Motion Boost Recording TestBy: Benros Emata, Published on Oct 11, 2011
Both motion and continuous based recording have downsides. In this test, we look at a feature that aims to rectify both of them simultaneously.
With motion based recording, you can miss relevant events that the camera or VMS does not detect. This exposes user to liability risks and is the most common reason why users object to motion based recording.
However, with continous recording, you almost certainly will be recording for significant periods where nothing is happening - a significant waste of bandwidth, and more importantly, storage. Users can often save 60 - 90% of storage costs by using motion based instead of continuous recording.
On the one side you save storage and money, on the other you risk liability issues.
The best solution to this issue is to use a feature often called 'boosting' or 'speedup' recording. This allows recording lower quality video when no motion is detected while increasing the quality when motion is detected. This ensures that video is always recorded, minimizing liability risk but with lower storage consumption when no activity is happening, saving storage.
The main concern is how well boost up recording works. With analog DVRs, this was not a risk nor concern as the encoder inside the DVR handled this completely. However, with IP cameras and VMS software, most systems require the VMS software to request the IP camera to adjust its quality settings when motion is detected. Because of this, risks exist in whether it works, how well it works and what length of delay is incurred in cutting over. In a recent 89 comment discusion with our Pro Members [link no longer available], this was frequently cited as a concern.
We devised a series of tests with 3 VMSes and 3 cameras with 2 different boost up approaches.
- The 3 VMSes were ExacqVision, Genetec and Milestone Corporate
- The 3 cameras were from Axis, Panasonic and Sony
- The 2 approaches were boosting up just frame rate (resolution remained the same) and boosting up both frame rate and resolution)
- Exacq, Genetec, and Milestone transitioned to the higher quality stream on motion in fairly quick and timely fashion when only the frame-rate was changed (as opposed to changing both fps and resolution)
- Changing both fps and resolution was generally less efficient than changing of only frame-rate
- Efficiency or speed of transition to higher resolution/fps stream during motion tended to vary based on the particular camera/VMS combination
- Timelapse, Boost, and Speed-Up features, while not as storage efficient as recording only on motion, provide material benefits to investigations.
- Timelapse, Boost, and Speed-Up performance/benefits are highly dependent on optimized motion detection. However, a motion false negative in a Timelapse, Boost, and Speed-Up environment is less detrimental than if in a motion only record scenario
- VMSes tend to vary in terms of complexity in optimizing motion detection
- Genetec very simple to configure 'Boost' feature, but difficult optimizing motion detection
- Milestone somewhat complicated to configure 'Speed-Up' feature, but motion detection worked acceptably for the simple indoor scene tested
- Using any of the vendor specific stream promotion features (Boost, Speed-Up, and Time-lapse) on motion can provide meaningful gains in storage efficiency and mitigate drawbacks of false negatives. We recommend promotion of only the frame-rate as this tended produce better and less problematic results
- Consider recording continously at a normal or maximum (e.g. 720p) resolution with a nominal frame-rate (e.g. 1 to 5 fps) and on motion record increase only frame-rate (e.g. to 15 - 30 fps). This will assure better odds of a quick and timely transition of the video stream as changing both resolution and fps tends to increase complexity and reliability risks
- If you are interested in doing a boost of both the frame-rate and resolution test the specific camera model and VMS combination as performance tends to vary on this basis
- If your nominal frame-rate requirements are higher than 1fps (e.g. 5fps) consider Genetec/Milestone VMSes as these allow a greater breadth of frame-rates to choose for the continuous record mode
- Exacqvision VMS 4.6
- Genetec Security Center 5.0 SR1
- Milestone Corporate 4.0
Here are the key findings:
In light of these findings the following are recommended:
Exacqvision VMS (Timelapse)
Exacqvision's Timelapse feature provides continuous recording at a maximum frame rate of 1fps. Upon motion detection it will increase the frame-rate to an administratively defined value (e.g. 30fps). A key difference separating Exacq from the other two VMSes tested is that there is no ability to change the resolution setting (e.g. from VGA to 720p) upon a motion event. Another key difference, of course, is that Timelapse allows a maximum of only 1fps for the continuous record mode - Genetec and Milestone generally allow many fps profiles (e.g. 5, 10, 15, 30 etc.).
Exacq's Timelaspe feature is very simple and straightforward to configure.
Below is a screenshot of 'Timelapse' configuration:
[link no longer available]
In Exacqvision's implementation, there is no VMS side or software based motion detection, and cameras must support their own motion detection in order for Exacq to record on motion events. For our simple scene, the default motion detection zone that the Exacq VMS sets up in each of the three cameras performs moderately well.
In this video we examine the playback of the motion event to note the changes made to the video stream and to better understand in practical terms what impact Timelapse recording may have on investigations. Each of the three cameras are configured to record continuously at 720p/1fps, using the Timelapse feature, and to change to 30fps upon motion (720p resolution does not change).
The subject's motion is detected in a timely manner across all three cameras and the switch from 1 to 30fps occurs with near seamlessness. The motion detection required little to no optimizations to work acceptably in this simple indoor scene.
Genetec Security Center (Boost)
Security Center's Boost feature provides continuous recording at various supported lower resolutions and frame rates (e.g. QVGA at 5fps), and upon motion detection it can increase both/either resolution and/or frame rate (e.g. 720p at 30fps).
Configuring 'Boost' functionality was simple and straight-forward to do. We discovered that when configuring motion detection for a 'Boost' scenario it is advisable to optimize motion detection at the lower (pre-boost or continuous record) resolution prior to turning on the 'boost' feature.
Below is a screenshot of the 'boost' configuration:
[link no longer available]
In Genetec's Security Center motion detection events are generally handled either on-board the camera or centrally via the VMS/Archiver. Obviously, optimizing motion detection is key to successful use of the 'Boost' feature. We found doing so quite challenging and were not able to get motion detection working effectively on all three cameras.
In this video we examine the playback of the motion event to note the changes made to the video stream and to better understand in practical terms what impact Genetec's 'Boost' recording may have on investigations. All cameras are configured to continuously record at QVGA and 5-7 fps. Upon motion detection the 'boost' feature increases both frame rate and resolution to 30fps and 720p, respectively.
The video highlights some of the negative consequences of not having well optimized motion detection. For example, the Sony CH140 did not pick up the subject until very late in the event/activity. When the 'boost' to 30fps/720p did finally occur, our subject was already out of view. Note that because of the continuous record component of the 'boost' feature, the downside (complete loss of video in a record on motion only approach) of a false negative is mitigated, as the system is still able to provide video - albeit with no increase in resolution/frame-rate nor indexing/flagging of the event in the video database.
It's important to note that the test demonstrated in the video involves the promotion of both frame-rate and resolution. In further testing, changing of just the frame-rate (no resolution change) from ~5fps to 30fps resulted in a near seamless and very quick transition to the promoted stream.
Milestone Corporate (Speed-Up)
Milestone Corporate's Speed-Up feature provides continuous recording at various supported lower resolutions and frame rates (e.g. QVGA at 5fps), and upon motion detection it can increase both/either resolution and/or frame rate (e.g. 720p at 30fps).
The configuration of 'Speed-Up' is rather involved as it uses 'Rules'. 'Rules' provide some interesting control and flexibility of the system, but may take some time and effort to learn. A key point to note is that a separate rule must be configured for continuous recording. Also a 'Speed-Up' rule is configured for each camera that will needs to be 'sped-up' upon a motion event.
Below is a screenshot of the 'Speed-Up' configuration using 'Rules':
[link no longer available]
In this video we examine the playback of the motion event to note the changes made to the video stream and to better understand in practical terms what impact Milestone's 'Speed-Up' recording may have on investigations. All cameras are configured to continuously record at QVGA and 5fps. Upon motion detection the 'Speed-Up' feature increases both frame rate and resolution to 30fps and 720p, respectively.
In the Axis Q1755 playback the switch to higher frame rate and resolution occurs at roughly midway through the subjects walk. The switch occurs slightly earlier in the Sony CH140. For the Panasonic SP-306, the switch takes place very late in the course of the incident and also appears to have lost some continuity of video.
It's important to note that the test demonstrated in the video involves the promotion of both frame-rate and resolution. In further testing, changing of just the frame-rate (no resolution change) from 5fps to 30fps resulted in a near seamless and very quick transition to the promoted stream.
Here are the three (3) VMSes used in 'Motion Recording Boost':
Here are the three (3) cameras used in our test: