1: Introduction dynamics
When taking astrophotographs on weak objects like nebulas you want the weak parts of the nebula of course be visible, but if you take long exposures at high ISO setting or gain the stars will be oversaturated with no color information. To me it look very boring.
A CCD camera normally have 16 bit ADC, Analog Digital Converter. With that you use more of the information that is stored in the sensor after an exposure. A DSLR or CMOS camera has 10 to 14 bit ADC, that limit the signal from the sensor, with high gain it clip all higher signals even if it's stored in the sensor.
Can we do anything about that? Modern CMOS sensors have very low readout noise. Then you can divide the exposure in shorter exposure time. It doesn't increase the total readout noise that much as it did in old CCD sensors. With many short exposures you don't over saturated the sensor and then have better dynamic in the data.
2: Dynamic weak part of a object
Here I have done two test photos. One with 16 x 120 seconds exposures, and another with 65 x 30 seconds exposures. Both have a total exposure time of of 2 minutes. My camera is a Canon EOS 6D and that camera has low read out noise. It depends on which ISO setting I chose. Higher ISO lower the noise because Canon has built in a pre amplifier that raise the signal above the ADC readout noise. But higher ISO limit the dynamic range. It has to be a compromise, I have the camera set to ISO 1600. At ISO 1600 it has a gain of 3, 3 ADU step for each electron it detect. The 14 bit ADC has a upper limit of 16000 ADU, subtract the bias and 140000 is left. Divide by 3 and you have the max limit of about 5000 electrons. The sensor can store about 60000 electrons so it's a shame that you can't use all of the sensor's signal. But still it works very good.
When reducing the exposure time from 120 seconds to 30 seconds you can detect 4 times higher signals. But will it affect your ability to detect weak signals? It depends very much how dark your site is and how well good camera is, especially important is that it has a low readout noise. With my setting above the readout noise is 3 electrons.
One reason I do this test is that I couldn't polar align my Star Adventurer good enough. With 120 seconds exposure you can see that the stars drift because of this. With shorter exposure time it will not be that problem.
The first photo is a stack of 16 x 120 seconds photos:
The second photo is a stack of 65 x 30 seconds photos:
The difference in color are most because of me, couldn't set it up exactly the same. Can you see any difference in the nebulosity? I have a gamma of 5 and the white point is set low to limit the range around the nebula's levels.
3: Dynamic high signal strength
The above example was when you want get the weak signal visible, but how about the strong signals we get from stars or a galaxy core? Same photos as above, but now I have raised the high clipping point to when the one with short exposures clip. And also set the color saturation to very high level.
If you don't recognize the object, it's Pleiades, Messier 45.
The last one has more of the color information undisturbed from clipping effects. One problem is of course how to process the image to both preserve the weak signal and the strong signals. The effect of bad tracking is also less in the second photo.
4: Line graphs and histogram
I often use the free software Fitswork when analyzing image data. Here I use it to do a line graph. In this case it's a line along the X-axis, the Y axis is the strength of the signal. The object is a strong star that is over saturated.
Here how it looks, the top of the star has been clipped because of limits in the dynamics of the camera.
Here you can see the star that's analyzed. There is also a histogram that I have set to correspond with the photo with many short exposures. At the bottom of the histogram you see how the signal only use about 1/5 of the dynamics compare to the many short exposures photo. All the weak signals and background is the bump most to the left.
This photo is the one built by many short exposures, you see that the star clip at a higher level (multiply the Y-axis scale by 4). The star is more narrow, but even in this case the signal is clipped, but less. In the histogram you see how the signal use more of the available dynamic, signal all the way to the right.
When stacking the photos I used the median method. That's why the scale of the axis not correspond to the difference between the photos.
I hope this give you something and not only confusing.