CCTV Recording, Frame Rate, Retention and Storage under the SPF VSS Standard

Installing cameras is only one part of a Video Surveillance System.

Someone, or increasingly some form of automated detection, still needs to determine when something important is happening.

The Singapore Police Force Video Surveillance System Standard for Buildings therefore considers not only cameras and recording, but also:

live monitoring, monitor design, camera selection, operator workload, video analytics and integration with other security systems.

This is an important distinction.

A building may have hundreds of cameras, but if an operator is simply presented with hundreds of passive images and expected to notice the important one at the right moment, the practical security value can be limited.

The VSS Standard instead encourages a system where operators can quickly retrieve relevant cameras and where integration and analytics can help bring important events to their attention.

The central principle is:

The objective is not to make the operator watch more video. It is to help the operator notice, understand and respond to the events that matter.

Where should live CCTV be monitored?

The Standard says live VSS images should be monitored by operators at locations such as:

Security Control Rooms, Fire Command Centres or other VSS viewing facilities within the building.

The exact arrangement will depend on the building and its operational requirements.

The important point is that VSS is not viewed solely as an unattended recording system.

The Standard recognises a real-time operational role.

Does every CCTV camera need to be permanently displayed?

The Standard does not say that every camera must always occupy a dedicated monitor.

Instead, it says operators should be able to:

• select any camera for display on any monitor;
• establish scanning sequences;
• adjust the dwell time of those sequences;
• rapidly select camera views with minimum manual effort; and
• use multi-view displays.

This provides flexibility.

The operator does not necessarily need 100 physical screens simply because the building has 100 cameras.

Should an operator be able to display any camera on any monitor?

Yes.

The Standard says that during an incident, each monitor within the viewing facility should be capable of displaying any camera within the building’s VSS.

This is operationally useful because the importance of a camera can change immediately when an event occurs.

A monitor normally used for general perimeter surveillance may need to display:

an emergency exit, loading bay, lobby or Critical Area

during a particular incident.

Can several operators view the same camera simultaneously?

The Standard says one user selecting a live feed should not prevent another user from selecting:

the same feed

or

another feed on the same system.

This matters in larger security operations where:

one operator may be handling the incident while another person, supervisor or responder also needs the same visual information.

Should CCTV cameras have clear names and numbers?

Yes.

The Standard says camera labelling and numbering, as well as the recording sequence, should be carefully planned to help operators retrieve recorded footage quickly.

It also requires displayed camera images to show:

camera ID, date and time.

Consultant’s perspective

Camera naming should therefore help the operator understand the location immediately.

For example:

L03-East-Lobby-Entrance

is operationally more useful than:

CAM047

unless the numbering system itself is clearly understood and mapped.

Why does camera naming matter during a real incident?

Because security personnel should not have to translate meaningless device numbers while under pressure.

Consider an alarm message:

Camera 146

versus:

Level 2 West Emergency Exit

The second immediately tells the operator where to look.

The VSS Standard does not prescribe one naming convention, but it does require planning the labelling and numbering so footage can be retrieved rapidly.

What information should appear on a live CCTV image?

The Standard says displayed camera pictures should include a single superimposition showing:

camera ID, date and time.

This information helps operators establish:

where the image comes from

and

when the event is occurring.

The overlay should still avoid unnecessarily obscuring important parts of the image.

Does monitor quality matter?

Yes.

The Standard says display monitors should provide colour images with suitable controls for matters such as:

contrast, brightness, sharpness and colour.

Displayed images should remain:

sharp, stable, accurately coloured and free from interference, ghosting or pulsing.

It also says the correct image aspect ratio should be maintained so the picture is not distorted.

So camera quality alone is not enough.

The operator must also be able to see the image properly.

Can a good CCTV image become useless on the wrong monitor?

Potentially, yes.

The Standard says monitor size and resolution need to be considered together with viewing distance.

It specifically warns that an operator seated too far away may not be able to perceive the detail available on a small high-resolution monitor.

This is an important control-room design issue.

A high-quality Identification camera is of limited real-time value if the operator cannot actually perceive that detail on the display being used.

How far should an operator sit from the monitor?

The Standard provides a viewing-distance calculation and also gives examples for 1080p displays.

Its table provides the following reference values:

These are values in the Standard’s example table rather than universal seating rules for every control room.

The underlying principle is more important:

Monitor size, resolution and operator distance need to be designed together.

Does a bigger monitor automatically mean better CCTV monitoring?

No.

A larger screen can help where the operator sits further away, but monitor size is only one factor.

The Standard relates:

display size, resolution, displayed video resolution and viewing distance.

A large screen displaying many small camera tiles may still present individual images too small for detailed observation.

Consultant’s perspective

The question should therefore be:

“Can the operator perceive the detail needed for this monitoring task from their normal working position?”

rather than:

“How big is the video wall?”

Do I need a CCTV video wall?

The Standard does not state that every VSS requires a large video wall.

It focuses on whether:

• relevant images can be displayed;
• cameras can be selected quickly;
• multiple views can be displayed where required; and
• the operator can perceive the necessary detail.

A large video wall may be useful for some control rooms.

It should not be treated as proof that the monitoring arrangement is effective.

How many CCTV cameras can one operator monitor?

The SPF VSS Standard does not prescribe one universal maximum number.

Instead, it says the exact number and presentation of images should be based on:

security, crime-detection and operational requirements.

This is important because a statement such as:

“One operator can monitor 16 cameras.”

or

“One operator can monitor 100 cameras.”

cannot be derived from this Standard as a universal rule.

What factors determine how many cameras one operator can monitor?

The Standard gives several factors:

• the risk if an event occurs and is not detected;
• the purpose of the observations;
• activity and targets within the image;
• expected frequency of incidents;
• how long the operator is likely to watch an event;
• other tasks performed by the operator; and
• operator competence.

That is a far more useful approach than assigning an arbitrary camera-to-operator ratio.

Why is the operator’s other work relevant?

Because CCTV monitoring may not be the person’s only responsibility.

An operator may also be expected to:

answer intercom calls, acknowledge alarms, control doors, speak to visitors, coordinate guards, answer telephone calls or prepare reports.

The VSS Standard explicitly says other tasks carried out by the operator should be considered when determining how many camera images are presented.

Consultant’s perspective

A person cannot realistically be treated as a full-time CCTV observer if their attention is constantly required elsewhere.

Does the risk level affect how many cameras should be shown?

Yes.

The Standard specifically includes:

the risk associated with an event occurring and not being detected

as a factor when deciding how many camera views should be presented to an operator.

This means monitoring design should reflect consequence, not simply camera quantity.

Missing an event in a general corridor may have a different consequence from missing:

an intrusion into a Critical Area.

Should the number of cameras per operator ever be reviewed?

Yes.

The Standard says performance evaluations should be undertaken periodically and when there is a significant change to:

the viewing task

or

the control-room setup.

This is important because surveillance operations can change.

For example:

20 cameras may later become 80 cameras.

Analytics may be added.

The operator may be assigned access-control responsibilities.

The original monitoring arrangement may no longer remain appropriate.

What is video analytics?

The Standard describes Video Analytics, or VA, as an important tool for detecting:

unauthorised intrusion

and

suspicious activity near the perimeter,

particularly where a single operator needs to monitor many cameras.

Video analytics can examine the video stream for programmed conditions and alert the operator when those conditions occur.

What types of video analytics does the Standard mention?

The Standard gives examples of rule-based events such as:

• crossing a virtual tripwire;
• suspicious people loitering in a prohibited area;
• suspicious vehicles loitering;
• unattended objects; and
• events associated with an emergency.

These are examples rather than an exhaustive list of every analytics function available today.

Does the SPF Standard require AI CCTV?

No universal AI requirement appears in the document.

The Standard discusses video analytics and refers to rule-based detection.

It does not say every VSS must use artificial intelligence, facial recognition or advanced machine learning.

Where we discuss newer AI capabilities separately, those would need to be clearly distinguished from what this particular source states.

Can video analytics replace CCTV operators?

The VSS Standard does not say that.

It describes analytics as a tool that can alert the operator, particularly where one operator is responsible for many cameras.

That framing is important.

The source presents analytics as support for the monitoring function rather than a replacement for human security decision-making.

What is the main value of video analytics?

Its value is not necessarily that the system “understands” the entire scene.

Its practical value can be:

drawing the operator’s attention to a potentially important event.

For example, instead of expecting an operator to continuously notice movement across 60 passive feeds, the system might alert them when:

a person crosses a predefined boundary after hours.

The operator can then examine the video and determine whether a response is required.

Can video analytics generate false alarms?

The cited Video Analytics section does not provide detailed false-alarm performance criteria.

However, because the Standard describes analytics as generating alerts for the operator, the effectiveness of those alerts will depend on how the rules are configured and how the surrounding environment behaves.

Consultant’s perspective

Analytics should therefore be commissioned against the actual scene and operating conditions, not merely enabled using factory defaults.

If every harmless activity creates an alert, the operator may eventually stop treating alerts as meaningful.

That implementation observation goes beyond the specific wording in the source.

Should video analytics be configured the same way for every camera?

The Standard says common rule-based violations may be programmed for each scene of interest.

This implies that analytics should reflect the purpose of the individual scene.

For example:

A virtual tripwire might make sense along a perimeter.

Loitering analytics might make sense around a prohibited area.

An unattended-object rule may be relevant in another environment.

The event rule should correspond to the security problem being addressed.

What is CCTV system integration?

The Standard discusses the advantage of integrating newer IP-based CCTV systems with technologies such as:

video analytics, intrusion detection and access control.

The objective is not simply technical connectivity.

Integration should improve security operations.

What can an integrated security system do?

The Standard gives a very practical example.

When a security alarm activates, an integrated system can:

notify the security officer → display the alarm location on a site plan → display the relevant camera for verification.

That is a useful model for security integration.

The operator receives:

event + location + visual verification.

Why is automatic camera pop-up useful?

Consider an intrusion alarm at a door.

Without integration, the operator may need to:

1. read the alarm;
2. determine where the door is;
3. find which camera covers it;
4. search the VMS;
5. display that camera; and
6. determine what happened.

With integration:

alarm occurs → relevant camera appears automatically.

The Standard specifically requires this approach for intrusion-alarmed doors associated with Sensitive Rooms.

That can reduce the time needed to verify an incident.

Does every alarm need to trigger CCTV?

The Standard gives specific examples and recognises the value of alarm/CCTV integration, but it does not say that every event from every building system must automatically produce a camera pop-up.

Consultant’s perspective

The best candidates are events where visual verification materially helps the security response.

Examples might include:

unauthorised door opening, perimeter intrusion or other location-specific security alarms.

The exact integration matrix should be designed around the Operational Requirements.

How should CCTV integrate with access control?

The source states that VSS can be integrated with access-control systems and gives the broader example of alarms generating operator notification, map location and camera display.

The Standard does not provide a complete access-control event matrix.

Consultant’s perspective

Possible project-specific integrations could include visually verifying events such as:

unauthorised door opening or other defined access-control alarms.

But those specific workflows should be identified as project design choices unless expressly stated by the source.

How should CCTV integrate with intrusion detection?

This is more explicit.

The Standard says newly designed systems can integrate with intrusion detection, and for intrusion-alarmed Sensitive Room doors it says activation should automatically show the associated camera.

The operational chain becomes:

Detect → Alert → Locate → Display → Verify → Respond

That is much stronger than treating the intrusion alarm and CCTV as separate systems.

Does integration mean all systems must be from the same manufacturer?

The VSS Standard does not say this.

It elsewhere promotes IP-system interoperability, including ONVIF or equivalent approaches for compatible physical-security products.

The integration section also says the System Integrator should provide an SDK supporting commands including:

Select Camera, View, Extract, PTZ and Playback.

This indicates an interest in systems being capable of interacting, rather than prescribing one single manufacturer.

What is an SDK and why does the Standard mention it?

SDK means Software Development Kit.

The Standard says the System Integrator should provide an SDK supporting CCTV commands including:

Select camera, View, Extract, PTZ and Playback.

In practical terms, an SDK can allow other authorised software or integrated systems to interact with the VSS.

That may support future:

integration, incident-management functions or custom interfaces.

Does the Standard support monitoring across adjacent buildings?

Yes.

Where adjacent buildings have the same proprietor, the Standard recommends that each building’s VSS have the capability to access images from the adjacent locations.

This can assist coordinated operations across a multi-building estate or campus.

The source does not prescribe the detailed network architecture required to achieve this.

Should emergency responders have access to CCTV?

The Standard recommends providing a reserve viewing terminal for Emergency Response Agencies to assist with incident management.

This is a notable design requirement because it recognises that CCTV can support more than routine security operations.

During a serious incident, external responders may also require access to visual information.

Does that mean emergency responders should use the operator’s workstation?

The Standard specifically refers to a reserve viewing terminal, which suggests separate viewing capability rather than simply taking over the operator’s normal station.

Consultant’s perspective

This can help maintain normal security operation while giving responding agencies access to the information they need.

The exact location and permission structure would be project-specific.

Should CCTV integrate with a site map?

The Standard specifically gives the example of integrated security systems showing the location of an alarm on a site plan.

This can be particularly useful in large or complex buildings.

An operator may not instantly know the location of:

Door D-07-145.

But a highlighted location on a floor plan can immediately establish:

where the event is occurring

and

which response route may be relevant.

What is better: showing every camera or showing cameras by exception?

The Standard does not use those exact terms.

However, it provides both:

general live monitoring capabilities

and

analytics/alarm-driven alerting.

Consultant’s perspective

This suggests two complementary monitoring approaches:

Routine situational monitoring
Operators view selected cameras, sequences and multi-view layouts.

Exception-driven monitoring
An alarm or analytic event draws attention to a particular location and relevant camera.

A mature control-room design can use both instead of expecting operators to stare continuously at every camera.

Does a multi-view display reduce image detail?

When several cameras are shown on one screen, each camera occupies a smaller portion of the display.

The Standard does not provide a specific maximum number of tiles per monitor.

Instead, it requires the number and presentation of images to follow the monitoring task and operational requirements, while separately emphasising correct monitor size and viewing distance.

Consultant’s perspective

If an operator needs to make a detailed assessment, the relevant camera should be easily enlarged rather than leaving it as one small tile among dozens.

Is a 64-camera matrix on one screen useful?

The Standard does not prescribe or prohibit a 64-camera layout.

The real test is whether the operator can perform the required surveillance task.

The number of views should reflect:

risk, purpose, scene activity, incident frequency, viewing duration, other operator tasks and operator competence.

So:

“We can technically display 64 cameras”

does not prove:

“One operator can effectively monitor 64 cameras.”

Should operators manually cycle through cameras?

The Standard allows scanning sequences with adjustable dwell times.

However, it also expects rapid manual camera selection and supports integration and analytics.

So scanning is only one available method.

Consultant’s perspective

Critical cameras should not necessarily disappear from the operator’s awareness simply because an automatic sequence has moved on to the next scene.

The display strategy should reflect the risk and required monitoring objective.

What should happen when a video analytic alert occurs?

The Standard does not provide a complete response SOP.

It describes analytics as alerting the operator to security events.

A practical operational workflow could therefore be:

Analytic detects condition → operator receives alert → relevant camera displayed → operator verifies → applicable SOP followed.

This is consultant implementation guidance built around the source’s monitoring philosophy.

Should analytics automatically unlock or lock doors?

The VSS Standard does not state this.

It discusses integrating VSS with access control and intrusion detection but does not prescribe automatic door-control actions based solely on video analytics.

Any such function would require separate:

risk assessment, access-control design, life-safety coordination and operational approval.

We should not infer it from this document.

Can analytics make an incorrect decision?

The source does not provide performance thresholds or accuracy rates for video analytics.

It simply identifies VA as a useful detection tool and refers readers to Singapore Standards Technical Reference TR 69:2019 for deployment details.

So this article should not claim specific detection accuracy or reliability figures from the VSS Standard.

Does the VSS Standard cover facial recognition?

The sections used for this article do not prescribe facial-recognition analytics.

The Standard’s use of Recognition and Identification in its Field of View section refers to the amount of visual detail in an image, not necessarily automated facial-recognition software.

That distinction is important.

Identification-level CCTV ≠ facial-recognition technology.

Does “Recognise” mean facial recognition AI?

No.

In the VSS Field of View terminology, Recognise refers to an observer being able to determine with a high degree of certainty whether the person is someone they have previously seen.

It does not mean that facial-recognition software must be installed.

This distinction avoids confusing:

image-quality performance

with

biometric analytics.

What does an effective CCTV monitoring workflow look like?

A practical workflow could be:

Routine state

Operators monitor selected overview cameras and relevant security locations.

Event detection

The event may be detected through:

an operator, video analytics, access control or intrusion detection.

Alert

The system notifies the operator.

Location

The system identifies or displays where the event occurred.

Visual verification

The relevant camera automatically appears or is rapidly selected.

Assessment

The operator determines whether the event appears genuine and what is occurring.

Response

The appropriate security SOP is initiated.

Recording and investigation

The associated footage can later be retrieved and exported.

This is a consultant’s implementation model based on the functions described in the VSS Standard.

Example: intrusion at a Sensitive Room

Consider an NVR room protected by an intrusion alarm.

The Standard calls for:

Observation coverage of the external access area and, where the door has an intrusion alarm, automatic display of the associated camera when the alarm activates, with the door view at Identification level.

A well-integrated event might therefore operate as follows:

Door alarm → operator alert → room highlighted → Identification camera appears → operator verifies person/event → response initiated.

The operator does not need to search manually through the CCTV system while the incident is unfolding.

Example: perimeter analytics

Suppose a perimeter camera is configured with a virtual tripwire.

The Standard specifically identifies tripwire intrusion as an example of a video-analytics event.

A practical sequence could be:

Person crosses defined line → VA generates alert → perimeter camera displayed → operator checks whether entry is authorised → response follows if necessary.

The value of the analytics is therefore not merely generating another alarm.

It is focusing operator attention on an event that might otherwise be missed.

Example: loitering near a restricted area

The Standard also mentions suspicious person or vehicle loitering in prohibited areas as possible video-analytics rules.

The Operational Requirement might therefore be:

Alert the operator when a person remains within the prohibited area beyond the defined operational condition.

The exact duration, zone and response are project-specific and are not prescribed by the source.

How should camera layouts on monitors be designed?

The Standard does not prescribe one universal screen layout.

It says the number and presentation of images should follow:

security and operational requirements

and the characteristics of the monitoring task.

Consultant’s perspective

Useful display groups may correspond to operational zones such as:

Perimeter
Main Entrances
Loading Bay
Critical Areas
Car Park

rather than arranging cameras purely according to camera number.

This can reduce the mental effort required to understand where an event is occurring.

Should the CCTV viewing arrangement be tested?

Yes.

The Standard specifically says performance evaluations should be carried out periodically and when there is a significant change in:

the viewing task

or

the control-room setup.

This provides a basis for asking:

Can the operator still effectively perform the monitoring role after the system changes?

What changes should trigger a monitoring review?

The Standard does not give an exhaustive list.

Examples of significant operational changes could include:

• adding many more cameras;
• changing monitor layouts;
• adding analytics;
• assigning new duties to operators;
• changing the control-room arrangement; or
• changing the security function of the facility.

Those examples are consultant interpretations of the source’s requirement to reassess performance after significant changes.

CCTV monitoring is not the same as CCTV recording

This distinction runs throughout the VSS Standard.

Recording ensures information is available afterwards.

Monitoring helps the security operation act while the incident is occurring.

A system can be excellent at recording and poor at live response.

For example:

The camera records a perimeter intrusion perfectly, but nobody becomes aware of it until the next morning.

The footage may be useful for investigation.

But the opportunity for immediate security response has already passed.

Common misunderstandings about CCTV monitoring and integration

“One operator should be able to monitor a fixed number of cameras.”
The Standard provides no universal ratio. The number depends on risk, monitoring purpose, activity, incident frequency, other duties and operator competence.

“More monitors mean better security.”
Not necessarily. Monitor size, viewing distance, image presentation and operator workload all matter.

“If all cameras are shown on the video wall, they are being monitored.”
Displaying an image does not prove the operator can meaningfully observe it.

“Video analytics replaces the operator.”
The Standard describes VA as a tool to alert the operator, not as a universal replacement for the operator.

“Recognition means facial-recognition software.”
No. Recognition is a Field of View image-detail category in the Standard.

“Integration means putting CCTV and access control on the same network.”
Not necessarily. The useful integration described by the Standard produces operational outcomes such as event notification, map location and relevant camera display.

“Alarm integration is just a convenience.”
It can materially reduce the time required for an operator to locate and visually verify an incident.

“Every alarm should automatically perform security actions.”
The Standard supports information integration but does not prescribe automatic responses such as locking or unlocking doors based solely on CCTV analytics.

“A huge multi-view display is always better.”
The number of views should match what the operator can effectively use for the monitoring task.

Practical CCTV Monitoring and Integration Review

A project team can review the VSS using the following questions:

1. Where are live cameras monitored?
2. What security role does each operator perform?
3. What other tasks compete for the operator’s attention?
4. How many live images are normally presented?
5. Why were those particular camera views selected?
6. What is the consequence if an event on one of those cameras is missed?
7. Can any camera be selected quickly?
8. Can any camera be displayed on any incident-monitoring screen?
9. Can several users view the same feed simultaneously?
10. Are camera IDs, dates and times clearly displayed?
11. Are camera names meaningful to operators?
12. Can recorded footage be retrieved quickly using the same naming structure?
13. Are monitor sizes appropriate to their viewing distance?
14. Can operators actually perceive the detail required from their normal seat?
15. Are important cameras displayed at a useful size?
16. Can the operator enlarge a camera rapidly during an incident?
17. Are scanning sequences appropriate for the surveillance task?
18. Is the number of views per operator based on risk and workload rather than convenience?
19. Has operator performance been evaluated?
20. Has the arrangement been reviewed after cameras or operator duties changed?
21. Are video analytics being used where they solve a defined monitoring problem?
22. What event is each analytic rule intended to detect?
23. Does each analytic alert identify the correct camera and location?
24. Are intrusion alarms linked to relevant CCTV views where required?
25. Does access-control integration provide useful visual verification?
26. Can a site plan show where an alarm is occurring?
27. Is there unnecessary operator effort between receiving an alarm and seeing the relevant image?
28. Can adjacent-building cameras be accessed where the same proprietor requires coordinated monitoring?
29. Is a reserve viewing terminal available for Emergency Response Agencies as recommended?
30. Have monitoring workflows been tested using realistic security scenarios?

This is a consultant’s review framework derived from the VSS Standard rather than a checklist reproduced from it.

Conclusion

The SPF VSS Standard does not treat CCTV monitoring as:

“Put all the cameras on a wall and ask a guard to watch them.”

Instead, it recognises that surveillance effectiveness depends on:

what images are presented, how many images an operator is expected to handle, how quickly relevant cameras can be retrieved, what information accompanies the image, and whether alarms or analytics can help direct attention to the correct scene.

The Standard deliberately does not set a universal camera-per-operator number.

Instead, it asks designers to consider:

risk → monitoring purpose → scene activity → incident frequency → operator workload → operator competence.

Video analytics can then help focus attention on events such as:

tripwire intrusion, loitering and unattended objects.

And security-system integration can transform an alarm into:

Alert → Location → Relevant Camera → Verification → Response.

The central lesson is therefore:

The purpose of CCTV monitoring is not to maximise the number of images an operator can see. It is to minimise the time between an important event occurring and the security operation understanding what is happening.