720p vs 1080p Camera ShootoutBy John Honovich, Published on Oct 25, 2011
When choosing High Definition cameras, one of the most fundamental questions is what of the two common resolution levels one should choose: Do you use 720p or 1080p cameras? The fundamental alternatives are familiar to any user who has bought a TV in the last few years. However, how do you choose when buying network cameras?
The main case for 1080p cameras over 720p ones is that a 1080p camera has over twice the resolution (i.e., pixels) than a 720p one. A 1080p camera has a resolution of 1920 x 1080 (2.07 MP) while a 720p camera resolution is 'only' 1280 x 720 (.92 MP). Because of that, marketing people often conclude that a 1080p is equivalent to (2) 720p cameras.
In this test, we choose two very similar cameras from the same manufacturer - one that was 1080p (Sony CH210) and the other that was 720p (Sony CH110). As you can see in the sample image below, from the outside the cameras even look identical:
We wanted to better see how much performance would vary between two seemingly identical cameras with resolution being the main functional difference.
[Note: this is just the latest of our comparison tests of cameras with different resolution. We routinely test SD vs 720p vs 2MP vs 5MP etc. (for other comparisons, see our indoor camera shootout, our parking lot shoot, WDR camera shootout, etc.)]
The Test Process
We took these two cameras and we ran a series of head to head shootouts in a number of conditions:
- Indoor lighting as one might have in an office or retailer that represent ideal conditions for cameras
- Low lighting conditions (from 20 lux down to 7, 3 and 1 lux levels) that represent challenging common scenarios for cameras
- Outdoors for larger area monitoring
Expectations vs Results
Conventional wisdom sets two common expectations of megapixel:
- Higher resolution cameras provide much better image quality (2x, 3x, 4x, etc.) than lower resolution cameras.
- Higher resolution cameras are much worse at night than lower resolution cameras.
Here are the 3 key points learned from the shootout:
- The 1080p CH210 was far better in low light than the 720p CH110
- Both cameras delivered essentially the same image detail in an indoor medium FoV scene
- The 1080p CH210 provided modestly better image details in extremely wide outdoor areas
Relative to conventional wisdom, all 3 findings are surprising.
Key Differences Between the Cameras
While the cameras look the same on the outside, on the inside there are a few important differences:
- The 1080p CH210 camera uses a 1/2.8 imager while the 720p CH110 uses a much smaller 1/3.8 imager.
- The 1080p CH210 uses an EXMOR [link no longer available] imager while the CH110 uses an older generation generic CMOS imager.
- Minor: The 1080p has a fixed 88 degree FoV about 10% wider than the 720p with a fixed 80.7 degree FoV.
- Minor but beware: The 1080p has a default exposure time of 1/30s ('normal') while the 720p CH110 has a default exposure time of 1/8s ('aggressive/slow')
Within Sony's line, all the 720p X series (entry level) cameras (like the CH110) use the generic CMOS sensor. However, all the 1080p X series and all other HD mid level and premium level camears (like the E and V series) do use EXMOR. Finally, while the EXMOR-R [link no longer available], the latest version of EXMOR has gained attention, none of the production Sony network cameras support that.
This leads to a number of key recommendations:
- Be careful in checking camera details even when the cameras look the same and come from the same manufacturer. The guts of these cameras are materially different.
- Beware of smaller imager sizes. While it is not universal, cameras that use imager sensors at or close to 1/4" tend to deliver far worse performance than cameras with 1/3" or larger.
- Consider the sensor used. This is the most risky recommendation as many manufacturers champion the use of the Sony EXMOR chip with widely varying performance in our tests. However, it is likely that the EXMOR is the cause of the 210's superior low light performance.
- If considering between the CH210 and the CH110, you are essentially paying $150 more (~$450 online for the CH210 vs $300 online for the CH110) to get superior low light performance and modestly better wide area coverage. That noted, if you really want strong low light performance, you likely should not consider these cameras as neither support mechanical cut filters (i.e. day/night).
With that noted, let's look at the actual performance.
Even Indoor Light (Easy Case)
We start with an 'easy' scene with 250 lux and approximately 12 foot wide FoV where our subject stands.
Note that the FoV between the two cameras is slightly different because the 1080p CH210 has a roughly 10% wider FoV. However, the CH210 also has 50% more horizontal pixels. The effective advantage in pixels per foot for the 1080p is roughly 40%.
Zooming in to both cameras, we can see the level of details in the face and in the objects on the board are very similar.
The main difference noticeable is the 'trueness' of the colors. The CH210 more realistically captures the skin tone of the subject.
When looking at low light, it is useful to descend into decreasing levels of darkness. Let's start with 20 luxs which is visibly dim for humans but still a fair amount of light.
You can see in the comparison below that the 210 delivers a much clearer image of the subject's face than the 110. The 110 is obviously already aggressively using gain control as indicated by the sizeable noise displayed.
Now, let's drop the lighting to 7 lux. Looking at the image below, all of a sudden, the CH110 actually looks brighter and perhaps better than the 110. What's going on here is a 'trick' or, at least, a subtle variance in default camera settings. The CH110 with a default max exposure of 1/8s is taking in nearly 3x the light as the CH210 with a default max exposure of 1/30s.
Now, let's normalize the exposures putting them both on a more standard 1/30s exposure. Now, the CH110's image is clearly far darker and noisier, much worse than the CH210. Alternatively, we could have slowed the CH210's exposure to 1/8s to match the CH110's default. The net result would remain the same - as long as the exposure is equal the CH210's performance is better.
Now's let drop down to 3 lux which is starting to get pretty dark. Below, we start with default exposures again making the 110 look better:
However, after we normalize the exposure to 1/30s on both side, the 210 now is superior delivering some details of the face and objects on the chart while the 110 is much darker and noisier.
Finally, let's take a look at a 1 lux scene with normalized exposures. While the CH110 is almost pitch black and filled with noise, the CH210 shows an outline of the subject and some details of the objects.
NOTE: It is important to remember that while the performance of the CH210 at 1 lux is better than the CH110, almost all Day/Night MP cameras will outperform the color only CH210 in the same conditions.
Outdoors - Very Wide FoV
Finally, we did some tests outdoors to see what would happen in a very wide FoV . The image below is at roughly a 100 - 120 foot FoV. The 210 delivers more details than the 110. However, the increase in details are fairly modest.
Finally, in this last shot at about 200 foot wide FoV, the details are fairly similar but the 110 has some coloration issues (note the yellow tones) that might be a product of the sensor: