The future of Super35 sensors

[Etienne Vigneault]

Hi guys !

I'm about to hop on the anamorphic train with the 3-lenses RF-Mount set from Vazen for my Komodo. But I'm kinda wondering if non full frame lenses are a good thing to buy? Like I feel it will soon be the end of S35 sensor seeing that most new cameras announced from every companies are full frame. Like if there will ever be a "Komodo Gen2", or a new system a bit less expensive than the V-Raptor, will it be S35?

Thanks a lot !

 

[Phil Holland]

S35 as a format isn't going anywhere.

Likely a hot take and certainly a personal opinion, the general fickle mindset and uneasiness of having multiple formats on the market in the digital space has been difficult to watch really.

The market was much smaller when film was the primary medium, but most didn't get too deep in debating on S16, S35, VV, and 65mm glass taking over the world. The only thing that really happened in that time frame was Super 35mm became the most widely used format. Which at the moment is still very true industry wide.

And I'll say, I own 3 8K VV cameras now and still use S35 glass on them in 5K-6K resolution. In the world of RED these are indeed multi-format tools that work pretty effectively.

That said, yep, trend is now everybody is offering the FF35/LF or VV stuff now. But the trend months earlier were a bunch of S35 models. 2022 will likely bring a bit of both.

Impossible to say anything about any future cameras, nobody is going to give those heads up at this point in the game. The only thing that is always going to be true is new cameras are going to be made and released.

 

[Ignacio Aguilar]

ARRI is about to release a new S35 camera with a new sensor. That could be a game changer if it's half as successful as the Alexa.

 

[Karim D. Ghantous]

ARRI is about to release a new S35 camera with a new sensor. That could be a game changer if it's half as successful as the Alexa.

The only reason they launched the LF was because they needed time to develop a 4K S35 sensor. IMHO.

 

[Phil Holland]

The only reason they launched the LF was because they needed time to develop a 4K S35 sensor. IMHO.

Grossly incorrect fyi.

Several things actually going on were:
- Need for 4K/4K+ acquisition
- Desire for a larger format camera humans could purchase
- The recognition that many Alexa 65 projects were utilizing the 5K mode, which is a hair larger than the LF sensor

Arri, and to their credit, responded to market needs and demands.

Similarly what would transpire in an alternate universe if RED just kept to Dragon S35 and something like say..... the Venice 2 8.6K came out. Pressures would be felt to have things like 8K and a larger format sensor as that is part of the equation. In this universe it has been the other way around with RED doing this, mainly in the cameras used in our market first.

Expressively though, 8K was always in the cards for Sony. V2 is just the first camera to actually do it right in the face of a previous CineAlta release, the F65. Perhaps even inspiring RED to step on the gas to develop Dragon, Monstro, Raptor and Helium.

What's very interesting now is the pricing tiers of the cameras and how they may or may not relate to various types and tiers of productions. And there is the interesting trend that has been developing for the last 3-5 years. A trend that also may or may not mean much to some. But a trend nonetheless that means a lot to rental houses in particular. And honestly considering the newness of a good chunk of them, this all makes sense.

I've been trying to develop some sort of way to understand the current landscape with all the variables at play. Hard to really describe, but there's a whole lot of difference between companies who have 1-4 cameras in their inventory versus those who have 40-300+. There's a great deal of impact that this creates as a large chunk of productions have nothing to do with gear ownership, though some do. There's also a very wild entry level rental market that has nothing to do with servicing and maintaining equipment, which is virtually impossible to wrangle in that realm. Then there's a whole broad range of companies that tend to all of the things required for general professional support for projects big and small that live outside of rental house model or owner op model. And then there's studios and companies that do internal work.

Complicated stack of unwrangled cables. All deliver power. Not all the cables are reliable or even in good shape. Most people reach to the well coiled cable and hopefully return it in kind, probably willing to pay more for such things. And then there's some cables that aren't in that stack that you can never use, because they are always plugged in elsewhere.

 

[Etienne Vigneault]

And I'll say, I own 3 8K VV cameras now and still use S35 glass on them in 5K-6K resolution. In the world of RED these are indeed multi-format tools that work pretty effectively.

Thanks for your opinion Phil ! I never thought about the possibility of using S35 lenses on full frame with a crop. And doing mainly corporate videos for clients who don't even own 4K screens yet, I wouldn't mind using a cropped sensor on a future body to still be able to use my S35 Vazen lenses.

Thanks a lot for your time

 

[Phil Holland]

Thanks for your opinion Phil ! I never thought about the possibility of using S35 lenses on full frame with a crop. And doing mainly corporate videos for clients who don't even own 4K screens yet, I wouldn't mind using a cropped sensor on a future body to still be able to use my S35 Vazen lenses.

Thanks a lot for your time

Theory and theory in practice perspective here if you're curious. When I began in this industry I dealt a great deal with high resolution motion picture scanning. The general concept behind motion picture film is it has "one resolution" per each negative stock. To increase resolution and minimize grain you have a larger negative format size. Which is what you see very clearly when looking at 8mm, 16mm, S35, VV, 65mm, and 70mm 15-perf material.

There's a lot of interesting perspective regarding film scanning and resolution, often tied to the specific scanners used or even optics at play. Removing all things from the equation film ends up with somewhere between a 4-6 micron pixel pitch depending on the type of film.

There's been far too many advancements to even mention here, but digital is out resolving film these days on and one of the more common practices in motion picture finishing is actually softening the image, but that truly will depend on your intentions for your image. Not everything shot is a pure 100% film emulation. Often it's a little bit from here, a little bit from there, and hey what if we try this sort of thing.

RED with Dragon, Monstro, and V-Raptor built on a very intelligent 5 micron pixel size. Which allows for the 8K VV format as a whole to give you a pretty faithful analog (in mindset) to the film format workflow. In the VV sensor you have Super 8mm, Super 16mm, Super 35mm, and VistaVision. Nearly nobody shoots Super 8mm for traditional production mainly due to the image quality drop and resolution loss in the modern era, but "you can do this". However, S16, S35, and VV all look pretty nice off this "digital film stock".

This has been my perspective for years, it's actually nice to see RED actually marketing it in this way on V-Raptor. Though it's more then merely "dual format" in my mind.

Current Kodak Vision3 stock is offered in ASA 50 through 500 speed and in processing and exposure you can have an impact on visible grain. Also interestingly enough in printing as well. For digital you have pretty much the same idea. How far can you push or pull the ISO rating to create a pleasing image? There's tradeoffs to both mediums when doing this, mainly to how and where you are allocating dynamic range above and below 18% gray. Not mind blowing if you've seen my posts here over the years as this all works as expected in practice.

Interestingly enough, perhaps too much of an abstract take here, but my general perspective which is somewhat echoed by their use case is tying each unique sensor pixel pitch to a film stock and sensitivity and attempting to think of them in this manner. I've done some wild experiments on this front that have been rather interesting for some crossing the film to digital barrier or even vice versa.

This is the perspective and mindset I base a lot of things on. The tricky aspect of all of this is how digital is different from film. The first part being a cleaner medium in general on the higher end cameras by a considerable amount, especially if you process the imagery further. Film generally really doesn't like to be under expose, but reacts nicely to reasonable overexposure. You don't get the gains in cinematography that you have in straight photography due to our general use of rather confined shutter speeds. Digital on the other hand reacts fairly well to underexposure within reasonable limits. The key thing about all of this is when you stress both mediums the ugly warts begin to reveal themselves however. Generally for motion picture film you don't really exceed pushing or pulling 1 stop, 2 at the most extreme. This actually wasn't even done much due to the latitude of film being pretty low until about the 50s or 60s. But you can use it to have an impact on texture as well as saturation depending on your exposure practices and processing.

In general, it's always a good idea to start with more image information, but there's even things lost in film emulation that don't take into account things like the resolving power and magnification on the lenses themselves. A neat trick is shooting something relatively detailed with a S16 lens then finishing it out accordingly and adjusting for that via the format size and resolution gains of say the VV format.

Lots of hairy chatter between the pros and cons of different format sizes. Seems to set off a hot series of discussion among filmmakers, but the core concepts of how this has worked for motion picture film remain the same as it pertains to digital. Whether those differences are "meaningful" or not, that's I strongly feel a conversational point of individual opinion rather than how things actually are working.

That's been also a contested thought in modern times. Not shockingly filmmakers have very strong opinions and like to argue about things whether they know what they are talking about or not, or have been subjected to some localized experience, something seen, or word of mouth even. In more recent times it's been wild discussing topics surrounding tools and technology with people who have zero to little experience, but everybody has elbows it seems. Keeping an open mind, open eyes, and a turned ear is something that has benefitted me greatly in my career and I've been fortunate to learn from a lot of people who have built a great deal of the industry that surrounds us today.

Heh. Didn't mean for that to be a long post, but I did in fact eat breakfast while writing it. Slow bagel chomping today it seems.

 

[Karim D. Ghantous]

- Desire for a larger format camera humans could purchase

Yeah, but why? What does VV give you other than a clean 8K at ASA 1600 with terrific DR? If it doesn't give you that, why do you want it? I know, people are irrational. But is there a rationality behind this? I haven't seen it.

 

[Etienne Vigneault]

Heh. Didn't mean for that to be a long post, but I did in fact eat breakfast while writing it. Slow bagel chomping today it seems.

Really interesting read! You know, I come from a small town in northern Canada, far away from big cities. So I never had the opportunity to learn from people like you who carry a TON of experience. I'll definitely spend more time here on the forums! Thanks again for your time!

 

[Phil Holland]

Yeah, but why? What does VV give you other than a clean 8K at ASA 1600 with terrific DR? If it doesn't give you that, why do you want it? I know, people are irrational. But is there a rationality behind this? I haven't seen it.

Yes, ISO 1600 say at full 8K versus 1600 at 5K or 3K from the same sensor and finishing out to 4K or 2K, the image noise and texture will be smaller obviously with the 8K source. As well as downsampling improvements on details. On the very technical side of things, an 8K sensor when it comes to a classic Bayer Pattern Color Filter Array also does "nice things" technically when going to 4K or 2K from a pure color perspective.

Somewhere between the FUD, mud, and actual truth you'll find the reasons people choose all formats big or small. You can match FOV across different formats and even squarely match blur circles/bokeh/out of focus image shape. But you can intrinsically never match the subtle way glass works on each format.

Some fascinating examples. Take a large format lens and use it on it's native format through smaller format, shine a light into it and create a lens flare. The smaller format will never "see" the rest of the projected image image circle and that will look different. Small thing, but a thing.

Also taking that same lens and understanding that the smaller the format the more magnified in on the glass itself. With that comes slightly larger artifacts. Small thing, but a thing.

The glass, say a VV 58mm lens at T5. To match the bokeh and FOV on S35 you'd be at about T2.8 on a 35mm prime. S16 would need about a T1.6 on an 18mm prime to do the same. This gets hairier if you shoot at say T1.5 or T2 on VV, because there aren't fast enough lenses to get there. Also looking at those T stops and how they pertain to optimal image quality, diffraction, and again magnification in the optics themselves; differences. Those differences btw hold merit for each of those formats as well. But it should be noted that the widest rectilinear lens you can get for VistaVision in 2021 is 10-12mm. The equivalent FOV on Super 35mm would be a 6-7mm prime which doesn't really exist. Same FOV on S16 is a 3mm lens at the widest btw. On the flip side you get more punch from a smaller sensor on the longer side of glass, which has real value when it comes to shooting certain subject matter. For instance on S35 a 300mm gives you the same FOV of about a 500mm prime on VV. In this specific case you can find rather fast 300mm primes, but not as fast 500mm for instance. Subtle and notable difference.

These are all qualities that have been known to the filmmaking community when film was the dominating medium. I don't know exactly what transpired, but there was marketing and often a very defensive stance taken on large format or whatever format really as it was becoming a growing industry trend in terms of cameras being manufactured. Some chose the pathway to discuss some of the reasons they chose each format, others chose the path to "if you just do this, this, and this these look the same and there is no difference between them!". Some quantified some of those topics with a dangerous term of "meaningful difference" not understanding or acknowledging that some discerning filmmakers make very deliberate choices on what they choose to tell stories based on very small differences and preferences or not. One persons "meaningful" is not another's.

Down in the grit of it, the real grit of it. If you want to use something like the Zeiss Otus 55mm f/1.4, that lens and specific focal length only works one way on each format size you use it on. Similarly if you like the Leica Summilux-C T1.4 primes on S35, you can't just use them on VV and even get a projected image circle that covers. You can use expanders and focus reducers (speed boosters) but this comes at the expense or benefit of light loss or light gained, but always comes at the cost of adding additional optics into the chain that do have a visual impact on image quality over when using said glass for the it's intended given format.

The things I'm mentioning here aren't crazy ideas or anything. It's just how this all works. It's always been how this all works.

And one thing I'll mention, mostly because I've observed it, and this is going to be the only "wild" thought there. I've witnessed first hand by my own work and how other work with each of these formats and I've been able to identify a few things that separate the formats, how you place the camera, how far it's from talent, etc. Whether those things are meaningful or not, hard to tell, but It's just an observation nonetheless.

Something not related to format size that I've also seen influence the way people work with cameras is a somewhat obvious one. The physical size of the camera itself. But that's off topic.