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

Sony A7s - New Low-Light Camera - Wow!

This topic has been archived. This means that you cannot reply to this topic.
698 replies to this topic

#376 mclewis1

mclewis1

    Thread Killer

  • *****
  • Posts: 18880
  • Joined: 25 Feb 2006

Posted 21 September 2014 - 02:15 PM

I love the idea of using the A7s (or any camera for that matter) for EAA. I just don't think that the ability to take shorter exposures and better looking images on their own makes for an EAA appropriate setup (although it certainly helps). Don't get me wrong, I think the A7 can be part of a great EAA setup (those great IIE images and videos are good examples). I just think most of the discussion going on about the A7s so far hasn't been EAA oriented, just more about getting better images with shorter exposures. I thought EAA is primarily about viewing not imaging. IMHO the imaging part should be about accurately capturing what you are viewing.

 

Let's see live viewing, constantly updating software that constantly displays an image that I can take a good close look at or display to others viewing with me. No, it doesn't have to be video, just something that will keep displaying the current image from the camera without having to fiddle with the software.

 

And yes I know that there's a bunch of video camera software that is going in the opposite direction by adding more and more complexity and behind the scenes image manipulation ... but at least they still keep displaying the most up to date image in a fashion that I can study without having to do anything to refresh it.



#377 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003

Posted 21 September 2014 - 03:06 PM

For the most part I've been ignoring the posts that confuse the issue between imaging and semi live viewing on this thread. I don't think it's a big deal personally. If someone prefers to image they should probably start a thread or join an existing one on the A7s in one of the many imaging sub forums.

There have been those that have broadcast in video mode on NSN for example. After having seen that, I believe that it's just a matter of time IMO before we get the right support/software to use this awesome tech for much more flexible EAA. As has been said before, because what we do is a niche we make due with what's available. Well this particular camera has immense potential for EAA and I would encourage everyone not to give up so easily/quickly.

Just a little while ago I was looking for a 100% digital solution to EAA, now we have Astrolive and if I had waited I might have had that instead of the analog/digital camera I do now. So don't give up!

#378 RandyC

RandyC

    Viking 1

  • *****
  • Posts: 665
  • Joined: 01 Apr 2013

Posted 21 September 2014 - 03:55 PM

Something still doesn't add up with the pixel sizes because I calculate the ExMor CMOS has twice the number of pixels in an equivalent ICX418 area. If ICX418 has 380k pixels and ExMor CMOS has 12,200k pixels. Based on ICX418 area of 44square mm (7.4x5.95), the Exmor CMOS should have 625,193 pixels in the same area. This is based on Exmor total pixels of 12,200k pixels/859sq mm=14,199 pixels/square mm. (35.8mm*24mm=859.2) Then multiply 14,199 pixels/sq mm * 44sq mm = 625,193 pixels. 


Edited by RandyC, 21 September 2014 - 04:15 PM.


#379 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003

Posted 21 September 2014 - 04:12 PM

Per:

http://store.sony.co...?MobileOptOut=1

The A7s has a "gapless on chip design" presumably allowing for no gaps between pixels. Could this be the difference is chip density your calculating?

#380 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10302
  • Joined: 03 Jul 2004

Posted 21 September 2014 - 06:43 PM

A7s uses square pixel 8.4um x 8.4um.

The active image area has 4240 x 2832 (12,007,680) pixels, occupying  35.62mm x 23.79mm area.  <-- not counting OBP and dummy pixels.

 

ICX828 uses 7.4um x 9.8um pixel. The active image area has 768x494 (379,362) pixels, occupying 5.683mm x 4.84mm.

 

 

Per pixel size: A7s' is tiny bit smaller (97.3%) 

Total pixels: A7s vs fine resolution type-1/2" NTSC CCD  31.65:1  (almost 32 times)

Total area: A7s vs fine resolution type-1/2" NTSC CCD 30.81:1 (almost 31 times)

 

Clear Skies!

 

ccs_hello



#381 RandyC

RandyC

    Viking 1

  • *****
  • Posts: 665
  • Joined: 01 Apr 2013

Posted 21 September 2014 - 09:54 PM

Yes and within the same area of 44sq mm, I count 380k pixels versus ExMor CMOS of 615k pixels. So 60% more in A7s.


Edited by RandyC, 21 September 2014 - 09:55 PM.


#382 Chris A

Chris A

    Vendor Affiliate - Mallincam

  • -----
  • Posts: 1157
  • Joined: 03 Feb 2007

Posted 21 September 2014 - 09:58 PM

A7s uses square pixel 8.4um x 8.4um.

The active image area has 4240 x 2832 (12,007,680) pixels, occupying  35.62mm x 23.79mm area.  <-- not counting OBP and dummy pixels.

 

ICX828 uses 7.4um x 9.8um pixel. The active image area has 768x494 (379,362) pixels, occupying 5.683mm x 4.84mm.

 

 

Per pixel size: A7s' is tiny bit smaller (97.3%) 

Total pixels: A7s vs fine resolution type-1/2" NTSC CCD  31.65:1  (almost 32 times)

Total area: A7s vs fine resolution type-1/2" NTSC CCD 30.81:1 (almost 31 times)

 

Clear Skies!

 

ccs_hello

ICX828 pixels are 8.4 x 9.8 um

 

Chris



#383 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10302
  • Joined: 03 Jul 2004

Posted 21 September 2014 - 10:13 PM

Thanks Chris.

Now redo the calculation:

 

----------------------------------------

 

A7s uses square pixel 8.4um x 8.4um.
The active image area has 4240 x 2832 (12,007,680) pixels, occupying  35.62mm x 23.79mm area.  <-- not counting OBP and dummy pixels.

ICX828 uses 8.4um x 9.8um pixel. The active image area has 768x494 (379,362) pixels, occupying 6.45mm x 4.84mm.


Per pixel size: A7s' is a bit smaller (85.71%)
Total pixel count: A7s vs fine resolution type-1/2" NTSC CCD  31.65:1  (almost 32 times)
Total area: A7s vs fine resolution type-1/2" NTSC CCD 27.15:1 (almost 27 times)

Clear Skies!

ccs_hello



#384 RandyC

RandyC

    Viking 1

  • *****
  • Posts: 665
  • Joined: 01 Apr 2013

Posted 21 September 2014 - 10:25 PM

Yes and within the same area of 44sq mm, I count 380k pixels versus ExMor CMOS of 615k pixels. So 60% more in A7s.

Actually ICX418 active pixel area is 39.5 sq mm (not 5.683mm x 4.84mm=27.5). ICX418 effective pixels=380k. Exmor CMOS in same 39.5mm area is 552k, or 45% more. Of course, Exmor total full frame pixel count is 32 times more.


Edited by RandyC, 21 September 2014 - 10:26 PM.


#385 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10302
  • Joined: 03 Jul 2004

Posted 21 September 2014 - 10:51 PM

Randy,

ICX418 / ICX828's active image area is 6.45 x 4.84 = 31.22 square mm.

On the same area basis, A7s will squeeze in 16.67% more pixels than ICX418's.

Clear Skies!

ccs_hello

#386 RandyC

RandyC

    Viking 1

  • *****
  • Posts: 665
  • Joined: 01 Apr 2013

Posted 21 September 2014 - 11:16 PM

Randy,

ICX418 / ICX828's active image area is 6.45 x 4.84 = 31.22 square mm.

On the same area basis, A7s will squeeze in 16.67% more pixels than ICX418's.

Clear Skies!

ccs_hello

How do you get 6.45 x 4.84 = 31.22 square mm? I am looking at the data sheet:

h: 6.9mm (based on 93% effective of 7.4mm)

v: 5.8mm (based on 97% effective of 5.95mm)

so I calculate 39.9 sq mm and 558k Exmor pixels or 47% more.



#387 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10302
  • Joined: 03 Jul 2004

Posted 21 September 2014 - 11:29 PM

Randy,

Chip size does not accurately reflect the active image area (less H-registers, charge pump, sense amplifier, Optical Black Pixels, dummy pixels, bonding pad area, etc.)

Best method is to multiply pixel size by the "active pixel" count.

Clear Skies!

ccs_hello

Edited by ccs_hello, 21 September 2014 - 11:52 PM.


#388 RandyC

RandyC

    Viking 1

  • *****
  • Posts: 665
  • Joined: 01 Apr 2013

Posted 21 September 2014 - 11:41 PM

Randy,

Chip size does not accurately reflect the active image area (H-registers, charge pump, sense amplifier, Optical Black Pixels, dummy pixels, bonding pad area, etc.)

Best method is to multiply pixel size by the "active pixel" count.

Clear Skies!

ccs_hello

but the pixels are different sizes.



#389 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10302
  • Joined: 03 Jul 2004

Posted 21 September 2014 - 11:43 PM

BTW, 6.45mm H x 4.84mm V is
8.064mm diagonal. This is why type-1/2" is also known as diag 8mm sensor.

Clear Skies!

ccs_hello

#390 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10302
  • Joined: 03 Jul 2004

Posted 21 September 2014 - 11:50 PM

Randy

Use ICX418 as an example,
pixel size is 8.4um x 9.8um.
In the active image area, there are 768 pixels horizontally and 494 vertically.

8.4um x 768 = 6.45mm (horizontal)
9.8um x 494 = 4.84mm (vertically)

Hope this helps.

Clear Skies!

ccs_hello

#391 RandyC

RandyC

    Viking 1

  • *****
  • Posts: 665
  • Joined: 01 Apr 2013

Posted 21 September 2014 - 11:52 PM

BTW, 6.45mm H x 4.84mm V is
8.064mm diagonal. This is why type-1/2" is also known as diag 8mm sensor.

Clear Skies!

ccs_hello

Data sjeet says, Chip size: 7.40mm (H) × 5.95mm (V) that's total size.

effective size: 7.0 x 5.8

 

so roughly if icx pixels is 17% larger, knock 17% off my 49% more to get 41% more.

 

so approx. 24% more, accounting for larger pixels.


Edited by RandyC, 22 September 2014 - 12:00 AM.


#392 RandyC

RandyC

    Viking 1

  • *****
  • Posts: 665
  • Joined: 01 Apr 2013

Posted 21 September 2014 - 11:55 PM

Randy

Use ICX418 as an example,
pixel size is 8.4um x 9.8um.
In the active image area, there are 768 pixels horizontally and 494 vertically.

8.4um x 768 = 6.45mm (horizontal)
9.8um x 494 = 4.84mm (vertically)

Hope this helps.

Clear Skies!

ccs_hello

no, cause you're not accounting for space between the pixels.



#393 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10302
  • Joined: 03 Jul 2004

Posted 22 September 2014 - 12:00 AM

Randy,

Please read and understand what I have said. Sony data sheet is not the easiest one to understand. Please study it thoroughly.

BTW, this is an A7s thread. Not a sensor study thread. For image sensors, I have few threads that you can read at your leisure.

Clear Skies!

ccs_hello

#394 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10302
  • Joined: 03 Jul 2004

Posted 22 September 2014 - 12:03 AM

Re: count space in between the pixels
I really have to bail out.
Thanks.

#395 RandyC

RandyC

    Viking 1

  • *****
  • Posts: 665
  • Joined: 01 Apr 2013

Posted 22 September 2014 - 12:04 AM

okay, but the whole reason Exmor can fit in more is because of more efficient use of space. Otherwise we'd simply calculate without accounting for any space. That's why Relativist pointed out a "gapless on chip design" on A7s.

 

i'm saying Exmor 41% more pixels in same area.

 

Answer: correct answer.


Edited by RandyC, 22 September 2014 - 12:04 AM.


#396 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10302
  • Joined: 03 Jul 2004

Posted 22 September 2014 - 07:23 AM

Per:

http://store.sony.co...?MobileOptOut=1

The A7s has a "gapless on chip design" presumably allowing for no gaps between pixels. Could this be the difference is chip density your calculating?

Curtis,

 

Gapless microlens design has nothing to do with the pixel size.

 

Pixel size is simply defined as the image area divided by the number of pixels.

 

E.g., if a sensor has the image area of 6mm x 4mm, while the pixel size is 10um x 10 um, then there will be 600 pixels across and 400 pixels down, forming a 240,000 pixel sensor.

 

Naturally, for each pixel area, not all such real estate are photo-sensitive (there are many structure underneath: circuits, interline transfer buffer, etc.  <-- even a BSI sensor cannot get away form that).

The ratio of the photo-sensitive area vs. the pixel real estate space is  called the "fill factor".  It will never be 100% (BSI needs an isolation guard band as borders separating pixels).

Here the microlens comes to rescue to try to focus the incoming light toward to photo-sensitive area of each pixel.  Gapless microlens is just a revised design to increase its effectiveness.

 

 

 

BTW, chip size  >  image area size (due to other support circuitry, bonding pads, etc. not counted in)   >  active image area (optical black pixels and dummy pixels not counted in)

 

Clear Skies!

 

ccs_hello


Edited by ccs_hello, 22 September 2014 - 07:37 AM.


#397 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003

Posted 22 September 2014 - 08:04 AM

Here's a thread that goes into depth for those interested. http://www.dpreview....s/post/52351544

 

Sony isn't the only one using this, see also: http://www.cameratec...g-at-iso-50000/

 

Sony is not highlighting this feature of the A7s just because, it's true, there have traditionally been gaps between the microlenses on each 'pixel'. I'm confused by the different area claims for the 1/2 format chips that are compared, but I do not find the matter relevant myself. The fact that Sony is using such techniques in this sensor is all the more reason to use it in EAA.



#398 Moromete

Moromete

    Viking 1

  • *****
  • Posts: 711
  • Joined: 15 Feb 2012

Posted 22 September 2014 - 01:31 PM

I love the idea of using the A7s (or any camera for that matter) for EAA. I just don't think that the ability to take shorter exposures and better looking images on their own makes for an EAA appropriate setup (although it certainly helps). Don't get me wrong, I think the A7 can be part of a great EAA setup (those great IIE images and videos are good examples). I just think most of the discussion going on about the A7s so far hasn't been EAA oriented, just more about getting better images with shorter exposures. I thought EAA is primarily about viewing not imaging. IMHO the imaging part should be about accurately capturing what you are viewing.

 

Let's see live viewing, constantly updating software that constantly displays an image that I can take a good close look at or display to others viewing with me. No, it doesn't have to be video, just something that will keep displaying the current image from the camera without having to fiddle with the software.

 

And yes I know that there's a bunch of video camera software that is going in the opposite direction by adding more and more complexity and behind the scenes image manipulation ... but at least they still keep displaying the most up to date image in a fashion that I can study without having to do anything to refresh it.

 

The A7S is more than adequate for EAA, besides taking nice pictures, because you can it in Video mode with a shutter speed of 1/4s and ISO 400.000 with little noise if S-Log2 is selected.

 

In this Video mode you can literally see in the dark with cheap 50mm F1.8 lens.

 

Now, just think about how powerful is the A7S in this Video mode when combined with Hyperstar! I tell you this is Nirvana! ... Ok, you can stop thinking now and watch this  http://www.youtube.c...fllZDo0A-tXdoyw . Please note the following: C14 Edge, SS 1/4s in movie mode, FL=665mm, F1.9 for some of the images and the A7S is astro modified too! Other images are filmed at F7!

 

In short, using the A7S in Video mode with a 1/4s shutter speed is more real time viewing than any Mallincam like cameras which need several seconds to stack a bunch of images.



#399 Chris A

Chris A

    Vendor Affiliate - Mallincam

  • -----
  • Posts: 1157
  • Joined: 03 Feb 2007

Posted 22 September 2014 - 02:54 PM

 

I love the idea of using the A7s (or any camera for that matter) for EAA. I just don't think that the ability to take shorter exposures and better looking images on their own makes for an EAA appropriate setup (although it certainly helps). Don't get me wrong, I think the A7 can be part of a great EAA setup (those great IIE images and videos are good examples). I just think most of the discussion going on about the A7s so far hasn't been EAA oriented, just more about getting better images with shorter exposures. I thought EAA is primarily about viewing not imaging. IMHO the imaging part should be about accurately capturing what you are viewing.

 

Let's see live viewing, constantly updating software that constantly displays an image that I can take a good close look at or display to others viewing with me. No, it doesn't have to be video, just something that will keep displaying the current image from the camera without having to fiddle with the software.

 

And yes I know that there's a bunch of video camera software that is going in the opposite direction by adding more and more complexity and behind the scenes image manipulation ... but at least they still keep displaying the most up to date image in a fashion that I can study without having to do anything to refresh it.

 

The A7S is more than adequate for EAA, besides taking nice pictures, because you can it in Video mode with a shutter speed of 1/4s and ISO 400.000 with little noise if S-Log2 is selected.

 

In this Video mode you can literally see in the dark with cheap 50mm F1.8 lens.

 

Now, just think about how powerful is the A7S in this Video mode when combined with Hyperstar! I tell you this is Nirvana! ... Ok, you can stop thinking now and watch this  http://www.youtube.c...fllZDo0A-tXdoyw . Please note the following: C14 Edge, SS 1/4s in movie mode, FL=665mm, F1.9 for some of the images and the A7S is astro modified too! Other images are filmed at F7!

 

In short, using the A7S in Video mode with a 1/4s shutter speed is more real time viewing than any Mallincam like cameras which need several seconds to stack a bunch of images.

 

I personally thought the images were very cool considering the video exposure was 1/4 sec, however, you still need to use a fair size scope  (C14) and a very fast focal ratio at F1.9 for those fainter DSO's. It is too bad that the video mode cannot go a bit higher in exposure say 1/2 to 1 full second in order to reveal much more of the faint areas that are missing and to reduce the noise. Looks like it is improving but the technology is still just not there yet.

 

Chris A



#400 chasing photons

chasing photons

    Mariner 2

  • -----
  • Posts: 287
  • Joined: 19 Mar 2013

Posted 22 September 2014 - 04:33 PM

Although the maximum 0.25 second exposure in video mode is truly impressive with the Sony A7S, I anxiously await more demonstrations of its capabilities in still mode with single exposures under 30 seconds and also with the noise reduction or wide dynamic range functions enabled depending on the target DSO. I want to see jpeg stills with any in camera processing that improves the output, not raw images.

Video is not required for near real time observing of DSOs so why cripple the Sony A7S output in this fashion? I think it will truly shine for stills under 30 seconds with OTAs in the f/4 to f/6 range with moderate apertures and moderate ISOs.

Also what is missing so far is a clear explanation of workflow for performing the above. Ideally, I would like to use the Sony A7S via Wi-Fi with a tablet for both control of camera settings and display of the current still image. Alternatively, I would like to eventually control it with software like AstroLive, but that will be a LONG time coming, if ever, I am afraid.

Anyway, more technology like the Sony A7S is needed in this endeavor, and I am happy to see products like this develop and become popular. It can only help!

Edited by chasing photons, 22 September 2014 - 04:35 PM.



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