technology.  Each generation gets better.

 

 

 

Edited January 16, 2022 by CCD-Freak

[Lee_P]

5 minutes ago, CCD-Freak said:

The newest CMOS cameras have little or no amp glow and I have never had the "green cast" using Astro Art 8.  Yes you do calibration a bit differently than CCD but I don't find it any more difficult.  The main thing I find is that I use shorter subs to keep from saturating the sensor and shoot way many more to get a long total integration.   Processing 300 or more subs can be a bit daunting if you computer is not a "power house".

Definitely agree. This is 660 x 120 seconds, ASI2600MC PRO, no filters, under Bortle 8 skies. (@ollypenrice and most of the data were taken with a bright Moon nearby in the sky!)

 

Edited January 16, 2022 by Lee_P

[ollypenrice]

5 minutes ago, CCD-Freak said:

I was a skeptic at first too but after getting an ASI-533MCP and recently an ASI2600MCP I have been "assimilated".  8^P   

The newest CMOS cameras have little or no amp glow and I have never had the "green cast" using Astro Art 8.  Yes you do calibration a bit differently than CCD but I don't find it any more difficult.  The main thing I find is that I use shorter subs to keep from saturating the sensor and shoot way many more to get a long total integration.   Processing 300 or more subs can be a bit daunting if you computer is not a "power house". 

Sony no longer produces CCD sensors and there are only a few suppliers of scientific grade CCDs but the prices are way above my "pay grade".

Work with the new camera and figure out the things that are different and enjoy the CMOS technology.  Each generation gets better.

 

 

 

The Flame is a very good test of colour and this passes muster with full marks in my view.

Olly

[ollypenrice]

1 minute ago, Lee_P said:

Definitely agree. This is 660 x 120 seconds, ASI2600MC PRO, no filters, under Bortle 8 skies. (@ollypenrice and most of the data were taken with a bright Moon nearby in the sky!)

 

OK, my pessimism about the moon seems to be misplaced!

Olly

[Lee_P]

6 minutes ago, ollypenrice said:

OK, my pessimism about the moon seems to be misplaced!

Olly

PixInsight has some good processes that help, in particular NormalizeScaleGradient. Nowadays I don't even factor the Moon's position or phase into my imaging, even when shooting broadband targets. But I do aim for long integration times, typically 20+ hours.

[CCD-Freak]

7 minutes ago, ollypenrice said:

The Flame is a very good test of colour and this passes muster with full marks in my view.

Olly

I should have used a UV-IR filter for that shot.  I use one all the time now when shooting OSC to block IR which makes for better color.

I have not had the ASI-2600MCP out yet but I am really looking forward to turning it on the sky.

My CMOS cameras and more processing experience have produced my best images yet.

 This is my lastest project with the ASI-533MCP / SS15028HNT combo.

 

[centroid]

23 minutes ago, Lee_P said:

I think that a big benefit of modern CMOS sensors is their ability to function well with short subframe lengths. This has a few benefits, such as putting less pressure on your mount's performance. Also note that the newer generations of CMOS sensors, such as those used in the camera I use (ASI 2600MC PRO) have no amp glow.

I agree, processing skills make or break an image, regardless of sensor type used.

I'm sure that before long you'll be producing great pics with your new camera. But if not, you could always go back to CCD; just because it's old technology doesn't mean it's bad. All these cameras are tools for us to use, choose whatever suits you best and you enjoy using 

 

In fairness Lee, the 294c is the only cmos based astro cam that I have owned/used, and from what I read, the 294 sensor has been around for a whiie now,  so almost certainly old tech compared to the 2600.

Technology moves on at a pace, and with it things usually improve.

Once I get to understand the 294 and its quirks better, I will get more from it.

As you say, I can always go to ccd, but its early days yet.

[centroid]

38 minutes ago, DaveS said:

What was the SX CCD you were using? the image is square, so a Kodak 4022 sensor?

Well spotted Dave 🙂. Yes it was the Kodak sensor in an SX SXVR H16, and of all the Sony sensor based cameras I had, up to H9 size, it was the best,.

Far more hot pixels than the Sony ccd sensors, but easily removed in processing.

[centroid]

44 minutes ago, ollypenrice said:

Are you comparing OSC with mono here? OSC will be very hard-hit by the moon. The fairly recent dual- or tri-band filters would open up the sky to your OSC CMOS.

I've just dipped my toes in the waters of CMOS and think it's fine, though there are other big variables when I compare results from both. Most obviously the CMOS is in an F2 system as opposed to F5 and F7 with my CCDs.

Give it a go without the moon, I'd say.

Olly

Yes it is OSC Olly, and would never have bothered imaging DSOs in the presence of Moonlight. However,  having returned to imaging at the end of last September,  i have had to jump a number of hurdles, hardware, firmware, and software.  After dealing with each one, this week was the first chance I have had to use the gear in  anger, even if it was just a short test session.

So, moonlight or not,  I had to give it a go.

Unlike my 'plug and play' CCD days, I have found the need to experiment with gain and black level settings. Thank goodness for the Histogram.

Interestingly, having taken both 15s and 90s images, the blackpoint setting used for one exposure, was way out for the other exposure. Not a problem objects not requiring the combination of two different exposures, such as M42. Never a problem with ccd.

I will give it a fair try Olly, but does seem from what I am hearing, the 2600 is a much better option than the 294.

 

[ollypenrice]

1 minute ago, centroid said:

Yes it is OSC Olly, and would never have bothered imaging DSOs in the presence of Moonlight. However,  having returned to imaging at the end of last September,  i have had to jump a number of hurdles, hardware, firmware, and software.  After dealing with each one, this week was the first chance I have had to use the gear in  anger, even if it was just a short test session.

So, moonlight or not,  I had to give it a go.

Unlike my 'plug and play' CCD days, I have found the need to experiment with gain and black level settings. Thank goodness for the Histogram.

Interestingly, having taken both 15s and 90s images, the blackpoint setting used for one exposure, was way out for the other exposure. Not a problem objects not requiring the combination of two different exposures, such as M42. Never a problem with ccd.

I will give it a fair try Olly, but does seem from what I am hearing, the 2600 is a much better option than the 294.

 

I've only used the 2600 so can't compare them.

Olly

[Whirlwind]

1 hour ago, centroid said:

I will give it a fair try Olly, but does seem from what I am hearing, the 2600 is a much better option than the 294.

From what I can tell each CMOS camera has its own unique 'oddities' that have to be dealt with.  From what I've read the 294 needs careful flats because of the way its pixels are structured.  At the moment there isn't one comprehensive 'manual' to ultimately calibrate them.  The 2600/268 are most 'CCD-like' in this regard from what I understand.

[MartinB]

I've been doing a widefield 9 panel mosaic of Orion this week largely to get myself used to the ASIair.  The results have been poor thanks largely to the moon being so close by.  Your's is a fine image given a short exposure time but that moon is  doing some damage.

For over 10 years I imaged with a QSI 532 CCD camera.  It was a NABG chip and with a max QE of over 80%.  It achieved this using microlensing with a much greater capacity to produce star artefacts than any of my 3 CMOS cameras.  I thought I would continue to use the QSI but no, I've not been tempted to get it out again.  For me the main advantage  of CMOS, apart from price for a given amount of chip real estate, is the low read noise.  Even under Bortle 6 skies I was still barely over the read noise hump using Baader 7nm NB filters and 30 min exposures.  I lost so much exposure time thanks to the odd passing cloud.  I now use 5 minute exposures for NB which makes the preprocessing routine slower but that's not really a problem.  I love having control of the gain level which allows you to fine tune your camera depending on whether you are imaging NB or broadband.  My ASI 1600 is 12 bit.  I thought this would limit how much stretching I would be able to do but in the real world it has been a non issue and I have managed to tease out the faintest of OIII.  Yes, for your 294 you will need to use flat darks matched to your flat exposures instead of bias and make sure you clear any "optimize darks" settings,  but that isn't much of chore.  My 294 does show some impressive amp glow but this is very deceptive and is only revealed by a massive auto stretch, it calibrates out perfectly and I have never been able to see any adverse effect on my final images.  I also have an ASI 2600 MC which I am using with 135/200mm fast lenses and an NBZ filter.  I get a green cast which I assume is down to the RGGB bayer matrix.  It takes about 5 seconds to get rid of.  The build quality of the 2600 is way way better than my old SX 25 OSC although that's a different subject!  I am sure you will learn to love CMOS in time Dave!! 

[centroid]

"I am sure you will learn to love CMOS in time Dave!! "

I don't think that I have much of a choice Martin, other than go back to CCD, or give up 😅

I has been an interesting discussion, with some very useful input, your own included.

I can't see me getting back into NB, or LRGB imaging again, I did that 'back in the day'. As I said previously, astro is now very much a secondary interest to my photography, which I can pursue, day, night, cloud or no cloud, and not be bothered the Moon 😀.

I only came back into astro imaging, as it was an itch that never went  away, and I needed to scratch it 🙂.  Roger of course, also now a keen photographer (wildlife), has zero desire to return to astro imaging.

I will be happy if I can get the occasional decent DSO image,  with the  CMOS OSC camera, to maintain the interest, and to justify the money I have spent buying new kit (I sold everything back in 2014). Now if only I had kept the SXVR H16, as an ornament or such like. ☹️

As you say, your 294 "does show some impressive amp glow", a bit of an understatement in "my book"😉. Having the TEC cooled model, it should not be a problem in matching the Lights, Dark, and Flats, although I have seen reports of it being somewhat difficult to eliminate in exposures over 240s.

These CMOS sensors are very much re-purposed devices, never designed for astro imaging, as was the case with CCD in amateur astro cameras also, but the CCD, although less sensitive, coped far better. As least in my opinion.

CMOS is great technology for digital cameras (terrestrial photography), and the CMOS sensor in the EOS 5d MklV is superb. I can photograph virtually noise free, up to ISO 12800. Would I hang it on a £1200 telescope, no way, too expensive, and far too heavy.

The green cast, is as you say, apparently down to the RGGB matrix, and the lack of white balance I guess. However, it is easily removed using the "Hast la Vista Green"  plug in for photoshop.

It is early days yet Martin, and still enough winter nights left,  to explore what the 294c is capable of, quirks and all.

Then come the Spring, it will be time put it away, get the Motorhome out of storage, and get out and about again. Who knows, we might cross paths, when you are out and about with yours. The Peak District is very much a favourite of ours, and of course it is your part of the world.

Edited January 16, 2022 by centroid

[tomato]

 CMOS cameras have made large format sensors affordable to me and as previously mentioned, the latest offerings do not suffer from amp glow. I started out with a little OSC CCD which IMHO is definitely not in the same league as the latest CMOS OSC cameras.

Having said that, CCD mono cameras remain first class Astro cameras and it’s a great time to be a fan of them, as the second hand market has loosened considerably.

As for putting the Astro imaging gear away in the Spring, that’s not for me, it’s Galaxy Season!

[centroid]

2 hours ago, tomato said:

,As for putting the Astro imaging gear away in the Spring, that’s not for me, it’s Galaxy Season!

Back in the day, it was not unknown for me to stay out in my then obsy, all night, but I'm not that enthusiastic now, and a lot older.

[Stuart1971]

On 15/01/2022 at 15:46, centroid said:

The first astro image in the 8 years since I gave up the hobby, with a somewhat disappointing result

 Just a test run under a Moonlit sky, when I would not have normally bothered with imaging deep sky objects, but after weeks of all sorts of problems, both hardware, and software, I had to give it a go.

 Just 15 x 90 second exposures, combined with 15 x 15 second exposures, and processed in Astroart 8, and Photoshop.

 Way below the quality that I would normally accept, but a new type of camera technology for me. Whereas I used to use CCD based astro cameras, I now have a 294c CMOS based camera.  More sensitive, and relatively cheaper than CCD, but they have some odd 'quirks', that take bit of getting used to dealing with.

 Yes, it could have done with a lot more subs, but even so a pretty lack lustre result, and one that I had to 'push' hard in processing the get that.

 I have included another image, of the same subject, that I took with an SX ccd camera, back in 2012.

 Early days yet, but the temptation to go back to ccd, while they are still available, is growing 😉

You are in the exact same position as me, and feeling the same as me, I have used CCD for years and decided to move to the QHY268c  and I too am a little disappointed, my images are full of noise and not pretty at all, and yet I read how much more sensitive these CMOS cameras are but I don’t see it at all, the quirks and trying to get correct gains and offsets, it driving my nuts, and I too am thinking of a return to CCD, this CMOS lark is just not what it was advertised as, unless I am doing something seriously wrong, which I doubt, I would love to know where and how all the superb images on here were created, it does make me wonder how much of them were created in the processing more so than the camera..But I will stick at it after shelling out over £1600 on this new QHY268c camera, I guess this past week has not been the best time for imaging with the moon around every night…so maybe I was expecting a bit too much…time will tell…👍🏼

Edited January 17, 2022 by Stuart1971

[Phillyo]

23 minutes ago, Stuart1971 said:

You are in the exact same position as me, and feeling the same as me, I have used CCD for years and decided to move to the QHY268c  and I too am a little disappointed, my images are full of noise and not pretty at all, and yet I read how much more sensitive these CMOS cameras are but I don’t see it at all, the quirks and trying to get correct gains and offsets, it driving my nuts, and I too am thinking of a return to CCD, this CMOS lark is just not what it was advertised as, unless I am doing something seriously wrong, which I doubt, I would love to know where and how all the superb images on here were created, it does make me wonder how much of them were created in the processing more so than the camera..But I will stick at it after shelling out over £1600 on this new QHY268c camera, I guess this past week has not been the best time for imaging with the moon around every night…so maybe I was expecting a bit too much…time will tell…👍🏼

Stick with it. I moved from CCD to CMOS and much prefer CMOS. I use an ASI533 and it's great. The reason you're seeing more noise is probably short sub lengths compared to CCD. However, if you increase your total integration time to the same as you did with CCD you should see the noise drop and a LOT more signal come through. Imaging with a colour sensor with a bright moon is going to produce lots of noise unfortunately. Try it without the moon and give it lots of integration time, you'll get the results you want.