Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Sony Building SUPER REALITY SENSOR That Surpasses Human Vision

W isn't for white. W should also be green photosites which is typically where luminance values are extracted from. I assume this is what is represented here but could be wrong. Actually this kinda looks like they are lowering chroma resolution to get greater luma resolution... perhaps this is why they need an 8K sensor for a 4K image?

Lunimance isn't derived from the green channel, it is a product of red, green and blue values combined. White photosites imply no filtering, or at least neutral filitering of the light spectrum. Its an interesting bayer pattern, especially with the 2*4k oversample.

The reason they're using an 8k sensor is so that a 4k image can be constructed that contains a value from a blue photosite, a red photosite, and of course, two green photosites (depending on the bayer pattern design). That makes perfect sense, and will result in less colour artifacts at edges. Actually, you could construct a full colour 4k image using an 8k sensor without the debayer process and its subjective math processes .. but I kinda like the math/magic of debayering. Don't want to upset the fanboys, so I'll point out a 5k bayer should produce a lovely 4k frankly, again, with fewer artifacts, especially with higher redcode datarates.

..I'd rather have a 3k sensor with a global shutter personally, than an uber-high res sensor, but hey, I'm sure their new camera will be expensive and useful in equal measure, just like Red cameras, which are marginally cheaper, but prohibitively expensive for many/ most people. This may change
 
Last edited:
I don't think so that Sony is stupid as some people here can might think.

Just count on that the biggest box offices in our age the movies like Jim Cameron Avatar and all George Lucas Star Wars were shot with Sony (HD) cameras.

Also I'm not defending Sony here but trying to stay real...

No... I'm not saying that Sony is stupid as a company. I'm just saying it's stupid how they assume their customers are all brain-dead consumer zombies. Somebody who is in the market for a cinema-style camera is going to know the advantages/aesthetics of a large-sensor camera already... they don't need Sony pointing out a de-focused background with an arrow and pretending like it's a new technology. :rolleyes5: THAT is stupid.

As part of their "Super Reality" sensor program, Sony has apparently invented depth of field for moving pictures.

Yes, we should all feel blessed that Sony would bestow this amazing technology upon us. I can just hear all the producers and executives now...

Exec: "we want the new Super-Reality backgrounds that the Sony has..."

Director/DP: "What do you mean? This is just 35mm photography... this has been in movies for ages."

Exec: "yes, but that isn't the same as the Super-Reality backgrounds that the F65 has... I saw the diagrams."

Director/DP: *slaps face*
 
Lunimance isn't derived from the green channel, it is a product of red, green and blue values combined. White photosites imply no filtering, or at least neutral filitering of the light spectrum. Its an interesting bayer pattern, especially with the 2*4k oversample.

The reason they're using an 8k sensor is so that a 4k image can be constructed that contains a value from a blue photosite, a red photosite, and of course, two green photosites (depending on the bayer pattern design). That makes perfect sense, and will result in less colour artifacts at edges. Actually, you could construct a full colour 4k image using an 8k sensor without the debayer process and its subjective math processes .. but I kinda like the math/magic of debayering. Don't want to upset the fanboys, so I'll point out a 5k bayer should produce a lovely 4k frankly, again, with fewer artifacts, especially with higher redcode datarates.

..I'd rather have a 3k sensor with a global shutter personally, than an uber-high res sensor, but hey, I'm sure their new camera will be expensive and useful in equal measure, just like Red cameras, which are marginally cheaper, but prohibitively expensive for many/ most people

That kind of makes sense, except what they are showing here wouldn't really be a bayer pattern would it? Also I thought Sony said they would be using a Q67 pattern which is like Bayer, but tilted 45 degrees... Which based on the picture would mean they are using a line of green, then a line of unfiltered, then a line of half red half blue... interesting.

I don't think you'll offend anyone here by pointing out what we already know about the debayer on the Epic. That's one of the selling points of shooting 5K is to get a full 4K image after debayer, it's not really talked about because it's a known factor.
 
That kind of makes sense, except what they are showing here wouldn't really be a bayer pattern would it?

Its not Mr Bayers original design, but an extension of. And in as much as it wouldn't need to be debayered at all, you're right.

The pattern they're using contains true luminance, as well as colour samples. It is interesting for sure.

I'd like to see the difference between a theoretical f65 sensor grab that has been demosaic'd then downsampled (4 pixels averaged), and the same sensor grab were the 4k is derived from a simple weighting of the 8k sensor values.
 
Its not Mr Bayers original design, but an extension of. And in as much as it wouldn't need to be debayered at all, you're right.

The pattern they're using contains true luminance, as well as colour samples. It is interesting for sure.

I'd like to see the difference between a theoretical f65 sensor grab that has been demosaic'd then downsampled (4 pixels averaged), and the same sensor grab were the 4k is derived from a simple weighting of the 8k sensor values.

So this is how they're getting 24+ stops of dynamic range on their Super Reality sensor... Kinda got my Sony info crossed and assumed this was the sensor pattern on the F65 which I'm guessing it isn't. I know Graeme's opinion on the Q67 pattern, wondering what he has to say about the strips of luminance in this pattern.
 
Except it's not an 8k sensor which would need to have 37mp by standard Bayer math, or 80mp by the Sony 4k math. Either way, at 20mp, it is correctly sized for their 4k method and way too few for 8k by anyone's math.

Adding in the white (clear) pixel is great for very small pixel consumer cameras where you need more light hitting the sensor. I don't see an advantage for the larger pixels we use for digital cinema. The Sony papers talk about many things in their CMOS roadmap - not all are applicable to all applications of CMOS sensors.

Graeme
 
EDIT; Graeme posted it already (but much nicer) while I wrote this.

There are nothing remotely close to 8K resolution in the F65's 20 megapixel sensor except Sony hyperbole.

Please don't drink the Sony "cool aid" and continue referring the Sony marketing departments claim that the F65 is a 8K camera.

It has a 20mpx sensor which contain some "exstra pixel magic thingy",
but can only deliver 4Kx2K resolution from that sensor (figure that out! :bigear:).

Until some tests can confirm that this can deliver some "magic benefits" for the 4K format, it is just hyperbole marketing.

8K resolution (over 30mpx in all conventional counting) this sensor will never ever deliver.

Same goes for the "SUPER REALITY SENSOR" hyperbole claim! I mean really!!!. :puke:

Hope this thread gets deleted soon! :emote_hanged:
 
I have to admit, I just assumed when Sony said 8k they were being "numerically honest" .. I guess marketing people are as good at math as they are ethics. I feel raped
 
Except it's not an 8k sensor which would need to have 37mp by standard Bayer math, or 80mp by the Sony 4k math. Either way, at 20mp, it is correctly sized for their 4k method and way too few for 8k by anyone's math.

Still, its 2 x 10mp ..
 
I just wanted to make it clear - I'm not attacking Sony's inclusion of 'w' photo sites - for all I know, they could be the best thing to ever happen to CMOS sensors. Their inclusion of them in the diagram, without any explanation of what they were, however, left me with questions about 'what' they are.
 
But don't ya just love whats coming... :)))))) for all parties.

If RED did not push the envelope, then sony would sit on this shit for years... milking jo public with there XD cam 2/3 inche chip shit.

It only takes one to show it can be done.

Jim is like Chuck Yeager of the camera world.
 
Jim is the Howard Hughes of 4K (and now 5K).
 
Except it's not an 8k sensor which would need to have 37mp by standard Bayer math, or 80mp by the Sony 4k math. Either way, at 20mp, it is correctly sized for their 4k method and way too few for 8k by anyone's math.

Adding in the white (clear) pixel is great for very small pixel consumer cameras where you need more light hitting the sensor. I don't see an advantage for the larger pixels we use for digital cinema. The Sony papers talk about many things in their CMOS roadmap - not all are applicable to all applications of CMOS sensors.

Graeme

It really looked like they would need a LOT of pixels with their sensor layout to pull off what they are claiming, I'm just not smart enough about sensors to do the math. I guess kudos to Sony if they can pull it off, and for even trying. Something tells me that if RED had an 80mp sensor it would debayer to something much much higher than 4K...
 
Back
Top