future of film cameras

[Helge Abrahamson]

It's a really interesting question.

Imagine if a completely new camera was designed, using modern technology, what a radically different camera it would be possible to make.

Imagine that the ground glass of the optical viewfinder was replaced with a CMOS sensor. 
One would not only get a much better viewfinder and assist system, but it would be possible to actually use the videotap for focus and exposure. 
And by having predefined profiles on the CMOS sensor, it would be possible to exactly imitate the sensitivity, colour balance and grain structure of the selected film stock. One could also select simulation of different development and processing methods on the assist system, making it truly a "what you see is what you get" system.

Imagine how much less mechanically complicated a film camera would be, if mirror, movement and magazine drive was made with highly precise digitally controlled, direct drive, stepper motors, all being kept in perfect sync electronically.

Imagine using a stepper motor for the movement and a gate that was electronically adjustable. One would be able to select 2, 3 or 4 perf by the push of a button.

Imagine how much lighter it would be, if structural components and magazines was made of carbon fiber instead of metal.

I would love seeing someone having a go at designing a modern film camera from the ground up.

[Uli Meyer]

42 minutes ago, Helge Abrahamson said:

a "what you see is what you get" system.

What is the fun in that? Part of the excitement of shooting film is that you don't. For me at least 😉

[John Rizzo]

David Sekanina from Zurich who is a member of this forum 

he is working on a super 16mm camera with a  anticipated delivery July 2025

 

https://www.yolk.org/about

 

https://www.yolk.org/

 

also Logmar is working on another Super 8 mm camera   https://logmar.dk/gentoo-s8/

which hopefully be  available by the end of this year.

 

[Helge Abrahamson]

37 minutes ago, Uli Meyer said:

What is the fun in that? Part of the excitement of shooting film is that you don't. For me at least 😉

I get you point, even though I think it must sometimes be more nerve wracking than fun for the director. 
But think of it a bit like composing a piece of music with the comfort of using a good synthesiser and then later, having it played back to you on a Steinway concert piano.

[Uli Meyer]

5 minutes ago, Helge Abrahamson said:

I get you point, even though I think it must sometimes be more nerve wracking than fun for the director. 
But think of it a bit like composing a piece of music with the comfort of using a good synthesiser and then later, having it played back to you on a Steinway concert piano.

I'm a director and I always trust my DP to know what he is doing. We discuss the lighting and atmosphere and how to achieve it. Going to see the rushes the next day and discovering that it looks often better than imagined is magical. Before digital, nobody would even think twice and nerve wrecking has only become an issue since digital "what you see is what you get" systems exist.

[Geffen Avraham]

I’d love a WSIWYG film camera, but at the same time I do feel like the reason I shoot film is that the limitations make you a better cinematographer.

Edited June 1, 2024 by Geffen Avraham

[Geffen Avraham]

3 hours ago, Aapo Lettinen said:

I regularly shoot at 12800 iso on S5 + atomos and that is almost 4 years old mediocre camera, nothing high end at all. And have used the end results for cinema release on 4k dcp (local run + festival stuff), works fine.

 

You’re right, my one concern is that the S5 (and my R5) are full frame cameras with larger pixels than most.

On the other hand, the GH5s delivers good results in low light, and has an even larger pixel pitch than the R5, although significantly smaller than the S5.

The question is whether low light MFT sensors are commercially available off the shelf.

Edited June 1, 2024 by Geffen Avraham

[Simon Wyss]

6 hours ago, Uli Meyer said:

nerve wrecking has only become an issue since digital "what you see is what you get" systems exist

Nothing new, Bell & Howell Co. advertised in the 1920s What You See, You Get. 🙂

[Uli Meyer]

1 hour ago, Simon Wyss said:

Nothing new, Bell & Howell Co. advertised in the 1920s What You See, You Get. 🙂

So they did! 😉

[Aapo Lettinen]

6 hours ago, Geffen Avraham said:

You’re right, my one concern is that the S5 (and my R5) are full frame cameras with larger pixels than most.

On the other hand, the GH5s delivers good results in low light, and has an even larger pixel pitch than the R5, although significantly smaller than the S5.

The question is whether low light MFT sensors are commercially available off the shelf.

I have both and the S5 is better in low light, something between one and two stops difference. the gh5s has better internal codecs (the 400Mbps h264 4:2:2 intra is perfect for documentary stuff and for backup+b-roll camera when shooting short films) which is why I usually shoot the gh5s when need to use internal codecs (one quick gimbal setup, when needing to shoot tons of material in compressed format) or need huge crop factor for tele lenses, and choose the s5 for other stuff and low light when having possibility to use atomos with it.

the exact sensors used on mirrorless cameras or video cameras are often not possible to get off-the-shelf, you might try ordering them if you can purchase thousands at a time (Z-cam uses same sensors on most of their cameras than Panasonic has on their mirrorless line so it is possible to get them. I don't know how many hundred k:s or millions you need to pay to get access to the sensors but it is possible if you have huge budget). They are expensive though so normal people or crowdfunding campaigns can't afford them.

Another issue with "just getting the sensor and building the camera around it" is that you need to be extremely good at FPGA programming to be able to manage the data streams from the sensor to usable image and to control the sensor readout and timing correctly. Like really, really good at FPGA stuff. If you happen to do that make-digital-camera-from-scratch for living and done it as profession like 10 to 15 years or more, then you might have a chance of getting it working in reasonable time (even a year could be possible). Otherwise I would not bother, that is insanely difficult and complex stuff and if you are capable of doing that you could just release your own "RED" or "Blackmagic" digital cinema camera in no time because that is basically what you would be doing by "just using a sensor and process data it to a usable image which could be recorded", building a digital cinema camera entirely from scratch

[Aapo Lettinen]

9 hours ago, Helge Abrahamson said:

It's a really interesting question.

Imagine if a completely new camera was designed, using modern technology, what a radically different camera it would be possible to make.

Imagine that the ground glass of the optical viewfinder was replaced with a CMOS sensor. 
One would not only get a much better viewfinder and assist system, but it would be possible to actually use the videotap for focus and exposure. 
And by having predefined profiles on the CMOS sensor, it would be possible to exactly imitate the sensitivity, colour balance and grain structure of the selected film stock. One could also select simulation of different development and processing methods on the assist system, making it truly a "what you see is what you get" system.

Imagine how much less mechanically complicated a film camera would be, if mirror, movement and magazine drive was made with highly precise digitally controlled, direct drive, stepper motors, all being kept in perfect sync electronically.

Imagine using a stepper motor for the movement and a gate that was electronically adjustable. One would be able to select 2, 3 or 4 perf by the push of a button.

Imagine how much lighter it would be, if structural components and magazines was made of carbon fiber instead of metal.

I would love seeing someone having a go at designing a modern film camera from the ground up.

Electronic film transport would be inferior to mechanically made one at normal sync sound speeds, let alone anything higher than 24fps. It would be noisier, less precise, very likely to jam ocassionally and would in most cases be more expensive to make too.

Electronic movement could work perfectly for time lapse use though. I think one could handle something like 2fps or 3fps reliably with electronic system no problem. Even a solenoid based could work at those speeds. It would just start to fall apart when speeding up and likely be completely unusable above like 5fps or 8fps or so.

It is, however, possible to electrically syncronize the mirror shutter and take out of the mechanical link between the shutter and the movement. The Arricams have that kind of arrangement and it is working fine. People making crystal sync motors are often capable of making that kind of stuff, for example I could make a electronically linked shutter no problem if there would just be enough time and budget for it as I am making similar stuff for other purposes already.

Magazine motors are separate on many cameras and that does not need anything special, just a motor that is running slightly "too fast" and a friction disc which slips to reduce the speed to the required (all the time varying as the roll changes size) speed the takeup needs. It does not need any fancy control though sophisticated control circuit would lower the power consumption if used.

I went through the complexities of getting a video image out of a bare imaging sensor on previous post, it is super compex and demanding stuff very few people are capable of doing. and takes long to get working even if finding someone capable of doing it.

------------

Of course if you have couple of million Euros of free cash it would be possible to arrange something, if being able to hire enough super experienced professionals to design and manufacture the camera for you then of course could be possible. but one would get a truckload of better existing film cameras at that price and still have enough left to shoot tons of film with them. So would be expensive hobby project with the end result not being any better than something which could already be purchased used for 50 or 100 times cheaper

[Steve Switaj]

On 5/30/2024 at 11:40 PM, Tyler Purcell said:

A beam splitter, but yes that would be the idea. Similar to the Canon RS series still cameras. That beam splitter is made of a very special and proprietary coating, but it works very well. 

 

Back in the 80's Fries Engineering did reflex Mitchell conversions and one of the options was a version using a pellicle beamsplitter. 

They were popular in certain parts of the VFX community because, unlike the spinning mirror conversions, the beamsplitter option preserved the original focal plane shutter, and didn't leak light when used for things like time-lapse or animation. 

They did result in a loss of a 2/3-stop of exposure, and even though the pellicle was very thin, you could still occasionally see hints of a vertically doubled image on contrasty subjects. We learned to live with this since it was mostly a viewfinder thing caused by light reflecting from both surfaces of the pellicle, and didn't really affect the transmitted photographic image, but it did really freak out the occasional DP who had never experienced it before.

Those Fries cameras were set up so that the mirror and ground glass was a unit that slid in to a cavity in the side of camera in front  of the focal plane shutter, and they usually came with a set of mirror options  - a 33% pellicle, a full 100% solid mirror for composing and previewing, and a full transmission block with no mirror.

I did motion control work in the 90's with these cameras and it was common to treat them like old rackovers. You would program and preview with the 100% mirror ( because video taps sucked back in the day ) then swap to the full-transmission block for shooting ( you didn't shoot with the pellicle to avoid losing even more light when you already have 4 second exposures ). I personally never got bitten, but I know of more than one day-long shot that was lost because someone forgot to swap out the 100% mirror and shot 50 feet of blank film instead of some dramatic miniature.

 

 

Edited June 1, 2024 by Steve Switaj

[Helge Abrahamson]

55 minutes ago, Aapo Lettinen said:

Electronic film transport would be inferior to mechanically made one at normal sync sound speeds, let alone anything higher than 24fps. It would be noisier, less precise, very likely to jam ocassionally and would in most cases be more expensive to make too.

Stepper motors are being used for lots of high precision applications - they are really fast, relatively noiseless, strong and mind bugging precise - don't see why they couldn't be used for a film movement...

But you are right, it would take a couple of million euros (possibly even more) to design and build a completely new film camera using today's technology - but imagine how great it could be.

Question, I guess, is if there's a demand for such a camera, or if the current "popularity" of the celluloid media is really only the last nostalgic swan song, before eternal oblivion.

[Tyler Purcell]

6 hours ago, Steve Switaj said:

They did result in a loss of a 2/3-stop of exposure, and even though the pellicle was very thin, you could still occasionally see hints of a vertically doubled image on contrasty subjects. We learned to live with this since it was mostly a viewfinder thing caused by light reflecting from both surfaces of the pellicle, and didn't really affect the transmitted photographic image, but it did really freak out the occasional DP who had never experienced it before.

Interesting, I think I'd have to go with something similar to what Canon developed for the EOS-1 RS camera. It's a special coating, which supposedly is available, which goes on the front side of the pellicle and reduces the issues with reflection quite a bit. We don't plan on using any ground glass and it doesn't need to send MUCH light to a modern digital imager, especially a good one. 

I hope once we get a prototype going, we can experiment with different optical units and perhaps come up with something that already exists on the market. Right now we're in limbo waiting for some movement parts to arrive, then we can start designing around them. 

Thanks for the info go, very interesting. 

[Tyler Purcell]

6 hours ago, Helge Abrahamson said:

Question, I guess, is if there's a demand for such a camera, or if the current "popularity" of the celluloid media is really only the last nostalgic swan song, before eternal oblivion.

If the camera can't be manufactured and sold for a price round $5 - 8k, it's not worth doing, no. Obviously if we go into a world war or major recession, the project will probably be put on hold as well. 

There is a pretty decent demand for a new camera, something supported, something more open source and something that can be built upon to make better. I think the reason why nobody has really built one, is simply because it's scary. It's a huge investment in time and resources, to perhaps make something a few people will buy. I have many ideas to keep the cost down AND best of all, because I've been a tech for a few years now, I understand the film path very well and how to make a good camera that works. No guessing game required, the tools are in my hands now. This helps a lot, it's not just starting from scratch ya know? 

I think the critical thing is a prototype that can be shown off, even if it's ugly. It's going to be all about the design philosophy and getting people happy about something new. Film isn't going to die anytime soon, but expendable money, especially on expensive older cameras, is dwindling. Something you can buy on a credit card brand new, for about the cost of a low-end modern digital cinema camera, opens up the market to new people as well. 

 

[Geffen Avraham]

13 hours ago, Aapo Lettinen said:

I have both and the S5 is better in low light, something between one and two stops difference. the gh5s has better internal codecs (the 400Mbps h264 4:2:2 intra is perfect for documentary stuff and for backup+b-roll camera when shooting short films) which is why I usually shoot the gh5s when need to use internal codecs (one quick gimbal setup, when needing to shoot tons of material in compressed format) or need huge crop factor for tele lenses, and choose the s5 for other stuff and low light when having possibility to use atomos with it.

the exact sensors used on mirrorless cameras or video cameras are often not possible to get off-the-shelf, you might try ordering them if you can purchase thousands at a time (Z-cam uses same sensors on most of their cameras than Panasonic has on their mirrorless line so it is possible to get them. I don't know how many hundred k:s or millions you need to pay to get access to the sensors but it is possible if you have huge budget). They are expensive though so normal people or crowdfunding campaigns can't afford them.

Another issue with "just getting the sensor and building the camera around it" is that you need to be extremely good at FPGA programming to be able to manage the data streams from the sensor to usable image and to control the sensor readout and timing correctly. Like really, really good at FPGA stuff. If you happen to do that make-digital-camera-from-scratch for living and done it as profession like 10 to 15 years or more, then you might have a chance of getting it working in reasonable time (even a year could be possible). Otherwise I would not bother, that is insanely difficult and complex stuff and if you are capable of doing that you could just release your own "RED" or "Blackmagic" digital cinema camera in no time because that is basically what you would be doing by "just using a sensor and process data it to a usable image which could be recorded", building a digital cinema camera entirely from scratch

No doubt it is hard, but luckily there are quite a few companies that do this all the time, and can be contracted for custom work. Often they will have COTS products built around available sensors, Ximea is one example. Prices can vary.

For example, for making a new Eclair Video Tap, one company sells an extremely thin 1080p SDI board camera for 250 euros.

It would be worth talking to the founder of Digital Bolex, their sensor is still highly regarded, despite them being a startup on a relatively limited budget. Digital Bolexes also sell on eBay these days for double their price 10 years ago - the only digital camera from that era to appreciate in value.

I think he's expressed interest in working on a new camera.

Edited June 2, 2024 by Geffen Avraham

[Geffen Avraham]

I believe the D16 used the KAI-04050, a CCD sensor which sadly is no longer made. It was developed by Kodak, then Truesense, then finally Onsemi, who of course also makes the ALEV sensors.

Onsemi's closest CMOS equivalent is the PYTHON5000, a global shutter sensor with slightly higher resolution and slightly smaller pixels. It costs about $200, and Basler makes COTS C-mount USB3 cameras with it that are available for about $1,000.

[Robin Phillips]

9 hours ago, Geffen Avraham said:

I believe the D16 used the KAI-04050, a CCD sensor which sadly is no longer made. It was developed by Kodak, then Truesense, then finally Onsemi, who of course also makes the ALEV sensors.

Onsemi's closest CMOS equivalent is the PYTHON5000, a global shutter sensor with slightly higher resolution and slightly smaller pixels. It costs about $200, and Basler makes COTS C-mount USB3 cameras with it that are available for about $1,000.

I think theres an argument to be made that if a turn key solution for accurate exposure readout to the monitor. I'd imagine something akin to what Andree is using in his HD taps, which can handle 12 stops of latitude in 10bit all in a small unit, would make the most sense. going overboard with something like the D16 is probably workable if you need a digital backup that is 100% usable, but thats a pretty expensive way to build a new film camera. 

[Aapo Lettinen]

32 minutes ago, Robin Phillips said:

going overboard with something like the D16 is probably workable if you need a digital backup that is 100% usable, but thats a pretty expensive way to build a new film camera. 

Digital backup ruins the whole idea of shooting on real film because people are lazy and would just use the digital all the way and not even load film in the camera. especially if the emulation is well working. could then just toss all the film movement parts as unnecessary and make it small digital camera with good film emulation features. All the vimeographers would buy those if it would be "cool enough design" so that they could get attention with it 🙂

Edited June 2, 2024 by Aapo Lettinen

[Max Field]

On 6/1/2024 at 6:41 AM, John Rizzo said:

David Sekanina from Zurich who is a member of this forum 

he is working on a super 16mm camera with a  anticipated delivery July 2025

 

https://www.yolk.org/about

 

https://www.yolk.org/

 

I hope he can get it off the ground with most of the quality of life features one could expect with modern cameras.

At the same time I don't understand where any of you are finding the money for color film when the middle-class production budgets have been shrinking more and more every decade.

[Geffen Avraham]

51 minutes ago, Aapo Lettinen said:

Digital backup ruins the whole idea of shooting on real film because people are lazy and would just use the digital all the way and not even load film in the camera. especially if the emulation is well working. could then just toss all the film movement parts as unnecessary and make it small digital camera with good film emulation features. All the vimeographers would buy those if it would be "cool enough design" so that they could get attention with it 🙂

on the other hand, the digital bolex itself was not a financial success. it looked cool and vintage. it made beautiful images, even without filmbox or cineprint16 or whatever.. it's highly regarded. but people did not buy it.

[Aapo Lettinen]

28 minutes ago, Max Field said:

At the same time I don't understand where any of you are finding the money for color film when the middle-class production budgets have been shrinking more and more every decade.

most people DON'T have the money at the moment, thus there is more time to spend on theoretical camera stuff on forums like this 🙂

[Robin Phillips]

3 hours ago, Geffen Avraham said:

on the other hand, the digital bolex itself was not a financial success. it looked cool and vintage. it made beautiful images, even without filmbox or cineprint16 or whatever.. it's highly regarded. but people did not buy it.

the story of the digital bolex is a bit more complicated, especially since it started its existence as a speciality low volume product that wasnt really meant to be sold as a commodity the way it ultimately ended up trying to be sold. Though I think once it went for sale more broadly the internal only storage design turned out to be a major error. 

[Steve Switaj]

20 hours ago, Tyler Purcell said:

If the camera can't be manufactured and sold for a price round $5 - 8k, it's not worth doing, no. Obviously if we go into a world war or major recession, the project will probably be put on hold as well. 

 

Ah... and here we hit the nail on the head.

The biggest technical problem with building a new film camera is that it's not a technical problem.

OK, it's not easy, but technically it's not an insane reach.

Yes, we're used to thinking about cameras as precision machines, but from a builders point of view they are actually fairly straightforward. There are only a few areas in the movement and around the mirror box where you have to hold really tight tolerances, and even at, say, +/-.0005” or .01mm that's not all that ridiculous in a good shop with modern equipment (after all, Mitchell and Bell and Howell managed it in 1930 with manual machine tools)

The rest of the camera is just film handling. Those parts have “normal” tolerances which should be pretty easy to hold with modern machine tools.

And, while modern CNC tools are not the magic bullet some people think they are, they do dramatically change the equation. For example, if you were building a camera in 1960 you would have to find a metal foundry to cast your camera and mag housings, a complex and expensive investment you better get right on the first try. In 2024 it's totally practical (if not especially economical) to turn your CNC milling machine loose to spend 12 hours hogging your housings out of a solid block of stock. And if you have that one client who needs something custom --- well, you can do that with minimal fuss.

Many people think the electronics are the most difficult element (I was an electrical engineer before I got into VFX, so I don't share this view), but regardless, there's no argument that in 2042 it's orders of magnitude easier to run servo motors that it has ever been, not to mention that there are all sorts of COTS camera options to use as a basis for a solid video tap

In fairness, there are some parts that are going to be harder to source in the modern world, like a ground glass, but overall I think it's doable.

And the idea that a small shop can build this stuff is not just theoretical. Just off the top of my head I can think of several people like Doug Fries (Fries Engineering), Jeff Williamson (WilCam), Marty Mueller (MSM) and Bruce McNaughton (Aranda Film) who built really solid cameras with a small organization.

But, the thing is, Doug, Jeff, Marty and Bruce had something that was critical – customers that were willing to pay to have a camera built.

As Tyler and Aapo note, that's what is missing in 2024.

It's possible have an original camera built to the exact specs you want – that's how MovieCam got started – but there are vanishingly few clients who are going to pay mid-5 figures for a new film camera when there's so much solid and familiar equipment on the used market.

That's always been the real reason we're unlikely to ever see a new camera for anything but the most niche applications, Like IMAX where there is just no other option, or super 8 where the stakes are low enough for the hipster market to support the project.

Edited June 3, 2024 by Steve Switaj

[Aapo Lettinen]

One possibility very few people talk about (because it is practical and economic but NOT interesting and "media sexy" so no one would care about the option) would be to make a "frankencamera" by taking the movement and some film transport parts or the movement+mirror+groundglass block from a existing camera model and take some other parts from elsewhere, then cnc machine and 3d print the rest of the necessary stuff to stitch together a complete working film camera.

The most practical solution is to use a existing camera which has the movement+gate+mirror shutter+groundglass as a single assembly which can be separated from the camera without ruining the tolerances completely. The most basic camera model for 4perf 35mm would be Konvas1m, they are easy to get and easy to open, easy to get spare parts for and you would get the central assembly out in 5 minutes to start working on the rest of the design. Then transferring the whole movement+optical block to the new camera chassis which has pl-mount, internal motor, newly made pressure plate assembly(though it could use film path parts like pressure plate from the Konvas magazine to reduce machining work needed) and newly made magazine (if you don't want to use Mitchell magazine with it). A sprocket drive to move the film, couple or rollers, belt drive to move the sprocket and that's about it.

But as said it is NOT 2-perf, it uses mostly existing parts, it is kind of frankencamera... and people would not buy it because:

A:  it is not "sexy" enough because it is not completely made out of scratch

B:  because it uses some existing parts, people would think they could just copy the design and make their own version of it for cheaper (or just for fun) instead of buying the original camera. So there would be million copycats and absolutely no customers.

Maybe I could make something like this on free time later on but probably won't bother, I won't specifically need it even when it would be much more silent than Konvas so could likely be somewhat sync sounc capable (at least similar style like a blimped mos camera they liked to use in the 60's) . I think it is much more interesting to get my super rare Soyuz US3N back to shooting condition and shoot some b/w high contrast stuff with it, has much more entertaining value to get attention (entertaining value is all people care about if really thinking about it :D )