New DSLR with Exmor R CMOS sensor

[bakubo]

Someone on dpreview posted the link to this new Sony page:

http://australia.sony.com.au/SonyTwilig ... prize.html

"This feast of football celebrates the launch of new Cyber-shot, Handycam and α DSLR models featuring the Exmor R CMOS sensor; and Sony's commitment to the world game as an Official FIFA Partner."

[agorabasta]

Most likely - nothing more than awkward wording.

Yet the backlit tech may hold something for the future DSLRs.
As the wiring is no obstacle anymore, the 'pixel' could be made of some tens of individual tiny photosites. E.g. imagine a square pixel of 49 photosites, of which 21 green, 14/14 for blue/red; all connected in parallel for given colour. Could be arranged in 7 stripes as RGBGRGB or RBGGGBR or any other config.

That would make demosaicing unnecessary, and provide a Foveon-like per-pixel resolution advantage.

[David Kilpatrick]

Although the Bayer pattern support camp claims that two greens per one red, one blue matches human vision it is more an excuse for a problem created by the grid pattern. The solution you suggest could be done with equal R G B elements, or with a very finely tuned mix - not necessarily 2:1:1. With the size of the photosites on small sensors like the new TX-1 and WX-1, you could fit three of those into the same space as a single sensel on an A900/700-ish sensor. Under one microlens for each cluster of three.

That would be amazing if it could be done, RGB direct from the sensor. You'll have to patent a few hundred variations of patterns...

David

[KevinBarrett]

Either you misquoted or they fixed it. The page now shows the following text:

"This feast of football celebrates the launch of new Cyber-shot, Handycam and α DSLR models featuring the Exmor CMOS sensor; and Sony's commitment to the world game as an Official FIFA Partner."

No "R."

[Javelin]

Sony was sensitive to this press spilling over into DSLR rumours. When exmor R talk first hit the DPR site there was a post from Sony on the blog about how upset they were that people were drawing their own conclusions on tech from other products. when they saw it happening again now they probably made them re-word their site.

it wasn't a misquote. they changed it a couple hours after the quote appeared on dpr..

[dkloi]

Most likely - nothing more than awkward wording.

Yet the backlit tech may hold something for the future DSLRs.
As the wiring is no obstacle anymore, the 'pixel' could be made of some tens of individual tiny photosites. E.g. imagine a square pixel of 49 photosites, of which 21 green, 14/14 for blue/red; all connected in parallel for given colour. Could be arranged in 7 stripes as RGBGRGB or RBGGGBR or any other config.

That would make demosaicing unnecessary, and provide a Foveon-like per-pixel resolution advantage.

In that situation, you might as well just read out the sub-pixel values individually and end up with 49 times the pixel resolution in the final image . If you want to achieve full RGB per output image pixel, you can simply downsample a Bayer image by a factor of 2 linear (4 area). By keeping the original pixel data, you can choose to either have high resolution, or downsample later and get good per output-pixel rendition.

It's interesting to note that the IMX061 EXMOR R backlit sensor what debuted in handycams actually uses a 1:6:1(RGB) Bayer array in a ClearVid (45 degree) arrangement (a la Fuji HR), but for what reasons I can only speculate. By covering the sensor with mostly green pixels, you maximise spatial resolution at the expense of colour resolution. The ClearVid array also increases horizontal and vertical resolution slightly (at the expense of diagonals). However, the video output is a downsampled version of the 6MP actual resolution so having only an eighth red and blue pixels probably doesn't impact that much on the HD output colour resolution compared to a conventional Bayer pattern.

From what I've seen, a Foveon sensor has about a 40% increase in linear resolution on realworld imagery over a 1:2:1 Bayer array. This makes sense since resolution is mainly determined by green channel data and adding extra blue and red coverage doesn't raise the overall spatial resolution as much. Hence, it would seem to me that a good tradeoff between spatial resolution, demosaicing artifacts (you can still get moire even with a Foveon sensor if there is image detail at higher spatial frequency than the Nyquist limit, you get foldback into the spatial sampling bandwidth), and RAW image size, would be to go towards a two layer sensor. The first layer would be all green sensitive, but allow red and green through. The second layer would be a red/blue chequerboard.

Stacked pixels (RGB) have been patented by Fuji (using dye-coupler technology developed from their colour film experience, US2005205903 (A1)) and in 2006 they demonstrated green dye pixels. http://techon.nikkeibp.co.jp/article/HO ... 15/123642/ I haven't found anything more recent unfortunately. A 2-layer architecture would give you the majority of the spatial and colour resolution of a full RGB sensor, but with 33% less information to have to cart off the sensor and process, and 33% less circuitry to produce (less fabrication steps) and go wrong. It would also boost sensitivity since more light is being turned into photoelectrons. In fact, I could envisage an array where one layer was a chequerboard of cyan and yellow, and the second layer a chequerboard of red and blue, with red underneath the cyan pixels and the blue under yellow. You could interpolate an RGB image from this, but this array would ideally turn all the incoming light into photoelectrons. This kind of pixel architecture would make ideal use of Fuji's dye-coupler technology .

Foveon uses the wavelength dependent absorption coefficient of silicon to perform colour separation which isn't that close to the CIE colour model, hence converting the three values recorded at each pixel site to a visual colour (colour is all in the eye/brain) is not straightforward. Using dyes with more selectivity at each pixel site would give better colour fidelity. I can imagine special Velvia CMOS sensors .

Cheers,
Daniel.

PS I came across Fuji patents using BSI. Sony has been the first to commercialise the technology in major mass production but I'm sure others will follow. Micron and OmniVision were in a race with Sony to develop BSI for cellphone camera sensors AFAIR.

[bakubo]

Either you misquoted or they fixed it.

I did a copy and paste. I just checked the site again and it still said what I posted, but then I did a refresh and got the new text. They changed it.

[alexramos7]

The ClearVid use a pattern like Fuji sensor from long time ago it is not new to Sony ClearVid sensor.

[dkloi]

The ClearVid use a pattern like Fuji sensor from long time ago it is not new to Sony ClearVid sensor.

I did actually say "ClearVid (45 degree) arrangement (a la Fuji HR)", wasn't claiming it was new at all. The intriguing thing was the use of a 1:6:1 RGB Bayer array, in addition to ClearVid.

Cheers,
Daniel.

[PhotoTraveler]

Finally, the first thread I have found on any of the 3 forums that someone stopped the madness and pointed out the advert doesn't say "R" in it anywhere.

I'm getting the impression it may have said it at one point in time. But clearly someone who typed it up just made a mistake. People really need to follow the rule of the right answer is the simple one. Sony made a typo, they fixed it. The twilight football is having handy cams and other cameras there, which is where Sony has been introducing R sensors. That's why a person most likely made a mistake making the ad.

Stop with conspiracy theories that Sony is trying to cover it up and there really is R APS/FF sensors coming. We have no basis for that at this time. It could happen, but it's not what we are seeing.

[bakubo]

Finally, the first thread I have found on any of the 3 forums that someone stopped the madness and pointed out the advert doesn't say "R" in it anywhere.

You are wrong. It said exactly what I quoted above. It has been changed.

I'm getting the impression it may have said it at one point in time.

Yes, see my post above.

But clearly someone who typed it up just made a mistake. People really need to follow the rule of the right answer is the simple one.

Please don't tell me what I should do. That is offensive and condescending. The world is full of examples where the right answer is not the simple one.

Sony had it on their website. They changed it. They screwed up.

[PhotoTraveler]

Give me a break, did you really need to respond to my comment where I stated it looks like it had been changed just to repeat what I had all ready said?

What I said is not insulting. I'm commenting on the insanity that one person finds this advert, the next thing we know the whole world of camera forums is going nuts with countless threads on this with all sorts of wild claims about R sensors and alphas. There are now countless thread post "DPR, Dyxum, etc" pointing to the advert and the folks going "look look" haven't even looked back to see it doesn't say what they think it says.

Now when the time comes, and there are no R sensors in the new alphas all we will see is countless threads of people moaning and saying sony tricked them, or has slimmy marketing, bla bla bla.

Sony makes mistakes on their websites all the time. Look how often lens information has been wrong, often for years with the wrong specs applied to a lens, or the name of a lens wrong (the 1.4/50 was listed as SAL50F18 for years on Sony sites).

Yes, the simple answer isn't always what happens, but it's a good starting point instead of going off on the R sensor parade just to have that never happen.

Your last line is correct, Sony screw up. But that's no reason to go all conspiracy on this and think they are trying to cover up their plans.

Sure, they could come out with R sensors in the alphas, that would be nice, if it happens, great. But lets wait for that to happen. Of course it making an APR or FF R sensor cost way more than a conventional CMOS sensor and reflects it in the cost of new models, folks won't be so excited then.

Finally, the first thread I have found on any of the 3 forums that someone stopped the madness and pointed out the advert doesn't say "R" in it anywhere.

You are wrong. It said exactly what I quoted above. It has been changed.

I'm getting the impression it may have said it at one point in time.

Yes, see my post above.

But clearly someone who typed it up just made a mistake. People really need to follow the rule of the right answer is the simple one.

Please don't tell me what I should do. That is offensive and condescending. The world is full of examples where the right answer is not the simple one.

Sony had it on their website. They changed it. They screwed up.

[KevinBarrett]

Hey guys, don't get catty. A useful thread to clear up a rumor has been blemished with a needless defensive exchange between two personalities. Clean it up or I will. Others might enter the terminology of the technology in a search and this page shows up in results as easily as any. Please represent the forum in a good light.

[agorabasta]

Quote DPR forum message by 'dougas' http://forums.dpreview.com/forums/read. ... e=32614265

After reading the Australian twilight football link that mentioned Sony Alpha Exmor-R cameras at the event I emailed Sony PR Office and got this reply this morning:

Hi Steve

Thanks for your email about Sony’s Twilight Football competition.

Unfortunately this was an error in the copy, there will not be Exmor R Alpha models at the event. There will, however, be Exmor models such as the A900.

Don’t hesitate to let me know if you need anything further.

Kind regards,
Kelly

Kelly Finlay
Public Relations Manager
Sony Australia

[01af]

Don't be silly everyone! The first Sony Alpha DSLR cameras with Exmor R sensors won't arrive before 2010 or 2011, I'm sure. The DSLRs currently announced to be released within the next weeks will have regular Exmor sensors, just like the current ones.

-- Olaf