Nano technology based ISO 50000 sensor is developed

Started Nov 10, 2005 | Discussions
Raxel
Regular MemberPosts: 407
Like?
Nano technology based ISO 50000 sensor is developed
Nov 10, 2005

Korean researchers in ETRI (Electronics and Telecommunications Research Institute) and venture company planet82 have developed a new type of sensor based on nano technology called SMPD.

Main advantage of SMPD is its huge sensitivity (500 times greater than other sensors.) In terms of ISO, its sensitivity is about 50000 ISO. During the demonstration, SMPD sensor can capture 30fps video at extremely dark condition (1lux, where human eye cannot see anything)

Furtheremore SMPD is much cheaper than CCD as it uses CMOS production line. (1/100 price of CCD)

Current chip is VGA resolution which is for video cameras and CCTV, but researchers say they will increase MP count soon.

Imagine tiny digicam that can goes up to ISO 10000 and is noiseless at ISO 3200. Imagine DSLR that goes up to noiseless ISO 25600. Isn't it nice?

related link:
http://www.planet82.co.kr/english/rad_cmos_06.asp

Related korean news link with video link:

http://news.naver.com/news/read.php?mode=LSS2D&office_id=052&article_id=0000098369§ion_id=105§ion_id2=230&menu_id=105

alanr0
Senior MemberPosts: 1,237
Like?
Re: Nano technology based ISO 50000 sensor is developed
In reply to Raxel, Nov 10, 2005

Raxel wrote:

Korean researchers in ETRI (Electronics and Telecommunications
Research Institute) and venture company planet82 have developed a
new type of sensor based on nano technology called SMPD.

Main advantage of SMPD is its huge sensitivity (500 times greater
than other sensors.) In terms of ISO, its sensitivity is about
50000 ISO.

Imagine tiny digicam that can goes up to ISO 10000 and is noiseless
at ISO 3200. Imagine DSLR that goes up to noiseless ISO 25600.
Isn't it nice?

I see SNR is specified as n/a (not available). There may be an SNR advantage in amplifying at each photosite, but I will reserve judgement until I see evidence that they can do better than the Poisson limit on photon statistics.
BTW, have they told Canon that CMOS sensors are 1 Mpixel or smaller?

With 1000x higher gain than conventional sensors, the dynamic range is going to be limited. The link specifies 80 dB, and I would be fascinated to hear how this is defined - possibly the AGC allows gain adjustments, so that 80 dB might not apply to dynamic range withing a single exposure.

Interesting stuff in any case.
--
Alan Robinson

Reply   Reply with quote   Complain
rander3127
Veteran MemberPosts: 4,628
Like?
Re: Nano technology based ISO 50000 sensor is developed
In reply to Raxel, Nov 10, 2005
-- hide signature --

Every other day someone comes up with technology that is not only 100x
better than what we have, but FAR cheaper. Too bad we never see
any of it. I'm still waiting (28 years after the fact) for roads that don't
wear out. NANO tech has great promise, but some of it is a huge money
pit and a scam.
-Rich

Reply   Reply with quote   Complain
Vik
Vik
Contributing MemberPosts: 705
Like?
Re: Nano technology based ISO 50000 sensor is developed
In reply to alanr0, Nov 10, 2005

alanr0 wrote:

I see SNR is specified as n/a (not available). There may be an SNR
advantage in amplifying at each photosite, but I will reserve
judgement until I see evidence that they can do better than the
Poisson limit on photon statistics.

Doesn't that mean they have 15-20 dB gain in this preamplifying circuit?

BTW, have they told Canon that CMOS sensors are 1 Mpixel or smaller?

There are also many errors on the site (f.x.see data mixed up in both tables)...

With 1000x higher gain than conventional sensors, the dynamic range
is going to be limited. The link specifies 80 dB, and I would be
fascinated to hear how this is defined - possibly the AGC allows
gain adjustments, so that 80 dB might not apply to dynamic range
withing a single exposure.

Sure resulting DR was measured with AGC circuit on. During a single exposure the gain applied to individual units (photosites) may just work as DR compressor, so we will get compressed output DR thus significally narrowed from much wider input DR. The question about S/N ratio still open - what would be the noise levels and how easy they can be controlled.

Interesting stuff in any case.

Yes, at least until we know more.

Reply   Reply with quote   Complain
Vik
Vik
Contributing MemberPosts: 705
Like?
Re: Nano technology based ISO 50000 sensor is developed
In reply to Raxel, Nov 10, 2005

Imagine small 400g P&S camera with ISO 64000, 1/8" sensor and x30 optical internal zoom (sealed lens) all in waterproof case for just $500.
No flash unit, just like EOS 5D.

Reply   Reply with quote   Complain
Shaun Williamson
Senior MemberPosts: 1,641
Like?
Re: Nano technology based ISO 50000 sensor is developed
In reply to Vik, Nov 11, 2005

Well its a nice idea, so lets wait and see if it leads to any real products.
I know of one UK university who have been working on holographic
storage devices for the past 20 year (promises 100s of terabytes on a 1
inch cube) but no real practical devices have ever arrived in the market.

Reply   Reply with quote   Complain
Just Looking
Veteran MemberPosts: 5,740
Like?
What does ISO 50000 mean?
In reply to Raxel, Nov 11, 2005

According to definitions that I'm aware of, that means the mean focal-plane illuminance times time for a proper exposure should be 10/50000 or 0.0002 Lux-seconds, or about 43 photons in a 4-micron square pixel (based on my recollection of 1 Lux being about 13,500 visible photons per second per square micron -- someone check me on this).

Shot noise (standard deviation of Poisson statistics) for 43 photons is about 6 or 7 photons, for a pixel SNR (assuming monochrome) of about 6 or 7. Since this is below the level needed for a "first acceptable" photo according to ISO standards, an ISO speed of 50,000 is not possible with pixels that small. Now if they were about 6 microns square, then it would be theoretically possible to hit SNR=10 for an "acceptable" ISO 50000 (if everything else is perfect). To get up to the 40 SNR needed for a "first excellent" photo, another factor of four in each dimension, or 24-micron pixel would be needed. For color, throw in at least a factor of three in pixel area.

Interesting possibilities . . . certainly any technology that lets you be limited only by photon statistics is the holy grail. Is this one?

Raxel wrote:

Korean researchers in ETRI (Electronics and Telecommunications
Research Institute) and venture company planet82 have developed a
new type of sensor based on nano technology called SMPD.

Main advantage of SMPD is its huge sensitivity (500 times greater
than other sensors.) In terms of ISO, its sensitivity is about
50000 ISO. During the demonstration, SMPD sensor can capture 30fps
video at extremely dark condition (1lux, where human eye cannot see
anything)

Did you ever read by candlelight? 1 lux is the illuminance provided by a standard candle (1 candela) at a distance of 1 meter. You can easily see in that. I think 30 fps at 1 lux, f/2 could be done with about ISO 5000, never mind 50000.

j

Furtheremore SMPD is much cheaper than CCD as it uses CMOS
production line. (1/100 price of CCD)

Current chip is VGA resolution which is for video cameras and CCTV,
but researchers say they will increase MP count soon.

Imagine tiny digicam that can goes up to ISO 10000 and is noiseless
at ISO 3200. Imagine DSLR that goes up to noiseless ISO 25600.
Isn't it nice?

related link:
http://www.planet82.co.kr/english/rad_cmos_06.asp

Related korean news link with video link:

http://news.naver.com/news/read.php?mode=LSS2D&office_id=052&article_id=0000098369§ion_id=105§ion_id2=230&menu_id=105

Reply   Reply with quote   Complain
Quill
Forum MemberPosts: 70
Like?
Re: Nano technology based ISO 50000 sensor is developed
In reply to Raxel, Nov 11, 2005

I'm not sure this is related or not, but it is sure interesting:

http://www.berkeley.edu/news/media/releases/2005/04/21_superlens.shtml

Reply   Reply with quote   Complain
Unda Covalava
Veteran MemberPosts: 7,154
Like?
Yes!
In reply to Shaun Williamson, Nov 11, 2005

I remember this too! I first heard about it in the early '90s. It was "a couple of years off" and there was a prototype already built. Not only 100s of terabytes per cubic inch, but also incredibly fast since it was all transferred via light (optics). Needless to say, that was a long time ago, and I'm still waiting.

It almost seems like the incentive is for incremental improvement.

Oh, and Foveon didn't quite pan-out either.

Shaun Williamson wrote:

Well its a nice idea, so lets wait and see if it leads to any real
products.
I know of one UK university who have been working on holographic
storage devices for the past 20 year (promises 100s of terabytes on
a 1
inch cube) but no real practical devices have ever arrived in the
market.

Reply   Reply with quote   Complain
Roland Karlsson
Forum ProPosts: 20,728Gear list
Like?
Re: Yes!
In reply to Unda Covalava, Nov 11, 2005

Unda Covalava wrote:

Oh, and Foveon didn't quite pan-out either.

Technically - Foveon is a success really. There has only been made one serious sensor - and it is competitive. With its 3.4 Mpixel is as sharp as an 6-8 Mpixel camera depending on who you ask. The pictures looks clear and fresh. There are disadvantages and advantages.

Compare this to the first Bayer sensors and you can see that technologically the Foveon technique show great promise.

Roland

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma DP2 Quattro RX100 III Pentax K-3 Pentax smc DA 15mm F4 ED AL Limited +8 more
Reply   Reply with quote   Complain
Olivier GALLEN
Contributing MemberPosts: 516
Like?
'ISO 50000' doesn't mean a lot...
In reply to Raxel, Nov 14, 2005

It just means "put that Gain up!"... for Signal as well as Noise.
...and is the Noise (especially Read Noise) detailled here? No.

Bottom-Line: take any recent sensor, increase the Gain beyond sanity... and there you are -> you can take noisy videos with little light, just as they did. Is this useful for photography: no. Is there anything new here: probably not.

Moreover, I would be careful about a company that advertize an Aperture Ratio of 1.87% as an advantage!!! ( http://www.planet82.co.kr/english/rad_cmos_03.asp )

Olivier

PS: and don't expect to much... Physics say that Light has Noise (Photon Noise = square root of Signal) and there is nothing Technology can do about this one...

Reply   Reply with quote   Complain
PalmsWestPhoto
Forum ProPosts: 10,003
Like?
Re: Yes!
In reply to Roland Karlsson, Nov 14, 2005

problem is that the only followup so far was a huge mess with lousy image quality and after that we havent seen or heard anything.

I had hopes for foveon in video cameras replacing 3CCD cams and having pixel binning seemed to be a good start for them but possibly there are issues as well.

-- hide signature --

Michael Salzlechner
http://www.PalmsWestPhoto.com

Reply   Reply with quote   Complain
dciobota
Contributing MemberPosts: 758
Like?
Hmmm...
In reply to Raxel, Nov 14, 2005

Sony VGA chips already claim sensitivity of

Daniel

-- hide signature --
Reply   Reply with quote   Complain
Roland Karlsson
Forum ProPosts: 20,728Gear list
Like?
Re: Yes!
In reply to PalmsWestPhoto, Nov 14, 2005

PalmsWestPhoto wrote:

problem is that the only followup so far was a huge mess with lousy
image quality and after that we havent seen or heard anything.

Now - I am not a 100% Foveon fan - but this does not sound correct.

The first Sigma camera (SD9) was indeed a lousy camera. It sampled the picture in small dots and had no anti alias filter. Therefore you get strong aliasing for sharp lenses. Lots of horrible pictures out there.

But the SD10 has micro lenses that works as a good enough anti alias filter, and it produces very nice pictures. There are lots of pictures on the web - good looking ones.

I had hopes for foveon in video cameras replacing 3CCD cams and
having pixel binning seemed to be a good start for them but
possibly there are issues as well.

It is true that the promise of video cameras with Foveon is not realized. Don't know why. Maybe it was not enough time for Foveon or interest from any camera maker. Don't know.

Roland

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma DP2 Quattro RX100 III Pentax K-3 Pentax smc DA 15mm F4 ED AL Limited +8 more
Reply   Reply with quote   Complain
rander3127
Veteran MemberPosts: 4,628
Like?
Re: Nano technology based ISO 50000 sensor is developed
In reply to Raxel, Nov 15, 2005

Raxel wrote:

Korean researchers in ETRI (Electronics and Telecommunications
Research Institute) and venture company planet82 have developed a
new type of sensor based on nano technology called SMPD.

Main advantage of SMPD is its huge sensitivity (500 times greater
than other sensors.) In terms of ISO, its sensitivity is about
50000 ISO. During the demonstration, SMPD sensor can capture 30fps
video at extremely dark condition (1lux, where human eye cannot see
anything)

Yes, sure. 1 candle at 1 meter won't allow the human eye to see
anything. Under what blind person's measuring system???

Or, sorry! Here's a 0.1 lux camera for $200.

http://www.bronzepoint.com/color-board-camera.htm

Good luck, Korea.
-Rich

Reply   Reply with quote   Complain
Just Looking
Veteran MemberPosts: 5,740
Like?
strange point of view
In reply to Roland Karlsson, Nov 15, 2005

Roland Karlsson wrote:

The first Sigma camera (SD9) was indeed a lousy camera. It sampled
the picture in small dots and had no anti alias filter. Therefore
you get strong aliasing for sharp lenses. Lots of horrible pictures
out there.

The only truth in what you say is that the SD9 was somewhat more likely to show a little aliasing than the SD10. But still it was seldom seen, and certainly I don't know where these "lots of horrible pictures" are.

Some users actually prefer the SD9 over the SD10 because its smaller sampling apertures make it a little sharper; the aliasing that comes with the sharpness is seldom noticable.. It is certainly capable of excellent pictures, as the many examples in the SD9 users gallery demonstrates:

http://www.pbase.com/sigmasd9/user_home

But the SD10 has micro lenses that works as a good enough anti
alias filter, and it produces very nice pictures. There are lots of
pictures on the web - good looking ones.

Same is true of SD9.

j

Reply   Reply with quote   Complain
photoaddict
Contributing MemberPosts: 666
Like?
Sony already done .1 lux sensor?
In reply to rander3127, Nov 16, 2005

"Or, sorry! Here's a 0.1 lux camera for $200.

http://www.bronzepoint.com/color-board-camera.htm "

What caught my attention was that it said it uses 1/3" Sony Super HAD CCD, S/N ratio over 48db.

What's up with that?!?! If Sony sensor could detect so well in the dark, why hasn't it increased its ISO in its consumer cameras?

-jeff

Reply   Reply with quote   Complain
Roland Karlsson
Forum ProPosts: 20,728Gear list
Like?
Re: Sony already done .1 lux sensor?
In reply to photoaddict, Nov 16, 2005

photoaddict wrote:

"Or, sorry! Here's a 0.1 lux camera for $200.

http://www.bronzepoint.com/color-board-camera.htm "

What caught my attention was that it said it uses 1/3" Sony Super
HAD CCD, S/N ratio over 48db.

What's up with that?!?! If Sony sensor could detect so well in the
dark, why hasn't it increased its ISO in its consumer cameras?

The resolution is 400 lines -- hmmmm .. that should mean leass than 100K pixels. But pinhole? Actuall pinhole? Or do the mean very small lens?

BTW - this kind of cameras has a lousy picture quality. It has been found that the quality is so bad that the result is useless for identfying people. So - maybe they don't have all that high quality demands when defining ISO?

Roland

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma DP2 Quattro RX100 III Pentax K-3 Pentax smc DA 15mm F4 ED AL Limited +8 more
Reply   Reply with quote   Complain
Vik
Vik
Contributing MemberPosts: 705
Like?
Re: Sony already done .1 lux sensor?
In reply to Roland Karlsson, Nov 16, 2005

Probably small lens attached.

Even if it is VGA 300k pixels, due to quite wide angle and/or cheap lens the face resolution would be quite bad.

Reply   Reply with quote   Complain
Leon Wittwer
Forum ProPosts: 12,808
Like?
Tough sell...
In reply to Raxel, Nov 16, 2005

To use this sensor in the 1Ds MK II and maintain the same signal to noise at ISO 50000 that you now have at ISO 1600, the sensor would have to have less than about 500,000 pixels. If the pixels size is not increased the signal to noise at ISO 1600 would increase from about 0.025 to 0.14. At ISO 50000, the signal to noise would be about 0.8, which is not very good. The signal to noise is approximately one over the square root of the number of photons collected in the pixel. Of course, if you make the 1Ds MK II sensor about 32 times larger in area with the same number of pixels, you are home free. That would be a difficult camera to carry. I welcome others to check my figures.
--
Leon
http://homepage.mac.com/leonwittwer/landscapes.htm

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads