Is traditional long exposure imaging dead?

[Rob Farmiloe]

Is long exposure imaging now old tech?

Interesting debate here.

https://forums.sharpcap.co.uk/viewtopic.php?t=1507

What is the reality?

[wxsatuser]

My goto for Ha is 600 to 900secs @ f/2 Atik 383L

[Rob Farmiloe]

The theory being presented is that a requirement for long exposures is a 'myth'.

That CCDs required this because they are noisy and inefficient.

That lots of very short exposures using a Cmos sensor is the future.

I know which side of the fence I am current but would welcome the wider opinion.

 

 

[wxsatuser]

This most likely true for modern CMOS but I know which fence my CCD likes. ?

[Louis D]

If by traditional long exposure photography you mean using hypersensitized film, then yes.

[Knight of Clear Skies]

Certainly not dead, no, but the new low read-noise CMOS cameras do have their advantages. Short exposures make the hassle and learning curve of guiding optional, mean fewer subs are lost in less-than-ideal conditions and their small pixels are well suited to short focal lengths. However, the sensor of my 1600MM Cool is smaller than many CCD cameras, 4/3rds" rather than APS-C or full-frame, so it's not using the full image circle of most optics.

[andrew s]

You still do better, even with CMOS, with the longest exposure your system can manage without saturation. 

Regards Andrew 

[pete_l]

An interesting lecture.

The take-aways (for me, at least) are that with CMOS sensors there is much less that a manufacturer can add in terms of image quality, since so much is built in to the camera. (Though that doesn't necessarily tell you anything about component reliability or customer support)

Second: nothing beats a dark sky (but we knew that, anyway)
It looks like F/5 or F/6 scopes give very acceptable exposure times
Exposure times of no more than a few minutes are all that's needed in urban environments. Unless you have a nice dark sky when longer will give better results.

[Rob Farmiloe]

1 hour ago, andrew s said:

You still do better, even with CMOS, with the longest exposure your system can manage without saturation. 

Regards Andrew 

This is the method I use.  Expose long enough so your histogram stretches to the right but without clipping.

Robins theory recommends much shorter exposures. Often less than 10s from what I gather.

As a method to overcome poor tracking and cheap cmos cameras with lots of amp glow / starburst I can appreciate it will work.

What i am waiting for is images that show this is actually a better overall method than using longer (120 - 900sec) exposures. 

Especially ones that match or better a CCD or when narrowband is required.

[carastro]

While the principle of Shorter exposures seems a big advantage, the amount of data needed filling up computers is a big disadvantage.  Certainly my clapped out computers would not be able to cope with it.

Also I have seen a number of threads, on here and elsewhere where people are having a variety of problems with these cameras.  I am yet to be convinced that the results compare well with CCD cameras, though there have been a few people who seem to be doing very well with them. 

Carole 

[tomato]

This was an interesting talk and I for one can’t disagree with the scientific principles that sit behind Robin’s conclusions. 

However, before committing to another camera purchase (and I am quite happy to make the £1000+ saving by buying CMOS),  I would like to personally demonstrate this with some actual images, I’m currently trying to experiment with an ASI 120 on galaxy targets, but the UK weather  is not co-operating.

One thing is certain, if it’s cloudy, you don’t need to calculate any kind of exposure time...

[The Admiral]

I found this thread an interesting read and I think provides some input to the question, IIRC.

Ian

[Rob Farmiloe]

So the theory of stacking lots and lots of short exposures is not really new.

My question is,

Where are all the amazing images?

Astrobins is packed with stunning images, mainly taken with CCD cameras. Some to be fair with ASi1600 cmos etc.

DSLRs still contribute some great images. 

What I am not seeing is a raft of short exposure cmos images that are up there with the best?

As already pointed out above, I also see a lot of posts by frustrated people struggling to match the sales literature claims with the results.

[Grogfish]

Firstly I should say that everything I do is very short exposure by most standards (typically 25s or so) because I’m Alt-Az. And given my vary modest setup I’m pretty pleased with the results - stacking 100 frames does considerably reduce noise. (Except that with Alt Az I have frame rotation, and that means the number of frames being stacked drops dramatically near the edge - thus cropping is needed (or shutting up and putting up ?)). Also I’m a firm user of SharpCap Pro, so I greatly appreciate Robin!

However, whilst I felt Robin explained noise in some detail, he didn’t really seem to touch on the question of sensitivity... And a short exposure will be less sensitive than a long one, no matter how much you stack. So for some objects (or for the particular appearance you desire), you will simply need a certain amount of exposure. I’m a rank amateur, but my experience has already shown me that, for example, there’s a certain amount of nebulosity I can image in M45 using 26s no matter how big the stack. If I could image with say 45s, I’d expect to see much more nebulosity.

The comparison images he used were of M42, but actually I don’t think that was a great object to choose because it’s so bright. If for example he’d picked M33, I’d have eaten my hat if 100x1s would show anything of the galaxy at all, whereas at 4x25s you certainly do see something.

Increasing bit-depth of the camera would presumably also be a way to improve the effective sensitivity (via stretching).

Ahh, to be able to take subs with triple digit exposures!

 

[Alien 13]

55 minutes ago, Rob Farmiloe said:

So the theory of stacking lots and lots of short exposures is not really new.

My question is,

Where are all the amazing images?

Astrobins is packed with stunning images, mainly taken with CCD cameras. Some to be fair with ASi1600 cmos etc.

DSLRs still contribute some great images. 

What I am not seeing is a raft of short exposure cmos images that are up there with the best?

As already pointed out above, I also see a lot of posts by frustrated people struggling to match the sales literature claims with the results.

I think some of the answers might be because a lot of the top end CCD cameras get used on top end refactors at dark sites, I think the comparison regarding CCD/CMOS needs to be done in normal city type locations.

Alan

[Grogfish]

49 minutes ago, Rob Farmiloe said:

Where are all the amazing images?

....

What I am not seeing is a raft of short exposure cmos images that are up there with the best?

Check out “The No EQ DSO Challenge!” thread - because it’s all Alt-Az the exposures are all sub 1min (or near enough). Worth a look through.

 

I also suspect there’s something of a self-fulfilling prophecy regards image quality. If you’ve got great gear, then you have a high resolution, high bit depth  sensor and can track for multiple minutes. Plus with some lovely LP filters. And given you own all that kit, that’s precisely what you do. Hence fantastic images from  “long exposure”. And as it works so well, why try anything different.

On the other hand most people that are doing “short exposure” are doing it because they can’t afford any better tracking. That also tends to mean you can only afford cheaper sensors with lower bit depth and lower resolution (like my GPCAM IMX224), and lesser (or no) anti-LP.

So the “short exposure” pics are taken at say 1280x1024 12bit (that’s me!) compared to perhaps 5000x3500 14bit for “long exposure”, which is another reason why some images look better than others.

 

Gosh, could really sound like I have chip on my shoulder, but I really don’t ?

[symmetal]

I don't think the talk indicated that long exposure imaging is dead, but seemed to confirm the reasoning (that I've mentioned in the past ) that there's little point in exposing beyond the point where the sub sky background swamps the read noise. For really dark skies, even with CMOS, this still means exposures of several minutes or more and it's only in badly light polluted skies where exposures around of around 20 to 30s are considered 'optimal'.

The formula I use is where the sky background reaches 10 * RN^2, (that is 10 * (Read Noise squared)), which fits in nicely to the equations at the end of the talk before he skipped all the gain related info. For my ASI1600 at unity gain and offset 56, 10 * RN^2 = 1386 ADU which is what I aim for. For my rural skies this equates to 45s for L, 120s for RGB and a lot longer for Ha. Actually to get 1386 ADU for Ha takes over an hour so I usually stick to 600s Ha as a compromise.

For my Atik One 6.0 CCD, 10 * RN^2 = 1226 ADU (estimated offset) which equates to around 240s L, 600s RGB and almost forever for Ha. I limit Ha to 1200s usually.

The talk graphs indicated stacking 1200 5s exposures  won't give you a similar image to stacking 20 images of 300s (same integration time), as cameras aren't 'perfect' and there is a significant minimum exposure time that is necessary, even for CMOS.

So I think I'll continue as I have been doing. 

Alan

[Tomatobro]

I very much enjoyed Robins talk and its clear he knows this subject. I have watched the video of his talk several times. It was a shame that the end was cut short due to time constraints. I have to take the data presented at face value but I it looked good to me. Keeping an open mind and questioning everything is the key.

  Not all have ready access to dark sky's which was one of his key points as I recall.

[The Admiral]

11 hours ago, Grogfish said:

And a short exposure will be less sensitive than a long one, no matter how much you stack. So for some objects (or for the particular appearance you desire), you will simply need a certain amount of exposure. 

That would be true in perfectly dark skies and with a perfect camera. Surely the issue though is that however long your exposure, or however many subs you stack, it is going to become harder and harder to resolve objects buried in the sky background. Dark skies always win! NB imaging is one way to effectively reduce your sky background of course, and allow you to increase exposures and thus see deeper, but then one needs more than a DSLR and Alt-Az mount I'm afraid :<(

Ian

[Rob Farmiloe]

My light pollution is B6/7.

I have tried a few LP filters but for now I am finding dealing with the graidents after gives better results than a filter.

I had a 183 sensor. Tried all sorts of combinations but in the end decided it was inferior to my DSLR.

I sold it at a fair loss. Instead of taking the sales claims for cameras i studied Astrobins for a long time with an open mind.  I looked at the capture equipment and methods used for images I liked.

From this I decided a CCD was for me despite the great amount of negative comments regarding 'old noisy technology'.

I purchased a used Atik One 6.0

Overnight all the headaches and disappointments the 183 sensor gave me disappeared. It just works.

Now the challenge is me and the weather not the camera.

I am currently using around 240s for L

2x2 bin for colour dependent on scope or lens.

860s for Ha

My scope was £350 used. A few camera lens I already have. A HEQ5 mount.

My biggest expense was wasting time and money on the wrong gear rather than buying what works for me.

I have an Astrotrac as need to travel for better skies.  If short imaging works it would be welcome.  I remain unconvinced as my experience so far is that if the histogram is low then the results are inferior.

Hence I am waiting for the images to demonstrate the method does give an advantage in the end results. 

 

 

 

 

[Scooot]

I found the talk very interesting.

My light pollution is quite bad, probably around B6. His formula with a light pollution /shot value of 15 (a guess) calculates to 30 seconds at F4 with my camera. Through trial and error, setting the histogram about a quarter in from the left, I usually image between 30 seconds and a minute so it seems as if it’s quite a good guide.

[Rob Farmiloe]

Setting the peak of the histogram about 1/4 to 1/3 in is an approximate way of exposing to the max duration avoiding clipped highlights?  That is what I would describe as the traditional method.

30 to 60s at f4 is like 60-120s at f5.6.

Depending on the target, M42 , M45 etc I would not categorize those as particularly short exposures.

Talking to people who have tried the Sharpcap Brain function it has recommended exposures much shorter than this such as 9s regardless of the intended object.

 I have not tried it myself.

I pay for and use Sharpcap for polar alignment and to run my Allsky camera.

It does both of these tasks well so is a bargain 

I do not use it for DSO though as it has no sequencer. 

 

[Astrobug]

When they will sell a big mono cmos (full frame or more) you will see more cmos pictures on astrobin with high end setups. Also experience how to proces images is also important. Most 1600mm user are still in the early (like me) learning curve. 

[Scooot]

31 minutes ago, Rob Farmiloe said:

Setting the peak of the histogram about 1/4 to 1/3 in is an approximate way of exposing to the max duration avoiding clipped highlights?  That is what I would describe as the traditional method.

30 to 60s at f4 is like 60-120s at f5.6.

Depending on the target, M42 , M45 etc I would not categorize those as particularly short exposures.

Talking to people who have tried the Sharpcap Brain function it has recommended exposures much shorter than this such as 9s regardless of the intended object.

 I have not tried it myself.

I pay for and use Sharpcap for polar alignment and to run my Allsky camera.

It does both of these tasks well so is a bargain 

I do not use it for DSO though as it has no sequencer. 

 

I wasn’t clear enough, I wasn’t trying to suggest 30 to 60 seconds is a particularly short exposure. Only that in my case his formula for calculating the optimum exposure given the light pollution was an approximate match to the traditional histogram method. In other words any longer wasn’t going to glean much of a benefit.

[sloz1664]

Can you bin with a cmos sensor? Because no one has mentioned it.

Steve