Why do DSLRs have shutters?

Since DSLRs have digital chips why do they need shutters? Couldn’t the chip just be activated to start capture light and then deactivated when it’s done? Is it so they’d go click like a film camera when you push the button? Is it to keep the digital sensors from being exposed to intense, direct light for extended periods? Is it because that the sixth floor guys/gals share the same lab with the SLR design dept while the 5th floor people share the same lab with the camcorder/CCTV video camera design dept?

There’s a lot of online speculation as to the reason for the shutters. We wanted a definitive answer so we went to camera manufacturer Eastman Kodak where Keith Wetzel provided the following answer:

DSLRs have shutters to prevent motion artifacts caused by movement in the scene during CCD or CMOS image sensor readout. Since several different sensor technologies are employed in DSLRs today, the technical details vary depending upon the sensor technology used.

For example, some DSLR manufacturers use interlaced interline CCD technology which basically reads the image out into two pieces: an even field (consisting of even lines) and an odd field (odd lines). By using a shutter, both even lines and odd lines are captured simultaneously and then readout "in the dark" after the shutter is closed.

Other DSLR manufactures use full-frame technology in which the whole surface of the CCD is photosensitive. Full-Frame CCDs need to be readout in the dark regardless of whether there is motion in the scene or not. In a full-frame CCD, the image is captured then transferred line through the image area line by line to a single line readout structure. Since the image area continually captures light, contamination of the image occurs during readout as each line is transferred through the image area. Reading a full-frame sensor out after the shutter is closed mitigates this effect.

Other DSLR manufacturers use CMOS image sensors and typically CMOS sensors use a "rolling shutter" capture architecture. For example, Kodak uses a "rolling shutter" capture architecture in its CMOS image sensor and one line is captured while another is being readout. The readout is typically destructive and clears the pixel. As long as there is no motion, this "rolling shutter" produces no image artifact. But if there is motion, then one line is capturing a portion of the image at a different time than the other lines. If an object in the scene moves, this produces a smearing in the image or distortion in the moving object which is objectionable. To allow image capture without motion artifacts, a shutter is used to allow full image capture, then readout occurs after the shutter is closed (readout occurs in the dark).