Sensor Split Fix Opens Up New Camera Possibilities?

[Phil Holland]

Anyway...whatever...still amazed by, and challenged to achieve, the high quality images that can be captured by the currently avilable RED cameras.

I'm here with you on this and my main "wants" this generation are primarily software/firmware related to empower filmmakers to do more and more effective work with the tools at hand. Gio Scope for example, being back in the camera, is one of those major useful things that I am using literally everyday and nearly every shot. Off topic tip, you can really inspect two areas right now with this as RED's IPP2 Highlight Extension tech combined with both Gio Scope and a Waveform Monitor truly tells you "the edge" if you catch my drift.

My distant wants are related to other topics like larger formats and higher resolution imaging, but I'm actually pretty excited to see how they build on DSMC3. I've found the love in this new system in both the small and larger XL bodies, though I still am a DSMC2 body enjoyer. In particular the small body is extremely powerful in all it can do and does well.

Related to future things. Right now, I don't think you'll see what we consider digital cinema cameras focusing on anything lower than about a 3.2 micron pixel pitch. I think that reveals sort of the sweat spot for our industry between 3.2 and up to about 6.4 microns. Raptor 8K VV (as well as Dragon and Monstro) are very intelligently built on a 5 micron pitch. Hard work coming up with novel and improved upon pixel designs, sensor stacks, etc. But so much is moving fast now and I do think there's more goodness to be had. Everything I've tested with smaller pitches produce rather uncomfortable compromises and results. Not saying there isn't a good sub-3.2µm design on the horizon, but to date nothing "wild" that we would want.

Post image processing will continue to evolve, but the question there is the line between purity and truth versus result. I dance that line often with various things. I have some projects where no Noise Reduction or literally any image processing beyond pure color are expected. Others I can go wild with all sorts of processing methods. Luckily with RAW that's all after the fact, but I could see beefing up REDCINE-X Pro a bit and maybe some added features to the SDK. RED's already done well there.

 

[M Harvey]

More processing that has no effect on the RAW capture should be possible.

At the very least, it seems as if we'd be able to do things like real-time in-camera removal of chromatic aberration, artificial background blur ("portrait mode"), and adjustment of spatial distortion (e.g. reduce field curvature)-- things we've seen for years now in still cameras. The addition of depth sensors and other ancillary measurement devices (such as accelerometers) could make these capabilities more professionally useful as you could better calibrate the effect and have it respond more organically to the scene, rather than treat it like an instagram filter. Maybe the next level of that would be lens emulation (Kowa anamorphics but with the bokeh of s4s)? Not sure how desirable some of these capabilities would be on a professional production, might get them first on consumer video devices...

 

[Les Hillis]

P.S.

Looking further into it, there are still advances that can be made, at the sensor and in front of and behind it, which could result in better cinema format cameras, and larger format cameras too if there was a sufficient market demand.

At some point though it's limited by the fundamental nature of pixel technology itself.

So I looked at metamaterials and nanolattice tech, thinking that was further along in development and practical use than it seems to actually be. Looks like the cross-over to using something like photo-responsive nanolattice arrays will take a while, similar to the transition from film to digital. Not just because of the development of the materials but in having to deal with the nature of light and energy itself at that level.

Which again just adds up to appreciating and making the most of what we can do now with what we've already got, while the technology side of it slowly moves on.

 

[Björn Benckert]

Hej Björn, wouldn't mind seeing some comparisons if and when it's possible?

I post something later on, but project is still under NDA for some time forward. But yes will be interesting to see. The Arri S35 is good but so far I don't really see it holding up to the hype.

 

[Zack Birlew]

It's been a while since this thread was made, the 360 degree option seems to have been taken over by the new Big Sky Sphere 18K camera which uses special fisheye lenses to capture a massive picture and while Fuji hasn't made the full move to full medium format sensor readout in 8K or higher on their GFX line, they at least have a scaled full frame 8K option and much nicer full sensor 4K performance. The Blackmagic OLPF 12K came out and made an already great camera even better and far more useful for advanced virtual production. On top of that, Sony has been making some of the nicest 4K specialty cameras out there and they now have another 8K camera in the Burano.

Lately I have been warming up to the Raptor cameras, they do tick many boxes and some of the RF lens developments have gotten more interesting for keeping things simple and light. However, I feel that now more than ever RED is in a good point to shake things up, particularly if it's something in the Big Sky and Achtel 9x7 level of cameras or even just something to knock the URSA 12K OLPF down a peg. I'm not sure what though as we've had some discussions around about troubles with going medium format and more than likely a FF35 scaled option would have to be present as a definite "must have" for many. There's already rumors of Blackmagic Design getting ready to shake things up again even though the 12K OLPF and 6K FF Cinema have been making waves so I'm wondering if NAB will be a bit of a surprise this year. Thoughts?

 

[Les Hillis]

I don't think much has changed over the last few months.

Imo RED still has the edge over everyone else when it comes to providing a high-resolution dedicated cinema-camera system.

Not saying the RED options are perfect, but everything else seems like it requires more compromises or pita work-arounds.

If it were me I'd be looking at the V-Raptors and ways to lens, light and compose that play towards that 'medium-format' look.

At the other end of the spectrum, but still talking about potential new sensors, was playing around with a 500mm lens and 2x tele extender the other day using the Helium @ 8K FF (17:9) and was impressed by the quality of the 8K image.

Mentioning it because it also got me thinking about sensor sizes.

The 8K V-Raptor VV & S35 are both 35.39 Megapixels (@ 8K 17:9 - 8192x4320)

If the V-Raptor S35's pixels are 36% smaller than the VV ones, that means if RED made a VV sized sensor using those S35 pixels, its resolution would be 11141x5875 (8192x4320 + 36% = 65.45 Megapixels)

Which would be getting very close to this (in the horizontal) -

70mm 15-Perf
11586x8658 (1.34:1AR)
100.31 MP

A sensor pushed out to 12K using the same sized pixels as in the V-Raptor VV would also be interesting.

Would be good to be able to crop in to get an 8K frame, without having a FF resolution that gets you into the potential weirdness of actual medium and large format acquisition, especially with the emphasis still being on the requirement for capturing images in motion.

 

[Phil Holland]

My general perspective and position on digital cinema hasn't changed much since the early 2000s.

My main point back then in the pre-RED days was digital was going to evolve into a system that gave us the same format and resolution flexibility of motion picture film as well as still film. That pretty much is how this has unraveled with some moderate surprises and good adaptation.

With RED, I like their strategy of pushing to the extents of the format in terms of image circle/diagonal. Readding back here I used to refer to Dragon 6K for instance as S35+ because the image circle was a hair larger than S35 4-perf spec. VV followed the same route more or less compared to 135 format, or what we call FF35, by being nearly 5mm wider of a format size.

Another "good thing" RED has done is adhering to the DCI Aspect Ratio of 1.9:1 (called 17:9 now) generally, with only a couple of exceptions. This has made a lot of sense in terms of vertical coverage as well as providing the wider dimensions that is often more appealing to cinematographers. Hope this generally continues.

In terms of how I want this to play out for me personally, I want digital cinema solutions for:
- Super 16mm
- Super 35mm
- VistaVision
- 65mm 5-Perf
- 70mm 15-Perf

With modern quality digital cinema cameras being multi-format as you crop, which also often nicely helps with frame rates, I don't think we'll see S16mm being explored as the latest mobile phone sensors are alarmingly close to that format size.

S35 and VV, and by proxy, LF and FF35 are likely going to be the mass market breadwinners in terms of sales as well as the majority of production use. Also these formats provide a huge range of glass possibilities.

The two major shockers to me in terms of digital formats are Micro Four Thirds, which is the same approximate format size as 110 film, a format that died. But it has opened up the entry level market and lives nicely between S16 and S35.

The other is the more recent popularity of Middle Format, 44x33mm sensors. For Phase One owners, including myself, we skipped over the entry level offerings and focused on the larger 645 equivalent backs (about 53.4x40.4mm). It didn't gain traction in sales for them, so much so they didn't release a new model IQ4 with that format size. But Fuji and Hasselblad have made that format size more approachable and developing on that format. This is still mainly a still market trend, but motion plays a role on these formats.

As for larger than 70mm 15-perf digital cinema, likely a few niche cases, but expensive and difficult to work with in both size and power needs. Also more practical things like working f/Stops/T-Stops would get difficult. Long, long ago, Kodak was exploring a new ~100mm film format to provide "the ultimate" while also leaning into the worldwide adopting trend of widescreen, but it never went down. Much like VV and 70mm 15-perf, still a horizontal transport, but with a wider format size. That would have been awesome, but I do think perhaps not worth the squeeze as the full system would need to be looked at harder than ever and the cameras would be large.

In all of this is the market demand and various filmmaker's needs, wants, and desires. I know from tickling the bushes that I am in a very rare air space surrounding the tools I'm looking for down the line. Thankfully many of the main trends I was after have been latched onto and are even being iterated on. I have two major hopes where future formats go for motion picture use, beyond what is covered today, but they will be costly and difficult in some ways.

But I'll circle back, S35 and VV are not going anywhere in my mind. VV in particular, at least RED's take on it, is wonderful for my personal filmmaking needs. I still shoot S35 and pretty much purchase dedicated cameras to work in that format if I'm not cropping on Raptor 8K VV, Komodo-X being a good example and works great with the S35 glass.

 

[Karim D. Ghantous]

There's no point unless you're getting significantly more resolution. DR will be slightly more but not enough to be worth it by itself.

Film and digital are very different. I estimate that digital sensors need 1/4 of the surface area of film to match film on a technical level. Also, the relationship between sensor size and image quality is elastic for digital. Want proof?

RED Helium 8K DxOMark Sensor Score: 108 — A new all-time-high score!

RED’s prototype Helium 8K sensor has shattered our previous top DxOMark sensor score. With a dramatic jump to 108, it leaps past the only other sensor to have broken the 100 barrier – RED’s own Epic Dragon prototype (101 points). RED has made a name for itself by producing high-performance...

www.dxomark.com

The Dragon and Helium sensors scored as good as, or better than, larger sensors in Phase One and Sony cameras in terms of DR.

 

[Les Hillis]

If I'm reading the resolution, image-circle and acquisition format numbers correctly, then 70mm 15-perf and 6x7 medium formats would be covered by 12K resolution variations of the current V-Raptor sensors.

In the case of the larger pixels in the VV camera, you'd end up with sensors similar in physical size to their film equivalent, while obviously with the smaller pixels of the V-Raptor S35, you'd get an equivalent resolution in a sensor with smaller physical dimensions, which could be useful if it enables the use of a wider range of lenses.

I think RED would be able to sell more than a couple of those.

I could also see RED taking a page out of their own book and doing different sized cuts of the same sensor and putting them in different bodies that have different kinds of processing capabilities, to re-unify their entire camera range more around a particular sensor (or two) again.

With every new camera they've brought out since the DSMC2 Monstro, they've made improvements that could have been applied to the previous camera/sensor.

DSMC4 would give them an opportunity to implement those changes, and more, across an entirely new camera range, from their smallest 'crash-cam' to their most capable flagship camera.

It'll be interesting to see where they go with it.

 

[Brian Timmons]

Just saw this video in my YouTube recommendations.

It looks close to the size and weight of the V-Raptor XL (maybe bigger).

As far as potential large format motion cameras, I have to imagine that
AR/VR might be just as important at driving development of these cameras
as cinema...perhaps even more.

Brian Timmons
BRITIM/MEDIA

 

[Les Hillis]

I remember when RED effectively broke the measuring scales with the DxOMark Sensor Score that Karim linked to above. I thought it was cool that it was a camera/sensor primarily designed for capturing motion pictures that did it.

I like the idea that RED could also potentially make the best performing Medim Format camera as well, as long as it's just incidental to their making the best Digital Cinema camera's they can.

Doing a search for images made with Medium and Large Format camera's and lenses, it's quickly clear that just like any other medium, it's what people do with the equipment and subject-matter that counts and that, non-technically speaking, there's nothing intrinsically better about an image just because it's made with those tools.

I'd even go so far as to say, that the "Medium Format Look' is actually a novelty effect, similar to 3D or stereoscopic images. When it works and is done well, the effect looks really good and engages the viewer, but it's still the content of the image that really matters, especially when the novelty of the effect wears off through repetition or familiarity.

It could be argued that high resolution and accurate colour fidelity are also a novelty effect, and even that all audio-visual mediums are, but you've got to draw the line somewhere, or you risk throwing the baby out with the bath-water and dismissing the importance of communication mediums altogether.

I think there's still room there, in terms of resolution and dynamic range, for RED to keep advancing into, as a way to facilitate better communication and more expressive creativity.

 

[Zack Birlew]

Quick note, I have recently gotten into a Fuji GX680 setup, seems like a nice camera system, majorly loud slap when taking a photo on it, but adapting it to digital is tricky for sure. It's interesting that there are "Franken-rig" efforts to make it a much more digital system but unless you use an expensive digital back or, as I've been looking at, experimenting with that new tiny industrial Sony box camera, it's been quite a challenge. However, the GX680 isn't the only game in town for digital adaptation experiments and it is a good example to show that people do want digital medium format and RED putting out a FF35 4K+ video capable digital back could be just the thing to bring that market full circle and open RED up to a new market.