Why is there a mechanical shutter in a digital camera?

Question

I understand that in a camera the diameter of shutter opening — the aperture — controls how much light comes in, and this affects the resulting exposure.

But I don't understand why, in a digital camera, the shutter needs to close and open when taking a picture or making continuous shots. Can't the limitations on framerate (frames per second) or fastest shutter speed (for example,¹/₃₂₀₀) just be a property of the electronic sensor?

I ask this because my new new camera can't make more than one shot per second in continuous mode. 1 FPS is ridiculous in a 2011 camera, don't you think? (It can do 30fps for HD video.)

Answer 1

Video does not use the mechanical shutter. The shutter remains open. Something called rolling shutter is used to scan the sensor. This takes time, and is not instantaneous like a 1/4000th exposure with the mechanical shutter where all sites on the sensor record at the same instant. At 30 fps, each frame is effectively 1/30th of a second, so this type of capture is really not the same as when you click a single frame using the mechanical shutter. You also get reduced resolution, as the camera can only process so many pixels and write them to media.

So if you did away with your mechanical shutter you'd end up with lower resolution and more blur, not to mention warping and other effects explained in the reference on rolling shutter.

Answer 2

Probably the reason for using mechanical shutters is that their disadvantages are easiest to live with; competing technologies are not clearly superior.

There have been electronic shutters used for higher speeds on some older Nikon DSLRs, like D1 or D70. On these cameras, grid-like patterns were reported to sometimes appear on plain tonal areas with shutter speeds that used electronic shutter.

One major problem is that electronic shutter can only be implemented on CCD sensor, while CMOS is preferred on new DSLRs. CCD is commonplace in compacts, and they often use electronic shutters.

Another problem is that shutter circuitry takes room on sensor which could instead be used for light sensitive components.

I suspect you only had the commonly used focal plane shutter in mind; leaf shutter is another mechanical design for a shutter. Its main benefits are quietness and ability to sync flash at any speed, because the shutter always fully opens. But leaf shutter either needs to be located just where aperture diaphragm is (i.e. in every lens), or needs specially designed lenses that have a nodal point at certain distance between lens and image plane. The first option is expensive, used in many medium format systems; the other is restrictive for lens design, but has been used in some old SLR models (e.g. Topcon Auto 100).

On your Canon SX30IS, the shutter (which can go at 1/3200s) is most likely not the limiting factor for burst speed. The speed is more likely inhibited by data bandwidth - even when you shoot low-res, camera still reads all the 14MP from sensor to give maximum image quality. In video, image quality is less important than frame rate, so the camera reads only selected rows and columns off the sensor.

According to specifications, turning off LCD should help you reach 1.3 fps. Or if you'd like to trade image quality for burst speed, just shoot video and extract frames later.

Answer 3

What rfusca said. A D3 and 1D4 shoots 9-10fps with a mechanical full frame shutter. The fullframe shutter can easily cycle in 1/125s second so that's not the issue for framerate.

Processing a 10+ mp dump from a sensor into a jpeg and then writing it down to a card isn't a menial task to accomplish.

Taking video is a wholly different story than high speed photography because the data is streamed and the hardware and codec is matched to the task at hand.

1080p is about 2 megapixels, and the compression scheme (aside from already being highly optimized, lowish rate on a consumer camera) may even be done partly in hardware.

What you describe is more like a RED camera, which leans more towards high speed camera with its 10+ megapixel cmos sensor that captures compressed "raw" which is still in essence a lossy X>1:1 ratio jpeg like scheme.

What this says is that there's two problems, one is interdevice/device to storage bandwidth, and the second one is post processing feasibility, which Red sort of addressed by processing the raw in hardware before storing on a media to be edited later.

A Red is gigantic and has an active cooling system with heatsink & fans. This is not your typical milliwatt embedded chip in something one person can carry with two hands.