WDR Megapixel Camera Shootout

AL
Antony Look
Published Jul 31, 2011 00:00 AM

Dealing with direct sunlight is one of video surveillance's toughest problems. Avoiding the sun is frequently impossible. Making things worse, when a scene has direct sunlight creating both dark and bright areas, image quality can suffer dramatically.

In video surveillance, the technical term used to describe a camera's ability to handle these conditions is WDR or Wide Dynamic Range. In this context, range refers to the variations of light levels that a camera can capture and deliver a quality image. The greater the range, the more likely the camera can handle both very bright and dark areas (e.g., sunlight on a person's face, shadow on the car in the corner).

Earlier this year, we did our first WDR shootout with (2) SD and (2) MP cameras. One of the clear, yet surprising, results was that cameras with more pixels (i.e., megapixel) tended to outperform (i.e., capture more image details) than standard resolution cameras, even if the SD cameras were marketed as supporting WDR.

Given those results, in this test, we wanted to learn more about the differences in megapixel camera WDR performance. To do so, we tested 6 Megapixel cameras from Arecont Vision (AV1315), Axis (P1344), Panasonic (WV-SP306 and WV-NP502), Sony (CH140) and Vivotek (IP8151P) to see who was the best and worst at handling WDR scenes.

[UPDATE 2012: We have tested the Axis Q1604 against the best performing WDR cameras in this group.]

We picked 2 common real world scene:

  • Doorway entrance facing outdoors - this is a common pain point for users who cannot properly identify individuals coming inside for hours a day as the sun faces the door
  • Sun setting in the line of sight of an outdoor camera - this is a frequent problem for parking lot and street surveillance. Accidents or thefts occurring during those times of day can be obscured by the sun.

In addition to testing core WDR performances, we did additional testing of 2 advanced features to see what benefits they provide:

  • Face WDR: Panasonic has an optional feature that can detect faces and then adjusts exposure to maximize the image quality of the detected face.
  • Exposure Regions: A number of cameras, including Arecont and Axis, support exposure regions that allow lighting to be optimized for a specific area of the scene (e.g. region specified for where faces normally appear).

Below is a preview of our test results from the doorway showing fairly dramatic differences among the 6 megapixel cameras tested:

Inside the Pro section, we share full results and rankings of each camera's doorway performance.

Below is a preview of our test results from the outdoor setting sun scene showing major performance differences among the 6 megapixel cameras tested:

Our tests revealed a fairly consistent pattern of performance variations. Based on observable differences in image quality captured, we rank the 6 camera's performance into 4 major categories:

  • Superior: Sony CH140
  • Above Average: Panasonic WV-SP306 and WV-NP502
  • Below Average: Arecont AV1315 and Axis P1344
  • Inferior: Vivotek IP8151P

Below, of course, are the raw video clips and the prepared image comparisons so that you can judge for yourself.

Key Findings and Recommendations

Here are the key findings:

  • WDR performance was generally better in cameras specifically marketing or stating WDR enhancements/capabilities - namely Panasonic WV-SP306, Panasonic WV-NP502, and Sony CH140. The exception was the Vivotek IP8151P
  • Cameras demonstrating good/strong WDR performance (indoor hallway), tended to handle harsh direct sunlight conditions effectively as well
  • Direct harsh sunlight does not necessarily produce a lighting differential that is characteristic of a WDR scene, but WDR capabilities appear to improve image results nonetheless
  • The Sony CH140 consistently outperformed the field throughout the WDR tests
  • The Panasonic WV-SP306 and WV-NP502 generally outperformed the field, except the Sony CH140, throughout the WDR tests
  • The Panasonic WV-SP306 performed similarly to its Panasonic counterpart and predecessor the WV-NP502 in the majority of WDR tests. This is despite marketing claims that MegaSuperDynamic (WV-NP502) is more advanced than SuperDynamic (WV-SP306)
  • The Panasonic WV-NP502 exhibited delay 'turning on' WDR when conditions changed suddenly
  • The Panasonic camera's face WDR feature did not demonstrate any practical benefits in our tests.
  • Exposure regions (such as with the Arecont) were generally or overall detrimental to image quality for a majority of the WDR test scenarios

In light of these findings the following is recommended:

  • Consider cameras with proven WDR capabilities for both conventional WDR scenes (e.g. bright doorway or window) and scenes with direct sunlight. Such cameras can significantly improve image quality under these difficult lighting conditions
  • If you are considering a Sony camera for a WDR application, note that WDR is only available in their premium line (V series, i.e., the CH140 but NOT the CH120)
  • If your WDR application undergoes abrupt changes in lighting (e.g. door opening suddenly) consider a Sony WDR over a Panasonic WDR camera. If not subject to sudden changes in lighting, you may wish to consider the Panasonic WV-SP306 at ~$350 less than the Sony CH140 online
  • In general, do not expect or rely on exposure regions to solve issues in WDR type scenes

Download Video Clips

To review the raw video, download the sample clips from each test run:

  • Indoor WDR Hallway (Door Propped) [link no longer available]
  • Indoor WDR Hallway (Door Closed/Open) [link no longer available]
  • Outdoor Direct Sunlight [link no longer available]

Indoor Hallway (Door Propped)

This is a prototypical WDR scenario of a long hallway with a propped open doorway creating an intense backlight relative to a dimly lit hallway. The lux differential peaks to roughly 20x when the subject is positioned very near to the cameras.

Bright (~1500 Lux)

[link no longer available]

There is quite a spectrum of reactions by the various cameras to the intense WDR/backlight in this scene. At one end is the CH140, which renders strong details within the backlit regions. The WV-SP306 and WV-NP502 are able to produce modest level of details within the backlit regions. At the other end of the spectrum is the IP8151, in which the image is nearly completely washed out and is of little utility at this point.

Dark (~150 Lux)

[link no longer available]

At this point in the scene the lighting differential is ~10x, and the two Panasonics and the Sony appear to handle the variation quite effectively. In contrast the rest of the field exhibit some level of over exposure or characteristic white haze effect.

Darker (~75 Lux)

[link no longer available]

Here at a ~20x lighting differential is where differences become much more apparent between cameras. In the 1315DN and IP8151 the subject's face is severely under exposed - almost a complete silhouette. The Panasonics and the CH140 provide a relatively better level of exposure of the subject. Interestingly, the P1344's image of the subject is somewhat over-exposed.

Indoor Hallway (Door Closed/Open)

This scenario is a variation of the same hallway scene, except that the door is now opened and closed. The key differences to note here are that the backlighting is introduced 'suddenly' into the scene when the subject opens the door; and furthermore, as the door closes passively behind the subject the lighting differential decreases rapidly - eventually returning to a non-WDR scene.

Bright (~1100 Lux)

[link no longer available]

As the door is opened the scene experiences an abrupt increase in WDR character. Interestingly the WV-NP502 appears slow to react. This is evidenced by the subject appearing slightly more washed out in this scene versus the previous scene where the door is propped open, and also by the fact that the sign on the door - '5' - is completely obscured.

Dark (~120 Lux)

[link no longer available]

When the subject arrives at this position the backlit region is considerably less than the ~1100 lux at initial opening. The lux differential is therefore only ~5x, and across the board cameras appear to provide better image quality compared to the previous 10x 'dark' scene.

Darker (~25 Lux)

[link no longer available]

At this point the door has completely closed and the scene is no longer one with WDR characteristics. As a result even the 1315DN and IP8151, which were completely under-exposed in the previous 'propped' counterpart scene now provide relatively well exposed images of the subject.

Outdoor Direct Sunlight

In this scene harsh sunlight is directed into the cameras and poses a particularly challenging lighting situation. The first composite is fully zoomed out to demonstrate the overall effect the difficult lighting has on each camera.

Overview Composite

[link no longer available]

The same cameras performing well in the indoor WDR scenes (namely, the WV-SP306, WV-NP502 and CH140) appear to do well in this outdoor scene as well . The converse appears true as well with the P1344, 1315DN and IP8151 showing signs of difficulty with the harsh lighting.

Digitally Zoomed Composite (Wide FoV ~50ft)

[link no longer available]

The P1344 and IP8151 are severely washed out compared to the field. They also have a distinct pinkish tinge or haze blanketing their images. The Panasonics and Sony are relatively better than the rest of the field, but blurriness and lack of sharpness still limit facial details at this wider FoV.

Digitally Zoomed Composite (Narrow FoV ~20ft)

[link no longer available]

At this nearer FoV details begin to emerge from the Panasonics and Sony images. There is little sign of lighting difficulties present in their images as well. The WV-NP502 exhibits particularly good contrast and facial details. In contrast the P1344 and IP8151 continue to provide poor subject facial details and show considerable signs of adverse lighting effects.

Additional WDR Techniques

Face WDR (Panasonic)

Both Panasonics tested support a feature called Face WDR. It requires that an analytic feature called face detection be enabled. After a face is detected within the scene, the camera purports to use WDR enhancements around the face to better resolve facial details.

Below is the WV-SP306 with a Face WDR off/on comparison:

[link no longer available]

Below is the WV-NP502 with a Face WDR off/on comparison:

[link no longer available]

The foregoing comparisons assume that Face WDR was indeed triggered at the point of the 'ON' snapshots. However, as this event is monitored through a separate interface (the camera's web interface), exact confirmation was difficult to ascertain. Based on preliminary testing however, the subject is in a region where the face detection showed a high probability of being triggered. Through these initial tests it was also noticed that face detection/face WDR events did not correspond to any meaningful improvements in facial detail or clarity.

With these reservations in mind, the images depict no clear advantages to using Face WDR in this particular test scenario.

Exposure Region (Arecont)

In this battery of tests, we configure exposure regions on the bright areas of both the indoor hallway scene and the outdoor harsh sunlight scene. The overall effects of this strategy are evaluated via the composite below:

[link no longer available]

Placing an exposure region around the door frame when the door is propped open, decreases overall exposure on the subject when in the bright area. This tends to provide a better overall image of the subject at this point. However, the rest of the scene is essentially 'blacked-out', and the subject is totally undetectable throughout most parts of the hallway interior.

In the next scenario when the door is initially closed, the abrupt opening causes the subject to become under-exposed with the same exposure region in effect. The subject is still considerably under-exposed as he moves inside and to the left. As the subject progresses, and the door shuts behind him, the image obviously improves (not shown) but this is due to an evening out of lighting (loss of WDR character) in the scene and not due to the exposure region.

In the outdoor scenario an exposure region placed near the bright horizon causes the entire picture to darken. As a result, the majority of the scene is heavily under-exposed, significantly obscuring subject and other object details.

Methodology

Here are the six (6) cameras used in the WDR Mega-Shootout:

  • Arecont AV1315DN (online $460) - 1.3MP D/N; 1/2.7" CMOS; MPL4-10; 0.1/0 (Color/BW)
  • Axis P1344 (online $759) - 720p D/N; 1/4" CMOS; F1.2 Computar ; 0.05 Lux (BW)
  • Panasonic WV-SP306 (online $550) - 1.3MP D/N; 1/3" MOS; WV-LZA62/2 lens; 0.3/0.05 lux (Color/BW)
  • Panasonic WV-NP502 (online $950) - 3MP D/N; 1/3" CCD; WV-LZA62/2 lens; 1.0/0.08 lux (Color/BW)
  • Sony CH140 (online $800) - 720p D/N; 1/3" CMOS; F1.2 Fujinon; 0.1 Lux (BW)
  • Vivotek IP8151 (online $754) - 1.3MP; F1.2 Computar; 0.04/0.001 (Color/BW)

All camera lenses were configured at a uniform lens angle (~45 degrees) in order to make fair comparisons. Video was simultaneously recorded and analyzed at their highest VMS supported resolutions. WDR was enabled on all cameras supporting it (WV-SP306, WV-NP502, CH140, P1344). Note the IP8151 markets as a WDR enhanced camera, but apparently does not require it be enabled manually; also the 1315DN is not marketed as having an explicit WDR capability.

Here are the 3 key scenarios:

  • Indoor Hallway (Door Propped)
  • Indoor Hallway (Door Closed/Opened)
  • Outdoor Harsh Direct Sunlight

Each scenario involved two (2) recordings. The first involved basic WDR turned on, while the second involved turning on Face WDR on the Panasonic cameras and an exposure region configured on the AV1315DN.