double-layer sensors... a new norm to come?

[Emanuel]

 

[ND64]

It benefits mostly smartphone sensors, with increasing the photo diode area. This area is already large in big sensors, even 60mp big sensors. 

[kye]

1 minute ago, Eric Calabros said:

It benefits mostly smartphone sensors, with increasing the photo diode area. This area is already large in big sensors, even 60mp big sensors. 

I disagree - it will benefit all camera types.  This is because the second layer is in the same area of the sensor, and could be used to capture at a different amount of gain, which could then be used to extend the dynamic range of every pixel.  

Imagine a Sony sensor where it could have identical performance but use a second layer to extend DR by many stops...  

[ND64]

4 hours ago, kye said:

Imagine a Sony sensor where it could have identical performance but use a second layer to extend DR by many stops...  

Sorry thats not how it works. All these developments are about increasing the full well capacity, which is very low in small sensors, and decrease noise, which is relatively high in small sensors, to extend the DR, maybe 1/3 of a stop. 

Here is the formula: 20*log(full well capacity÷noise). You can play with this to estimate how much room to improve is still there. Capacity is already capped at 3500 electron per one um². And noise is already very low. A1 is at 1.8e now.

 

The only way to extend DR by several stops in conventional CMOS sensor is multi exposure solutions. Nikon is playing with a dual layer idea, but its on photo diode section, not electronics. They want to place a CMY layer on top of RGB layer, then set the upper layer exposure differently from the lower layer exposure. 

 

[kye]

9 minutes ago, Eric Calabros said:

Sorry thats not how it works. All these developments are about increasing the full well capacity, which is very low in small sensors, and decrease noise, which is relatively high in small sensors, to extend the DR. 

Here is the formula: 20*log(full well capacity÷noise). You can play with this to estimate how much room to improve is still there. Capacity is already capped at 3500 electron per one um². And noise is already very low. A1 is at 1.8e now.

The only way to extend DR by several stops in conventional CMOS sensor is multi exposure solutions. Nikon is playing with a dual layer idea, but its on photo diode section, not electronics. They want to place a CMY layer on top of RGB layer, then set the upper layer exposure different from the lower layer exposure. 

I had a closer look at the picture, and yeah, it's not what I thought.

Funnily enough, I thought "dual layer" meant two layers.  No.  Those words in the tweet are complete bullshit.

There is already two (or more) layers, and they're making it three (or more), but it's still a single sensor.  They've been gradually taking things that were taking up real estate on the front of the sensor (and therefore blocking light) and putting them behind the photo site for years.  This is an incremental upgrade that some numpty decided was the ultimate breakthrough and somehow none of the previous ones mattered or count in this "FIRST DUAL LAYER" spasm they posted.

I thought the implication that 'one became two' implied a second photodiode under the first, presumably with the ability to be operated independently.  Sadly, no.

[kye]

I don't know why I read the headlines any more.  This isn't the first time I've misinterpreted one that turned out to include a completely false statement of fact.

Maybe somewhere in the back of my mind I thought technical discussions on a forum such as this would have the slightest modicum of filtration from the "post truth" reality the rest of the world now seems to live in.  

I guess not!

*sigh*

[Emanuel]

No worries, kye, you're not the only one : ) reason why these forums serve I guess, people come here to share and put the discussion in progress ; ) I also inferred the same and looked like some interesting add-on to obviously benefit "all camera types" as you wrote. Although, if there is who already is playing with the idea even if only on photo side as Nikon, I think we can estimate it will happen to electronics as well, just a matter of time.

The tweet is coming from this Sony's press release from last Thursday:

https://www.sony-semicon.co.jp/e/news/2021/2021121601.html

Sony Develops World’s First* Stacked CMOS Image Sensor Technology
with 2-Layer Transistor Pixel
Widens Dynamic Range and Reduces Noise by
Approximately Doubling Saturation Signal Level

* As of announcement on December 16, 2021.

 

[Emanuel]

...a matter of time or as further business toss-up, well, for some reason people drive their questions through interrogation marks ; )

 

Here's petapixel's guys by Jaron Schneider:

«Sony doesn’t specify when it plans to manufacture sensors at scale using this technology but does say it will continue to iterate on the design to further increase image quality in sensors large and small.»

source

 

We tend to immediately read one after the other for Sony's large sensor size tradition but seems we can't separate the waters for much time longer... ;- )

[techie]

45 minutes ago, Emanuel said:

The tweet is coming from this Sony's press release from last Thursday:

https://www.sony-semicon.co.jp/e/news/2021/2021121601.html

Sony Develops World’s First* Stacked CMOS Image Sensor Technology
with 2-Layer Transistor Pixel
Widens Dynamic Range and Reduces Noise by
Approximately Doubling Saturation Signal Level

* As of announcement on December 16, 2021.

 

That's not a minor development:

"Sony’s new architecture is an advancement in stacked CMOS image sensor technology. Using its proprietary stacking technology, Sony packaged the photodiodes and pixel transistors on separate substrates stacked one atop the other.
In conventional stacked CMOS image sensors, by contrast, the photodiodes and pixel transistors sit alongside each other on the same substrate. The new stacking technology enables adoption of architectures that allow the photodiode and pixel transistor layers to each be optimized, thereby approximately doubling saturation signal level relative to conventional image sensors and, in turn, widening dynamic range."

[herein2020]

8 hours ago, Eric Calabros said:

Sorry thats not how it works. All these developments are about increasing the full well capacity, which is very low in small sensors, and decrease noise, which is relatively high in small sensors, to extend the DR, maybe 1/3 of a stop. 

Here is the formula: 20*log(full well capacity÷noise). You can play with this to estimate how much room to improve is still there. Capacity is already capped at 3500 electron per one um². And noise is already very low. A1 is at 1.8e now.

 

The only way to extend DR by several stops in conventional CMOS sensor is multi exposure solutions. Nikon is playing with a dual layer idea, but its on photo diode section, not electronics. They want to place a CMY layer on top of RGB layer, then set the upper layer exposure differently from the lower layer exposure. 

 

I have no clue regarding most of what you said, but just looking at the images, wouldn't this enable higher resolution sensors for a given sensor size and potentially higher color sensitivity in all sensors since with this architecture 100% of the sensor real estate is dedicated towards light gathering (I am assuming based on the graphic that current sensors lose 50%  of their real estate to the sensor transistor circuitry) since they sit side-by-side in current sensors?

This DPReview article kind of says the same thing but they are looking at it from a light gathering standpoint vs sensor resolution. Like I said, I know nothing about how sensors work but a graphic that shows me something is doubling in its ability to perform some task while taking up the same real estate as a previous version seems like a good thing in my mind.

 

12 hours ago, kye said:

I disagree - it will benefit all camera types.  This is because the second layer is in the same area of the sensor, and could be used to capture at a different amount of gain, which could then be used to extend the dynamic range of every pixel.  

Imagine a Sony sensor where it could have identical performance but use a second layer to extend DR by many stops...  

The Canon DGO sensor already does this by using dual gain circuits but still tops out around 16 stops. When I first heard about the DGO sensor I thought it was going to be some unheard of breakthrough but as usual it was mostly marketing.

[kye]

Technology really is wonderful, I'd just rather they use it for better pixels rather than having so many of them.  However, I'll bet that they'll take the doubling of light gathering and just use it to deliver more pixels of the same quality we have them at now.

[Emanuel]

2 hours ago, herein2020 said:

This DPReview article kind of says the same thing but they are looking at it from a light gathering standpoint vs sensor resolution. Like I said, I know nothing about how sensors work but a graphic that shows me something is doubling in its ability to perform some task while taking up the same real estate as a previous version seems like a good thing in my mind.

 

The Canon DGO sensor already does this by using dual gain circuits but still tops out around 16 stops. When I first heard about the DGO sensor I thought it was going to be some unheard of breakthrough but as usual it was mostly marketing.

Yeah, I read it : ) And basically the same has been said a little bit everywhere... To my view, aside the marketing thing, I guess we have no reason to be pessimist beyond the actual trend the many the better :- )

[herein2020]

1 hour ago, Emanuel said:

Yeah, I read it : ) And basically the same has been said a little bit everywhere... To my view, aside the marketing thing, I guess we have no reason to be pessimist beyond the actual trend the many the better :- )

Speaking of pessimism, the way I am reading all of this is that this is just going to bring cell phones one step closer to taking all of our jobs. I see a bleak future where cell phones have 30 stops of DR, eliminating the need for auxiliary lighting, an app that turns any video shot with a cell phone into an Arri Alexa masterpiece, GoPro's hypersmooth technology to eliminate the need for a gimbal...and the ability to shoot the next Titanic all while texting, taking selfies, and checking Instagram at the same time and all in the amount of time it takes us just to set our exposure triangle properly.

[Emanuel]

I think it can work for computational acquisition and all the remaining stuff... : )

But they still have to work hard to mimic our exposure triangle properly as you fairly say. They're a pain to use today other than the pocket small form factor : X

Let alone the production values mantra which will stand no matter the tech evolves ;- )

[ND64]

20 hours ago, herein2020 said:

have no clue regarding most of what you said, but just looking at the images, wouldn't this enable higher resolution sensors for a given sensor size and potentially higher color sensitivity in all sensors since with this architecture 100% of the sensor real estate is dedicated towards light gathering (I am assuming based on the graphic that current sensors lose 50%  of their real estate to the sensor transistor circuitry) since they sit side-by-side in current sensors?

They claim this increase in area gives them one full stop extra DR, but its just a claim. However, even if remains true in production, they will use this advantage to make even smaller pixel and keep the same DR as current sensors. They're working on pixels smaller than the wavelength they're supposed to absorb! 

[greenscreen]

20 minutes ago, Eric Calabros said:

However, even if remains true in production, they will use this advantage to make even smaller pixel and keep the same DR as current sensors.

How do you know that?

[kye]

On 12/20/2021 at 11:55 AM, herein2020 said:

Speaking of pessimism, the way I am reading all of this is that this is just going to bring cell phones one step closer to taking all of our jobs. I see a bleak future where cell phones have 30 stops of DR, eliminating the need for auxiliary lighting, an app that turns any video shot with a cell phone into an Arri Alexa masterpiece, GoPro's hypersmooth technology to eliminate the need for a gimbal...and the ability to shoot the next Titanic all while texting, taking selfies, and checking Instagram at the same time and all in the amount of time it takes us just to set our exposure triangle properly.

I absolutely completely disagree with this, with one caveat.

Firstly, to provide some context, I have a background in computing and am very optimistic about how good the tech will become, however, I don't think it will get that good that quickly.

The caveat is that technology will gradually nibble away at the bottom-end of the market.  In this case, cell phones will gradually have better and better image quality and will more and more replace having a "better" camera.  Essentially, this just means that it's cheaper to get the same image quality.

But, since when did having a cheaper camera "take jobs"?

Cameras are spectacularly better than they have ever been, but there are probably more people making their living as professionals who contribute to video creation than ever before, and this is even if you don't count social media influencers and you tubers.  I'm talking about more and more people doing corporate, weddings, events, advertising / marketing / PR, etc.

In fact, the improvement in the screen of the cell phone has probably created more jobs than the camera has killed because the total time watching video content for the average person is through the roof because of mobile.

Everyone knows that the best camera in the hands of the worst operator looks awful, and that doesn't even account for lighting, set dressing, hair, makeup, etc...  let alone writing, story, acting, etc.

[Emanuel]

I guess our mate up there had focused on the impact the smartphone evolution may produce on this craft. The fact is new competition brings both a fresh way of producing motion pictures either on new players looking for the same audiences but not an equivalent content nor production values or the consequences coming from there and the whole impact in the labour force participation and employment ratios of current industry workers.

 

https://www.instagram.com/reel/CWqMmvYgTL1/

In Portuguese, the woman says she is a lovely woman because of her eyes, she is still able to attract any man in her country...

:- )

 

[herein2020]

15 hours ago, kye said:

I absolutely completely disagree with this, with one caveat.

Firstly, to provide some context, I have a background in computing and am very optimistic about how good the tech will become, however, I don't think it will get that good that quickly.

The caveat is that technology will gradually nibble away at the bottom-end of the market.  In this case, cell phones will gradually have better and better image quality and will more and more replace having a "better" camera.  Essentially, this just means that it's cheaper to get the same image quality.

But, since when did having a cheaper camera "take jobs"?

Cameras are spectacularly better than they have ever been, but there are probably more people making their living as professionals who contribute to video creation than ever before, and this is even if you don't count social media influencers and you tubers.  I'm talking about more and more people doing corporate, weddings, events, advertising / marketing / PR, etc.

In fact, the improvement in the screen of the cell phone has probably created more jobs than the camera has killed because the total time watching video content for the average person is through the roof because of mobile.

Everyone knows that the best camera in the hands of the worst operator looks awful, and that doesn't even account for lighting, set dressing, hair, makeup, etc...  let alone writing, story, acting, etc.

 

I was being a bit comical in some of the predictions but since you mention it...cell phones have absolutely taken many jobs and replaced many technologies; photography and video has been just as equally hard hit as any of them. As a working professional who has been doing this for over 10yrs even in that relatively short period of time I can tell you that many photography and videography opportunities for professionals have evaporated thanks to cell phones. 

It is well documented that the camera industry has been devastated by cell phones, just like the mobile GPS market, lets not even talk about the PC market, and the list goes on and that is just in the HW segments alone. Lets not even talk about the endless apps that do things in one click that people used to pay professionals for. Every one of those contractions represent jobs lost. 

You mention that more and more people are doing these content production jobs but where are the numbers to back that up which show actual job creation? What I have seen first hand is a serious contraction in the prices clients are willing to pay to do those jobs and in many cases they ultimately determine that an employee placing their cell phone on a little tripod stand is good enough for their social media accounts. Lets not even forget the endless Kim K pictures that get millions of views vs the professional who spent decades perfecting their craft creating works of art that now no one ever sees (or is willing to pay for).

Sure there's more "content creators" than ever before, but that does not mean they are paying professionals to do the creation or even that the process was anything more than pressing record on their cell phone. The value of professional video and photography has absolutely plummeted; the average cost for a wedding video which could easily require $20K+ worth of equipment nationwide in the USA is now around $1400USD whereas it used to easily be $7K+. 

The reality is, this absolute saturation of low quality content has lowered people's expectations to the point to where they do not value the time, equipment, or expertise needed to separate the great from the average and their content creation budgets reflect that. Almost every photographer and videographer I know is now doing it on the side due to passion vs full time because it simply will not pay enough to quit their day jobs. And lets not forget the plethora of mind numbingly boring movies coming out of Netflix, Amazon, Apple, and others that literally feel like cookie cutter corporate clones of each other with no plot, storytelling, or suspense whose only purpose is to gain more subscribers to their service.

So yes, I do believe that as cell phones continue to get better, the devaluing of professional content creators will continue as well which will in turn continue to lead to job loss. Just because more content is being created than ever before does not mean the job statistics for professional photographers / videographers is following the same trajectory. Ok, end of rant, how about those double stacked sensors 😀

[Emanuel]

Yeah, I'm afraid the whole transformation of this business is putting too many heads upside down. On the other hand, I think arthouse has gained much more than losing jobs anyway and affordable technology is kicking asses and taking names as ever before : ) You don't need to kiss someone's ass anymore to shoot at the mainstream film look as it was then.

The time has come with or without double stacked sensors on small form factor devices ;- )