Exposure time v number of subs

[Dealy (Kev)]

Thanks Rick

That's a good comparison. In the original link in this thread the bloke who wrote it said there will be little difference in SNR after 20 subs. Your examples clearly show that's not the case. So the more subs/darks/lights etc the better to reduce noise.

There would of course be a limit to this, but you've shown it's a lot more than 20 subs.

The signal looks about the same from 20 to 40 subs, so he may be right by saying the only way to increase signal is to take longer exposures, not more subs.

If that's the case then the resulting image from 10hrs of 15min subs will produce a very clean and smooth 15min exposure, whereas 5hrs of 30min subs will produce a somewhat noisier image but should show more detail ie more faint nebulosity for example.

Is that right?

I realise taking very long subs causes other problems like bloating, over exposure of bright areas, and tracking becomes more difficult. So there is obviously a limit to this based on the target, equipment used and the skill of the imager.

I haven't read all the articles that have been referred to in this thread yet but I will. There's a lot of info to get my fading brain around.

Regards, Kev


.

[RickS (Rick)]

Kev,

I started writing a long reply but it's probably better if you read some of the articles. They will say the same stuff but more articulately

Once you understand shot noise, read noise and sky limited exposures you'll be well-equipped to decide how long your subs should be and how many you need.

Cheers,
Rick.

[Dealy (Kev)]

Thanks Rick

[sil (Steve)]

Kev, I think you're thinking in terms of photography as many people do. longer exposure time in practice can reveal more distant (ie faint) objects in the signal. Stacking improves the SNR which gives you extra flexibility for teasing out faint signal from the data before the noise overwhelms it. Each source frame I take is rarely exposed anything like what I end up with, which is what you aim for in photography. The way I approach my deep space AP is capture as much signal as I can and understand what various tools I can use to make best use of the signal. There are some amazing things possible with PixInsight such as removing the stars from an image so you can work on the faint signal to make it visible then add the stars back. Most astrophotography people talk about are just photos, they rely on someone's personal opinion of what the image should look like and may not be accurate to be able to take scientific measurements.
In any image you capture there is a lot of signal that may not be apparent or wanted. Increasing exposure time will get you more signal (even more of the signal you aren't aware of) but the quality of the signal is also effected by tracking quality to avoid smearing it. You can of course over expose and reduce the usable signal. More frames can improve the SNR which can give you a cleaner/smoother image of a well exposed target but also let you do more processing to bring out fainter signals. It all depends on what you are trying to show with the shot. One of the things I enjoy with my AP is just exploring the data, I plate solve everything and see if anything of interest is within the field of view then I can try to reveal the faint fuzzies. Since my stroke a year ago I can't use my telescope and my AP is mostly just with a DSLR on a tripod, or a high end "compact" on a vixen polarie and I'm constantly amazed at what I can capture. The Southern Pinwheel Galaxy is one of my most proud shots using DSLR +500mm lens and its not visible in any of my shots, but once registered and integrated there is plenty of signal to pull out the spiral galaxy. If you try to make the most of your capture hardware and have a bit of an understanding of your data and what you want to achieve with it you will be amazed at what you can get, buying more gear might not be the answer to improving your imaging. I'm no expert and plenty of people with better gear and good understanding make better shots but then again I've seen plenty of shots from people with more expensive gear that are just garbage compared to what I can achieve.

So i guess, exposure time vs more subs = depends

[Dealy (Kev)]

Thanks Sil

Very informative

The part that had me most confused I guess was why people were taking many, many hours of exposures when in the link on the original post the writer of the article stated there wasn't any benefit of more than about 20 subs.

Rick proved that was incorrect with his examples.

I'll do some more reading I think, and keep practicing.

Regards, Kev

[Amaranthus (Barry)]

Subs = number of subexposures
Integration = total length of capture

So 20 x 30 sec subs gives 600 sec (10 min) of integration
Whereas 20 x 10 min subs gives 12,000 sec (200 min) of integration

Same number of subs, but the latter gives a vastly higher signal-to-noise ratio

[Shiraz (Ray)]

Originally Posted by Dealy

Thanks Sil

Very informative

The part that had me most confused I guess was why people were taking many, many hours of exposures when in the link on the original post the writer of the article stated there wasn't any benefit of more than about 20 subs.

Rick proved that was incorrect with his examples.

I'll do some more reading I think, and keep practicing.

Regards, Kev

good idea Kev - I suggest that you look at alternatives to the tome that you linked to - I think that the author has misunderstood quite a few of the basic ideas behind image stacking.

a few comments:
- provided there is enough signal that the shot noise overwhelms read noise, 20 subs of 5 minutes will produce the same result as 100 minutes in a single exposure. Provided read noise is controlled, the important measure is the total number of photons you detect and they don't care if you use lots of exposures or an equivalent single long one.
- stars do not "bloat" from long exposure with an anti-bloom camera (not even sure what "bloat" means). The cores may saturate, but the star sizes will be determined by the seeing and the optics, and will be the same for either a stack of multiple exposures or one equivalent long exposure.
- there is no magic number of subs beyond which you get no benefit - every sub helps. For example, last night I took an image of Saturn that required 22,000 subs, of which I stacked about 4000. Even though I could have removed any one sub without seeing a major difference in the result, nonetheless every selected sub added a tiny little bit to the outcome.

Stick with experts like Craig Stark or the Starizona website.

Following on from Ricks post, the attachment shows how noise decreases going from 1, 20, 100 subs - it always improves with more data.

[Dealy (Kev)]

Thanks Ray,

It's all beginning to make sense