#181  
Old 30-05-2016, 10:05 AM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by gregbradley View Post
I've been following this thread with interest.

It seems to me the following are critical with this camera:

1. Highlights will blow out if the gain is too high. I think gain is the same as ISO on a digital camera so then when you increase the gain you lower the dynamic range. Also typically there is usually a sweet spot in the ISO range of a digital camera. Where the signal is boosted enough to raise shadows but not introducing much noise and not lowering dynamic range too much.

So think in terms of ISO and you would aim for the lowest ISO to get the shot as noise is lowest at lowest ISO and dynamic range is greatest. If you look at dynamic range graphs of digital cameras you'll see its not a straight line drop with increases in ISO. It tends to worsen more quickly the higher you go.

From what I have seen so far this seems to be the most critical point as 2000x 1 second subs at high gain of M51 with a totally blown core and surrounds is never going to cut it against a high quality CCD camera.

2. Again the low read noise is probably down with the low gain as the higher gain probably amplifies all the noise including the read noise.

One way of working out the optimum gain (ISO) for this sensor would be to look at the DXO scoring for the EM1 this sensor came from. They measure what is the highest ISO you can use before image degradation starts. I haven't looked it up but its probably around ISO1200 for a small sensor like this. So trying to relate that to the gain settings you might be able to find a table of gain and ISO from someone who tests various cameras that way.
It would be valuable to know.

Most DSLR imagers use around ISO800 when stacking lots of images so whatever that equates to would be good to know also. Its probably quite low.

It seems that gain control is a potential image wrecker and should be treated with a little is a lot type of approach and less is better.

Greg.
nope, the read noise goes just the opposite way Greg - it reduces with increasing gain and is lowest at highest gain. However, the effective wells are not very big at high gain, so you need short subs - which you can do because of the low read noise.

The graphs posted earlier and available on the ZWO website show exactly how the various parameters vary with gain. ISO is not used as a measure of gain with these cameras, actual amplifier gain is used. This chip has made a complete transition from the world of DSLRs to the world of CCD measurements.
Reply With Quote
  #182  
Old 30-05-2016, 10:30 AM
Slawomir's Avatar
Slawomir (Suavi)
Registered User

Slawomir is offline
 
Join Date: Sep 2014
Location: North Queensland
Posts: 3,240
I also have been following this thread with a great interest. Regardless of knowing the exact numbers up to 4th decimal place for gain, QE, dynamic range etc, it appears that amateur astrophotographers have indeed a bright future, not only in terms of light pollution, but also in terms of having access to relatively inexpensive and very capable astro-cameras.

If I did not have my current camera, I would be all over this one. But it is good to know that when the day will come that my camera will retire, I will be possibly able to afford a replacement camera that hopefully will have better parameters at a lower cost :-)

I think it is not long before other manufacturers such as Atik, SX and QSI, just to name a few, will also be offering CMOS cameras.

Now, if I only could 3D print an AP mount...
Reply With Quote
  #183  
Old 30-05-2016, 10:42 AM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by glend View Post
Well Ray, good to hear it has arrived. I will be expecting some images tomorrow morning! Have fun.
PS, it would be nice to get your test results on M8 to compare settings, processing, etc.
thanks Glen. Just unpacked it and fired it up.

Whaaat the!! it just instantly started working. that was nice change - no software fuss at all.
Reply With Quote
  #184  
Old 30-05-2016, 10:51 AM
trent_julie's Avatar
trent_julie
Registered User

trent_julie is offline
 
Join Date: Nov 2008
Location: Canberra
Posts: 581
Silly Question,
Given the sensitivity of these cameras; Could the colour variants overcome the advantages of monos when considering filters and the like?

Ie could I use a colour camera with a narrowband filter (something which was not recommended in the past).

Trent
Reply With Quote
  #185  
Old 30-05-2016, 11:13 AM
codemonkey's Avatar
codemonkey (Lee)
Lee "Wormsy" Borsboom

codemonkey is offline
 
Join Date: Jul 2013
Location: Kilcoy, QLD
Posts: 2,058
Quote:
Originally Posted by Shiraz View Post
naah, keep it. The QEs of the Sony CCDs still give them a significant edge in conventional imaging with medium size scopes. The big Kodak chips also still have their place because they are best matched to large, long focal length scopes or widefield with scopes like the FSQ106.

However, the 1600 opens up some entirely new ways of doing things with medium size scopes and even camera lenses (including high resolution DSO, solar and planetary imaging) and doesn't appear to give up much/any performance in conventional long-sub DSO imaging, if you want to use it that way.

Mine just arrived at the local PO, so now begins the process of changing cameras and sorting out all the little issues like focal plane tilt, vignetting, software etc - groan. One nice thing though is that the back focus is almost identical to that of the 694, so, for initial testing, I will just be able to unscrew one camera and screw on the other (that's the theory anyway)
It's all about trade-offs. I've been wanting better sampling, so the 3.8u pixels are attractive, vs the 4.54u of the 674. I've been wanting a bigger FOV for nebulae. The lower read noise will offset the reduced QE to an extent, but it's obviously impossible to say how well because no one knows what the QE is.

The other thing to remember is that by virtue of having a much bigger sensor, scaling down for presentation has the effect of averaging the image, improving the effective SNR.

Unless something big becomes apparent about these cameras, I'll probably sell the 674 to finance an ASI1600.
Reply With Quote
  #186  
Old 30-05-2016, 11:27 AM
glend (Glen)
Registered User

glend is offline
 
Join Date: Jun 2013
Location: Lake Macquarie
Posts: 7,033
Lee, John Rista (CN) has published a graph on the CN thread comparing the QE of the KAF8300 and the ASI1600MM. They are very close, in some part of the spectrum ASI is higher, peaking above 60, and in others like the red end the 8300 has a slight edge. Don't ask me how he did it.
Reply With Quote
  #187  
Old 30-05-2016, 11:36 AM
codemonkey's Avatar
codemonkey (Lee)
Lee "Wormsy" Borsboom

codemonkey is offline
 
Join Date: Jul 2013
Location: Kilcoy, QLD
Posts: 2,058
Cheers Glen. He's plotted the relative QE, which is definitely useful info, but it's still not absolute QE.

My 674 has a peak absolute QE of 77% (from memory), but it's much flatter at the peak, so the shape of the relative QE tells me that I'll be losing out in red and blue, but I still don't know by how much because it doesn't tell me that the peak absolute QE of the 1600's sensor is X%.

It's estimated to be about 60%, but last I looked ZWO still didn't know exactly what it was.
Reply With Quote
  #188  
Old 30-05-2016, 11:48 AM
glend (Glen)
Registered User

glend is offline
 
Join Date: Jun 2013
Location: Lake Macquarie
Posts: 7,033
Lee, I thought he said he had calculated it. The chart is in Post #1115 here:

http://www.cloudynights.com/topic/53...a-test/page-45

and on page 47, Post #1159, he explains it all - not that I understand the math.

Honestly the QE is so much better than my DSLR that I won't split hairs on the absolute value. I don't know that the intention was to kill off 674's market.

As to Jon Rista work, I won't call him ' The Oracle' but he is doing some good work on this topic. He has now revised his 'Sweet Spot' recommendation, quoted here in a new post:

"There is definitely more testing needed to find the sweet spot for this camera. My original guess of a gain setting of 60 was close, but not dead on to maximize DR and minimize noise. I tested a gain of 70 last night, with bias offset of 12, and that seems to be closer but still not right on the ball. "
I will try 70 and 12 tonight on various sub lengths, and targets.

And thanks to Jerome for reprocessing my M8 image to demonstrate that the core is not blown out - it was just my hacker processing that caused it.
Reply With Quote
  #189  
Old 30-05-2016, 11:58 AM
codemonkey's Avatar
codemonkey (Lee)
Lee "Wormsy" Borsboom

codemonkey is offline
 
Join Date: Jul 2013
Location: Kilcoy, QLD
Posts: 2,058
Nah, the graph is the plotted relative QE. The absolute is not published, and I believe it's non-trivial to determine? That's what ZWO have said anyway.

Jon's post on page 47 says If we assume we gather about the same charge per unit area, then we might gather 20.2e- total (object and skyfog) per 5.4 microns square in a short exposure of say 60 seconds.

That assumption essentially says "if we assume they have the same QE". The 1600 is assumed to be pretty much bang on with the kaf-8300, with different relative differences, but it's still assumed, no one actually knows.

Thinking about the whole SNR thing though... read noise adds in quadrature in an image. If we take zero gain on the 1600, we have a read noise of about 3.6e-/s. Squaring this we get 12.96. I measured my 674 at ~5.45e-/s, which gives me a squared value of 29.7

If my understanding is correct, that means the absolute QE would have to be (12.96/29.7 =) less than 0.43 of the 674 for me to be "worse off" in terms of SNR. So I think I'd actually still be ahead here, unless the estimated QE of the 1600 is waaaay off. Keen for someone to correct me if I'm wrong here!
Reply With Quote
  #190  
Old 30-05-2016, 12:29 PM
glend (Glen)
Registered User

glend is offline
 
Join Date: Jun 2013
Location: Lake Macquarie
Posts: 7,033
Certainly won't be me correcting you Lee.

In other activities, I have managed to reprocess my M8 Ha 10 sub stack to eliminate the core blow out. Here:

http://www.astrobin.com/250826/

Big Screen view here:

http://www.astrobin.com/full/250826/None/

I have to pay more attention to processing.
Reply With Quote
  #191  
Old 30-05-2016, 01:03 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by codemonkey View Post
Nah, the graph is the plotted relative QE. The absolute is not published, and I believe it's non-trivial to determine? That's what ZWO have said anyway.

Jon's post on page 47 says If we assume we gather about the same charge per unit area, then we might gather 20.2e- total (object and skyfog) per 5.4 microns square in a short exposure of say 60 seconds.

That assumption essentially says "if we assume they have the same QE". The 1600 is assumed to be pretty much bang on with the kaf-8300, with different relative differences, but it's still assumed, no one actually knows.

Thinking about the whole SNR thing though... read noise adds in quadrature in an image. If we take zero gain on the 1600, we have a read noise of about 3.6e-/s. Squaring this we get 12.96. I measured my 674 at ~5.45e-/s, which gives me a squared value of 29.7

If my understanding is correct, that means the absolute QE would have to be (12.96/29.7 =) less than 0.43 of the 674 for me to be "worse off" in terms of SNR. So I think I'd actually still be ahead here, unless the estimated QE of the 1600 is waaaay off. Keen for someone to correct me if I'm wrong here!
I think that all applies if read noise is the limiting factor. If sub exposures are long enough that shot noise buries the read noise, then QE wins out. The 674 is still a very effective chip, especially for Ha - it just needs much longer subs than the 1600 (sub length varies as RN^2).
Reply With Quote
  #192  
Old 30-05-2016, 02:27 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
just in case anyone is interested in analysing data from these cameras, just did an inspection of a FITs sub from Nebulosity and all numbers that came up were divisible by 16 - conclusion: that the lower 4 bits are padded to give 16 bit numbers using a 12 bit ADC. So if trying to work stuff out in electrons, divide the ADU by 16 before taking the gain into account.
Reply With Quote
  #193  
Old 30-05-2016, 05:23 PM
lazjen's Avatar
lazjen (Chris)
PI cult member

lazjen is offline
 
Join Date: Dec 2012
Location: Flaxton, Qld
Posts: 2,064
Quote:
Originally Posted by glend View Post
Certainly won't be me correcting you Lee.

In other activities, I have managed to reprocess my M8 Ha 10 sub stack to eliminate the core blow out. Here:

http://www.astrobin.com/250826/

Big Screen view here:

http://www.astrobin.com/full/250826/None/

I have to pay more attention to processing.
I prefer the previous version - it seemed more silky smooth to the sharp grains on this one. Maybe it's just me though.
Reply With Quote
  #194  
Old 30-05-2016, 05:41 PM
glend (Glen)
Registered User

glend is offline
 
Join Date: Jun 2013
Location: Lake Macquarie
Posts: 7,033
Quote:
Originally Posted by lazjen View Post
I prefer the previous version - it seemed more silky smooth to the sharp grains on this one. Maybe it's just me though.
It is not just you Chris, the original is still up on the Astrobin Big Wall.
Reply With Quote
  #195  
Old 30-05-2016, 05:48 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,871
Quote:
Originally Posted by Shiraz View Post
nope, the read noise goes just the opposite way Greg - it reduces with increasing gain and is lowest at highest gain. However, the effective wells are not very big at high gain, so you need short subs - which you can do because of the low read noise.

The graphs posted earlier and available on the ZWO website show exactly how the various parameters vary with gain. ISO is not used as a measure of gain with these cameras, actual amplifier gain is used. This chip has made a complete transition from the world of DSLRs to the world of CCD measurements.
I wasn't really referring to read noise there Ray but rather overall noise. But yes I thought read noise dropped with higher ISO. Not sure why but I have heard that before.
As far as gain and amplifier in digital cameras there is amplifier gain after a certain point in the ISO in a digital camera. For example the Fuji Xpro 2 starts using amplifier gain at ISO1600. Not sure where the gain comes from below that point, I suppose its digital gain in the ADC.

I don't know about a complete transition as the images I have seen so far seem to have similar constraints - blown out highlights.

A typical DSLR image differs from an Astro CCD image mainly in the stars and highlights. Typically DSLR images have blown stars that are all white with no colour data.

Its worth being aware that this sensor may respond similarly and work around that limitation to avoid images with white stars and blown highlights.

Anyway that is just conjecture and your testing will show these things more clearly.

Greg.
Reply With Quote
  #196  
Old 30-05-2016, 06:35 PM
codemonkey's Avatar
codemonkey (Lee)
Lee "Wormsy" Borsboom

codemonkey is offline
 
Join Date: Jul 2013
Location: Kilcoy, QLD
Posts: 2,058
Quote:
Originally Posted by Shiraz View Post
I think that all applies if read noise is the limiting factor. If sub exposures are long enough that shot noise buries the read noise, then QE wins out. The 674 is still a very effective chip, especially for Ha - it just needs much longer subs than the 1600 (sub length varies as RN^2).
Yep, I think you're right Ray. I used my trusty SNR spreadsheet and it backs up what you said (not a surprise).

By the way, I'm absolutely in agreement that the 674 is still a very effective chip. Really, it's better than the 1600. The 1600 has, however, just the blend of features that make it especially attractive to me with my specific conditions / gear, so that I think I would be happy to make the switch, as long as the 1600 does perform.

I'll be very interested to see how you find it in comparison to your 694.
Reply With Quote
  #197  
Old 30-05-2016, 06:38 PM
Atmos's Avatar
Atmos (Colin)
Ultimate Noob

Atmos is offline
 
Join Date: Aug 2011
Location: Melbourne
Posts: 6,980
I just want a 694/814 chip the size of the 16803 Perfection! Plus, only worth maybe $40k
Reply With Quote
  #198  
Old 30-05-2016, 06:42 PM
Slawomir's Avatar
Slawomir (Suavi)
Registered User

Slawomir is offline
 
Join Date: Sep 2014
Location: North Queensland
Posts: 3,240
Quote:
Originally Posted by Atmos View Post
I just want a 694/814 chip the size of the 16803 Perfection! Plus, only worth maybe $40k

Plus the cost of getting an expensive telescope with larger imaging circle, 4" sturdy focuser, 50mm filters...that's why I (read: my wallet) like small ICx814
Reply With Quote
  #199  
Old 30-05-2016, 06:43 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,871
Quote:
Originally Posted by Atmos View Post
I just want a 694/814 chip the size of the 16803 Perfection! Plus, only worth maybe $40k

Yes that would be heaven as long as it had deep wells.

Mind you some implementations of the KAF16200 sensor run at 6 electrons or less red noise. That is much the same as the 694 and "only" double the ASI1600 in its most likely setup.

Greg.
Reply With Quote
  #200  
Old 30-05-2016, 06:45 PM
codemonkey's Avatar
codemonkey (Lee)
Lee "Wormsy" Borsboom

codemonkey is offline
 
Join Date: Jul 2013
Location: Kilcoy, QLD
Posts: 2,058
Quote:
Originally Posted by Atmos View Post
I just want a 694/814 chip the size of the 16803 Perfection! Plus, only worth maybe $40k
Coming from a 674, I'd settle for an 814 at the cost of the 1600 ;-) Seriously, an 814 would be awesome, but alas, the price point is too high for me at this point (i.e. the wife says no).
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 03:45 PM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Bintel
Advertisement
Testar
Advertisement