Pages: [1]   Go Down

Author Topic: Sony and Hasselblad developing a new sensor?  (Read 3387 times)

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Sony and Hasselblad developing a new sensor?
« on: November 28, 2013, 01:24:52 am »

Hi,

Our friend Stefan Steib, of Hartblei.de fame, published a link to a Sony rumor: http://www.sonyalpharumors.com/sr4-hasselblad-and-sony-to-make-a-joint-sensor-announcement/

The rumor comes from three sources, and Andrea classified it as SR4, mean highly probable.

Sonyalpharumors usually has excellent info. My guess is that Sony leaks info as a marketing channel.

Best regards
Erik
Logged
Erik Kaffehr
 

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Re: Sony and Hasselblad developing a new sensor?
« Reply #1 on: November 28, 2013, 01:53:30 am »

If the "all sensor can have different shutter speed" rumor is true, this would probably have to mean it is a Foveon like True RGB sensor.

I don't see how they would manage with Bayer demoisacing otherwise.

This could mean many things:
- local shutter per pixel?
- different sensitivities to light?
- I heard rumors in the past that Sony was working on non linear/logarithmic sensor with an ability to take much higher saturation levels in order to manage highlights better. They would have had working prototypes already some years ago. They may have become able to control this on a pixel by pixel basis?

Cheers,
Bernard

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Sony and Hasselblad developing a new sensor?
« Reply #2 on: November 28, 2013, 02:05:33 am »

Hi,

I would guess they could have two different groups of pixels. On group receiving a short exposure and one group a long exposure, doing fusion in camera. Having two modes, extended dynamic range or high resolution.

I don't think it would be "true RGB".

Best regards
Erik

If the "all sensor can have different shutter speed" rumor is true, this would probably have to mean it is a Foveon like True RGB sensor.

I don't see how they would manage with Bayer demoisacing otherwise.

This could mean many things:
- local shutter per pixel?
- different sensitivities to light?
- I heard rumors in the past that Sony was working on non linear/logarithmic sensor with an ability to take much higher saturation levels in order to manage highlights better. They would have had working prototypes already some years ago. They may have become able to control this on a pixel by pixel basis?

Cheers,
Bernard

Logged
Erik Kaffehr
 

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Re: Sony and Hasselblad developing a new sensor?
« Reply #3 on: November 28, 2013, 03:14:28 am »

I would guess they could have two different groups of pixels. On group receiving a short exposure and one group a long exposure, doing fusion in camera. Having two modes, extended dynamic range or high resolution.

That would be a mere copy of what Fujifilm has been doing for 5-6 years.

Cheers,
Bernard

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: Sony and Hasselblad developing a new sensor?
« Reply #4 on: November 28, 2013, 03:52:06 am »

If the "all sensor can have different shutter speed" rumor is true, this would probably have to mean it is a Foveon like True RGB sensor.

I don't see how they would manage with Bayer demoisacing otherwise.
Why not? If they somehow expose each sensel for a different time (hopefully one that is a function of the light that hits it) and keeps a separate record of that sensels exposure time, then the two pieces together should be enough to recreate a mosaiced image with larger DR than what each sensel is capable of in itself. Store it as 16-bit integers and run regular demosaic algorithms to obtain a full-resolution image.

It would be exciting if one of the big sensor manufacturers did a modern MF sensor, it nothing else than bragging rights. One has to wonder if revolutionary cmos technology will be introduced in the (relatively small scale) MF world, where the norm seems to be that small sensors gets the goodies first.

-h
« Last Edit: November 28, 2013, 03:55:48 am by hjulenissen »
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Sony and Hasselblad developing a new sensor?
« Reply #5 on: November 28, 2013, 09:48:11 am »

Something wrong with that?

Best regards
Erik

That would be a mere copy of what Fujifilm has been doing for 5-6 years.

Cheers,
Bernard

Logged
Erik Kaffehr
 

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Re: Sony and Hasselblad developing a new sensor?
« Reply #6 on: November 28, 2013, 12:06:16 pm »

It sounds to me like electronic shutter in the pixel circuitry which automatically turns off at full well capacity. It would make ETTR history. You now pick exposure time based on the minimum for action blur relying on the circuit to prevent clipping.
Logged

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Re: Sony and Hasselblad developing a new sensor?
« Reply #7 on: November 28, 2013, 12:10:45 pm »

It also wipes out the cost of the whole metering system. As an electronic the exposure could be any multiple of a 1/1000 of a second. Forget the approximation of exposure doubling.
Logged

lelouarn

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 88
Re: Sony and Hasselblad developing a new sensor?
« Reply #8 on: November 28, 2013, 01:49:49 pm »

I know it's possible to do multiple non-destructive reads of each individual pixel in a CMOS image sensor. So basically each pixel is read, while the detector is still being exposed to light. If you store flux in a pixel as a function of time you can reduce noise (by doing a linear fit of the flux). Also, it's possible to increase dynamic range, since you can expose until the last pixel saturates (because from previous reads - where the pixel was not yet saturated - you can get the unsaturated value).
It requires a lot of processing, but is technically feasible.
Logged

Rob C

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 24074
Re: Sony and Hasselblad developing a new sensor?
« Reply #9 on: November 28, 2013, 02:32:53 pm »

It also wipes out the cost of the whole metering system. As an electronic the exposure could be any multiple of a 1/1000 of a second. Forget the approximation of exposure doubling.


Never mind shutter bounce!

Rob C

uaiomex

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1211
    • http://www.eduardocervantes.com
Re: Sony and Hasselblad developing a new sensor?
« Reply #10 on: December 02, 2013, 02:13:41 pm »

And the mirror would have to go once and for all.

Eduardo



Never mind shutter bounce!

Rob C
Logged

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: Sony and Hasselblad developing a new sensor?
« Reply #11 on: December 03, 2013, 04:35:07 am »

The "individually exposed sensel" sort of mirrors the "individually backlit LCD pixel" display. It seems to me that you just move the problem from one place to another (potentially a good or bad thing).

In the case of LCD tvs, having one backlight per pixel seems to not be cost-effective; the LCD panel would no longer be needed if you go all the way to the subpixel, and what you have is effectively a LED-tv, not a LCD-tv. What is cost-effective, though (or at least quality-boosting) is to have individual LED backlights operating on groups of LCD pixels. Since many/most images are kind of smooth/lowpass (with low-contrast, high-frequency local modulation vs high-contrast, low-frequency large-scale modulation, or at least our vision seems limited to such an interpretation), it makes sense to approach the high-resolution/high-contrast ideal with this compromise where you get either high-resolution or high-contrast (but not both at the same time).

The GND-filter does something similar: part of the scene is made darker, thus, the DR is compressed before entering the sensor. This comes at the cost of reduced total exposure (unless you compensate by increasing time/aperture). One could expand this concept by inserting a sort of LCD panel in front of the sensor, and control it using the liveview output (attenuate light in areas that are brighter).

Without specific knowledge of sensor technology: if individual readout/reset of sensels is "hard", might it be possible to make them auto-reset? Once the well has reached its capacity (no sooner or later), its charge is dumped (into "the garbage"), and it starts accumulating once again. When the integration period is over, you read every sensel like today, but their values would be different: sensels that previously would have been saturated, are now "wrapped around" like a phase value. By using clever 2-d unwrapping, one might have an estimate of the true sensel values (assuming that most scenes are smooth, just like highlight recovery).

-h
« Last Edit: December 03, 2013, 04:43:27 am by hjulenissen »
Logged

ondebanks

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 858
Re: Sony and Hasselblad developing a new sensor?
« Reply #12 on: December 03, 2013, 07:55:13 am »

Without specific knowledge of sensor technology: if individual readout/reset of sensels is "hard", might it be possible to make them auto-reset? Once the well has reached its capacity (no sooner or later), its charge is dumped (into "the garbage"), and it starts accumulating once again. When the integration period is over, you read every sensel like today, but their values would be different: sensels that previously would have been saturated, are now "wrapped around" like a phase value. By using clever 2-d unwrapping, one might have an estimate of the true sensel values (assuming that most scenes are smooth, just like highlight recovery).

That's similar to how the photon-counting Faint Object Camera (decommissioned in the 2002) on the Hubble Space Telescope wrapped its pixel intensities around:

"Unfortunately, due to memory constraints (Section 4.8 ), the largest formats (512×1024 and 512×1024) only have 8-bit deep pixels, i.e. they can only count up to 255 counts per pixel (Table 7). For a point source with a count rate of 0.25 counts sec-1 pixel-1, the peak pixel of the star will count up to 255 in 1000 seconds then reset to 0 on the next count. The image has lost track of those counts, resulting in a loss of photometry. When using these formats, short exposures should be taken, then co-added, as long as the sources in the image maintain count rates that are linear."

It was very difficult if not impossible to reconstruct the correct counts for a wrapped-around FOC point source pixel. Might have been doable for an extended source.

Ray
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: Sony and Hasselblad developing a new sensor?
« Reply #13 on: December 03, 2013, 02:28:16 pm »

I know it's possible to do multiple non-destructive reads of each individual pixel in a CMOS image sensor. So basically each pixel is read, while the detector is still being exposed to light. If you store flux in a pixel as a function of time you can reduce noise (by doing a linear fit of the flux). Also, it's possible to increase dynamic range, since you can expose until the last pixel saturates (because from previous reads - where the pixel was not yet saturated - you can get the unsaturated value).
It requires a lot of processing, but is technically feasible.
I have read of sensors used in security cameras (where high dynamic range is very important) that do something like this: checking multiple times during the exposure period whether a photosite has filled up, and for those that do, noting the level at the last reading before it filled, and the time of that reading: then the value can be linearly extrapolated to the count that would have been received in the full exposure period. This involved ADC at each photosite, which is probably better suited to low-resolution devices with large photosites, like those security cameras.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Sony and Hasselblad developing a new sensor?
« Reply #14 on: December 04, 2013, 12:55:31 am »

Hi,

My guess is that a sensor developed for photography would be careful with how silicon estate is used. I guess that the design would emphasize on simplicity, something that could be done with a minimum of gates.

My guess is the sensor would go the way some of the Fuji sensors work. A bank of pixels using short exposures and another bank of pixels for long exposures. Problem is that I don't know how to do it without causing a mess in demosaic.

Best regards
Erik

I have read of sensors used in security cameras (where high dynamic range is very important) that do something like this: checking multiple times during the exposure period whether a photosite has filled up, and for those that do, noting the level at the last reading before it filled, and the time of that reading: then the value can be linearly extrapolated to the count that would have been received in the full exposure period. This involved ADC at each photosite, which is probably better suited to low-resolution devices with large photosites, like those security cameras.
Logged
Erik Kaffehr
 

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: Sony and Hasselblad developing a new sensor?
« Reply #15 on: December 04, 2013, 03:18:10 am »

Hi,

My guess is that a sensor developed for photography would be careful with how silicon estate is used. I guess that the design would emphasize on simplicity, something that could be done with a minimum of gates.

My guess is the sensor would go the way some of the Fuji sensors work. A bank of pixels using short exposures and another bank of pixels for long exposures. Problem is that I don't know how to do it without causing a mess in demosaic.

Best regards
Erik
What about a largeish photography sensor optimized for moderate DOF and none-WA? If you know that there is going to be _some_ diffraction going on, you can perhaps reduce the OLPF filtering. If you know that light is going to enter the sensor at quite normal angles, you can afford to have very small active light sensing area, coupled with micro-lenses that focus a large area into the small relative sensel area. Then you have a lot of silicon real-estate that can be used for supporting electronics (increase well capacity, per-sensel ADC,…)

-h
Logged

Phil Indeblanc

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2017
Re: Sony and Hasselblad developing a new sensor?
« Reply #16 on: December 04, 2013, 09:19:01 pm »

Quote
I have read of sensors used in security cameras (where high dynamic range is very important) that do something like this:

Yes, Sony has this tech in the HD and Dynaview cameras. they are for when there is bright light in one part of the image while the rest fall in shadow. it is "timed" to expose this shadow area rather than meter for the light fall. Not sure how it works, but they are great, yet max at 600lines.

Logged
If you buy a camera, you're a photographer...
Pages: [1]   Go Up