limit of dslr exposure times

[Andywilliams]

Greetings all,

A question for you all.

I am wondering what is the longest useful time for DSLR exposure is before the sensor warms up and causes too much noise?

I am using my DSLR (canon 1000D) on a star adventurer but the more I read, the more I realise that there must be an upper exposure time limit before either sensor noise or coma become issues.

I usually use iso 800-1600 range and am currently in summertime and do not cool my DSLR.

Also I use a stock 70-300 zoom lens (usually at 170 focal length) that has a reasonable high f stop.

All thoughts and expertise are very welcomed.

cheers,

Andy

[Adam J]

As you suggest it totally depends on the temperature. In my experience, by the time you are at 0c sensor temperature 20min exposures at ISO800 are possible with a 1000D.

I would not use 1600ISO with the 1000D as its a boost setting as opposed to a native ISO.

Extrapolating from that, given that dark current doubles every 7c:

7c -> 14c = 10min

14c -> 21c = 5min

21c -> 28c = 2.5min

>28c....personally I would not bother with a DSLR but it depends on what you are happy with in terms of results.

 

[Andywilliams]

1 hour ago, Adam J said:

As you suggest it totally depends on the temperature. In my experience, by the time you are at 0c sensor temperature 20min exposures at ISO800 are possible with a 1000D.

I would not use 1600ISO with the 1000D as its a boost setting as opposed to a native ISO.

Extrapolating from that, given that dark current doubles every 7c:

7c -> 14c = 10min

14c -> 21c = 5min

21c -> 28c = 2.5min

>28c....personally I would not bother with a DSLR but it depends on what you are happy with in terms of results.

 

Thanks Adam......that is a concise answer to just what I was looking for .

I am glad to see that in the 14C-21C range (my current night temps) that I can get 5 minutes

I am very interested in your advice about not using 1600 iso for the 1000D...presumably the boost setting would increase noise for no benefits?

anyhow, thanks heaps for your advice.

cheers,

andy

[PeterCPC]

If you dither you can increase the exposure times and remove a lot of the noise generated.

Peter

[Delboy_Hog]

Great question Andy, I've been trying to figure out the same thing recently, and was planning on asking the very same thing.  It was helpful to see Adam's suggested times as that matches my very unscientific experiments more or less. 

With my 1100d, I've found I can do 15 - 20 min subs (under good dark skies) on cool nights (so under 10 degrees) at ISO 800.  They're *very* noisy though, so I haven't tried anything longer, and I don't yet know how many I'd need to stack before I got a useable image at the end of it.  That's the next stage in my investigation!  I've also wondered about whether it is necessary to let the camera cool down for a while after a long exposure.  With the noise, fear of losing a long sub to guiding issues, wind etc, I tend to opt for a larger number of slightly shorter subs to 'play it safe' - but then I've not been chasing anything faint enough to really need 20 minutes as yet.

As Peter mentions above, dithering (moving where on the screen the target object sits by a small amount) between frames has helped a lot with these shorter subs, so I'm looking forward to testing it on the longer subs.

On warmer nights (so 20+ degrees), I've seen some pretty ugly 5 min frames at ISO 1600 - so again, that matches pretty well with Adam's figures.

I'll be looking to pick peoples' brains on here about ways to cool a DSLR once we get to spring time here in the UK!

All the best,

Derek

[SteveNickolls]

Hi Andy,

Thanks for your thread. An interesting question you set. Now I haven’t a Canon 1000D to directly compare with you but I do use a Star Adventurer (SA) mount for imaging and which I use with two Canon DSLR’s, a 600D and 700D (modified), so here goes.

Since I got the SA last April I have been progressively increasing the lens FL used and exposure times and always my exposures are limited not by the camera or mount but by the light pollution conditions here. In November I began using an Astronomik clip–in UHC filter (as I image in a quite light polluted suburb just 3 miles out of Nottingham) which allowed for 5 minute exposures. Last Christmas I obtained a similar 12nm Ha filter to obtain better (and much longer exposure) images of nebulae.

The longest exposures so far have been 15 minutes using the Canon 85mm lens in my signature plus the Ha filter. These were taken at ISO 200 and f/2. I use BYEOS to image and the exposure histogram after 15 minutes showed a red channel peak below the first division indicating I can expose for longer another time. Atmospheric conditions also affect performance and exposure duration night to night. As we are in our winter season these images had an exif temperature of around 10 degrees C (or around 10 degrees above the ambient air temperature). In warmer months expect greater thermal noise to affect the camera sensor.

There are things you can do to help reduce noise such as switching off ‘Live View’ when on the imaging run and allowing a long pause between images (I currently use 60 second pause between images) which gives a pretty stable temperature profile to the images for later processing. I also try to match as best I can the dark frame temperatures to the light frames as this has an effect on overall picture quality. You also need to take enough dark frames else run the risk of actually adding noise.

If you drop your ISO from 800-1600 you will be able to image longer and if you use lenses with faster f/ratios can collect light quicker, so time while important is only one influencing factor.

Out of all of this the SA mount has performed marvellously and while at some point I must find a limit to its capability I have not yet found it at 85mm. Used within its design the mount is a really excellent and portable mount and keeps on producing.

I hope this is helpful but you see it isn’t just duration that affects the image or sensor temperature.

Good luck with your future imaging.

Cheers,
Steve

[Andywilliams]

12 hours ago, PeterCPC said:

If you dither you can increase the exposure times and remove a lot of the noise generated.

Peter

Thanks Peter,

I believe I can use dither with PHD and the star adventurer ? but its a new thing for me so its safe to say i'm on the beginning of the learning curve with dither !

[Andywilliams]

10 hours ago, Delboy_Hog said:

Great question Andy, I've been trying to figure out the same thing recently, and was planning on asking the very same thing.  It was helpful to see Adam's suggested times as that matches my very unscientific experiments more or less. 

With my 1100d, I've found I can do 15 - 20 min subs (under good dark skies) on cool nights (so under 10 degrees) at ISO 800.  They're *very* noisy though, so I haven't tried anything longer, and I don't yet know how many I'd need to stack before I got a useable image at the end of it.  That's the next stage in my investigation!  I've also wondered about whether it is necessary to let the camera cool down for a while after a long exposure.  With the noise, fear of losing a long sub to guiding issues, wind etc, I tend to opt for a larger number of slightly shorter subs to 'play it safe' - but then I've not been chasing anything faint enough to really need 20 minutes as yet.

As Peter mentions above, dithering (moving where on the screen the target object sits by a small amount) between frames has helped a lot with these shorter subs, so I'm looking forward to testing it on the longer subs.

On warmer nights (so 20+ degrees), I've seen some pretty ugly 5 min frames at ISO 1600 - so again, that matches pretty well with Adam's figures.

I'll be looking to pick peoples' brains on here about ways to cool a DSLR once we get to spring time here in the UK!

All the best,

Derek

Hi Derek,

thanks for your reply.

It seems to me looking on here and Astrobin that with DSLR people seem to use not much more than 300 seconds (5 minutes) and the multiples of that stacked, maybe this is the optimum?

Thanks for approaching the other important question that I forgot to ask, namely the optimum period of time to allow the sensor to cool....i'm afraid my knowledge of these electronics is lacking so all advice is welcomed.

cheers,

Andy

[bobro]

To echo the comments above, my 1000D is good for 10 min subs at circa 10 C without much noise. As exposure increases in excess of that time, grey spots of noise start to appear. You can see these in the greatly zoomed in part of an image below (960 sec sub, 7C @ ISO800). For scale, the blue dot is a noisy camera pixel. Not too bad really and stacking should help eliminate the noise. (The long exposure was due to an Ha filter).

I believe the Star Adventurer only supports guiding in the RA direction. So to maximise the dithering effect the mount could be dithered in RA for a few exposures, stopped, manually moved a little in DEC, and restarted. That will help randomise the dithering effect and prevent streaks which can happen with RA only dithering.

A great noise reduction feature that depends on dithering is Kappa Sigma clipping - DSS supports this. This ignores pixels which are removed from the average and so eliminates noise such as hot pixels.

I don't know where the camera sensor is on the 1000D, but my camera doesn't register a temperature rise with long exposures.

Hope this is useful.

Bob

 

[Andywilliams]

1 hour ago, bobro said:

To echo the comments above, my 1000D is good for 10 min subs at circa 10 C without much noise. As exposure increases in excess of that time, grey spots of noise start to appear. You can see these in the greatly zoomed in part of an image below (960 sec sub, 7C @ ISO800). For scale, the blue dot is a noisy camera pixel. Not too bad really and stacking should help eliminate the noise. (The long exposure was due to an Ha filter).

I believe the Star Adventurer only supports guiding in the RA direction. So to maximise the dithering effect the mount could be dithered in RA for a few exposures, stopped, manually moved a little in DEC, and restarted. That will help randomise the dithering effect and prevent streaks which can happen with RA only dithering.

A great noise reduction feature that depends on dithering is Kappa Sigma clipping - DSS supports this. This ignores pixels which are removed from the average and so eliminates noise such as hot pixels.

I don't know where the camera sensor is on the 1000D, but my camera doesn't register a temperature rise with long exposures.

Hope this is useful.

Bob

 

Many thanks for your interesting answer Bob.

I am really not up to speed with dithering at all.....I cant work out how I can move the mount manually in declination without spoiling my polar alignment?

10 minute subs sounds great !

cheers,

Andy

[Andywilliams]

My question seems to have elicited some interest......out of interest maybe some of you would like to post your longest single subs on DSLR?.....so that I can see what I can expect out of my camera?

clears skies and thanks again to one and all

[bobro]

12 hours ago, Andywilliams said:

I cant work out how I can move the mount manually in declination without spoiling my polar alignment?

Dec is the 'tilt' adjustment just below the camera attachment. It doesn't affect polar alignment. Don't rush to implement dithering - enjoy the journey!

[bobro]

12 hours ago, Andywilliams said:

maybe some of you would like to post your longest single subs on DSLR?

Well, here are 2 single subs of the California Nebula, taken at 8C, ISO800 with my astro modified 1000D. Both converted to jpg from raw. The first is an RGB of 480 secs and the second an Ha of 960 secs. Scope is 130/650mm reflector f5 with coma corrector. I think it will be difficult to make a direct comparision with your setup and sky, but the 1000D is a great camera!

[Tangoringo]

On 10/01/2018 at 03:42, Andywilliams said:

Thanks Adam......that is a concise answer to just what I was looking for .

I am glad to see that in the 14C-21C range (my current night temps) that I can get 5 minutes

I wish I could add some technical knowledge here....but alas I can’t....and even if I could with your night time temp statement I would have deleted it anyway as I’m currently sat outside in 5 degree drizzle getting thoroughly miserable.  Jealous doesn’t begin to cover my feelings right now!

[dph1nm]

On 10/01/2018 at 02:00, Adam J said:

I would not use 1600ISO with the 1000D as its a boost setting as opposed to a native ISO.

 Really - I am surprised. My 1000D is much better at 1600 than 800 in terms of banding and read noise. I read a long time ago that Canon use different circuitry on the 1000D at 1600 to improve the noise - however I have no idea whether this is true either!

NIgelM

[Andywilliams]

On 1/12/2018 at 02:42, dph1nm said:

 Really - I am surprised. My 1000D is much better at 1600 than 800 in terms of banding and read noise. I read a long time ago that Canon use different circuitry on the 1000D at 1600 to improve the noise - however I have no idea whether this is true either!

NIgelM

Thanks for your interesting reply Nigel.

i would be interested in a definitive answer on the iso performance of the 1000D ......anyone from Canon in here ? ?

[Adam J]

On 1/11/2018 at 13:42, dph1nm said:

 Really - I am surprised. My 1000D is much better at 1600 than 800 in terms of banding and read noise. I read a long time ago that Canon use different circuitry on the 1000D at 1600 to improve the noise - however I have no idea whether this is true either!

NIgelM

Its not about noise its about signal to noise. The noise at 1600 becomes more random so looks better visually but its not better in reality. Its so much more noisy that in reality the amp glow at the bottom right of the image gets hidden under the read noise. 

[newbie alert]

On 10/01/2018 at 20:06, Andywilliams said:

Thanks Peter,

I believe I can use dither with PHD and the star adventurer ? but its a new thing for me so its safe to say i'm on the beginning of the learning curve with dither !

If you're delaying 60 secs between exposures then you will get a natural dither anyway

[dph1nm]

On 1/26/2018 at 12:00, Adam J said:

Its not about noise its about signal to noise. The noise at 1600 becomes more random so looks better visually but its not better in reality. Its so much more noisy that in reality the amp glow at the bottom right of the image gets hidden under the read noise. 

Not so - the read noise (in electrons) is lower at ISO1600 that at lower ISO for the Canon 1000D. 

NigelM

[Adam J]

4 hours ago, dph1nm said:

Not so - the read noise (in electrons) is lower at ISO1600 that at lower ISO for the Canon 1000D. 

NigelM

Every single DSLR ever made has a lower read noise as the ISO is increased with most having their lowest read noise at their highest ISO, it does not mean that you should be using the camera in its highest ISO every single time.

I mean really its 6.4e at IOS1600 and 5e at ISO800 (so rubbish or slightly less rubbish but still really rubbish), but while most of the noise at ISO800 calibrates out well with bias the more random signal at ISO1600 does not calibrate out well, in any case would you really want to lose half your dynamic range for a 22% reduction in read noise given that ISO800 is already unity gain? Because of the generally high read noise and larger pixels on the 1000D its a camera much better suited to longer exposures than many short exposures and in that case you want dynamic range and read noise becomes less significant. Not the best idea to half your max exposure for a totally insignificant reduction in read noise as if you have to do twice as many subs then thats not a great idea with a high read noise camera.

 

[Andywilliams]

6 hours ago, Adam J said:

Every single DSLR ever made has a lower read noise as the ISO is increased with most having their lowest read noise at their highest ISO, it does not mean that you should be using the camera in its highest ISO every single time.

I mean really its 6.4e at IOS1600 and 5e at ISO800 (so rubbish or slightly less rubbish but still really rubbish), but while most of the noise at ISO800 calibrates out well with bias the more random signal at ISO1600 does not calibrate out well, in any case would you really want to lose half your dynamic range for a 22% reduction in read noise given that ISO800 is already unity gain? Because of the generally high read noise and larger pixels on the 1000D its a camera much better suited to longer exposures than many short exposures and in that case you want dynamic range and read noise becomes less significant. Not the best idea to half your max exposure for a totally insignificant reduction in read noise as if you have to do twice as many subs then thats not a great idea with a high read noise camera.

 

Hi Adam,

Thanks for your information.

So to recapitulate your main points here.....less subs in amount but longer subs and shot at a lower iso....this will give better results with the 1000d, yes?

In that case, does the same apply for the 40d ?.....my other camera is a modified 40d.

cheers and clear skies,

Andy

[Adam J]

22 minutes ago, Andywilliams said:

Hi Adam,

Thanks for your information.

So to recapitulate your main points here.....less subs in amount but longer subs and shot at a lower iso....this will give better results with the 1000d, yes?

In that case, does the same apply for the 40d ?.....my other camera is a modified 40d.

cheers and clear skies,

Andy

Yes for high read noise cameras you are better off with longer subs conducted at lower ISO so as to take advantage of the higher dynamic range and avoid accumulating read noise over large numbers of images. The only negative to that with the 1000D is the 12-bit a/d which will mean need more subs / intergration in total to smooth out the histogram when performing a heavy stretch...but its an old sensor, you cant have it all. 

The 40D has almost exactly the same sensor as the 1000D, but it actually has a 14bit analogue to digital converter and lower read noise in comparison to the 1000D, but looking at the read noise and dynamic range charts for it I would still chose ISO800, however the 14-but a/d and the slightly lower read noise means that I would use slightly shorter subs with the 40D than the 1000D.

The 40D is most likely the better camera if both are astro-modified. 

 

[Adam J]

On 1/26/2018 at 12:11, newbie alert said:

If you're delaying 60 secs between exposures then you will get a natural dither anyway

Not exactly, what you end up with is something called walking color noise and its not good. A proper large scale dither is the way to go. 

[Adam J]

On 1/10/2018 at 20:13, Andywilliams said:

Thanks for approaching the other important question that I forgot to ask, namely the optimum period of time to allow the sensor to cool....i'm afraid my knowledge of these electronics is lacking so all advice is welcomed.

 

The thinking behind letting the sensor cool between subs is flawed in my opinion. The reason for that is that the camera takes the sensor temperature reading at the start of the exposure and so suckers you into thinking that the temperature for the entire exposure has been lowered. Though some experimentation in this area I can tell you that after only 60 seconds or so the temperature is back up to where it started and so if you are doing a 5 min sub you are gaining a insignificant reduction in thermal noise and integrated over only 1/5th of the total exposure time. Yet you are losing large amounts of total integration time as a result of the pauses. 

This is worse in the case of the 1000D which is one of the coolest running cameras canon ever made. 

[tooth_dr]

2 hours ago, Andywilliams said:

Hi Adam,

Thanks for your information.

So to recapitulate your main points here.....less subs in amount but longer subs and shot at a lower iso....this will give better results with the 1000d, yes?

In that case, does the same apply for the 40d ?.....my other camera is a modified 40d.

cheers and clear skies,

Andy

I have a modded 1000d and a modded 40d too! Great thread this