When purchasing equipment for normal use, datasheets for two similar products can look like they are being marketed by competing departments within the same company. However, when you put them side by side and actually test them for normal use, like checking if the garden gate has been closed after bin night, there can be one specification which describes higher resolution, a second which highlights the gain and a third which describes a detailed test rig. While the first two specifications describe parts of the viewing chain, the latter simply describes a different part of it and does not enable a simple comparison of the two viewing chains as a whole.
Start with the scene, not the headline figure
A useful comparison for viewing applications requires a task description such as viewing across dimly lit gardens, identifying familiar objects near the gate or distinguishing between a thin branch and a cable etc. A high resolution figure enables viewing of fine details, but moving and low contrast objects require more than high resolution.
Higher gain settings will make objects in very dark scenes more visible, but also more noisy. Very bright objects like lamps, reflective surfaces and vehicle headlamps can produce 'blooming' or 'haloing' which are temporary 'overexposures' of the image and can also mask surrounding detail. Some devices are able to automatically control the gain to stop such 'overexposure' occurring. However, the user can also manually set the gain to suit their own requirements.
Each specification should be read as the answer to a specific question. Thus, the resolution of a system describes the amount of detail it is able to reproduce. The gain describes how much the light of a scene is amplified by the system. The test conditions describe how the above numbers were determined.
What resolution measures in practice
Resolution – as is usual – expresses the ability to resolve fine details of adjacent objects. In datasheets it is often specified in line pairs per mm and thus allows for a comparison between different devices. The higher the value for resolution the better the image intensifier will preserve fine details of the object under test.
Resolution can also be limited by parts of the viewing chain outside the intensifier itself (such as the objective and eyepiece) and will typically decline with defocus. Thus even an image intensifier with high resolution as specified will fail to deliver that resolution if it is concentrated in the center of the image and the viewing system is slightly out of focus.
Resolution typically is specified for the center of the image. The values at the edge can be significantly lower. In addition, average values for test targets are not necessarily the same as values for the center.
Thus for the garden gate check view of taking the bins out down the garden a higher resolution may reveal further detail however this must be viewed in context of a moving scene at some distance and so may require higher magnification and/or focusing down. And, in practice, focusing down may also require adjustment of the lens's focus. So, while higher resolution may be of use for viewing the gate's latch, a loose strap or even a branch, it is not a 'plug and play' solution.
Gain is amplification, not a promise of clarity
Note the type of gain (fixed, auto, variable) since higher gain does not necessarily mean a better image (it just makes darker scenes visible, but can also make more image noise visible). Mention the nominal gain as well as the gain range as indicated on the datasheet, and check for any related performance values (e.g. auto control behavior, contrast behavior).
Increasing gain may provide the ability to see into very dark scenes, but it can also increase the appearance of noise in the image. Bright objects such as lamps, reflectives, or vehicle headlights can cause the image to 'blooming' or produce halos around the objects. These effects can be reduced by utilizing an Automatic Gain Control (AGC), but the image will then be limited by the optics and the noise characteristics of the image intensifier.
Note also, gain is often described by a number (e.g. Gain 4x) and thus can easily be confused with a pure brightness setting. This setting determines how much signal amplification is applied to input light but different devices of apparently equal gain may have completely different amounts of noise as well as a different contrast and a different behavior of automatic gain control etc. The possibility of manual gain adjustment is also mentioned in some datasheets but this in no way means that one can simply compare the numbers.
So for the backyard example, rather than focusing on how much gain the images could be amplifying, the key question would be how to make the gate area as readable as possible when it is very dark and there is a very bright light in the scene. A high gain image with lots of noise could well be worse than an image with lower gain and less noise.
Test conditions change the meaning of a number
Always verify the test conditions for performance figures, such as amount and wavelength of light, test objects, optical system, ambient temperature, focus, test method etc. Whether the figure represents an average value or the minimum/maximum value is usually indicated by terms such as "typical", "minimum", "maximum" or "at…".
Minor differences in setup can have big differences in results. High contrast test targets can look very different from images of real scenes with a lot of dark foliage against dark backgrounds. Measurements of gain taken under controlled illumination do not automatically translate to dark or unevenly illuminated scenes such as under clouds or near street lights.
Check for words such as "typical", "minimum", "maximum", "measured at", "specification", "at specified illumination". Transfer all the information that you would record as uncertainty to a "Unknown" cell and require the manufacturer to complete for you.
Similarly, there can be large differences in figures between generations of technologies or between very different test methods. So, a figure of a certain value from one manufacturer does not necessarily mean that the same value for that parameter from another manufacturer (even on the same data sheet) was determined by the same test method.
Build an apples-to-apples comparison
Review products with the same attention as you would compare apples and oranges. First write down all relevant information for every data sheet, then make three different groups of this information.
Then separate the data into three groups:
– Directly comparable figures measured under substantially matching conditions.
– Useful indicators that point in the same direction but use different methods.
– Unknowns – one needs a clarification by the manufacturer, a sample image or even more cautions interpretation possibilities.
It's important to distinguish between specifications for components and specifications for complete systems. Sometimes even one component is described, whereas the complete system is actually intended for sale – or the other way round. So it is worth checking whether the specifications refer to the intensifier alone or to the whole system. And also whether the gain is controlled manually or automatically.
When reviewing seller sample images check if the lens, magnification, lighting and any image processing is noted on the image. An image that has been supplied by the manufacturer can look very impressive but until you have details of how the image was actually taken it does not provide much useful evidence of the night vision device's performance.
Turn the datasheet into a buying decision
We distinguish between a simple observation of objects at a certain distance and the level of detail that is visible. Furthermore we take into account the amount of light that is available, other light sources that could cause glare, and also movements of people or animals in the scene to be observed.
So be wary of papers that quote single figures. Create a short list and then check for a balanced specification of relevant parameters. The quality of the information in a datasheet is more important than the information itself. So check that the figures quoted relate to the complete night vision device and not to some internal component.
At the least, make a note of missing information such as the amount of illumination used, information regarding center versus edge resolution, the nature of gain control and the make-up of the optics. Such omissions do not necessarily indicate poor performance but do demand greater caution in your comparison.
The datasheet is a useful document which gives information of what a night vision device has been tested to do in a controlled environment. This information can then be used to estimate how it will perform in other situations, for example at the garden gate, down a dark path or near a street light. The key is to match the datasheet test situation with your own needs.
