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 :)

The next gen sensor

Christoffer Glans

Christoffer Glans

Well-known member
Joined
Jun 11, 2007
Messages
4,993
Reaction score
36
Points
48
Age
42
Location
Stockholm, Sweden
I was having a discussion with a friend about the future of sensors and we started speculating about the next gen Red sensor and what it needs to be in order to compete, especially if Arri comes out with a new 6K sensor in the near future.

What do you think it will be? Both speculative, educated guesses and as a wish list? I think there are three major corner stones for the next sensor tech and non of them are resolution;

- 20+ native DR with better shadow detail quality and even better highlight handling.
- Double base ISO at 800 and 5000 with an almost clean ISO5000.
- Better color science with better color accuracy. Not necessarily "filmlike" but more accurate to the human eye (which I think is what Red is better at compared to film-emulating sciences like Alexa).

As for resolution, 8K would be nice even for us with Weapon MG, but that is not important in my opinion. Keeping it at 6K with the above improvements would make a serious change to the quality we already have. Resolution hunt is over, time to get those colors and sensitivity :smiley:

I'm speaking about real, logical and realistic projections about the future, not wishes that are complete fantasies.
 
Since 8K is so hard to manufacture and ramping production for the 8K Dragon sensor might take a lot of time (probably fall/end of this year) than when will it make sense to realease new sensor tech?
I don't think the 8K guys will be very happy to hear that there is a better sensor released just a few days after they get their hands on the long awaited 8K (now outdated Dragon) sensor :)
 
I would be perfectly happy if they didn't change anything and just gave us a true global shutter, that would make the camera complete and perfect for me.
 
I doubt Arri has6k being released soon. SXT with same old sensor and new uprez nonsense isn't shipping til June... So by counter point how many days does that camera last if 6k Arri is imminent? Plus there obligated to all the new ish XTs who get free upgrades? This is the current roadmap cuz no one can get either model so your asking what's the next model basically while guys like me just want this one ironically lol? Just because my payments a year old dont make my 8k sensor old? And even funnier red made their release dates (besides 8k) and Arri didn't!!!! It's a whole new world! Loving it

but to my wish list


65 or IMAX size or high speed...
 
Last edited:
Andrei Goaga said:
Since 8K is so hard to manufacture and ramping production for the 8K Dragon sensor might take a lot of time (probably fall/end of this year) than when will it make sense to realease new sensor tech?
I don't think the 8K guys will be very happy to hear that there is a better sensor released just a few days after they get their hands on the long awaited 8K (now outdated Dragon) sensor :)
Click to expand...

I'm not talking about the near future here, I'm talking about the future. A new sensor will happen, that's just a fact.

Bobby Farokhzad said:
I would be perfectly happy if they didn't change anything and just gave us a true global shutter, that would make the camera complete and perfect for me.
Click to expand...

So no improvements to colors, DR, noise, base-ISO etc? Global shutter is good, but global shutters also holds back other potentials with a sensor and I rather have more DR than a global sensor.

Brian Merlen said:
I doubt Arri has6k being released soon. SXT with same old sensor and new uprez nonsense isn't shipping til June... So by counter point how many days does that camera last if 6k Arri is imminent? Plus there obligated to all the new ish XTs who get free upgrades? This is the current roadmap cuz no one can get either model so your asking what's the next model basically while guys like me just want this one ironically lol? Just because my payments a year old dont make my 8k sensor old? And even funnier red made their release dates (besides 8k) and Arri didn't!!!! It's a whole new world! Loving it

but to my wish list


65 or IMAX size or high speed...
Click to expand...

Like I said in the top of this post, I'm talking about the future, not the near future. A next gen sensor from Red will happen, that's just a fact and when it does, what would be key?
I don't think a 65mm will be good for everything, I actually feel that the 6K Dragon is a very good physical sensor size. I like the Vista Vision, but it CAN be a bit shallow in it's DOF and it's hard getting lenses for that, even harder for 65mm. A special 65mm camera from Red would be awesome, but the physical size of the sensor isn't really what I'm talking about here. They could make a 65mm Dragon now if they bump up the processing speed and do a larger cut of the sensor, so speaking of 8K, Vista Vision, 65mm etc. isn't really about the next gen sensor.

The next gen sensor is about how it performs outside of those parameters. The big thing about Dragon wasn't the 6K or the larger FOV, it was about the color science. So the next gen sensor NEED to be about better sensitivity and even better colors. Getting HDRx-like DR without using HDRx would in my opinion also be an extremely welcome feature to the world of digital cinema.

High speed improvements I can dig, but it would really cut down on other things. It would also mean we need new processors in our cameras due to the larger bandwidth it needs. As of now I'm looking at what kind of sensor would work in already existing bodies like the Weapon 6K MG.
 
I would speculate that there may be new tools and techniques available that could make for radically new sensor designs and backend camera systems:

+) 3D femto laser writing system to create optical phase masks -> which could help shape the light as it enters each pixel of the sensor.

+) Nano structured Waveguides (created using electron beam lithography + dielectrophoresis) that very efficiently convert a BAND of light (eg R,G,B) into electricity -> ie no need for light sapping coloured filters & the sensor would now have a Serial stream of electric voltages (representing the photo stream).

+) At this stage the Sensor would have a SUPER high frame rate image - that could be downsampled to 24 Fps depending on the frequency of Accumulation.


I guess what happens next is dependent on How much local memory is available, how long it takes to store data, and the ENERGY requirements.
+) Why not use Phase change memory in Deeper layer in the Sensor Store the Accumulated data : eg something like on-sensor 3D Xpoint to store the Globally shutter image?

+) Rather than Use energy to compress the data into R3D ... perhaps writing to a local 6TB block of 3D-Xpoint memory block in the camera.

+) Transmission using 7Gb/sec over 60Gz band of Current frame to external systems : Option of Compression at this point. Any spare bandwidth used to Sync the external system.



... With the exception of HFR however - I doubt any of us could want more than an 8K Dragon sensor!


AJ
 
If Red can make a global shutter without sacrificing any of the dynamic range and the color science, that by itself would be a huge achievement. To me, the rolling shutter and all the artifacts associated with it is a plague that needs a cure. As far as the 5000 iso base sensitivity goes, or a dual base sensitivity like the one in the Panasonic Varicam, I would be very happy if I can have a really clean 2000 iso and I'm being told by a few Weapon users with the LLO OLPF that we are not too far.
 
Personal, I like to se an autofocus to kill for. Its proven now days that autofocus can work solid for some type of shots by other manufactures. For fun, why not a "siri" to change framerate, iso, frameguids, ad metadata and what not.
 
Christoffer Glans said:
Getting HDRx-like DR without using HDRx would in my opinion also be an extremely welcome feature to the world of digital cinema.
Click to expand...

I do wonder if HDRx would be double the Datarate?

...I do know the Sensor has a set sensitivity and you're recording raw that you can manipulate in post.

You always need to expose for the RAW Sensor

Red could find a way to reduce noise and create an HDR effect and bring more information out of shadows. But, lifting unexposed areas(blacks and such) that don't have any detail/information is kinda pointless, because you are limited by the sensor data, not a baked in image like another camera system
 
Jeffery Anderson said:
It's your lucky day

http://www.red.com/store/products/dsmc-red-motion-ti-pl-mount
Click to expand...


That is pretty far from being a global shutter. Its an ND that goes darker for fraktions of a frame but it does not go black. And you got light loss and other artifacts.

Personally I would love if red went back and made a 6/4" front of lens Motion filter. that where driven via a CTR cable from the camera so we could control it on the touch screen but add and remove it to a mattbox just like any other filter. To cramp it into the mount comes with to many complications I think.
 
Not sure if this is due to some undercurrent of global economic slowdown, thus causing fewer buyers, but I get the sense that exponential growth in R & D is beginning to slow down. This could lead to a period of improving on current tech.

I bring this up to suggest that RED may stay with what they currently know and what they may have been working on for the past few years. And I think that could be something akin to a 645 sized sensor with slightly larger pixels in an 8k resolution release. This should solve both the 8k standard and better low-light performance.

We've seen in another thread how many loyal RED shooters were left behind in the last upgrade go-'round... so I think any cutting edge expensive tech will have to have a very high price on it to justify a smaller production due to fewer sales. Thus I think incremental changes like the 645 sensor could come in at a price point to keep many of us on board.

This is not to say there isn't funding for R & D in these bigger companies... just that with a shrinking market for cutting edge tech, there isn't a justification to spend there.

Hope I'm wrong and that new things continue to explode so fast we can't keep up with the new offerings, but my sense tells me things may just slow down for a bit.
 
Imagine how hard a 645sensor would be given the difficulty in 8k dragon
 
A 64mm sensor would be a special camera that would cost very very much and not all productions can or should work with it. It's the case of people thinking "bigger is better", without regard to what the setbacks and struggles you would have with such a large sensor, like not being able to find lenses faster than T3.5 and even if they did they wouldn't be able to focus on anything under that value. The amount of light you need would make it so that the only way to get a descent focus pulling ability is to shoot at T8.0 and have a sensor able to produce clean ISO5000 images on such a large sensor.

So, I think Vista Vision is the best sensor size to have, there are a lot of FF lenses that work great and you can shoot with high speed lenses. But even Vista Vision can be a struggle, so I think the Dragon sensor really is a great size, we just need even better clean sensitivity, DR and color science. Dragon still gets most negative feedback on it's colors, even though it's good as it is. Color handling and high noise/inability for nice looking shadows.
 
Christoffer Glans said:
I think the Dragon sensor really is a great size, we just need even better clean sensitivity, DR and color science. Dragon still gets most negative feedback on it's colors, even though it's good as it is. Color handling and high noise/inability for nice looking shadows.
Click to expand...

I'm with Christoffer 100% on this and his other points about improving dynamic range, ISO etc.

Firstly, in my opinion, the Dragon sensor size is a better general dimension for cinema compared with FF35, Vista Vision, 65, etc sensor sizes. Because I enjoy the deeper focus and perspective depth of wide and ultra-wide focal lengths, but do not enjoy the perspective distortion at the edges, I much prefer sensor sizes in the S35 range much more than Vista Vision of FF35 or larger, with some exceptions such as the use of immersive FOV in THE REVENANT.

Secondly, the present fantasy of global shutter with 20+ stops of DR, dual base ISO such as 800 and 5000, improved color gamut and overall color science, etc are likely to become realized inevitably. My personal preference would be to have these features on a variety of sensor sizes, including Dragon and 65, more than I would prefer more resolution. Less compression (or no compression) at 6K and 8K / more bitrate would be a huge plus as well.

Cheers ~NAS
 
Elsie N said:
Maybe they could get Hasselbad to make it for them. '-)
Click to expand...

Or Sony, as the new MF-Sensors Hasselblad and Phase One are using are made by them.

According to rumors the successor of Sony's F65 is supposed to use sensor technology used in Sony's new 42MP Full Frame Sensor, as seen in the A7RII.
If they could manage to use that Sensor. or maybe an even better one, with a 60P readout in a new Flagship Camera, that would be a pretty great thing to see. Resolution and Low Light capabilities are amazing in the A7RII.

But it's just rumors.
 
Last edited:
I think Sony's 300mm fabrication at Kumamoto runs 24 hours a day to keep up with increasing sensor demand (until the recent earthquake).

Their defect rate is Low.
The lithography process they use is leading edge.
Their ability to shape the light accurately is determined by the advanced process they are on (improving the SNR compared to less advanced fabrication processes)

We may be in a situation that RED had a more impressive chip design, and Sony has a fabrication process with the ability to form finer chip layouts.

AJ
 
You must log in or register to reply here.
Back
Top