The person coming through the door is a black silhouette in the recording, and the car park behind them a bright white blotch. At building entrances, glass-fronted lobbies, loading docks with open doors and tunnel exits, a standard camera gives this image, because the sunny outdoor area and the face in shadow appear in the same frame. WDR (Wide Dynamic Range) exists for these scenes; but the dB value on the datasheet often doesn’t tell you the result on site. Below we explain how WDR works, how it differs from BLC and HLC, how to read the dB value, and the selection and test method for a backlit point.
The backlight problem
A scene’s dynamic range is the ratio between the light levels in its brightest and darkest areas. The difference between a sunny car park and the shadow inside the door easily exceeds the range a sensor can record in a single exposure. If the camera sets exposure by average brightness, one of the two ends is lost:
- If exposure is set for the bright area, faces inside go dark.
- If exposure is set for the dark area, the outside clips to white, and vehicles or people there become invisible.
The problem changes during the day. A west-facing entrance gets the sun directly opposite it in the afternoon; the same camera looks fine in the morning. In winter, the low sun deepens the backlight. A common mistake is doing the survey and acceptance test in the morning and never seeing the worst hour. The survey should take the worst hour and season into account.
True WDR, digital WDR, BLC and HLC
In marketing copy, these concepts are used interchangeably. Technically, they’re different mechanisms.
| Method | How it works | Strength | Limit |
|---|---|---|---|
| True (multi-exposure) WDR | Short and long exposures are combined | Detail at both ends | Motion artifacts, LED flicker |
| Sensor-based HDR | Dual gain or split pixel | Fewer motion artifacts | Depends on the sensor; not on every model |
| Digital WDR (tone mapping) | The shadows of a single exposure are lifted in software | Cheap, few artifacts | Blown-out areas don’t come back; noise increases |
| BLC | Exposure is set for a selected area | The person in the foreground is visible | The background is lost completely |
| HLC | Very bright spots are masked | Suppresses headlight glare | No information in the masked area |
Multi-exposure WDR takes two or more exposures of different durations for the same frame: the short exposure captures the bright area and the long exposure the shadow; the image processor combines them into a single frame. Some sensors reduce the time difference between exposures by using line-interleaved exposure (such as the method Sony calls DOL-HDR).
Digital WDR (“D-WDR”) redistributes the tones of a single exposure. It can’t recover an area the sensor has already clipped to white; it only lifts the shadows and makes the noise there more visible. It may be enough in a medium-contrast scene. It isn’t enough for real backlight.
BLC (Backlight Compensation) makes the area selected by the user the metering area. It can be used to see the person coming through the door, but it sacrifices the outdoor area completely. HLC (Highlight Compensation) was developed specifically against vehicle headlights at night; it suppresses bright areas above a certain threshold and helps detail around the headlights become visible. Reliable plate reading, however, requires a short shutter and IR lighting suited to plates more than HLC; other factors affecting accuracy are in the LPR license plate recognition accuracy article.
How to interpret dB values
Dynamic range is expressed in decibels: DR (dB) = 20 × log10(highest signal / lowest distinguishable signal). With this definition, 120 dB is a ratio of one million to one; about 6 dB equals one stop (twice the light).
The problem is that manufacturers don’t measure this value with the same method. Some datasheets calculate a theoretical value from the duration ratio of the combined exposures, some measure it on their own test setup, and most don’t state the method at all. Standards such as EMVA 1288 define how sensor dynamic range should be measured; such a reference is rarely found on surveillance camera datasheets. In practice, this means:
- Ranking different brands by their dB values is misleading.
- A value such as “120 dB” is a sign that the camera has true WDR support; it isn’t proof of performance.
- Models without a stated value, or labeled “D-WDR,” aren’t candidates for a backlit point.
Write a performance criterion into the specification, not a dB value: “At the entrance door, at the worst sun hour, the face of the person inside the door will be visible at recognition level and vehicles in the outdoor area at observation level.” This criterion can be verified on site. A dB value can’t. The EMVA 1288 standard can be used as a reference for the measurement approach.
Side effects of WDR
WDR isn’t free. Left on at every point, it creates its own problems:
- Motion artifacts: Because exposures are taken at different moments, double contours or colored fringes can appear at the edges of fast-moving objects. The long exposure component also increases motion blur; the shutter relationship is in the camera exposure settings article.
- LED flicker: A short exposure may fail to capture LED signs, traffic lights or LED headlights driven by pulse-width modulation; flashing or banding appears in the image.
- Flat, gray image: Tone mapping reduces local contrast. With a high WDR level, the image looks washed out and artificial.
- Noise and bitrate: Lifting the shadows raises the noise there; encoder efficiency drops and bitrate rises.
- Frame rate limit: On some models, maximum frame rate or resolution drops when WDR is enabled. This constraint is often in a datasheet footnote; read it before testing.
A selection table by scenario
The table gives the starting choice for typical backlit scenes. The site test makes the final decision.
| Scene | Recommended approach | Watch out for |
|---|---|---|
| Glass-fronted lobby, looking out from inside | True WDR + a change of position | WDR alone is often not enough |
| Loading dock with open doors | True WDR, lighting inside the dock | Artifact test at forklift speed |
| Car park entrance, facing headlights at night | HLC or a dedicated LPR camera | Consider a separate camera for plates |
| Tunnel / indoor car park exit | True WDR, a position angled to the exit | Contrast that changes during the day |
| Office with windows, corridor | Digital WDR or off | Unnecessary WDR produces noise |
Situations that can be solved with mounting
Backlight arises from the camera and the bright source being on the same axis. Changing the axis is often the most effective solution, and cheaper than changing the camera:
- Turn the viewing direction around: Instead of watching the door from inside looking out, put the camera on the outside and have it look at the person coming in. With a camera looking from outside, the face is lit by outdoor light.
- Choose an angled position: A camera looking at the door from the side at a slight angle, rather than straight on, reduces the bright opening to a small part of the frame. Angle limits for face visibility are calculated in the camera mounting height and angle article.
- Increase indoor lighting: Lighting at face level in the entrance hall lowers the contrast ratio; it’s often cheaper than changing the camera.
- Take direct sun out of the frame: Choose a tilt that doesn’t include the sky or glass surfaces reflecting the sun.
- Vestibule or canopy: If architecturally possible, a double-door entrance or a canopy reduces the brightness of the door opening.
Field testing
In enterprise camera systems, WDR performance is determined by the site, not by laboratory values. Apply these steps in a pilot test or at the acceptance stage:
- Determine the hour when backlight is strongest (the hour when the sun is closest to the entrance axis) and run the test at that hour.
- If possible, mount the candidate cameras side by side, at the same point and the same angle, and record at the same time.
- Have a person pass through the door at normal speed; review the recording frame by frame and assess the detail on the face and in the outdoor area.
- Repeat the same test with WDR off, low and high; compare motion artifacts and noise.
- Record day and night bitrate values; reflect the increase with WDR enabled in the storage calculation.
- In the night test, look at how WDR behaves with headlight glare and together with IR; assess the night image separately against the IR range and night image criteria.
Backlit points are usually a facility’s most critical entrances. They need the right position as much as the right camera. We handle the survey and testing of these points as part of IP camera system design work; for façade and perimeter points, you can also see the outdoor security cameras page.
Frequently asked questions
Is 120 dB WDR always better than 100 dB WDR?
No. Manufacturers measure dynamic range with different methods or calculate it theoretically; the dB values of two brands can’t be compared directly. A side-by-side test in the same scene at the same time is a more reliable comparison than the dB value.
Is it right to leave WDR on for every camera?
No. In a low-contrast indoor scene, WDR flattens the image, can increase noise and produces motion artifacts in multi-exposure mode. Turn WDR on only at points where very bright and very dark areas appear together in the same frame.
What is the difference between BLC and WDR?
BLC sets exposure for a selected area; the person in the foreground is lit, and the background blows out completely. WDR tries to preserve detail in both bright and dark areas of the same frame. BLC is a software setting; true WDR requires sensor and image processor support.
Can a backlight problem be solved without changing the camera?
Often partly. Moving the camera to look inward rather than toward the bright exit, increasing indoor lighting or adding a canopy in front of the door lowers the contrast ratio. These measures give more lasting results than WDR settings.