New sensor tech coming from Panasonic

Started Feb 4, 2013 | Discussions thread
Dheorl
Senior MemberPosts: 1,548Gear list
Like?
Re: New sensor tech coming from Panasonic
In reply to RicksAstro, Feb 5, 2013

RicksAstro wrote:

Dheorl wrote:

RicksAstro wrote:

DrWhom wrote:

SeanU wrote:

http://panasonic.co.jp/corp/news/official.data/data.dir/2013/02/en130204-6/en130204-6.html

-- hide signature --

Sounds interesting - I wonder what the net gain will be in terms of light collection. If the current Bayer array blocks 50-70% of light, AT BEST we can probably hope for 1/2 stop improvement (if I am understanding this correctly).

Which is a lot! We're getting close to hitting the physics barriers, so 1/2 stop is a major breakthrough.

Purely out of curiosity, what physical barriers are we getting close to hitting?

Basically, the read noise is getting very, very low, and microlenses are capturing light from most of the surface area. Right now, photon noise is the dominant noise source of most captured images, which is a physical limit. If read noise was zero, photon & quantization noise would still exist in our images.

The quantum efficiencies (ratio of incident photons to recorded photon/electrons) of the best of the latest chips is in the 50-60% range.

The best they could get is 100% with the same color tech (traditional bayer array), which is one stop or less. With traditional Bayer color filters, it's doubtful that we'll see anything close to that.

But the Bayer array itself isn't as efficient as possible. If you consider a given area, even that perfect Bayer sensor is only capturing 50% of green and 25% of red and blue per unit area.

The ultimate sensor is one that counts every photon and records it's wavelength and spacial position on the sensor, and has no upper limit to the number counted and has no read or ADC noise. The quantization noise would be as low as possible (one photon/electron) but photon noise would still exist.

But even this currently way far fetched approach would only gain 2-3 stops of noise over the best tech of today's production DSLRs.

Panasonic's approach is in between this, so it's going past what Bayer can give us given the same electronic tech.

I'm not an electronics genious, so I may have some of the numbers wrong, but I think I'm pretty close.

-- hide signature --

Thanks for the explanation.

So if sigma could perfect the foveon sensor tech (and I mean unrealistically perfect, but just a thought excersise) then it would be 3 stops better than the current best (100% at each efficiency in all 3 primary colours). So then the next level or perfection would be to get rid of the RGB layers all together and just have a (near) continuous spectrum of absorption levels going down through the device...

Why am I making organic solar cells for my dissertation and not doing this =). I guess I'm utlimately working towards a similar aim, getting the largest electron flow out of a given light source.

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow