Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
11-23-2017, 01:27 AM   #1
Pentaxian




Join Date: Dec 2011
Posts: 2,249
Camera design problem at heart of Sony A7RIII sensor issues (raw cooking not enough)

Interesting research on the reasons for the problems with Sony A7R3.
The new Sony a7RIII camera still eats stars | Photo Rumors

Here is a fun quote first:
QuoteQuote:
A7r: A New Hope
A7rII: Star Eater strikes back
A7rIII: Return of Star Eater
In short it seems that Sony is cooking raws quite extensively to make the noise look better and they have to cook their raws due to the chosen body design. And they think rather loose image detail than expose the actual sensor noise to photographers in "raw" files.

QuoteQuote:
  1. The smaller body design/size of the Sony A7RIII (and other A series cameras) results in a higher internal camera heat level which means in general a more aggressive noise reduction is necessary across the ISO range, but particularly important in the high ISO range and during long exposures because the accumulated effect of the activated sensor causes the heat to skew the signal/noise ratio and produces much noisier images. This is much more evident in higher resolution cameras like the A7R II & III because the more intense data processing generates more heat and the electronics of a higher resolution sensor means more dark noise in the signal/noise ratio. In order to resolve that a more aggressive noise reduction had to be applied at ISOs > 3200 which causes pixel level stars to be wiped. Pixel level doesn't mean actual pixel sized stars, the threshold is actually 4 pixels that make up an RGB block of the Bayer array since color info is needed to make noise reduction adjustment, so stars that are ~4 pixels or smaller are at risk of getting removed.
  2. The phase detect AF sensors built into the imaging sensor generate dark noise in the sensor readout, and that is also especially noticeable at ISOs >3200. In addition, even though the PDAF system is not active during the actual capture of a photo, the electronics are passively in a "standby" mode and that generates additional dark noise. Because of this, the PDAF hot spot areas follow a noise reduction algorithm that is different than the rest of the sensor and accounts for that in order to deliver a smooth gradation and even noise pattern.
  3. I was also told that currently because of the hardware limitations there is no way to resolve the issue and that it is not technically an issue at all, it is just a limitation of the hardware. Any fixes that would be applied via firmware would end up exposing the higher noise, which the engineering team determined would be a worse problem for users so it doesn't seem like a proper solution is in the works.


11-23-2017, 02:38 AM   #2
Pentaxian
Digitalis's Avatar

Join Date: Mar 2009
Location: Adelaide.
Posts: 9,537
some people who dig into raw files also discovered Canon's dual pixel equipped sensors also generate increased dark noise in images, they don't eat stars though. Also due to the design of the sony cameras it would easier to remove the housing and apply a form of active cooling to reduce the heat issue.
11-23-2017, 09:03 AM - 2 Likes   #3
Loyal Site Supporter
Loyal Site Supporter




Join Date: Mar 2015
Photos: Gallery | Albums
Posts: 6,297
"My God, it's full of... oh, wait."
11-23-2017, 01:27 PM   #4
Site Supporter
Site Supporter
RGlasel's Avatar

Join Date: Sep 2013
Location: Saskatoon
Photos: Gallery | Albums
Posts: 2,246
QuoteOriginally posted by Digitalis Quote
apply a form of active cooling
Which increases the size of the camera, decreases battery life and probably results in a body that looks like a conventional DSLR (at least if it is still intended to be handheld). That wouldn't have a positive impact on A7 sales.

11-23-2017, 02:58 PM - 1 Like   #5
Pentaxian
Digitalis's Avatar

Join Date: Mar 2009
Location: Adelaide.
Posts: 9,537
Have you ever seen a dedicated astro camera? they aren't even remotely ergonomic.
11-23-2017, 06:15 PM   #6
Site Supporter
Site Supporter




Join Date: Mar 2009
Location: Gladys, Virginia
Photos: Gallery
Posts: 19,961
You can't have your cake and eat it too. Small and cooling don't necessarily play well together and sensors need cooling -- particularly ones that have very fast read out speeds.

That said, I'd say that most folks are going to be fine with the A7r III and those who shoot astro photography will avoid it.
11-24-2017, 12:45 AM   #7
Pentaxian
Digitalis's Avatar

Join Date: Mar 2009
Location: Adelaide.
Posts: 9,537
QuoteOriginally posted by Rondec Quote
Small and cooling don't necessarily play well together and sensors need cooling
The engineers at Sony seem to be just figuring that out, while everyone else looks on shaking their heads.
11-24-2017, 03:32 AM   #8
Site Supporter
Site Supporter




Join Date: Mar 2009
Location: Gladys, Virginia
Photos: Gallery
Posts: 19,961
QuoteOriginally posted by Digitalis Quote
The engineers at Sony seem to be just figuring that out, while everyone else looks on shaking their heads.
I suppose. What is the benefit of mirrorless if it isn't decreasing the size (depth) of the camera?

11-24-2017, 06:25 AM   #9
Pentaxian




Join Date: Dec 2011
Posts: 2,249
Original Poster
QuoteOriginally posted by Rondec Quote
Small and cooling don't necessarily play well together and sensors need cooling -- particularly ones that have very fast read out speeds.
Physics. When I first measured the temperature of my cameras sensors I was surprised to see how much they heat up (50 degrees Celsius after 20-30 minutes at 20C room temp). If you have put your hand into 50 C water you know that is quite something.

That actually is a reason why I am not really keen on more pixels or more readout speed. That would not benefit my requirements, but actually hurt them.

Here the old reaction "you don't have to use it" is misplaced as there is a tangible downside to certain "improvements".
11-24-2017, 09:02 AM   #10
Pentaxian




Join Date: Jan 2010
Location: Michigan
Photos: Gallery
Posts: 2,207
QuoteOriginally posted by beholder3 Quote
In short it seems that Sony is cooking raws quite extensively
Quite extensively?

The Nyquist of the A7 111 is about 110 line pair per mm?
With Bayer it is about 60 / mm?

The graphs produced in the referenced report by a 3rd party, if accurately depicting the filter based on dark noise alone,
show a roll-off with shutter speeds at 30 seconds of approximately minus 0.8 dB/Octave.
That is 0.192 of a "stop" per octave of spatial resolution.

And what lenses and mounts and atmospherics are going to give 100+line pair/mm over 30 seconds on a 35mm camera?

I don't know anything about astro photography so I would be interested to hear about typical
resolution of high performance astro lenses and sensors.

Last edited by wombat2go; 11-24-2017 at 09:07 AM.
11-24-2017, 09:55 AM   #11
Pentaxian
photoptimist's Avatar

Join Date: Jul 2016
Photos: Albums
Posts: 3,372
QuoteOriginally posted by beholder3 Quote
Physics. When I first measured the temperature of my cameras sensors I was surprised to see how much they heat up (50 degrees Celsius after 20-30 minutes at 20C room temp). If you have put your hand into 50 C water you know that is quite something.

That actually is a reason why I am not really keen on more pixels or more readout speed. That would not benefit my requirements, but actually hurt them.

Here the old reaction "you don't have to use it" is misplaced as there is a tangible downside to certain "improvements".
Exactly!

This is why I'm appalled by all the demands that Pentax must provide 4k video and high FPS. It sounds like a fine idea until one looks at the thermal implications for the sensor and the DR implications of that heat. Sure some camera makers get both high DR and high FPS but only at a very high cost (and I'm not convinced they are actually delivering good DR if the tests don't use a pre-heated camera).

---------- Post added 11-24-17 at 10:09 AM ----------

QuoteOriginally posted by wombat2go Quote
Quite extensively?

The Nyquist of the A7 111 is about 110 line pair per mm?
With Bayer it is about 60 / mm?

The graphs produced in the referenced report by a 3rd party, if accurately depicting the filter based on dark noise alone,
show a roll-off with shutter speeds at 30 seconds of approximately minus 0.8 dB/Octave.
That is 0.192 of a "stop" per octave of spatial resolution.

And what lenses and mounts and atmospherics are going to give 100+line pair/mm over 30 seconds on a 35mm camera?

I don't know anything about astro photography so I would be interested to hear about typical
resolution of high performance astro lenses and sensors.
The roll-off they show is strictly that created from cooking the RAW file and is independent of the lens and scenery. It would be in addition to the roll-off created by the lens, atmospherics, etc.

The user of this camera would discover that their RAW images are unaccountably soft even if they have an extremely good lens, solid tripod, no atmospherics, etc. but have let their shutter speed slow to 4 seconds or longer.
11-24-2017, 04:57 PM   #12
Pentaxian
Digitalis's Avatar

Join Date: Mar 2009
Location: Adelaide.
Posts: 9,537
QuoteOriginally posted by wombat2go Quote
And what lenses and mounts and atmospherics are going to give 100+line pair/mm over 30 seconds on a 35mm camera?
With what the average photographer has to hand you would be lucky to see only a few arcseconds worth of resolution. With dedicated telescopes, appropriate filtration and precise collimation the major limiting factor for resolution is atmospherics. Which is why Hubble was such an important platform for astronomy, atmospherics didn't get in the way of observation.
11-24-2017, 05:39 PM   #13
Pentaxian




Join Date: Jan 2010
Location: Michigan
Photos: Gallery
Posts: 2,207
QuoteOriginally posted by Digitalis Quote
With what the average photographer has to hand you would be lucky to see only a few arcseconds worth of resolution. With dedicated telescopes, appropriate filtration and precise collimation the major limiting factor for resolution is atmospherics. Which is why Hubble was such an important platform for astronomy, atmospherics didn't get in the way of observation.
I read a bit about it today.
I am a bit interested to do some astro photos because there is a dedicated viewing area opened up on northern Michigan coast.
However I will wait till spring or so , not wanting to freeze out in the cold now.

The bare human eye resolves about 120 arcsec.
The Full Width Half Max (FWHM) of an extended source needs to be at least 3+ pixels for any sharpness from a camera.
- But I think there are not many extended sources out there ?

The A7R 111 has a Bayer Nyquist of about 60 line pair /millimetre
So putting the numbers together ( my calculations) , the A7R 111 with a 40 mm lens would need an extended source of about 270 arc sec for good resolution.

A pixel pair in the A7 R 111 is about 160 arcsec. So any point source ( provided the lens is not rolling down at 60+linepair/mm) will occupy maybe 2+ pixels.

I read the o/p's links a couple of times and i don't see how the very low roll off out to the Nyquist as shown by the authors, would cause their stars to disappear.
Maybe they have an unrecognised system problem like their post processing software or are using 8 bit monitors with low level images.
11-24-2017, 06:10 PM - 1 Like   #14
Pentaxian
photoptimist's Avatar

Join Date: Jul 2016
Photos: Albums
Posts: 3,372
Pixel size of the A7Riii is 4.5 m.
With a 40 mm lens, there's about 23 arc-seconds of angle between adjacent pixels.
The diameter of the Airy disk for a 40 mm lens that is diffraction-limited at f/5.6 is a bit under 20 arc-seconds.
A 20 arc-second circle can easily fall entirely inside the bounds of a 23 arc-second square.

A good lens can and will resolve stars that fall onto single pixels. And a "hot-pixel" noise remover will detect those stars as spurious and delete them.
11-24-2017, 06:32 PM   #15
Pentaxian




Join Date: Jan 2010
Location: Michigan
Photos: Gallery
Posts: 2,207
QuoteOriginally posted by photoptimist Quote
With a 40 mm lens, there's about 23 arc-seconds of angle between adjacent pixels.
Yes,
The Nyquist for monochrome is then about 46 arcsec and for Bayer cell is about 1.6 times that = 73 arc sec.

I used the diagonal out to the corner to get the 40mm AOV so my numbers were more conservative, using 80 arcsec for the Bayer.
Then FWHM of 3 pairs = 240 arc secs for smallest extended sources.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
astro, camera, dr, fps, heat, implications, lens, lenses, level, noise, photo industry, photography, pixel, reduction, sensor, sony, stars
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Sony a6000 or Sony a6300 or Sony a6500? LeDave Non-Pentax Cameras: Canon, Nikon, etc. 5 09-27-2018 03:45 AM
People If your pictures aren't good enough, you're not close enough, Robert Capa Kerrowdown Post Your Photos! 35 07-08-2018 12:46 AM
Pixel shift appears in the new Sony A7riii travelswsage Non-Pentax Cameras: Canon, Nikon, etc. 15 10-28-2017 07:18 AM
645Z owners: Will you switch to a Sony 75MP A7RIII? SeattleDucks Pentax Medium Format 79 04-09-2016 06:20 AM
FA Limited lens series won the 2010 Good Design Long Life Design Award Patriot Pentax SLR Lens Discussion 26 11-29-2010 06:16 AM



All times are GMT -7. The time now is 01:43 PM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top