Thermal Imaging Explained
Thermal NETD Explained: Why Sensitivity Matters Beyond Resolution
Publisher: Xhunter Australia | Reviewed: 27 October 2026 | Reading Time: 10 Mins
On a cool, clear night, plenty of thermal devices can produce a convincing image. The harder test is a damp paddock after rain, when the ground, scrub and animal you're watching sit close together in temperature. NETD helps describe how well the sensor can separate those small temperature differences—but it doesn't replace resolution, good optics or sensible expectations.
What Does NETD Actually Mean?
NETD stands for Noise Equivalent Temperature Difference. In plain English, it describes the smallest temperature difference a thermal detector can distinguish from its own noise under stated test conditions.
The figure is usually shown in millikelvins, written as mK. A lower published number generally indicates that the detector can separate smaller temperature differences. That's useful when the scene has weak thermal contrast: wet vegetation, foggy air, an overcast evening or an animal partly blending into a background that has cooled to a similar temperature.
Think of it as the sensor's ability to separate subtle shades rather than count fine shapes. A sensitive detector may show more tonal separation in a low-contrast scene. It doesn't automatically place more pixels on the target, make the lens sharper or guarantee that you'll identify what you're looking at.
What Is a Millikelvin?
One kelvin represents the same size temperature step as one degree Celsius. A millikelvin is one thousandth of a kelvin, so 20 mK equals 0.020 K and 18 mK equals 0.018 K.
That doesn't mean a thermal device is a laboratory thermometer accurate to 0.018°C. NETD is a signal-to-noise measurement. It describes how small a thermal difference the detector can separate from noise in a controlled test, not the accuracy of an on-screen temperature reading.
Lower is generally better—not infinitely better
A lower NETD figure is a positive specification when the values have been measured on a comparable basis. But a difference of a few millikelvins on a product page doesn't guarantee that one complete device will look better in every scene. Optics, processing, calibration, display and weather all have a say.
Why NETD Matters More on Difficult Nights
Picture two evenings on the same farm. On the first, the ground has cooled quickly and a warm animal stands against it. The temperature difference is strong, so the subject appears clearly even if the detector isn't especially sensitive.
On the second evening, rain has passed through and everything is holding a similar temperature. The image looks flatter. This is where lower thermal noise can help preserve small differences between the subject, grass, branches and ground.
Humidity, fog and rain can also absorb or scatter infrared energy between the scene and the device. NETD doesn't remove that atmospheric limitation. It may help the detector work with weaker contrast, but it can't restore information that never reaches the lens.
The same caution applies to distance. As a subject occupies fewer pixels and the atmosphere has more space to interfere, sensitivity is only one part of the problem. NETD isn't a distance rating.
NETD and Resolution Measure Different Things
Resolution tells you how many detector pixels are available to describe spatial detail. NETD tells you how small a temperature difference can be separated from noise. One is about where detail can be drawn; the other is about how subtle the thermal contrast can be.
| Specification | What it describes | What it does not guarantee |
|---|---|---|
| Sensor resolution | Number of thermal detector pixels, such as 256×192, 384×288 or 640×512 | Sensitivity, lens quality or identification distance |
| NETD | Ability to separate small temperature differences from detector noise | Spatial detail, detection range or temperature accuracy |
| Pixel pitch | Physical spacing of detector pixels, often shown in micrometres | Overall image quality by itself |
| Lens and field of view | How much of the scene reaches the detector and how wide the view appears | Detector sensitivity or native pixel count |
| Display resolution | Pixels available on the eyepiece screen | Extra thermal data beyond what the detector captured |
A 640-class detector with average sensitivity can preserve more spatial detail than a 256-class detector, especially when the subject is small in the frame. A sensitive 256-class detector may still present useful contrast in difficult conditions. Neither number cancels the other.
Our 256×192 thermal sensor guide looks more closely at that resolution class and the trade-offs buyers should expect.
The Thermal Image Comes From a Complete System
It's tempting to line up two NETD figures and declare a winner. Real devices aren't that tidy.
The lens controls what reaches the detector
Lens material, aperture, focal length and transmission affect how much thermal energy reaches the sensor. Field of view changes how large a subject appears in the frame. A good detector can't recover detail that poor optics never delivered.
Processing changes the presentation
Noise reduction, sharpening, contrast mapping and image modes can make an image look cleaner or more dramatic. Heavy processing can also smear fine texture or create hard edges that weren't present in the raw data. A punchy showroom image isn't automatically the most informative field image.
Calibration keeps the image usable
Thermal detectors need Non-Uniformity Correction, or NUC, to compensate for variation across the array. Calibration behaviour affects consistency. On the TNC4K219R, a brief one-to-five-second image pause during NUC is expected rather than evidence that sensitivity has failed.
The display is the final link
Brightness, contrast, eyepiece focus and display resolution affect what the user can see. A high-resolution display can present processed imagery smoothly, but it cannot turn a lower-resolution detector into native high-resolution thermal.
Using the TNC4K219R's ≤18 mK Figure Correctly
The TNC4K219R thermal channel uses a 256×192, 12 μm detector with a published NETD of ≤18 mK at 25°C and a 19 mm objective. The “≤” symbol matters: it presents a maximum threshold under the stated test condition, not a promise that every scene will behave like a controlled 25°C measurement.
The ≤18 mK rating suggests strong thermal sensitivity on paper for its class. That's useful context when a wet paddock or humid night offers less contrast than a dry, clear evening. It doesn't make the detector 4K, add pixels or guarantee the same recognition detail as a higher-resolution core.
The product's 4K label belongs to its separate 3840×2160 digital day/night channel. The thermal channel remains 256×192. Our ThermNight 4K thermal versus digital guide explains why those two resolutions shouldn't be blended into one claim.
Thermal Image Super Resolution and image processing may improve presentation, but they don't replace native detector samples. The honest reading is simple: the TNC4K219R combines a sensitive published thermal specification with a modest native thermal pixel count. Sensitivity is a strength; fine thermal detail remains the trade-off.
How to Compare NETD Between Thermal Devices
Use NETD as one line in a comparison, not the headline result. Check whether the manufacturer states a test temperature or other conditions, and be cautious when two brands don't explain how their figures were measured.
Then compare the parts that change the actual viewing experience:
- native detector resolution and pixel pitch;
- lens focal length, aperture and field of view;
- base magnification and how much digital zoom will be needed;
- image processing, palettes and calibration behaviour;
- display quality and eye relief;
- weather sealing, battery system, controls and recording;
- the buyer's likely distances, terrain and weather; and
- current Australian support and total purchase cost.
Also compare the intended job. A broad observation device, a compact handheld and an integrated multispectral unit solve different problems. The lowest NETD number doesn't automatically make one of them the best tool.
What NETD Cannot Tell You
NETD cannot provide a reliable animal-identification distance. It cannot tell you whether a target fills enough pixels, whether the lens is focused, whether branches block the view or whether the operator can make a responsible identification.
It also isn't a guarantee of better performance through glass or solid cover. Long-wave thermal devices generally don't see through ordinary glass as if it were transparent, and no sensitivity figure makes a solid obstacle disappear.
Most importantly, detection isn't identification. A small heat source can be visible without showing enough shape or context to say what it is. Lower NETD may help separate it from the background, while resolution, optics, distance and field conditions determine how much useful detail remains.
When comparing products, give NETD the credit it deserves—then keep reading the rest of the specification sheet. That's the difference between choosing a well-matched thermal system and simply buying the smallest number.