Don't Trust Lux Ratingsby John Honovich, IPVM posted on Jan 09, 2013 About John Contact John
Do not trust lux ratings. Do not use lux ratings to specify cameras. Period.
Lux ratings are widely, and unfortunately, used to define low light performance, with the lower the lux rating, the stronger low light performance. Here's what this commonly looks like on manfacturer specifications:
For instance, a camera with 0.02 lux is supposedly 'better' in low light than a camera with 0.05 lux. (Background - See our Lux / Lux Meter Tutorial).
However, they are so riddled with fundamental problems that lux ratings must be abandoned.
In regular IPVM assessments, we find easily half of surveillance professionals believe and use lux ratings as a viable metric for assessing low light performance.
More importantly, lux ratings are overwhelmingly used in RFPs to require specific low light performance. Cameras that do not meet the RFPs lux ratings specification are rejected.
The Fundamental Problems
Here are the fundamental problems:
- Unrealistic numbers
- No standard process
- No revelation of what image looks like
- Failure to disclose camera settings
- Gradual image quality decline
Most manufacturers have lux ratings that are incredibly unrealistic, with ratings of .001 lux or lower common. However, that is incredibly dark. Almost any camera truly in such conditions in the real world would capture nothing or be so dark and noisy as to be practically useless.
No Standard Process
Each manufacturer measures on their own, by themselves and with whatever 'standards' that they like. As such, it is impossible to compare the results of two manufacturers without more knowledge or testing of one's own.
No Revelation of What Images Look Like
No manufacturer ever releases images of what their cameras look like at their claimed lux ratings. Almost certainly, the images would be terrible.
Failure to Disclose Camera Settings
Many manufacturers use tricky settings such as using super slow shutters and fostering tricks like 'sens up'. Often they will obscure this in their specification, resulting in seemingly eye popping low lux ratings like 0.000001 lux.
Gradual Image Quality Degradation
The image quality of all cameras gradually declines as light levels fall below 50 lux, with falling signal to noise ratios, and increasing gain control levels. There is no magical point where quality turns from good to bad.
The lion's share of the blame goes to consultants who regularly specify cameras based on minimum illumination specifications though they rarely if ever test to verify that the ratings are accurate.
This creates an ugly system where even the most ethical and responsible manufacturers are trapped. The first manufacturer to 'tell the truth' about their low light performance will be disqualified from many large projects. Because of this, no one can afford to do so.
Worse, it rewards the most unscrupulous vendors who realize that they are rarely, if ever, called on their specs.
What To Do?
The best way is to test cameras head to head in the same conditions with the same fundamental settings (especially shutter speed). For example, see our MP low light shootout. Take your finalists and place them for an evening in your desired deployment location. I can almost guarantee that the results will not match what the lux ratings suggest.
Short of that, a few specifications do deliver:
- F Stop: While small differences in F numbers (1.2 vs 1.4) have limited practical impact, going from F1.2 vs 2.4 typically has a major impact on low light quality.
- True Day/Night: Cameras with mechanical cut filters consistently deliver notable increases in low light performance.
- Adding IR: If a scene is quite dark and you want to ensure maximum illumination, consider adding IR either through integrated IR or add-on illuminators.
There is no magic number. Unfortunately, lux ratings are voodoo, more smoke and mirrors than reality. Let's move past them and focus on better metrics for reliably specifying high quality low light performance.
Most Recent Industry Reports
Axis 4K Tested on Oct 29, 2014
Two years ago, Axis declared the megapixel race over. Now, they are among the first to release a 4K (8.3MP) camera. However, the question is simply: do we really need more pixels? We bought an Ax...
Top Manufacturers Gaining and Losing Ground on Oct 27, 2014
IPVM survey results of 100 integrators show what manufacturers are gaining and losing ground in 2014. Integrators answered 2 questions: "Added: This year, in 2014, what new product lines have you...
Getting Started With Your IPVM Membership on Oct 24, 2014
Here's how to get started and get the most out of your IPVM membership. Getting Started Video Presentation You can watch the 20 minute video immediately below or scan through the whole post for i...
Testing Arecont Omni on Oct 20, 2014
The Arecont SurroundVideo Omni is one of the most novel new camera offerings in the past few years. Like their 180 and 360 predecessors, the Omni supports 4 imagers / cameras inside. The major adv...
The Worst HD Quality Camera Ever (Samsung 1280H) on Oct 15, 2014
Samsung, what are you doing? Their new 1280H offering significantly stretches video and is one of the most bizarre offerings we have ever seen. Earlier this year, Samsung trumpeted that 1280H wou...
Winners Losers Fall 2014 on Oct 13, 2014
We are all losers, IPVM included. The historically poor new releases for Essen / ASIS make 2014 the worst year for surveillance innovation in, at least, a decade. (See our in depth manufacturer re...
Smart IR Panoramic Camera Tested on Oct 10, 2014
Panoramic cameras are overwhelmingly quite bad in low light. Now, manufacturers are starting to add IR to them. We tested the Hikvision panoramic IR earlier this year. Now, Vivotek has released t...
Testing Speco IP Intensifier on Oct 06, 2014
Chances are you have seen Speco's promotion of its Intensifier HD IP line, with ads online and in trade mags claiming color images down to an incredible 0.0005 lx. Speco funded a vacation f...
Testing IP Camera Latency on Sep 26, 2014
How much does latency impact IP cameras? We tested a number of combinations, like so: In this report, we break down: Average latency metrics in our test Key drivers of latency Variations i...