The ultimate proof that arbitrarily short exposures can work, if you have enough of them

[aparker]

Martin Meredith especially will find this thread on CN entertaining:

http://www.cloudynights.com/topic/535754-ngc-6543-20000x400ms-with-250mm-with-asi224mc/#entry7193028

 

[HiloDon]

How about this one.  Apparently, neither EAA nor AP want this.

http://www.cloudynights.com/topic/533802-m64-2400-x-1-second-16/

There are some interesting things happening in imaging, whether it's called AP, EAA or VA.

Don

Here was my response to this:

http://www.cloudynights.com/topic/533802-m64-2400-x-1-second-16/?p=7167345

[Martin Meredith]

Thanks Alex. It is an interesting situation for the EAA approach to find itself in the middle of these two very different approaches to AP -- the very brief and the very long subs.  I'd love to think that the very brief approach will become mainstream, as it will surely bring down the entry price and complexity of getting into AP and EAA.

As I see it, even with low read noise cameras there are some practical issues to solve before this approach can be incorporated into near real-time observing. These guys throw away a large proportion of their subs, which might be not be tolerable for near-real time viewing, and might be difficult to achieve online anyway if the selection criterion is based having a huge batch of them recorded beforehand. Better mounts would help. The quantity of stars recorded per exposure might not be enough in some fields to enable the use of robust stacking. One solution is to increase the aperture.

We need to see more explorations in the tradeoffs of aperture and tracking accuracy, and the limits on faint objects which succumb to the approach. We're ideally placed in EAA to explore this space. I tend to use the shortest subs I think I can get away with, and for things like globs that can be as little as 1s even with a moderate read noise camera like the Lodestar. Trading pinpoint, unsaturated stars for read noise is perfectly acceptable to me. Here's a recent M5 with comparing 5x10s and 59x1s subs. The fixed pattern diagonal lines in the second are barely noticeable to me (your screens may vary!). No darks were used in either thanks to Paul's automatic defective pixel removal, which makes this kind of multi-lenth exposure testing so much easier.

 

 

 

 

Martin

 

 

[shirva]

great topic,video capture is changing fast,we all know that,the cameras are getting better all the time,instead of us looking for a more sensitive cctv cameras,astronomy camera manufacturers are creating cameas for our needs,,the software makes the big differance,the fast tube telescope is the fundamental bit of kit and a good tracking mount,,its our capture technique that we need to refine and if we look more closely to how astrophotographers do it,we are on the right track,,,,its about  pardon the pun ( making every photon count) we are loooking to take shorter exposure to reduce noise,but i believe we need to combine these with some longer exposures,,the new software developed for the usb type cameras will allow us to do this..no big dramas having to mess around with analogue camera menu pads..he way things are we are just looking at speeded up astrophotography now.in near real time..weve got to keep moving with the times now lol.

[Stub Mandrel]

Take this with a pinch of salt...

It seems to me that there is interesting 'middle ground' - using short-ish exposures and DSLRs to take small or large numbers of images.

I think the ALT-AZ imagers are making rapid strides into the bottom part of this area, but I can't help thinking that the opportunity exists for people with unguided EQ or Alt-Az setups to take very large numbers of subs, say a few hundred or even thousands over multiple nights.

 

[shirva]

Totally agree stub mandrel, im moving into dslr capture using backyard eos,I messed around with the free trial version before buying the software,pairing this with astrotoaster for near real time video.the backyard eos has the feature to do night after night captures .

[aparker]

On 5/7/2016 at 04:01, Martin Meredith said:

Thanks Alex. It is an interesting situation for the EAA approach to find itself in the middle of these two very different approaches to AP -- the very brief and the very long subs.  I'd love to think that the very brief approach will become mainstream, as it will surely bring down the entry price and complexity of getting into AP and EAA.

As I see it, even with low read noise cameras there are some practical issues to solve before this approach can be incorporated into near real-time observing. These guys throw away a large proportion of their subs, which might be not be tolerable for near-real time viewing, and might be difficult to achieve online anyway if the selection criterion is based having a huge batch of them recorded beforehand. Better mounts would help. The quantity of stars recorded per exposure might not be enough in some fields to enable the use of robust stacking. One solution is to increase the aperture.

We need to see more explorations in the tradeoffs of aperture and tracking accuracy, and the limits on faint objects which succumb to the approach. We're ideally placed in EAA to explore this space. I tend to use the shortest subs I think I can get away with, and for things like globs that can be as little as 1s even with a moderate read noise camera like the Lodestar. Trading pinpoint, unsaturated stars for read noise is perfectly acceptable to me. Here's a recent M5 with comparing 5x10s and 59x1s subs. The fixed pattern diagonal lines in the second are barely noticeable to me (your screens may vary!). No darks were used in either thanks to Paul's automatic defective pixel removal, which makes this kind of multi-lenth exposure testing so much easier.

 

 

 

 

Martin

 

 

Nice exemplary images, Martin.  The 59x1 sec is clearly superior.  Probably a combination of less saturation from bright stars and reduced impact of seeing?  I'm impressed you are able to get SL3 to keep up with stacking at 1sec.  With Ultrastar-size images, I'm pretty sure it won't, even on a modern macbook.   I've found that letting the capture get ahead of stacking is the "kiss-of-death" for SL, for me.  I end up having to force-quit eventually.  I may try this with 2x2 binning, assuming it ever stops raining in Boston, as the stacking with binned images is much, much faster. 

[stash_old]

14 hours ago, shirva said:

Totally agree stub mandrel, im moving into dslr capture using backyard eos,I messed around with the free trial version before buying the software,pairing this with astrotoaster for near real time video.the backyard eos has the feature to do night after night captures .

"backyard eos has the feature to do night after night captures" I use BYEOS 3.1.6 what setting are you referring to?      You can of course use plate solving to position scope (assuming EQ) from the last nights image from a previous night - Astrotortilla. APT can also do this via Astrotortilla or its inbuilt connection to Platesolve - both work very well via DSLR and APT can support CCD but I have never tried this function with CCD.

[shirva]

I havent used the feature yet stash,if you go on there website it will explain it.I have been learning the imaging and planetary capture sections just now but will get round to it at some stage. I bought the full version on mid april

[stash_old]

On 5/8/2016 at 17:02, shirva said:

I havent used the feature yet stash,if you go on there website it will explain it.I have been learning the imaging and planetary capture sections just now but will get round to it at some stage. I bought the full version on mid april

Tried out Astrotoaster last night on a couple of objects - here's M51 - strange lines of red,blue all over the screen. Not sure if that is my DSLR or Astrotoaster or my settings. One thing it slows PC's down no end - not unexpected. Verdict stick out but its another string to the bow of EAA/video. This was a screen print of one try.

[stash_old]

However I have just run Astrotoaster again against the same files but this time I hit recalibrate stack button and the lines have gone so must be my settings last night. NOTE as I didn't know what i was doing I set up 300 secs 400ISO x 5 . some of the other objects,e.g. M13, were set against a lot shorter times (like 60secs).

[shirva]

Well impressed stash,cant wait till im up and running,, ive just bought an heq5 pro and have fitted ed80 with sw 0.85 focal reducer and canon 600d with meikie pro grip,,startravel 102,and a second canon 600d  ,orion mini guider with qhy5l11,waiting on hitec heq5 cable arriving from flo.. ascom platform,eqmod and cdc,,will also be running byeos  and sharpcap,,,st102 will be used with canon 600d or video cam,,various options on this set up,, still a lot to learn and test.

[Martin Meredith]

On 8 May 2016 at 14:17, aparker said:

Nice exemplary images, Martin.  The 59x1 sec is clearly superior.  Probably a combination of less saturation from bright stars and reduced impact of seeing?  I'm impressed you are able to get SL3 to keep up with stacking at 1sec.  With Ultrastar-size images, I'm pretty sure it won't, even on a modern macbook.   I've found that letting the capture get ahead of stacking is the "kiss-of-death" for SL, for me.  I end up having to force-quit eventually.  I may try this with 2x2 binning, assuming it ever stops raining in Boston, as the stacking with binned images is much, much faster. 

Yes, probably a bit of both as tracking was Ok that night and it was windless. 

Stacking in SL3 seems much more of a constant speed process than in earlier versions. I have a relatively new Air and it is noticeably faster than the older one. I've never had to force quit SL (I've had to drag myself away to get some sleep occasionally...). Ultrastar + 2x2 binning sounds like a good bet (for lots of reasons).

Martin