Understanding video surveillance resolution can be surprisingly difficult and complex. While the word 'resolution' seems self-explanatory, its use in surveillance is far from it.
In this tutorial, we will explain 5 critical elements:
What resolution traditionally means – seeing details - and the constraints of this approach
What resolution usually means in surveillance – pixels – and the limits of using this metric
Aside from these releases, there has been limited movement to 8K and higher resolutions. Even if more models were released in 2021/2022, they are likely to remain niche due to the larger size and high cost for the next few years.
Resolution: Seeing Details
In normal English and general usage, resolution means the ability to resolve details – to see or make them out. For example, can you read the lowest line on an eye chart? Can the camera clearly display multiple lines side by side on a monitor? It is a performance metric focusing on results.
Historically, video surveillance used a similar test chart approach. Analog camera resolution was measured with line counts, literally the camera's ability to display more lines side by side in a given area on a monitor.
If you could see more lines, it meant you theoretically could see more real world details: facial features, characters, license plates, etc.
For example, in our testing, we found 1080p cameras to be equivalent to roughly 900 'lines', 5MP equivalent to 1,000, etc. as shown in the image below:
Problem: TVL Only Tested In Ideal Conditions
However, TVL testing is only intended to be done in even lighting conditions. Introducing direct sunlight, backlight, or lower light levels, TVL measurements drop significantly. Additionally, this testing is done only on stationary test charts, not moving objects common in the real world.
Because of this, TVL measurements do not account for overexposure, encoding noise and artifacts, IR illumination effects, motion blur, or other issues present in the real world.
Resolution: Pixel Count
In marketing IP cameras, manufacturers do not attempt to measure performance. Instead, resolution has been redefined as counting the number of physical pixels that an image sensor has.
*** *******, * ***** ********** ****** is ******** ********* ** ****** *** (million ******) ********** ******* *** ****** used *** ~* ******* ****** ** it (*********** *,***,*** ****** ** **** is *** ******* ** **** ********** x **** ******** ******). *** ***** of * ****** ***** ***** **** example:
Great Report! I have been trying to teach "there's much more to imaging than resolution alone" for years. Still get the comment, "the user wants 4MP cameras, your 2MP doesn't have enough resolution." But when we do side by side under real world conditions, the 2MP wins hands down, particularly in lower light and WDR.
Hallelujah! This should be mandatory reading for every sales, marketing, designing and installing chump professional out there. Excellent piece will so many valauble links to real-world tests. As Scotty said, "Ya cannae change the laws of physics, Captain!"
Truman, you can't do it abstractly (i.e., without having the cameras) and not risk being off by a very significant percentage. I think most people (excluding better, more sophisticated integrators) just deal with it after the fact and make tradeoffs later (i.e., lower the frame rate, increase the compression, reduce the number of days storage, increase the hard drives, etc.).
Before this article I thought only megapixels affected the image quality (all things being equal). I did not know there was such a thing as resolution overkill and always thought "more is better" was the approach. Just enough detail to achieve the objective for surveillance is a key understanding in this field I realise.
In your Resolution Tutorial it states that "we found 1080p cameras to be equivalent to roughly 900 'lines', 5MP equivalent to 1,000, etc." Looks like the chart is indicating 1,100... not the 1K as stated. Am I missing something?
If you look at the zoom in of that chart, you can easily distinguish all 9 vertical lines at around 1,000 TVL, but one is a little unclear at 1,100. It's a bit of a judgement call, but I error on the side of the lower reading.
We have from time-to-time considered software analysis as part of our testing, but each time we come to the same answer: that it is a poor metric for how well things perform in the real world. Because testing standards focus on still images, with few exceptions, and most are performed under highly controlled lighting conditions, their applicability to actual surveillance applications would require a lot of interpretation.
That being said, we have recently begun talking about lab testing again, but not as the sole means of testing, just as an additional parameter. We have been doing deeper analysis of video with different stream analyzers lately and will be doing more in the future.
Good tip on the ISO 12233 chart, thanks. We'll take a look at SFR charts and may get one to include in testing. The slanted edges are an interesting test.
Great article! Compression & Bitrate are HUGE factors. I've always felt resolution to be a 1/3 of the pie. I'm not sold on higher resolution cameras which I feel can overload the processor depending on how high everything gets configured (frame rate, bit rate, resolution). I would love to see a deep dive comparison of NVR/DVR "processors" along with HDD write speeds which I believe dictate everything.
For me it’s a balance between identification, bandwidth and storage.
7 years ago I implemented a 2MP standard with some exceptions. Higher resolutions must be justified except for areas where pedestrians have not been identified by access control or visitor management systems. For example, parking lots, entrances and sometimes lobbies, depending on the size of the room.
I’ve refused many proposals where the integrator specified 5MP and higher cameras for the entire project for absolutely no reason. They just figured “higher resolution is better”.
IPVM conducts reporting, tutorials and software funded by subscriber's payments enabling us to offer the most independent, accurate and in-depth information.
Comments (23)
Undisclosed Manufacturer #1
Great Report! I have been trying to teach "there's much more to imaging than resolution alone" for years. Still get the comment, "the user wants 4MP cameras, your 2MP doesn't have enough resolution." But when we do side by side under real world conditions, the 2MP wins hands down, particularly in lower light and WDR.
Create New Topic
Undisclosed #2
That may be true, but these days people certainly associate resolution with pixels.
With their 1080p and 4k big screens, computer monitors, smart phones etc, they have been inundated with the concept.
Which makes even more gullible for the more pixels=better picture pitch.
Create New Topic
Simon Lambert
Hallelujah! This should be mandatory reading for every sales, marketing, designing and installing chump professional out there. Excellent piece will so many valauble links to real-world tests. As Scotty said, "Ya cannae change the laws of physics, Captain!"
Create New Topic
Truman HW
How do I measure this without buying every camera that potentially belongs in a comparison?
What metric can I use to to make this beyond complex question more manageable?
Thanks -- I'm hoping to hear from a few people... and of course, John. :)
Create New Topic
Sheldon Beaupierre
Wow.... Didn't know that pixels determine potential now quality. Great stuff
Create New Topic
Undisclosed End User #3
An informative write-up. Good
Create New Topic
Amitkumar Sirjusingh
Before this article I thought only megapixels affected the image quality (all things being equal). I did not know there was such a thing as resolution overkill and always thought "more is better" was the approach. Just enough detail to achieve the objective for surveillance is a key understanding in this field I realise.
Create New Topic
John Deaton
In your Resolution Tutorial it states that "we found 1080p cameras to be equivalent to roughly 900 'lines', 5MP equivalent to 1,000, etc." Looks like the chart is indicating 1,100... not the 1K as stated. Am I missing something?
JD
Create New Topic
Undisclosed #4
the resolution quiz;
q. What is the horizontal and vertical resolution of a 3 MP camera (Answer in the form: H x V)?
entering 2304x1296 is marked wrong?
Create New Topic
mark montano
Great article! Compression & Bitrate are HUGE factors. I've always felt resolution to be a 1/3 of the pie. I'm not sold on higher resolution cameras which I feel can overload the processor depending on how high everything gets configured (frame rate, bit rate, resolution). I would love to see a deep dive comparison of NVR/DVR "processors" along with HDD write speeds which I believe dictate everything.
Create New Topic
Dwayne Cooney
For me it’s a balance between identification, bandwidth and storage.
7 years ago I implemented a 2MP standard with some exceptions. Higher resolutions must be justified except for areas where pedestrians have not been identified by access control or visitor management systems. For example, parking lots, entrances and sometimes lobbies, depending on the size of the room.
I’ve refused many proposals where the integrator specified 5MP and higher cameras for the entire project for absolutely no reason. They just figured “higher resolution is better”.
There’s many other factors to consider.
Create New Topic
Bobby Vadnais
What does the acronym TVL stand for? This article uses it, but does not define it, nor is it defined in the glossary.
Create New Topic
Bobby Vadnais
When I visit the link in "Pixels Determine Potential, Not Quality", I get served a page that reads "This report has not been published. Check back later."
Create New Topic