A digital camera captures an image by converting light into electrical signals, then processing those signals into a file your screen can display. Every digital photo, whether taken on a smartphone or a professional camera body, follows the same core sequence: light passes through a lens, hits a sensor covered in millions of tiny light-detecting sites, and each site generates a small electrical charge proportional to how much light struck it. That charge is measured, converted into a number, and assembled into the grid of colored pixels you see on screen.
How the Sensor Captures Light
The image sensor is the heart of any digital camera. It’s a chip covered in a grid of photodiodes, each one corresponding to a single pixel. When a photon of light hits one of these photodiodes, it knocks an electron loose inside the semiconductor material through a process called the internal photoelectric effect. The brighter the light hitting a given photodiode, the more electrons accumulate there during the exposure. A dim shadow might free just a handful of electrons, while a bright highlight generates thousands.
CMOS sensors dominate the market today, accounting for about 93% of image sensor revenue in 2025. Older CCD sensors still appear in specialized fields like astronomy and medical imaging, where their extremely low electronic noise is worth the tradeoff of slower readout speeds. But in phones, consumer cameras, and professional bodies alike, CMOS is the standard.
How a Single Sensor Sees Color
Each photodiode on its own is colorblind. It can measure how much light hits it, but not what color that light is. To capture color, manufacturers place a mosaic of tiny colored filters over the sensor, one filter per pixel. The most common arrangement is the Bayer pattern: in every 2×2 block of pixels, two are covered with green filters, one with red, and one with blue.
Green gets double the coverage because human vision is most sensitive to green light, and green carries most of the brightness information our eyes rely on. After the image is captured, the camera’s processor fills in the missing color values for each pixel by looking at its neighbors. If a pixel only recorded green, the processor estimates its red and blue values from the nearest red and blue pixels around it. This interpolation process, called demosaicing, produces a full red-green-blue value for every single pixel. Simple methods just average nearby values, but modern cameras use smarter algorithms that exploit the natural correlation between color channels to reduce artifacts like false color along sharp edges.
From Electrical Charge to Digital Numbers
Once the exposure ends, each photodiode holds an electrical charge. The sensor reads out these charges row by row, converting each one into a voltage. An analog-to-digital converter (ADC) then measures that voltage and assigns it a number. In most cameras, this converter works at 12 or 14 bits of precision, meaning it can distinguish between 4,096 or 16,384 different brightness levels per pixel.
This is where the camera makes a critical fork. If you’re shooting in RAW mode, those 12- or 14-bit values are saved directly to your memory card with minimal processing. If you’re shooting JPEG, the camera’s processor steps in: it applies white balance correction, adjusts the tone curve, reduces noise, sharpens the image, and then compresses everything down to 8 bits per color channel. That means each channel gets only 256 brightness levels instead of thousands. The JPEG looks great straight out of the camera, but you’ve permanently discarded a lot of tonal information. RAW files are larger and require editing software, but they preserve far more headroom for adjusting exposure, recovering highlights, and correcting white balance after the fact.
What ISO Actually Does
A common misconception is that raising your camera’s ISO makes the sensor more sensitive to light. It doesn’t. A digital sensor has one fixed sensitivity, determined by how efficiently its photodiodes convert photons into electrons. When you increase ISO, the camera applies a gain (amplification) to the signal after the sensor has already captured it. It’s like turning up the volume on a recording rather than moving the microphone closer to the sound.
At ISO 100, the full range of the sensor’s output maps to the full range of the image file. At ISO 800, the camera amplifies the signal so that a much smaller amount of collected light is stretched to fill that same range. The pixel reports “maximum brightness” at a fraction of the light it would need at ISO 100. The downside is that any electronic noise present in the signal gets amplified right along with it. That’s why high-ISO photos look grainy: you’re not capturing worse light, you’re just turning up the volume on a weaker signal.
How the Shutter Controls Exposure
The shutter determines how long light reaches the sensor. In a traditional DSLR, a pair of mechanical curtains physically slides across the sensor to start and stop the exposure. Many mirrorless cameras and nearly all smartphones use an electronic shutter instead, where the sensor’s pixels are simply switched on and off.
Electronic shutters come in two types. A global shutter activates and deactivates every pixel row simultaneously, so the entire frame is exposed at the same instant. A rolling shutter, which is more common, starts each row’s exposure slightly after the one above it. This slight time delay between rows is usually invisible, but it can cause distinctive distortion when you photograph fast-moving subjects or pan the camera quickly. A spinning propeller might appear bent, or a passing car might look slanted. The faster the subject moves relative to the readout speed, the more pronounced these artifacts become.
The Light Path: DSLR vs. Mirrorless
In a DSLR, light enters through the lens and hits a mechanical mirror angled at 45 degrees. That mirror bounces the light upward into a pentaprism or pentamirror, which flips the image right-side-up and sends it out through the optical viewfinder. You’re seeing actual light from the scene, with zero electronic delay. When you press the shutter button, the mirror flips up out of the way, the shutter opens, and light hits the sensor for the duration of the exposure.
A mirrorless camera eliminates the mirror and prism entirely. Light passes straight from the lens to the sensor at all times. The sensor’s live feed is displayed on a small electronic screen inside the viewfinder or on the rear LCD. This design makes the camera body significantly thinner and lighter, since there’s no mirror box or prism housing to accommodate. It also means the camera can use the imaging sensor itself for autofocus, rather than relying on a separate dedicated module. The tradeoff is that the electronic viewfinder introduces a tiny processing delay and consumes battery power in a way an optical viewfinder never does.
How Smartphones Push Beyond Hardware Limits
Phone cameras face a fundamental constraint: their sensors are physically small, which limits how much light each pixel can collect. To compensate, modern smartphones lean heavily on computational photography. Rather than capturing a single frame, the phone often fires off a rapid burst of exposures at different settings, then merges them together. Some frames might be optimized for shadow detail, others for highlights, producing a final image with a dynamic range no single exposure from that sensor could achieve.
This same multi-frame approach powers features like night mode, where the phone captures and aligns dozens of dim frames to build up a usable image in near-darkness. Portrait mode uses depth estimation (sometimes from multiple camera modules, sometimes from a single lens using computational models) to separate the subject from the background and apply artificial blur. The processing happens so quickly that most users never realize their “single photo” is actually a composite assembled from multiple captures and shaped by sophisticated image processing running on the phone’s chip.