Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Atik 414EX a possible near live imaging camera for EAA

  • Please log in to reply
124 replies to this topic

#26 Don Rudny

Don Rudny

    Surveyor 1

  • *****
  • Posts: 1624
  • Joined: 21 Mar 2013
  • Loc: Mauna Kea, Hawaii

Posted 23 December 2014 - 01:53 PM

Curtis and Don,

 

Thanks for the feedback.  I understand you can go to larger optics.  But you can do that with the Lodestar, too.  You're at F5 with the 11" to get the desired image scale.  But if I use the Lodestar then I can go faster.  Didn't work the number out, but it might be something like F3.3.  Am I missing something?  I think you have to compare cameras in the same scope.  I know you can get higher res, but it's going to cost you in speed is all I'm saying.  It still may be a good camera for EAA.  Proof is in testing it.  Maybe Atik will loan one to someone.  It's in their interest to get the word out if it's good for EAA.  Tests need to be unbiased though.



#27 DonBoy

DonBoy

    Apollo

  • *****
  • topic starter
  • Posts: 1239
  • Joined: 06 Sep 2007

Posted 23 December 2014 - 02:14 PM

You'll notice that the Atik 414 with the 8" at f5 will yield more detail even if cropped to the same size as the Lodestar image because the 414 arcsec/pixel is finer due to the smaller pixels.  1.74 arcsec/pix vs. 1.33 arcsec/pix.

 

I'm not looking at the 414 as a sensor that will yield images in less time than the Lodestar (althought that would be nice), but it's a sensor with more resolution and hopefully can deliver near live images in a reasonable time of exposure.   So I'll be able to still use the 8" SCT to capture a reasonable image scale without going to a larger aperture scope and have this higher resolution and detail than I wouldn't get with a 8+um pixel sensor like the Lodestar.


Edited by DonBoy, 23 December 2014 - 02:28 PM.


#28 Don Rudny

Don Rudny

    Surveyor 1

  • *****
  • Posts: 1624
  • Joined: 21 Mar 2013
  • Loc: Mauna Kea, Hawaii

Posted 23 December 2014 - 02:28 PM

You'll notice that the Atik 414 with the 8" at f5 will yield more detail even if cropped to the same size as the Lodestar image because the 414 arcsec/pixel is finer due to the smaller pixels.  1.74 arcsec/pix vs. 1.33 arcsec/pix.

 

I'm not looking at the 414 as a sensor that will yield images in less time than the Lodestar (althought that would be nice), but it's a sensor with more resolution and hopefully can deliver near live images in a reasonable time of exposure. vSo I'll be able to still use the 8" SCT to capture a reasonable image scale without going to a larger aperture scope and have this higher resolution and detail than I wouldn't get with a 8+um pixel sensor like the Lodestar.

Yes, Don.  I think we're on the same page.  I would be willing to sacrifice some speed to get better res, but how much becomes the issue.  The longer integration will be ok for brighter DSO's, but when you go really deep and dim, it may be a problem.   Maybe having two cameras is the answer.  Just what our CFO's want to hear.

 

Have a Merry Christmas.  It is an interesting thread.  Hope someone gets to test the 414 soon.



#29 DonBoy

DonBoy

    Apollo

  • *****
  • topic starter
  • Posts: 1239
  • Joined: 06 Sep 2007

Posted 23 December 2014 - 02:49 PM

 

Yes, Don.  I think we're on the same page.  I would be willing to sacrifice some speed to get better res, but how much becomes the issue.  The longer integration will be ok for brighter DSO's, but when you go really deep and dim, it may be a problem.   Maybe having two cameras is the answer.  Just what our CFO's want to hear.

When the exposure takes to long then use 2x2 binning (should cut exposure in half) but will yield half the resolution, but no worse than the Lodestar.  

 

I'll have to wait and see how the Atik 414 does before I'll consider a purchase.  If I were interested in getting into CCD imaging I would take my chances and buy the 414EXc and if it didn't work as a NLI, then atleast I would have a good starting CCD imager.

 

Merry Christmas



#30 Dom543

Dom543

    Apollo

  • *****
  • Posts: 1099
  • Joined: 24 Oct 2011

Posted 23 December 2014 - 03:14 PM

I have a question about binning in case of color CCD's. It was stated above that the Atik 414 switches to monochrome, when binned.

On the other hand, there are color CCD's, e.g. the Lodestar C or the Mallincam Universe, that can bin and retain color.

 

But there is hardware binning and software binning and they are very different animals with substantially different benefits.

 

Can someone give a definite answer and explanation, if hardware binning is possible for color sensors? Or is color binning necessarily just a software trick? The pixels are read individually as without binning but then the software applies a more elaborate debayering algorithm combining information from 2x2 super pixels (i.e. 4x4 base pixels)?

If 2x2 hardware binning means reading each 2x2 super pixel as one entity, then I don't understand how the color information can be recovered.

 

Thank you,

--Dom



#31 David B in NM

David B in NM

    Apollo

  • *****
  • Posts: 1027
  • Joined: 05 Sep 2010

Posted 23 December 2014 - 04:59 PM

Dom,

 

Binning capability is dependent on the ccd used (circuits).  Some color ccds can bin in color and others can't.

 

David B in NM



#32 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 23 December 2014 - 10:10 PM

Don,

 

I don't agree that you have to compare cameras on the same scope, mainly because the cameras are not comparable. They are both different formats (sizes) and resolutions. Like the other Don said the Atik needs to be evaluated for it's EAA capabilities. If it's sensitive enough to form an image in a reasonable amount of time is the primary question. Once that's determined, lets say it is good enough. There will next be a question of how fast it is. Some cameras are so sensitive that they can reasonably be used on alt/az scopes, others require some more time and are better suited to a equatorial mount. In my mind, the ability to increase aperture for a highly sensitive higher resolution chip is a great advantage, because not only can you get the framing one wants, but the higher aperture could allow one to go slightly deeper. In my own case, if the camera works out, I can go from my 8" f/4 imaging newt, to a 12" f/4 imaging newt.



#33 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10245
  • Joined: 03 Jul 2004

Posted 29 December 2014 - 12:24 AM

Dom

 

For CCD image sensors, there is the real binning (hardware bin in charge domain) and there are wanna bes.

The latter I call them software summing and very rarely seen, hardware-based summing.  <-- In my book, they are not binning.  They do not give you S/N improvement benefit.  They just lower the spatial resolution and jack up the numbers to make the image looks brighter.

 

Without special circuitry (as used in some P&S CCD based cameras) or a 4-channel output CCD image sensor, there is no hope that a real binning (hardware charge-domain binning) can keep the color.

That is, your color CCD image sensor becomes a low resolution monochrome sensor (color data got destroyed.)

 

ICX285 is a very basic CCD sensor and does not have the special circuitry to do "hardware binning to keep the color".  <-- (aka like-color binning)

ICX825 info is limited in public.  Per its data brief, it only has 1-ch or 2-ch output mode, thus using hardware binning will destroy colors.

 

Lodestar C and C2 (ICX429 and ICX829 complementary-color Bayer, 1-channel out, interlaced) fall under the same fate.

 

QHY8, 8L, 8 Pro, Universe use ICX413 (pro and the original 8 use ICX453) which only has 1-ch out.  These will not have color if hardware binning is used.

 

People can find various ways to stretch the definition to help sell a product.  That's beyond the scientific dept to argue with.  Clever consumers should make a note on that.

 

Next time you hear it, ask to see the original manufacturer's spec sheet (prefer in writing) and also ask "what's your definition of binning..."

 

In the world of CCD-based imager, be suspicious.  Like-color binning is a rare beast.

 

Clear Skies!

 

ccs_hello


Edited by ccs_hello, 29 December 2014 - 12:35 AM.


#34 Dom543

Dom543

    Apollo

  • *****
  • Posts: 1099
  • Joined: 24 Oct 2011

Posted 29 December 2014 - 11:44 PM

Thank you css_hello for answering my question about stacking in color cameras.

 

That clarifies the issue that not being able to stack in color is not a shortcoming but rather the general rule for average price CCD cameras. And that most claims of being capable to stack in color are just a misuse of the word "stacking" for a software trick that doesn't offer the same benefits as real "hardware" stacking.

 

Thanks again and Happy New Years!

--Dom



#35 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10245
  • Joined: 03 Jul 2004

Posted 30 December 2014 - 12:20 AM

Dom,

 

"Binning" usually means combining pixels in spatial dimension, e.g., a 2x2 bin on a 640x480 sensor would yield a 320x240 image.

It is a single image operation.

 

"Stacking" usually means combining multiple images in temporal dimension, e.g., stacking ten 2-sec images to form a single image as if it has an extra long exposure.

The spatial resolution does not change in the stacking case.

 

Hope this helps.

 

Clear Skies!

 

ccs_hello



#36 Dom543

Dom543

    Apollo

  • *****
  • Posts: 1099
  • Joined: 24 Oct 2011

Posted 30 December 2014 - 12:48 AM

Thank you for the correction, css_hello. I meant to say "binning" and not "stacking". It just has been a long day... 

--Dom



#37 nytecam

nytecam

    Voyager 1

  • *****
  • In Memoriam
  • Posts: 12692
  • Joined: 20 Aug 2005
  • Loc: London UK

Posted 30 December 2014 - 11:46 AM

Hi ccs-hello - as always very helpful post. But for some reason both Lodestar C and C2 can produce full colour images in 1x2 vertical binning (called 'Fast') and I've used the former camera for many of my images in my gallery linked below for some years using SX s/w.  The resultant images are full size 752x580p but lack slighty in vertical resolution but this goes largely unnoticed.  The 'progressive scan' etc is full-res.

 

2x2 binning produces a mono image of 376x290p and s/w provides two options to rescale upwards in the live view mode.  Camera sensitivity is doubled approx.   The same capture options are available for both mono cameras M and Mx2.  My recent 'earthshine' pic posted on EAA was 2x2 binned via L-Mx2 in 1ms exp. :waytogo:


Edited by nytecam, 30 December 2014 - 11:53 AM.

  • ccs_hello likes this

#38 Dom543

Dom543

    Apollo

  • *****
  • Posts: 1099
  • Joined: 24 Oct 2011

Posted 30 December 2014 - 03:48 PM

That's very interesting what you are saying about the 1x2 binning in the Lodestar color, nytecam!

 

I was curious about the effects of that mode on color. If I understand it correctly, in the CYMG bayer matrix of the sensor, the Yellow and Magenta sensels are in the same odd columns and the Cyan and Green sensels are in the same even columns. I wonder, if some special debayering algorithm is used in this mode to synthesize colors from the two colors resulting from the two colums.

 

Interestingly the QE peaks of both the Y and M sensels are at around 625nm, close to Ha. The peaks of the C and G sensels are around 515nm, close to O-III and Hb.

 

--Dom



#39 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 30 December 2014 - 07:05 PM

So far I haven't found much data on the ICX825 which is the chip used in the Atik414. Though this PDF from ptgrey does give more data than I saw on Sony's site. (May have missed it f it is there)

http://www.ptgrey.co...5-vs-icx825.pdf

Edited by Relativist, 30 December 2014 - 07:33 PM.


#40 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10245
  • Joined: 03 Jul 2004

Posted 30 December 2014 - 07:32 PM

Please note PT Grey tends to emphasize on it can produce high speed readout (thus high frame rate) IEEE1394, USB3, and gigE cameras as opposed to classic astroCCD imager firms (slow down as much as possible and special drive patterns to squeeze out the last juice of the CCD for a better S/N.)



#41 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 30 December 2014 - 07:34 PM

That shouldn't change the QE graph though.

#42 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10245
  • Joined: 03 Jul 2004

Posted 30 December 2014 - 08:06 PM

Hi ccs-hello - as always very helpful post. But for some reason both Lodestar C and C2 can produce full colour images in 1x2 vertical binning (called 'Fast') and I've used the former camera for many of my images in my gallery linked below for some years using SX s/w.  The resultant images are full size 752x580p but lack slighty in vertical resolution but this goes largely unnoticed.  The 'progressive scan' etc is full-res.

 

Maurice,

 

Per your description, the 1x2 binning mode actually is the classic video CCD (interlaced CCD with complementary-color CMY + G) most basic read out method, known as field integration mode CCD read out.

See http://www.cloudynig...s/#entry1380277 <-- (note highly technical)

 

Interlaced CCD will have odd field readout and even field readout.  It is the design nature of such CCD, (as opposed to progressive readout mode CCD.)

 

In 1x2 binning, <-- actually called neighboring line charge summation or field integration mode, is the default readout method for non long-exposure shooting conditions. 

Odd field will be CCD line1+2, line 3+4, etc., forming the first field with only half of the vertical output (which is final image's line 1, 3, 5, 7, ..., 477, and 479.)

Even field will be CCD line2+3, line 4+5, etc., forming the second field with only half of the vertical output (which is final image's line 2, 4, 6, 8, ..., 478, and 480.)

 

In long exposure mode, (say the total integration time is 2.1 sec), the first field time will be 1.05 second, followed by the even field of 1.05 second, in alternating manner.

 

This mode emphasizes on "the flow of motion", which is what video is 

 

===============

 

Then what is the non-binning mode of such video CCD readout?  (Some call it 1x1 bin, which means no binning.  Such an awkward way of saying no action is done such that a fabricate operation name is given.

There is no neighboring line charge summing at all.

 

This is called the frame integration mode.  This is good when spatial resolution is more preferred than showing the "motion".  E.g., shooting long exposure astro objects.

Each CCD line has its full integration time in entire duration dedicated gathering photons, uninterrupted.

 

Clear Skies!

 

ccs_hello


Edited by ccs_hello, 30 December 2014 - 08:28 PM.


#43 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10245
  • Joined: 03 Jul 2004

Posted 30 December 2014 - 08:16 PM

That shouldn't change the QE graph though.

Curtis,

 

I was seeing other numbers, such as S/N dept.

 

Clear Skies!

 

ccs_hello



#44 mac4lyfe

mac4lyfe

    Sputnik

  • *****
  • Posts: 44
  • Joined: 18 Apr 2014

Posted 30 December 2014 - 09:37 PM

I just put in an order for this camera. Should be here in a couple of weeks. 


  • Relativist likes this

#45 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 30 December 2014 - 10:30 PM

I just put in an order for this camera. Should be here in a couple of weeks. 

 

Nice! did you pick B&W or color? also I'm assuming your going to be using it on a mac, do you know what software you will be using it with?



#46 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10245
  • Joined: 03 Jul 2004

Posted 30 December 2014 - 11:55 PM

I am not familiar with Atik 414ex features.

Does anyone know its gain is user adjustable thu camera MENU?  How high the gain (in unit of dB) it can be?

 

Clear Skies!

 

ccs_hello



#47 nytecam

nytecam

    Voyager 1

  • *****
  • In Memoriam
  • Posts: 12692
  • Joined: 20 Aug 2005
  • Loc: London UK

Posted 31 December 2014 - 03:53 AM

That's very interesting what you are saying about the 1x2 binning in the Lodestar color, nytecam!

 

I was curious about the effects of that mode on color. If I understand it correctly, in the CYMG bayer matrix of the sensor, the Yellow and Magenta sensels are in the same odd columns and the Cyan and Green sensels are in the same even columns. I wonder, if some special debayering algorithm is used in this mode to synthesize colors from the two colors resulting from the two colums.

 

Interestingly the QE peaks of both the Y and M sensels are at around 625nm, close to Ha. The peaks of the C and G sensels are around 515nm, close to O-III and Hb.

 

--Dom

Hi Dom - the CYMG bayer mix has long had me wondering about effective Ha sensitivity and how the RGB output is created. I presume SX [via CEO Terry Platt] has used the basic Sony CYMG algorithm.  I've pretty sure the binned Fast option uses a common algorithm for all colour conversion and not a unique one. 

ccs_hello has responded on this and I need to get my old head around it !

 

Perhaps of interest the saved raw/unprocessed images use a 1/4 [852byte] of the disc space as the colour converted images [3408 bytes] which is something I find convenient - the 2x2bin =  modest 213 bytes  :waytogo:   .



#48 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 31 December 2014 - 10:47 AM

That's a lot of good Lodestar info hidden in a thread about a completely different camera. :lol:

 

Here's a side by side (video by ptgrey) of ICX285 vs. ICX825, starts at about 1:44

 

https://www.youtube....h?v=usX4rWesWIg

 

Interestingly ptgrey talks about slowing down the pixel clock during long exposures to reduce the noise, I wonder if Atik has incorporated such a tactic.



#49 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10245
  • Joined: 03 Jul 2004

Posted 31 December 2014 - 12:05 PM

Curtis,

 

Atik (carry over from the Artemis era) is a firm specialized/roted from astroCCD heritage. 

Being in that niche, it needs to know few tricks.

 

See this post:  http://www.cloudynig...aphy/?p=6344582

 

Clear Skies!

 

ccs_hello



#50 mac4lyfe

mac4lyfe

    Sputnik

  • *****
  • Posts: 44
  • Joined: 18 Apr 2014

Posted 31 December 2014 - 09:21 PM

Hey Curtis - I'm getting the color version to go with my hyperstar 11. Most likely use it with a windows machine as I don't mind putting those out in the cold :-)

 

 

I just put in an order for this camera. Should be here in a couple of weeks. 

 

Nice! did you pick B&W or color? also I'm assuming your going to be using it on a mac, do you know what software you will be using it with?

 


  • RafaelP likes this


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics