See3CAM_CU27 – an ultra-low light 100 fps USB 3.1 Gen 1 camera based on the SONY STARVIS IMX462 sensor that comes with superior near-infrared performance. See3CAM_CU27 is ideal for night vision applications and medical microscopes owing to its ultra-low light and near-IR sensitivity features.
Check out this video to know more about the product:
Here are the key features & benefits of See3CAM_CU27 for your reference,
High Sensitivity: Capture images even in extremely low lighting conditions (0 lux).
IR sensitivity: Capture high-quality images in near-infrared regions.
USB 3.1 Gen 1 interface: Streams MJPEG video in Full HD @ 100fps and UYVY video in Full HD @ 60fps.
Inbuilt image signal processor: Facilitates accurate color reproduction with auto exposure and auto white balance functions.
Rolling shutters are almost useless for film scanning. Who do you think would actually use that? If you’re engineering a film scanner that is keeping the film steady enough where a rolling shutter won’t matter, then you’re not going to be buying the cheapest HD camera on the market for your scanner.
It’s also not even the right size for film scanning. That camera has a 16x9 sensor, you’d be wasting 1/3rd of the resolution and scanning to around 1440x1080 and its bayer as well so the true resolution is lower.
If you get the film completely steady in the gate sure. But that’s not how the Kinograph is designed and not how most DIY scanners work. The Sasquatch is a special case, I’d say it’s atypical of DIY machines. If it’s steady in the gate in a DIY machine that means either very slow scanning, or you’re using a projector which holds the film steady. A projector-based scanner will still need a global shutter really since the film isn’t completely steady, at least not without some serious engineering to make it so. The steadiness in the commercial scanners is a result partly of the film being tensioned perfectly, in a projector gate there’s little tension so the film can move a little bit freely.
Engineering around the limitations of the sensor is what the old machines did and why they cost so much. Now you can buy a sensor that does what you want and is actually suitable for film scanning as it is.
Sasquatch is different than Kinograph, yes. But it’s the same as all the other high end scanners (Scanity aside, but I don’t consider that a particularly good scanner, just an expensive one). Director, Xena Pin registered, Arriscan, and going further back Northlight & Imagica too, in that it’s intermittent motion. In fact, a projector is intermittent motion by design, so if a DIY scanner is being used as the mechanics were originally designed, a rolling shutter would be fine, if fast enough for the speed being used.
I would argue that an intermittent motion scanner is actually easier to build than one that runs continuously, because you’re less dependent upon timing and coordinating things to ensure you get the exposure and lighting at the right moment. It’s a very logical, sequential system, and that also makes programming the control software simpler.
You make good points about the aspect ratio of this camera, and about the spamminess of these posts in general, but there is plenty of room for rolling shutter cameras in the DIY world. They tend to be a lot less expensive than many global shutter cameras, too.
In theory it would be fine for a projector, but in practise it isn’t. A projector isn’t intermittent - it’s continuous-motion but the film in the gate is held steady enough for projection so the film is not quite as steady as you may think. Most other machines that you can convert into a scanner including old telecines would not have been designed with keeping the film completely steady in the gate in the first place.
If you don’t need a global shutter for your particular scanner, then sure you can buy a rolling-shutter camera. That will be the exception rather than the norm though, the norm would be you would either need to engineer your film transport to be perfectly steady or you just use a global shutter.
It’s also one more thing that can go wrong. Even a finely tuned machine that does what you need could develop an issue over time, maybe you increased the speed of the machine or made some other “upgrade” or just general wear and tear. You wouldn’t try making a DIY scanner with a line-sensor (like the Scanity) as that would take serious engineering to get it right. My point is that there’s no point in spending a fortune engineering your film transport and gate to get the film completely steady when you can just buy a camera that does what you need and will solve that problem.
Not all projectors are made equal. For example, the Canon S-400 S-8/8 moves the film with a claw mechanism, and the claw acts for less than 1/3 of the motor turn, the other 2/3s the film is still. I’ll say that’s a pretty good example of intermittent motion mechanism, even when it can run at 24 fps. I use the gate mechanism of that very projector to do a longer stop-motion, no change on the mechanics.
This is simply untrue. A projector is the very definition of an intermittent motion system. As you yourself say, the way a projector works is by momentarily holding the film in the gate. Without that still image, persistence of vision wouldn’t work, and we’d see streaks, not a moving image. And if it was as unsteady as you say, then in projection you’d see the frame moving (not from frame to frame, but while a single frame is up), because the projection will dramatically magnify any defects. This just isn’t a problem.
The time it takes to make an exposure is measured in single or double digit milliseconds (depending on your optics and light source of course). In the case of our camera, our typical exposure time is 8ms, which is 1/125th of a second. That’s significantly shorter than the time the film is held still in the gate during projection.
Rolling shutters become a problem if you’re trying to capture something that’s actively in motion. they will work fine in a projector-based system if the exposures are taken during the part of cycle where the film is held still. It is definitely held still enough that a rolling shutter will work, and even if there is some creep of the frame through the gate during the still phase of the cycle, if your exposure times are short it’s not a problem. And that’s not hard to achieve with inexpensive high-powered LEDs.
Where projectors vary is in the stability of the image from frame to frame – some will have better edge guiding and pressure plates that will hold the film steadier. Others will not, and from one frame to the next you might see some variation. But that’s going to be the case whether you have a rolling or global shutter and it’s a matter of post-processing that you’d likely be doing anyway, to make things stable later.
My whole point is not about whether you need a global shutter for a particular purpose, that’s missing the point. The point is that you can buy a camera that does what you need. If you need HDR scanning you can now buy cameras with dual-gain HDR. The Arriscan uses a rolling shutter, but Arri spent a fortune engineering the scanner to get the film steady enough to use those sensors. If you need 6K scanning you’re not going to do it the way Arri does it with a piezo-shift motor and a 3K sensor, you would just buy a 6K camera. You’re certainly not going to build a DIY scanner using line-array CCD imagers like the Scanity (which you seem to hate haha), you would just buy an area camera instead. If you need 24fps you can buy the cameras that are fast enough for that, etc.
The thing that people don’t appreciate (you would since you’ve had tons of experience) is that the old imagers in scanners were really limited when you compare with what is available now. You don’t need to spend a fortune on engineering something a camera can now do/solve for you.
I think you’re overstating the engineering work that goes into holding the film steady. I’ll grant you that Arri probably did sink a ton of money into their design, but they’re Arri. Of course they’re going to do that.
In the case of the Arriscan, Northlight and Imagica scanners, the gate is mechanically pin-registered. Yes, there’s engineering work in designing that, but you’re designing something from scratch, so there’s always going to be some amount of work. You don’t really need to do anything special to hold the film still here. If you look at the imagica and northlight gates, they’re a combination pressure plate and gate, with a registration pin. In those two scanners, the imager takes seconds to pass over the film so any movement would be catastrophic. Yet the Northlight II can resolve to about 160lp/mm. There’s really nothing special about either gate, and we’ve run film through our Northlight with the reg pin removed, and it’s been rock steady. It takes 4.5 seconds to do a 4k exposure on that machine.
The lasergraphics director 10k is sprocketless and has no registration pin, and uses piezo movement like the arri to get 10k. The film is held in place by the pressure plate and tension. And the images from it are outstanding.
A projector is going to achieve the same effect if you don’t modify it (in fact, on the Northlight, the gate assembly is referred to by Filmlight as “the projector” because the mechanism is so similar). It may not have a registration pin, but the pressure plate and a spring-loaded edge guide will certainly hold the film steady enough for a rolling shutter. There is no extra engineering work here. This is a non-issue if the motion is intermittent and the exposure is fast.
One more example: My Pentax DSLR has a piezo-actuated pixel shift sensor. I’ve used it for shots on a tripod (certainly less steady than a projector gate, what with wind and all), and the results are pretty spectacular.
Well it also depends on your formats I suppose. An Arri is a dual-format scanner - 35mm + 16mm. If it was dead easy to engineer there would be no reason not to make it tri-format or quad-format like the Lasergraphics Director. Projectors are, mostly, single-gauge. Dual-format exists of course (35mm + 70mm) but those are outrageously expensive second-hand I can only imagine how much they cost when new, I think the base cost for a new Cinemec 35mm projector is around €4-5K. $30K for a 35mm/70mm Simplex projector right now (and Newton has a pair of them priced at $35K for the pair).
The original Arri was designed like that, the current models let you remove the pin and have a completely sprocketless transport. The others you mention are quite dated now (so are Arris if it’s an old model), the Northlights and Imagicas use line-sensors and they move the sensor over the film while the film is held steady. It definitely takes engineering to do it right, the Scanity I believe does it the other way so too I think do the Spirits - where you fix the line-sensor in place and move the film over/past it. Again, there is no way in 2021 that you are building a DIY scanner using a line sensor - and certainly not like a Filmlight or Imagica, IF you were doing that you’d have to do it the way normal telecines etc did it with a fixed sensor and moving the film past it.
They’re all only dual-format 35mm/16mm which I feel is worth repeating. As soon as you add 8mm or 70mm the price for an equivalent quality machine escalates like anything. Even just adding 16mm to a base model 35mm scanner can be expensive as well. The Scanstation 65 is effectively double the price of the normal tri-format Scansation for example, and the base price on it is $200K (that’s without 35/16/8mm).
Yeah but it has a global shutter camera. Unless they’ve changed new ones to the Pregius S, the current models come with the Pregius IMX342 imager. Pregius S would theoretically let them offer single-flash HDR, although I say “theoretically” because I’m not sure whether you can get both piezo and dual-gain HDR in the same camera model and nor am I sure whether that would actually improve scanning speeds for it compared to 2-flash HDR.
I never said it was “dead easy” – I said that if you’re building a scanner from scratch, you’re already doing the basic, required engineering work. And if you’re Arri, you’re probably over-enginering it. Because, Arri.
My point is that you do not need anything especially complicated to hold film steady enough to get a good image with a rolling shutter, but you do need to hold it still for a fraction of a second. Latham loops, edge guiding, pressure plates, and registration pins are things that have been in cameras and projectors for decades, and are well understood. Motors that hold the film steady without a loop (such as we’re doing) do the same thing. The point is that you’re making it sound like the Arri is doing something exotic or unique to hold the film steady, and it’s not.
To make a rolling shutter (or a pixelshift) camera work with a scanner, the film needs to be held still. That’s it. You do that by not moving it. It’s really, truly, not that complicated to hold the film still for a fraction of a second.
It doesn’t matter if they’re dated, they illustrate my point perfectly. Take the registration pin out of an Imagica or Northlight scanner, and all that’s holding the film in place is the pressure plate. Both are using projector-style latham loops. And guess what? The scans are fine. And yes, they’re line sensors, but I brought that up specifically because it illustrates my point. Nobody is suggesting using a line array for scanning in a newly designed machine. But a line array like that in the Imagica or Northlight is effectively the same thing as a rolling shutter, only many orders of magnitude slower, which would result in a much worse image if what you’re saying about the gate requiring extra work was true, which it isn’t.
A rolling shutter compiles the image as it scans and it does so in under 10ms given enough light. A line array in those two scanners takes 2-6 seconds to do the same thing. And yet, they produce very nice images, even with the registration pin removed.
By the way, multi-gauge thing you keep bringing up has nothing to do with what we’re talking about here. Those decisions are driven by a completely different set of issues, and I bet marketing and budgets have as much, if not more, say as engineering does in those choices. Of course it’s more complicated to make a multi-format machines, but it’s not that much more complicated, especially with modern methods of registering frames (in software).
We’re doing it right now and will be able to scan 28mm through 70mm in a dozen variations (28mm, 35m 2p, 3p, 4p, 5p, 8p, over/under 3D, 65/5, 70/5, 70mm side-by-side 3D, 70/10, 70/15), plus a couple formats I can’t talk about right now because we’re under NDA. To make that work, you pop in a gate that costs about $100 to manufacture that’s specific to that format. In some cases the same gate works for two or three formats because the bulk of the work here is in software.
To design those gates took me (not a trained engineer) a couple days and they work great. The only reason we’re not doing 16mm or 8mm in this machine is that we are retrofitting an old chassis and don’t have the physical space for the optical assembly that would be necessary to make that happen. Given enough room, it wouldn’t be a particularly big deal to implement that.
That’s actually a VERY good way of putting it! But that’s if you do it the way those scanners do it, the way the traditional telecines work (and Scanity, Spirits) is by pulling the film past it, which isn’t the same thing as you can have movement and flex in the film itself causing distortion.
Well you say that, but I just spoke to someone considering buying a Blackmagic Cintel as a 35/16 scanner. What I pointed out is that it’s designed as a 35mm scanner, unless your intention is low-end access scanning or something like that it’s pretty useless for 16mm. It may not seem relevant but it is, and that remains the case if you’re building something.
The northlight moves the film past the line array. The film is held in place by a registration pin (which you can easily remove, and is recommended you do so for shrunken film in the Northlight manual) and a pressure plate. With the registration pin removed, it is effectively the same thing as a projector. There is a loop before and after the gate and the film is held in place with a pressure plate only.
The Imagica holds the film in place and moves the sensor past it. The gate mechanism is really nothing special here - it’s physically a bit different than, but similar in principal to the Northlight: registration pins and pressure plates to hold the film steady.
This is my whole point: There really doesn’t need to be much to hold the film steady for a line array that takes seconds to move past the film, let alone a rolling shutter, which does the same basic thing, for a few milliseconds. Unless the film is physically moving in the gate, which it is not, unless you have a particularly bad transport or malformed loop, you’re good to go with a rolling shutter.
The Golden Eye, Scanity, Shadow and Spirit have the line array fixed in place and the film moves past it. This introduces a new complication: splice bumps, where the film warps at splices. Both the Golden Eye and the Scanity “get around” this by bundling software the does a dewarping pass on the splice bumps. A pretty lame solution, if you ask me.
Huh? You brought this into a conversation that was about whether a rolling shutter is a good choice for a sensor in a film scanner. Multi-gauge scanning is an entirely different conversation but really has nothing to do with sensor style choice.
