Welcome to our community

Be a part of something great, join today!

Log in
  • 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

The presentation pointed out how Bayer masks are supposedly Red/Blue deficient. I must be missing something, because it seems like the ratio of G:R:B photosites remains unchanged in Q67. It seems like a twisted case of supersampling, with the disadvantages of smaller pixels to boot. But I suppose we will have to give Sony the benefit of the doubt and see how it translates to final images. Clearly, they seem quite confident about it, though I am not convinced (at least in theory).

Also, I don't really understand why the twist was required? Maybe someone can explain? Shouldn't a straightforward supersample suffice?

Frank Cueto said:
Imagine it this way...

pixel(photosite) = bucket
Light Photon = Rain drop

Bigger bucket catches more raindrops ;-)

if you have a smaller bucket, then you need to make it rain more!

jejeje...
Click to expand...
It looks very much like the smaller bucket will overflow.
 
Gunleik Groven said:
RED has showed that that is doable...
Click to expand...
There is obviously WAY more involved in sensor design than pixel size but taking their simple graphic at face value I thought that it illustrated no advantage and some possible issues.
Marketing and Engineering could communicate better in this one case.​
 
Graeme Nattress said:
Yes, you can shrink the pixel if you make the pixel better. Or you can keep the pixel the same size and improve the pixel, like M to M-X. But all other things being equal, a smaller pixel will lead to some compromise in DR. Sony themselves show that in one of their graphs.

Graeme
Click to expand...

True, but... isn't the thing that counts actually the photosite area used for an individual end pixel?

In other words, is there a DR difference, if a pixel in final image is made from one large, or four small pixels (rotated or not), if the active area is the same?

Still in other words, if, say, Red One had a 2K big photosite sensor instead of 4K smaller photosite one, would there be a benefit in DR if comparing resulting images in 2K resolution (naturally properly scaled down in the 4K sensor case)?

If sensors are simplified as just buckets, can't multiple small buckets gather just as many photons as one big one?
 
Because each bucket needs circuitry, and the size of that circuitry is such that on a smaller pixel it takes up a greater proportion of the available area.

Graeme
 
I love how Sony compares an 18 MP Q67 with a 9MP Bayer to try and make it look better. They must already know that it can't compete on an even playing field.
 
Thanks Graeme for explaining that. I have to side with Peter Jackson, I don't know how it works I just like the image it makes. Of course, I also like the size, dynamic range, frame rates, resolution, RAW capabilities, frame rates, price, modularity.... did I mention frame rates?

Competition makes everyone better and leads to better products.

Keep fighting Red Team. I'm glad to be on your side.
 
Hehe...just thought...of all the ways to talk about being disingenuous regarding 4k...

These are some real "Marketing Pixels".
 
Anyone know the real reason that tape is still being talked about in 2011 - besides the fact the oil is a key ingredient in tape plastics?

It has something to do with Sag/Aftra TAPE contracts, evil/smart producers and corporate honchos, and the residuals they don't have to pay...

Talk about wicked evil...
 
Daniel Browning said:
I love how Sony compares an 18 MP Q67 with a 9MP Bayer to try and make it look better. They must already know that it can't compete on an even playing field.
Click to expand...

Precisely!

Graeme
 
I'd just like to share my personal point of view on anything I read from Graeme Nattress:
He is first and foremost an engineer, and a damn good one at that. You might think he's bashing Sony because he works for RED, but I just think he's pointing out the engineering truths that are being washed aside by marketing hooey. That's the thing about engineers, they just can't help themselves when it comes to correcting technical inaccuracies and misconceptions.
I'm pretty sure that if RED ever started spewing marketing hooey that was technically inaccurate, Graeme would be on this forum and many others correcting it and explaining the technical facts behind it all.
In the end, even if you doubt his motives, you can't argue with the technical facts he's bringing up, because he's quite simply right.
Sony's own diagrams show the Q67 is a bayer pattern rotated 45 degrees and with a doubled pixel count. They just don't want to call it that. Graeme is calling it that. There's no technical difference, the only difference is... marketing hooey.
 
M Most said:
Sounds to me like you're calling the Sony future product a scam.

Graeme, I know this is RedUser, but neither you nor anyone else here has seen what Sony is proposing to build. And just as Red doesn't want people drawing unsubstantiated conclusions about it's future technology and unreleased products, you really should be granting Sony the same courtesy. When they release something that can then be tested, and all of these things either proven or unproven, then it will be fair to analyze the results. Until then, it's all speculation, and you know well how valuable that usually turns out to be (I.e., not very).
Click to expand...

This is true, but the big take-away is: without Red providing substantial market pressure, Sony would feel zero impetus to even entertain a 4K solution. So, even when we expect them to succeed, they could at least tip their hat to what Red has accomplished. I mean, it's a two-way street, right?
 
Well, it doesn't sound to me like Graeme is calling the future Sony product a scam.
And I don't think he deserves this BS.

On another note, the Sony reps seem to be implying - through some warped sense of marketing logic- that by inserting the breadth of the diagonal of their pixel into the bayer pixel sample, that they have found a way to more effectively tile the plane, to more effectively fit in more pixels without acknowledging that they are using smaller pixels. As if one method was more physics friendly than the other. Yes, it seems that absurd to me.
 
Sorry, maybe I am missing something? Graeme?

By adding megapixels and using Bayer at 45 degrees with a less clever decoding algorithm, they are trading off dynamic range and sensitivity for:

1. Easier processing - so that they can output true 4K signal on the fly for broadcast without crazy ASICs.

2. Good resolution and anti-aliasing for 4K - in both luma and chroma domain.

Sure, a 6K Bayer RED camera will resolve higher:
- eg 5K in the luma domain (with a bit of chroma aliasing) if the OLPF is aggressive...
- or 4.5K (with no chroma aliasing) if the OLPF is conservative...

But the target is 4K, not 5K.

If Sony can can compensate for the dynamic range and sensitivity losses by having a superior manufacturing process (which they seem to be investing billions in) it really doesn't seem like a bad tradeoff to me, especially if your primary target is 4K broadcast?

Bruce Allen
www.boacinema.com
 
Bruce: But that would apply no matter what the orientation of the pixel - bayer pattern or otherwise. And it is the "superior diagonal orientation" of Q67 pixels that is the issue. Granted, it's a silly issue.
 
Bill Anderson said:
Bruce: But that would apply no matter what the orientation of the pixel - bayer pattern or otherwise. And it is the "superior diagonal orientation" of Q67 pixels that is the issue. Granted, it's a silly issue.
Click to expand...

AFAIK, the orientation is important: Sony is using the 45 rotation to enable their simpler, lower-quality "debayering" (deQ67ing?).

If you felt like it, you could "deQ67" Red's sensor by working diagonally, except:
- since you're working diagonally, your algorithn couldn't be able to work on a nice rectangle of pixels - you'd be working on an elongated diamond thing
- you'd have to rotate the whole image 45 degrees afterwards in post (making things softer)
- the OLPF wouldn't be right for it
- Q67 process is still lower-quality anyway, so why?

Likewise, you could do a true "proper" deBayer to the Sony image. But then you similarly have to make the algorithm work on a non-rectangular image. And you'd have to rotate in post, which would result in a softer end product. And the OLPF would be too strong.

Am I nuts?

Bruce Allen
www.boacinema.com
 
M Most said:
Sounds to me like you're calling the Sony future product a scam.

Graeme, I know this is RedUser, but neither you nor anyone else here has seen what Sony is proposing to build. And just as Red doesn't want people drawing unsubstantiated conclusions about it's future technology and unreleased products, you really should be granting Sony the same courtesy. When they release something that can then be tested, and all of these things either proven or unproven, then it will be fair to analyze the results. Until then, it's all speculation, and you know well how valuable that usually turns out to be (I.e., not very).
Click to expand...

well at least sony gave a fair representation of the competition. very courteous of them don't you think? this is the biggest load of garbage i've heard in a while. i didn't used to have anything against sony. i don't care that they try to trash arri or red or whoever, it's the fact that they're lying through their teeth that bothers me.

and for the record, they did just propose what they're building. there are theoretical and practical limitations of their design structure. but i don't care about the cameras, it's their dishonesty that bothers me.
 
"A Streetcar Named 4K Desire" is slowly but surely filling with the following respectful competitors:

- RED will show fully working Epic 5K at NAB.
- Canon already showed 4K prototype camera in 2010.
- Sony just announced 4K acquisition system roadmap.
- Aaton should finally show fully working Penelope ∆ at NAB.

What about Arri?

I heard some of rumors from Germany/Bayern that Arri is working on 4K camera project code named

"Alexander The Great" and it should be 4K successor of Alexa.
 
You must log in or register to reply here.
Back
Top