M51 - CC

[Herzy]

This is my latest image, and I was VERY surprised by the result. I wasn't expecting the details that I got from only 2hours. I was going for 4 hours, but my camera battery died and I had to come inside. Now I just wonder how much would've been revealed from 4 hours. This target is a nice break from the extreme gradients I get when I was imaging to the south (M16). That is where Oklahoma City is, and it is literally a huge orange dome.

Anyways, here is the image. Comments or critiques are welcome and encouraged:

 

[wimvb]

Nice, I like it. Longer exposure sub frames will reveal more detail. More sub frames help reducing the noise. Either way, more is always better. But generally, I would say that longer sub frame exposures win over more and shorter exposures.

In your case, going for 4 hours would definitely make a difference.

[Herzy]

46 minutes ago, wimvb said:

Nice, I like it. Longer exposure sub frames will reveal more detail. More sub frames help reducing the noise. Either way, more is always better. But generally, I would say that longer sub frame exposures win over more and shorter exposures.

In your case, going for 4 hours would definitely make a difference.

I don't think that is the case. 60s subs will have twice the signal as 30s subs. The difference is that I have twice the amount of 30s subs so it averages to the same amount of signal gathered. Longer exposures don't capture more detail, longer integration does. 

[Jokehoba]

18 minutes ago, Herzy said:

I don't think that is the case. 60s subs will have twice the signal as 30s subs. The difference is that I have twice the amount of 30s subs so it averages to the same amount of signal gathered. Longer exposures don't capture more detail, longer integration does. 

Certainly longer integration yields better detail - all things being equal. But they are not. Noise builds as the square root of the exposure time. The integrated signal may be the same with your shorter exposures but the signal-to-noise ratio will be worse. Longer sub-frames are definitely better, hence why we use guiding...

[Herzy]

2 minutes ago, Jokehoba said:

Certainly longer integration yields better detail - all things being equal. But they are not. Noise builds as the square root of the exposure time. The integrated signal may be the same with your shorter exposures but the signal-to-noise ratio will be worse. Longer sub-frames are definitely better, hence why we use guiding...

So I should bump my 30s to 60s? My mount can take it but it is just more work where as 30s is much easier to handle and yields better stars (no trails).

Wouldn't more subs be better because your giving the stacking program more random noise to work with. With more examples of random noise it can remove the noise easier, right?

[wimvb]

Jokehoba is correct; longer subs gather more signal and more detail in the faint parts.

The reason you can see more detail in the processed image is that the image with lower noise can be stretched more. It's the stretched image that can show more detail. On the other hand, the image with longer exposure subs, doesn't need as much stretching to begin with.

In short, the longest integration time, with the longest single frame exposures wins.

 

[Jokehoba]

1 hour ago, Herzy said:

So I should bump my 30s to 60s? My mount can take it but it is just more work where as 30s is much easier to handle and yields better stars (no trails).

Wouldn't more subs be better because your giving the stacking program more random noise to work with. With more examples of random noise it can remove the noise easier, right?

Longer exposures gather more photons and hence more accurately represent faint objects. With short exposures, faint detail gets lost in the noise and the stacking program will simply treat it all as noise. 

With a good polar alignment you could push your mount to 2mins or more without resorting to guiding. Even if you get slight trailing, do it as an exercise and compare the results.

[Herzy]

18 minutes ago, Jokehoba said:

Longer exposures gather more photons and hence more accurately represent faint objects. With short exposures, faint detail gets lost in the noise and the stacking program will simply treat it all as noise. 

With a good polar alignment you could push your mount to 2mins or more without resorting to guiding. Even if you get slight trailing, do it as an exercise and compare the results.

Those photons are still captures but just less per picture. Once stacked the same amount of signal is present. Is it just more noise present as well?

[wimvb]

Yes. You can think of it this way;

a faint target will give you, say, 1 electron every 30 seconds on average from your sensor. Read noise will give you 1 electron per image (on average).

Then, each 30 second sub will have 1 signal electron and on average 1 noise electron.

Each 60 second sub will have 2 signal electrons and 1 noise electron

Each 120 second sub will have 4 signal electrons and 1 noise electron.

 

In stacking we average the pixel values of subs

the 30 second subs will have an average signal of 1 electron. The 120 sec subs will have an average signal of 4 electrons.

Only if you have extremely low read noise, will shorter exposures let you get away with the exchange of single frame exposure time against number of frames.

This is the case with the new generation of CMOS sensors, as in ZWO cameras. High sensitivity in combination with extremely low read noise allows for short exposure times with lots of subs.

Then, if your subs can be very short exposures, you can also take advantage from variations in seeing conditions, and select only the sharpest images for stacking. What you have then is "lucky  imaging". But this only works well on bright targets. To get faint detail, you still need longer subs to collect the signal.