Disappointing

[Dave_S]

I'm still sceptical about the sensitivity of CMOS.

 

After 12 years (2002-2014) of imaging with SX CCD cameras, and now having returned to the hobby, with a 294c Altair CMOS camera, after reading the hype about CMOS being so much more sensitive, I am sceptical.

 

Some previous ccd images here:

 

https://davesimaging.wixsite.com/mysite

 

Last night, with a clear sky and no Moon, I had a 'crack' at the Rossette, and the result was disappointing, to say the least.

 

115mm f/7 APO with a 0.8x reducer, thus f/5.6, and a 294c camera (binned 1x1). 18 x 4 min exposures.

 

Not a lot of exposure time I accept, but even so, from previous CCD experience I would have expected better.

 

I have seen an image this object posted, and while an excellent example, it was 4 hours worth of 5 minute subs😮

 

If this is the sort of exposure time now required, I think I will retire back to my primary hobby of photography.

 

Maybe its because I am that much older now, and have 'been there, and done that', that I don't seem to have the enthusiasm for astro imaging that I once had. My initial results have done little to encourage any of the old enthusiasm. 😟

 

[Cumbrianwolf]

Still worth persevering, I say, as I am sure it still can be rewarding, and you just need to find a niche that works for you. I am just stepping into these waters of having my set-up go all digital with remote operation from the house, rather than freeze the extremities whilst standing outside with the scope, camera and intervelometer oh and not to forget the off axis guider. Apart from 35 mm film of which I moved over to digital photography I cannot comment on the sensors, as most of my pre digital was via 35 mm film, albeit not too good but fun, and it set some knowledge and skills I have to transfer them to astro imaging.

[Dave_S]

I have just seen a Rosette image on another forum, taken with the same camera, and just 14x5 min exposures, and the difference between that, and my 18x4 min image, is like "chalk and cheese". Maybe my 294c needs to go to Spec Savers.

 

Here's my image, unprocessed, as there is not enough in the image to push in processing. Ignore the dust motes, as I didn't bother wasting time on it with Flats.

Edited January 28, 2022 by Dave_S
added photo

[Cumbrianwolf]

Just now, Dave_S said:

I have just seen a Rosette image on another forum, taken with the same camera, and just 14x5 min exposures, and the difference between that, and my 18x4 min image, is like "chalk and cheese". Maybe my 294c needs to go to Spec Savers.

Maybe that is the case, I am not that clued-up on the dedicated astro cameras as for now as I am a DSLR user with an aim of finding my footing before I spend over a grand on a dedicated astro camera. I am currently aiming to fine-tune the imaging and processing of the images before I commit, and if only the weather would give us a break.

[Dave_S]

Here's the image after some VERY heavy processing, too heavy as you can see, but still very Lack Lustre, and way below to the quality I got with my SX CCD's. Maybe my 294c was a "Friday Afternoon" job 🙄.

 

I will persist with it a while longer, then maybe look to going back to a CCD camera while they are still available.

 

[Cumbrianwolf]

3 minutes ago, Dave_S said:

Here's the image after some VERY heavy processing, too heavy as you can see, but still very Lack Lustre, and way below to the quality I got with my SX CCD's. Maybe my 294c was a "Friday Afternoon" job 🙄.

 

I will persist with it a while longer, then maybe look to going back to a CCD camera while they are still available.

 

Is the camera still under warranty as it may just be faulty you never know, or maybe the seeing is not in your favour? I would still be happy with the above image.

[Dave_S]

It is only 5 months old, and little used. However, if returned to base, it would be found to be working, and only a live sky test woukd show it has having lower than expected sensitivity. 

 

If as someone suggested, it would need around 4 hours of exposure to achieve a descent Rosette image, then fotget it, just don't have that level of enthusiasm now. 

 

I never needed to resort to those exposure lengths, back my ccd imaging days, all 12 years of them.

[Cumbrianwolf]

5 minutes ago, Dave_S said:

It is only 5 months old, and little used. However, if returned to base, it would be found to be working, and only a live sky test woukd show it has having lower than expected sensitivity. 

 

If as someone suggested, it would need around 4 hours of exposure to achieve a descent Rosette image, then fotget it, just don't have that level of enthusiasm now. 

 

I never needed to resort to those exposure lengths, back my ccd imaging days, all 12 years of them.

What about using a focal reducer to allow quicker acquisition or a much faster OTA, I have an F/4 with the coma corrector reducing that to around F/3.5 mind you it is also large and heavy being a 254 mm Newtonian, but it is a pure light bucket and can get exposures very quickly. I am only surmising what you could do, as you no doubt have more experience than myself. I am sure others can or will chip in and hopefully help you out, it could be something so simple.

[Dave_S]

I am using a 0.8x reducer/flattener, which takes my 115mm scope down from 805mm FL and f/7, down to 644mm FL and f/5.6. 

 

Although back in the day I used SCTs up to 12", I also used a 110mm APO, with SX ccd cameras, and got much better results than I am seeing from this 115mm APO and 294c camera. Albeit, I'm sure that is down to the camera, and not, the scope.

Edited January 28, 2022 by Dave_S

[Dave_S]

Now that has put a bit more "life" into it . I dug out my copy of Lightroom, hadn't used it in years, and had a quick play with the above image in there. Well 'overcooked' now, but it has certainly extracted some more detail, and lifted the lack lustre appearance. It called "artistic licence" . I have never had to 'push' am imager that far before 🙄

Edited January 28, 2022 by Dave_S

[Dave_S]

Last 'play', and the final cleaned up image. Still below 'par' but the best of a bad job. 

[MartinS]

For me I have had far more image productivity since switching my SX36 for a QHY600M. The move from 15-30 min frames to 5-10 min frames have delivered so much more data to work with and can complete a data set in a good night. Now fitted to the RC the SX is still a loverly Camera.

 

May I ask what Gain and offset you were using? Noting you are using a colour camera.

[paul]

Keep at it - take more shots, you are getting good results.

 

With my F4.8 OTAs I use a cooled mono camera and the minimum I try for is 1 hr per filter. So for my taste; RGB shots are a minimum of three hours to ensure the target is rendered OK and background noise is tolerable - I grab more if I can. 

 

A one-shot colour camera gets you out imaging more often and every shot is a complete image, but despite the convenience you will have to take lots of images. A 1hr session with a colour camera is not equivalent to 3 hrs with a mono camera (assuming sensors with similar mono properties). Like for like mono cameras grab 4x more light per filter.

 

I've seen a lot of images by owners who started with CCD then switched to CMOS, in almost all cases the CMOS images have brought out more detail and greater signal to noise on the same OTAs.

[Dave_S]

On 1/30/2022 at 9:35 AM, MartinS said:

For me I have had far more image productivity since switching my SX36 for a QHY600M. The move from 15-30 min frames to 5-10 min frames have delivered so much more data to work with and can complete a data set in a good night. Now fitted to the RC the SX is still a loverly Camera.

 

May I ask what Gain and offset you were using? Noting you are using a colour camera.

Back in the day (2002-2014) I enjoyed successful imaging with both SX OSC (H9C), and SX Mono (H9, and H16). I switched over mono  because I wanted to do NB imaging, and invested in a 7 position motorised, usb controlled filter wheel wih L,R,G,B Ha, Olll, and Sll filters. My imaging moved up a gear when I bought the SXVR H16, with its large 7.4 x 7.4um light grabbing pixels. It was £2000 back in 2011.

 

Then in 2014,  sold everything, an stayed out of the hobby for 8 years, concentrating my love of photography, especially wildlife.

 

On reading all this hype about how much more sensitive these CMOS cameras supposedly are, compared to CCD, I thought I would return to the hobby with a more sensitive OSC camera. I no longer have the enthusiasm to go back to mono.

 

I chose the 294c based on its 4/3rds size sensor, and 4.6 x 4.6um pixels, as opposed to the 183 with its small sensor, and tiny 2.4 x 2.4um size pixels.

 

Sadly, it has yet to impress me, but then I am still trying to get my head around its strange quirks, compared to the 'plug & play' CCD camera., which IMO produced far better colour, no green tinted images, and no faffing about with offset and gain.

 

Gain setting with CMOS cameras is very much  a 'suck it and see' operation, with different manufacturers, and software developers quoting different scales, and numbers for what equates to the same setting.

 

For example a gain setting of 900 inSharpcap, is exactly the same as gain of 9 AA8 with ASCOM control

 

The gain setting scale in Sharpcap goes from 100-72000. on AA8 with ASCOM its from 1-50, and Altair Capture 1-160. Logical definitely not, confusing definitely yes.

 

I ran a sensor analysis  in Sharpcap, and the optimum gain/red noise cross over point was 900, so translating that into AA8, which I use for capture, guiding, calibration and stacking, I get a gain setting of 9.

 

The Offset/Blavck point is easy. I set it to what I think will be right, take an image, then look at the Histogram to make sure it isn't clipped at the black enf, and that there is a small gap between the black end, and where the image data starts.  I find that 30 is about right.

 

Trying to get a definitive answer to what the gain setting should be, provides conflicting answers. ZWO publish a manual for the cameras, with an informative set of graphs, which show a recommend gain setting for 294 of 121.. But 121 on whose capture software scale.

 

A bit of nostalgia, an image of IC434 that I took back in 2012, with the SX SXVR H16, and a WO 110 Triplet. 🙂

 

 

 

 

Edited January 31, 2022 by Dave_S
added image

[Dave_S]

Sensor analysis on the Hypercam 294c Pro camera.

 

Now having a good quality Light Panel, with a variable output, primarily for 'Flats', this afternoon, I ran a Sensor Analysis using the application in Sharpcap. I did this when I first bought the camera, but with a less suitable light source.

 

Today the setup was more refined, and controlled.

 

The resulting table showed an e/ADU of 1.68 at a gain setting of 100 (the lowest setting on the Sharpcap gain scale), and an e/ADU of 0.95 at a gain setting of 177.

 

By interpolation, this puts unity gain (1.0) at a gain setting of 172.

 

However the gain/read noise graph shows a sharp drop in read noise at a gain setting of around 900 (again on the Sharpcap scale). The level of read noise remaining low, and flat from that point on upwards.

 

This raises the question, do you run at unity gain with higher read noise, as apparently some folk do, or do you opt for the higher gain setting, and lower read noise.

 

The next part of the conundrum is, Sharcap's gain scale goes from 100 to 72000, whereas for example,  Astroart's (ASCOM) goes from 1 to 50,  and Altair Capture from 0 to 160. All very logical to the end user of course, NOT!!.

 

I do know that Sharpcap uses a multilpier of x100 for its gain figures.  However, I like my imaging software in a single comprehensive package, and Astroart does my capture, guiding, stacking and calibration etc, both effectively, and quickly. Leaving me just to process in Photoshop, PixInsight, or even Lightroom.

 

If I want to use a capture software other than Sharpcap, which I do, I need to transpose Sharpcap's 100-72000 gain scale in Asrtoart's  1-50 scale. If my thinking is correct, then a gain of 900 for example, in Sharpcap, equates to a gain of 9 in Astroart (ASCOM), or does it?, I have yet to find a definitive answer to this question. Imaging was relatively simple by comparison when we were using CCD cameras.  No gain or black level settings to get right, and no green tinted images 😉. Ok, CMOS is cheaper, and more sensitive, but there is no such thing as a 'free lunch'

 

Edited February 2, 2022 by Dave_S
added image

[mechanoid]

It seems as though your misgivings about CMOS cameras may have some grounds.

 

If I'm right in assuming that your camera has a 14 bit ADC, then your 1 electron/ADU is equivalent to 0.25 e/ADU for a 16 bit ADU.  My 16 bit ATIK camera with a Sony ICX694 CCD has a quoted gain factor of 0.27 e/ADU.

 

Based on this, it would appear that at least your particular CMOS camera is no more sensitive than my CCD one !  Perhaps the supposed better sensitivity of CMOS is a myth. Or maybe its just your camera. As far as I can find out, the quantum efficiency for both sensors peaks at around 75% in red.

 

I presume that the lower e/ADU at higher gain values will come at the expense of lower "full well" depth. So this isn't really "more sensitive", just more amplification.

 

Incidentally, my QHY10 OSC CCD camera requires me to set gain and offset values, so this is not exclusive to CMOS. This isn't actually of much use, except to mess up my images if I get the offset wrong.

 

 

[Dave_S]

Yes the 294c is 14bit.

 

The sensor analysis shows that the full well depth (FWD) is at its maximum at a gain setting of 900 (sharpcap), a relative gain of x8.89. Going up in gain from there, sees the FWD drop off, as it also does going down in gain from there. So the optimum performance in terms of read noise, and FWD would seem to be at 900 gain (Sharcap scale), and that is what I have been using.

 

If I were to show the analysis results the to Altair, or Toupek who make the camera, I'm sure they would say it is within spec.  So I don't think there is anything wrong with the camera.

 

Having imaged for some 12 years with CCD,  I was  both happy, and comfortable with their performance. It may of course be me, but as yet, I have not been able to equal those results, let alone better them, with this 294c CMOS camera.  

 

Have just seen a very impressive CMOS NB image of the Rosette, posted on the Altair FB page, but it was a total of over 8.5 hours of exposure. I never in all those years of imaging, had to go to those extreme exposure lengths with CCD, whether it be OSC,  Mono, or Ha, Olll, Sll  with a 110mm APO. I'm now using a 115mm APO.

 

I'm not giving up on this camera yet, and maybe I can drag some better results from it, time will tell.

 

I have now invested in an Optolong Enhance filter, which will up the exposure time even more (lol)

 

I note that some the best CMOS images seem to be coming from the 2600, so maybe the 294 now being somewhat a dated, isn't in the same league. Am I going to buy one, no chance. Another SX CCD, maybe.

 

EDIT:- have just seen another nice NB image (N Ameircan Nebula) taken with a ZWO 1600 CMOS. But!!, with some 13 hours of exposure. Where is this greater sensitivity over CCD. 🙄

Edited February 5, 2022 by Dave_S
Typo

[Dave_S]

To add to the confusion 😅

 

This afternoon I had a 'play' with the light panel, and taking some 'flats. Not on the scope but set up on my desk in the study.

 

With the light panel at its lowest setting, I had to cover it with six sheets of 80gm copy paper, to give me a Flat of around 2/3rds of Full Well capacity at for an exposure of 0.6 secs. As understand it, but correct me if I am wrong, Full Well capacity for a 14bit CMOS sensor is circa 17000 ADU.

 

I am led to believe that things get a bit non linear at exposures of less than 0.3 seconds.

 

The bottom line is, this is showing the sensitivity of the 294c sensor to be good, which contradicts what I have found so far when imaging. 

 

Definitely more night sky testing required, but that is of course dependent on the weather.

 

"watch this space".😉

Edited February 3, 2022 by Dave_S

[mechanoid]

I'm not sure how you can come to the conclusion that "the sensitivity is good" since you don't know the illumination intensity through six sheets of paper.

 

For me, the only sensitivity that matters is the quantum efficiency, and there is nothing remarkable about that. Not surprising since CMOS and CCD technologes both use Silicon substrates.

 

I know that other things matter too, such as dark current, readout noise and what area of a pixel actually is photosensitive.

 

I agree with you - in my opinion CMOS is not noticeably better than CCD.  Just cheaper and the only game in town once CCD stocks are exhausted.

 

I find that trying to compare exposure times is generally meaningless since astrophotographers rarely quote the focal ratio of their telescope. Four hours for them may be the same as one hour for me. There are not all that many scopes faster than f4, although people using camera lenses can go as low as f1.2  Given the small fraction of clear nights here in the UK, the faster the better for me.

[Dave_S]

1 hour ago, mechanoid said:

I'm not sure how you can come to the conclusion that "the sensitivity is good" since you don't know the illumination intensity through six sheets of paper.

Just a subjective assumption.

 

My background is in Radio Frequency, and Electromagnetic Compatibility Engineering, both at R&D, and practical levels. Certainly not imaging sensor technology. So you will have to forgive my assumption if not technically sound.

 

I made the assumption, based on my CCD experience of some 8+ years ago.

 

Back then I used a light panel for my Flats. How the light intensity was, compared to this one, I can only from memory make a 'wet finger in the air' guess, but it wasn't wildly different. I never had the equipment to measure it,  and neither do I now, but then how any amateur astro imagers do.

 

With the CCD cameras, I only ever had to use a single sheet of copy paper to bring the light down to a suitable level. 

 

All very unscientific, and purely subjective, but based on what would seem to be a very much reduced light level, required to achieve a suitable level of ADU, the CMOS would seem to be more sensitive. Which as I said, contradicts what I have found under imaging conditions.

 

For me, "the proof of the pudding is in the eating", and this CMOS camera has yet to impress, but maybe further down the line it will.

 

If it is going  to require hours of image integration, e.g. up the 13 hours, as mentioned earlier, then unlike back in the day, astro imaging is not for me this time around. I'll sell up, and stick to my primary intertest of photography. 

 

Looking back with fondness on the enjoyable 12 years (2002-2014) I spent imaging with SX CCD cameras.

[Dave_S]

Have just seen another quite average NB image (HA, Olll, Sll) posted on another forum, almost 19 hours of integration with ZWO ASI 1600 CMOS camera 😮.

 

I looked back in my archives. and found this, what to me is quite an acceptable, NB image (Ha, Olll), that I took back in 2013 with an SX H16 SXVR CCD camera (Kodak sensor). Although I don't have the FITS data now, I don't think it would have been more than an Hour or so through each filter. Certainly no more that 3 hours in total.

 

I know that many of today's imagers, will have come into the hobby with CMOS cameras, so have nothing to compare them with, but there are still plenty of "old Timers" around.

 

To sensor manufacturers like Sony, the amateur astro imaging market is miniscule compared to the digital camera (DSLR, Smart phone,  and industrial/scientific applications) markets.

 

CCD has now been surpassed by CMOS, its cheaper by comparison compared to CCD, and readily available, but is it better suited to our application. Some would say yes, but at least as far as the 294 is concerned, I am yet to be convinced.

 

From what I am seeing in images posted forums, the 2600 is a much better choice, albeit more expensive, but having paid £900 for the 294c, that is what I am stuck with for now. 

 

If CMOS has done one thing, it has allowed people with a limited budget, to get into astro imaging. I paid £2000 for my SX SXVR H16 way back in 2011, but never regretted it.

 

Hindsight is a wonderful thing, but I wish I had done my CMOS camera homework, before buying the 294c.

 

 

 

 

[GazAstro]

A bit of a curveball but I wouldn't mind betting that light pollution has worsened massively since 2002 ?

Hence the need for longer integration time maybe.

[Dave_S]

No doubt that light pollution has got worse Gaz, even more so with the move to the more difficult to filter out LED street lighting.

 

But, filter technology has moved on a long way from the old CLS filters, but 19 hours of integration with a supposedly more senitive camera technology, is to me, extreme. 

 

I live in a village, not a town or city, so my light pollution is low by comparison.

 

I'll keep plugging away with this 294c for time being, but to me it is a disappointment. Probably because I was expecting a leap in sensitivity over ccd.

 

Edited February 4, 2022 by Dave_S
Typo

[paul]

I too have seen many posts (not in this forum😉) where the author is majoring on his marathon run of imaging but when the image is published it looks it looks like they wasted most of the data with aggressive noise reduction and applied crayon colours. Some of them are fighting very aggressive light pollution of course. Having said that others produce excellent detailed low noise images (several excellent examples in this forum).

 

But I'm not sure your comparisons are really apples to apples. The H16 SXVR sensor pixels are approaching 2x the size of the ASI1600. So the CCD image is lower resolution so would be brighter if all other considerations are equal. Another way to look at it is the CMOS pixels are smaller but have very similar performance per pixel vs the CCD but are seeing fewer photons because they are smaller. CMOS is only going to get better.

 

There are other practical advantages to CMOS. The readout noise is lower, you can afford to take many short images versus fewer necessarily longer ones for CCD. That helps with more usable frames, you beat clouds, satellites and tracking issues etc. The CMOS cameras are also multifunction, being useful for solar system and deep sky. Widefield shots with CMOS tend to be more detailed as a result of their smaller pixels, 

Edited February 4, 2022 by paul

[Dave_S]

Hi Paul

 

I agree comparing the H16 with the 294c is not a fair comparison. As you say the H16's 7.6 x 7.6uM pixels (Kodak KAI4021M), are much larger than than those of the 183, and the 294. And of course the 294c is OSC, even though the H16 was behind RGB or NB filters, each full pixels was exposed to the incoming light, and not divided by a Bayer Matrix.

 

However, the H9c was OSC, and had 6.45 x 6.45uM pixels, so a little big bigger than the 294c's 4.63 x 4.63uM pixels. So, given the supposed greater sensitivity of the CMOS sensor, I would have expected to at see results equal to the H9c, or as I had hoped, better. Something as yet I am not seeing.

 

One the reasons I chose the 294c was its 4.63 4.63uM pixels. as opposed to the much smaller pixels of the 183.

 

What I am seeing from the 294c, that I didn't see with the H9c, are some very strange colour OSC images posted. ignoring the green tint issue. Maybe that is down to processing ability of the user of course.

 

Maybe as i get to understand the 294c better, I will be able to extract more from it. Last night being clear, I had another test session, and I gave it an hour each of some easy targets. M45, M1, and M42, with M42 divided into long and short exposures to capture the core of M42.

 

The RAWs subs do look more encouraging, but I will reserve judgement, until they are calibrated, stacked, and processed.

 

On a positive note, the HEQ5 Pro is now operating more like my NEQ6 did.  The Altair Starwave 115 Triplet I am very happy with, and I would put it on par with the WO 110 Triplet that I used to have. Astroart 8 is doing a great job in capture, guiding, and calibration. So nice to have all the functions in a single package, as opposed to one for capture, another for guiding, and yet another pre-processing. Back in the day, I used Maxim DL, but Astroart is equally as good, and much more sensibly priced.

 

Light pollution wise, I would put my location as Bortle 4, and even better when the street lights go off at 23.30.

 

 

Edited February 5, 2022 by Dave_S
typo