Total integration vs individual subs.

[nickp87]

Hi all.

Which of the 2 is more important?

Are 1000 5 second subs going to give you the same data as 80 60 second subs?

It would seem to me that the longer subs could capture fainter image data but I could be wrong. 

Thanks. 

[sharkmelley]

You will capture the same amount of light with both approaches.  But the stack of 1000 subs will have 1000 contributions of read noise whereas the stack of 80 subs will have only 80 contributions of read noise.  Almost certainly the stack of 1000 subs will be noisier therefore making it a bit more difficult to detect fainter objects.

[Budgie1]

Also think about the amount of storage space you would use/need.

My ZWO ASI1600MM Pro camera produces subs of around 30mb in size, no matter how long the exposure is. So, if I take 1000 subs, that's 30Gb of data I have to store to produce one stacked image. If I only take 80 longer subs, that's just 2.4Gb of data for the same integration time.

Not only that, the more subs you give your stacking software, the harder it has to work and longer it will take to produce the stacked image.  

[tomato]

Longer subs of several minutes will be more prone to guiding errors, and in the UK, passing cloud. These days I nearly always lose some subs to cloud, so shorter exposures minimise the pain. I used to do 5 min LRGB  subs with CCD cameras but use 2 minutes with the CMOS cameras. 
 
This is a good compromise on data storage and processing power, thousands of subs per final image would be challenge, that’s for sure.

[nickp87]

12 hours ago, sharkmelley said:

You will capture the same amount of light with both approaches.  But the stack of 1000 subs will have 1000 contributions of read noise whereas the stack of 80 subs will have only 80 contributions of read noise.  Almost certainly the stack of 1000 subs will be noisier therefore making it a bit more difficult to detect fainter objects.

Thanks, i hadn't thought about noise.

 

11 hours ago, Budgie1 said:

Also think about the amount of storage space you would use/need.

My ZWO ASI1600MM Pro camera produces subs of around 30mb in size, no matter how long the exposure is. So, if I take 1000 subs, that's 30Gb of data I have to store to produce one stacked image. If I only take 80 longer subs, that's just 2.4Gb of data for the same integration time.

Not only that, the more subs you give your stacking software, the harder it has to work and longer it will take to produce the stacked image.  

Thanks Martin

Storage space isn't a big issue, but your right it is  lot more data to store, my camera puts out about 19mb for a typical 10s exposure so would run at around 20GB. 

 

32 minutes ago, tomato said:

Longer subs of several minutes will be more prone to guiding errors, and in the UK, passing cloud. These days I nearly always lose some subs to cloud, so shorter exposures minimise the pain. I used to do 5 min LRGB  subs with CCD cameras but use 2 minutes with the CMOS cameras. 
 
This is a good compromise on data storage and processing power, thousands of subs per final image would be challenge, that’s for sure.

Thanks all for the replies. I thought longer was going to capture more/fainter data which appears incorrect, looks like I am going to have to find a good compromise of sub length and minimizing issues, I also have to work within the constraints of my az gti mount  

[vlaiv]

11 minutes ago, nickp87 said:

I am going to have to find a good compromise of sub length and minimizing issues

There is a point where going with longer subs gives you diminishing returns because of the way the noise adds.

It will depend on your conditions and equipment. As long as you have noise source that "swamps" read noise after some time - going longer is not feasible.

There are two major noise sources that can swamp read noise. Thermal / dark noise (noise associated with dark signal) and LP noise (noise associated with sky background glow or light pollution).

If you use cooled astronomy camera - first one is non issue and what remains is LP. Even if you use DSLR - in most cases LP noise will be larger of the two (unless you are somewhere really dark and work at higher sampling rates).

Look up how to determine max useful exposure length from one of your subs. There are several videos on Youtube that describe this and you can also find threads here on SGL that do the same (I believe that SharpCap also has functionality to do this automatically for you).

After that - decide if you want to go even shorter because of other factors (mount performance, chance of ruined subs and so on), or perhaps longer to save storage space (but that won't contribute to image quality).

[ollypenrice]

To pick up on sharkmelley's point about the number of doses of read noise, an important factor is the level of read noise of your camera. Excellent as they were and still are, CCD cameras had considerable read noise and really did best with long subs, guiding and sky permitting. Modern CMOS cameras have low read noise and don't, therefore, benefit from such long exposures.

Olly

[ian_bird]

On the other hand - if you are using a 3nm Ha filter - longer exposures are your friend indeed!

Edited December 12, 2023 by ian_bird