Testing StarDot's 5MP Camera (SD500BN)
While not well known in the security market, StarDot is one of a few manufacturers providing 5MP cameras. Though StarDot has been producing megapixel cameras for the tourism, construction and entertainment markets for years, they only recently entered security. Indeed, many, including myself, first heard of StarDot last August when integration with Milestone [link no longer available] was announced.
Though StarDot has been somewhat of a mystery, it was clear from discussions with industry people that they were gaining ground and being specified into some large projects.
- Low light video quality (sub 1 lux) was very good - impressive for a megapixel camera
- MJPEG Bandwidth consumption was modest and was lower than Arecont's H.264 during high motion and low lux scenes
- For complex scenes, the frame rate at 5MP was low
- Camera defaults should be changed to settings more appropriate for surveillance use
- Configuration requires modest training to understand settings
Product Overview
The camera tested was the SD500BN with a 4-13mm lens (LEN-5MV4513CS). Mechanically, the camera is fairly straightforward. The main feature it had that some others lack is a coax output for spot monitors. Main features it lacked (mechanically) was SD card support and an integrated auto-zoom lens.
The MSRP of the SD500BN is $999 and comes with a 4mm lens. The MSRP of the Lens tested with this camera was $259. There's also a 4-10mm lens (lower resolution rating) that has an MSRP of $99.
StarDot is supported by a modest number of NVR partners including Milestone, OnSSI and Exacq (see NVR support page [link no longer available]).
While this camera does not support H.264, StarDot plans to add H.264 support before the end of 2009 (up to 5MP). Demonstrations of the new camera line are scheduled for ASIS.
Examining Video Quality
The screencast below demonstrates and provides commentary on the video quality produced by the StarDot 5MP camera in a variety of conditions. In the following section, you can download the actual videos.
The visible resolution was higher than 1080p cameras.
The frame rate is rated at 10fps. For indoor testing, day and night, the frame rate visibly looked at or close to 10fps. However, for our intersection test with a wide field of view, trees, etc., the frame rate appeared to be closer to 2-5 fps. I cannot tell the cause of that, whether it was the camera, my PC, the VMS software, etc.
The low light performance was quite good - with many details made out at a measured .2 lux with very minor motion blur (only from me waving my arms quickly).
Download Sample Videos
Download a zip file of all 6 sample videos [Note: 65 MB total]. The sample videos play in the wrapped Exacq player allowing for digital PTZ and frame by frame control.
The sample videos include:
- Daytime indoors 250 lux
- Daytime outdoors intersection
- Nighttime outdoors intersection
- Nightime indoors 1 lux
- Nighttime indoors .2 lux
- Nighttime indoors .2 lux with low speed, high gain setting
General Configuration
While configuration was not hard, it was somewhat tricky. The primary factors were the use of non-standard surveillance terminology and the setting of defaults that are not optimized for surveillance. The screenshot explains these items which include:
- The JPEG quality setting is too high (default is 80). I tested various settings and settled on 40 - same video quality, lower bit rate.
- black/white mode is called IR and is off by default. If you are using this in low light, make sure to switch it to 'auto'
- Exposure is referenced by dividing into 48000. If you set an exposure value of 1600, that is 1/30 second exposure. [Note: StarDot will be updating this display in the next release to show this]
- Exposure pre-sets either limit exposure to 1/20 second or 1/3 second. I manually set an exposure limit of 1/10s (or in their terms 4800) for the optimal level for outdoor viewing.
Exposure Configuration
As explained above, I needed to change the exposure level. This screencast explains the default options and how I optimized the exposure configuration.
Bandwidth Tests
This was interesting. Daytime bandwidth consumption ranged about 12-14 Mb/s. Nighttime bandwidth consumption ranged about 9-11 Mb/s. The level of motion in the scene did not have a material impact on bandwidth consumption. The resolution was at the full 5MP, the frame rate was set to 10 and the quality level was set at 40.
For a 5MP camera, I was expecting significantly higher bandwidth utilization - 20Mb/s or more. Setting the quality level at 40 had an impact (though as I mentioned previously, I don't think this had an impact on visible resolution. The lower apparent frame rate also likely impacted bandwidth.
Most manufacturers agree that MJPEG 'efficiency' can vary. StarDot claims its efficiencies come from "DNR algorithms (on low light / night shots), cross sampling the image processing pipeline (reduces image file size) and hardware designed to minimize virtually all electronic noise in the image." I do not have the expertise to comment on these points.
The screencast below demonstrates bandwidth utilization in 4 different scenarios, night and day.
Recommendations
In general, the camera worked well with low light performance and modest bandwidth consumption standing out as 2 key test results. Additionally, for a 5MP camera, a $999 MSRP is quite competitive.
Given that outdoor surveillance (parking lots, courtyards, etc.) are a prime use to cut down camera costs by using megapixel cameras, the low light performance of the StartDot camera is attractive. Furthermore, with night time bandwidth consumption at or under 10Mb/s, even though its MJPEG, storage costs will be lower for the StarDot than for the Arecont H.264 cameras at night (with better low light performance).
On the other hand, while daytime image quality is good, it is likely similar to other cameras (including the Axis Q1755 and the Arecont H.264 cameras). Given StarDot's daytime bitrate of approximately 10Mb/s more than H.264 multi-megapixelcameras, using StarDot's current 5MP camera in primarily indoor or controlled lighting scenarios may result in hundreds of dollars or more of storage costs.
Finally, while StarDot is supported by a moderate amount of VMS/NVRs, its more established surveillance competitors are supported by many more. Buyers should check that StarDot's cameras are supported by the specified VMS.