What is the practical answer?
Use H.265 when your recorder, live-view clients and exported evidence all support it and a scene test confirms worthwhile savings. Keep H.264 when compatibility matters more. Compare matched image quality, not codec labels or promised storage percentages.
| Decision | H.264 / AVC | H.265 / HEVC |
|---|---|---|
| Compatibility | Usually the safer baseline for older equipment; verify exact clients | Verify recorder, browser, app, decoder and export player |
| Storage planning | Measure representative bitrate | May lower bitrate at comparable quality; measure the saving |
| Live grids | Often easier on older viewing hardware | Check simultaneous decode capacity, not just recording support |
| Evidence handover | Test native and portable exports | Confirm recipients can play originals; retain originals when transcoding |
| Upgrade decision | Keep if the workflow already meets retention and quality | Pilot one channel before applying to all cameras |
H.265 can reduce the data needed for comparable video, but a codec name does not guarantee a storage percentage or better evidence. Scene movement, noise, rate control, keyframes, decoder support and the camera’s own implementation determine the real result.
What changes between H.264 and H.265?
H.264/AVC and H.265/HEVC are video coding standards, not fixed quality settings. H.265 uses more advanced prediction and block structures to encode video more efficiently, especially at higher resolutions, but it generally demands more processing and may be less compatible with older recorders, browsers, phones and evidence players. The ITU lists current in-force editions of both standards; neither has become invalid simply because another is newer. In a CCTV project, choose the codec that every critical part of the workflow can record, display, export and replay. A small storage saving has little value if an investigator cannot open the file on the required machine.
How much storage will H.265 actually save?
Do not promise ‘half the storage’ from the codec label. Savings vary with scene complexity, implementation and target quality. A quiet corridor with a stable background compresses differently from rain, foliage, traffic, sensor noise or a constantly moving PTZ. Manufacturer-enhanced modes may use dynamic frame rate, dynamic group-of-pictures and region-aware compression to save more in surveillance scenes; those figures are vendor measurements and should be validated on the exact model and firmware. Run parallel test profiles where possible, capture representative day and night periods, and compare both measured bitrate and moving-subject quality before changing an entire estate.
Should you use CBR or VBR?
Bitrate control determines how the encoder spends data. Variable bitrate (VBR) allows demand to rise for complex scenes and fall when little changes, which usually protects quality efficiently but creates storage and network peaks. Constant bitrate (CBR) targets a steadier rate, useful where links have hard limits, yet may waste data in quiet scenes or lose detail when complexity exceeds the budget. Maximum bitrate caps can make capacity planning safer, but a cap set too low forces visible damage exactly when rain, crowds or motion make the scene important. Use a quality target plus a defensible maximum, monitor real percentiles and leave headroom rather than sizing links to the average alone.
How do you calculate retention from bitrate?
Storage arithmetic starts with measured aggregate bitrate. Approximate daily storage in gigabytes as bitrate in megabits per second multiplied by 10.8. A camera averaging 4 Mb/s therefore uses about 43.2 GB per day before container, database and operational overhead; eight such cameras use about 346 GB per day. Multiply by retention days, then add free-space and resilience allowances. Motion recording changes the average only after the real activation ratio is measured across weather, lighting and seasonal activity. Calculate the main recording stream separately from substreams, audio and redundant copies. Never plan from resolution alone.
Why do keyframes affect playback and recovery?
GOP structure affects recovery, seeking and failure behaviour. An I-frame is independently decodable; following predicted frames depend on other frames. Longer intervals can improve efficiency but may increase the time to a clean image after packet loss and make precise seeking or export less responsive. Some smart codecs extend the GOP dramatically when a scene is quiet, then shorten it around motion. That can work well, provided the NVR and player understand the stream and event footage has been tested. For evidential systems, verify playback at the start of exported clips, frame stepping around the event and recovery after a simulated link interruption.
What should you compare after dark?
Image quality must be judged on movement at the evidence zone. Pause a walking face, moving plate or hand at the till and compare it with the operational requirement. Blockiness, mosquito noise and smeared texture can come from too little bitrate, aggressive noise reduction, slow shutter or a combination; increasing bitrate cannot recover detail already blurred by exposure. At night, sensor noise consumes data and can make a previously stable VBR stream surge. Tune lighting and exposure first, then codec and bitrate. Keep the original resolution and stream settings documented so a later firmware update or factory reset cannot silently change retention or quality.
How can you migrate without breaking playback?
A safe migration is staged. Confirm that cameras, NVR/VMS, workstation decoders, mobile apps, export tools and any integration all support the selected H.265 profile. Enable it on one representative channel, record through day and night, test live grids, remote viewing, event search, export and third-party playback, then compare seven-day bitrate and storage with the H.264 baseline. Retain H.264 for devices or recipients that require it, and consider a lower-resolution substream for multi-camera viewing. Choose H.265 when measured savings survive the complete workflow; choose H.264 when interoperability and low decode load are more valuable.
Which settings belong in the test record?
Document the codec decision as a measured baseline. For each test channel, record resolution, frame rate, shutter limit, codec profile, rate-control mode, quality level, bitrate cap, I-frame interval, audio setting and smart-codec features. Capture average, 95th-percentile and peak bitrate during comparable time windows, then calculate expected retention with the same storage reserve. Keep short exported clips of quiet daylight, busy movement and difficult night conditions and verify them on a computer that is not the recorder. This evidence makes future tuning safer: if retention suddenly falls after a firmware update, the maintainer can compare settings and scene behaviour instead of reducing quality blindly. Re-run the baseline when lighting, foliage, traffic or camera position changes because the scene is part of the compression system.
What does maximum bitrate reached mean?
A recorder warning such as ‘maximum bitrate reached’ means the aggregate incoming streams have exceeded the NVR’s stated or practical decoding budget; it does not automatically mean every camera is configured incorrectly. Add the configured or measured bitrate of active channels, then compare it with the recorder’s incoming-bandwidth limit and any separate live-view or playback limit. Reduce one representative stream deliberately—prefer VBR with a sensible cap, an appropriate frame rate and H.265 only when the complete workflow supports it—then confirm moving evidence before applying the change elsewhere. If the recorder lists channels to close, redistribute cameras, lower non-critical stream demand or move to a recorder with verified headroom instead of suppressing the warning by cutting every channel blindly.
How do you run a fair codec comparison?
A useful comparison starts with a worksheet and a repeatable event. Choose a camera that represents the difficult part of the installation, such as a gate with moving foliage and passing vehicles. If the camera supports simultaneous profiles from the same source frames, record both codecs with matched resolution, frame rate, exposure and quality goals. Otherwise run repeated controlled walks in comparable light and acknowledge that the scenes are not identical. Record which hardware and software decoded each export. Inspect faces, carried objects and the moment a subject enters the frame; attractive empty-scene screenshots do not answer those questions. Do not call a lower bitrate a saving until the detail requirement survives. Keep the clips and configuration files so a second person can repeat the comparison after a firmware change.
When is the measured saving worth the change?
Here is a planning example, not a measured product result: suppose four cameras average 4 Mb/s each with one profile and 3 Mb/s each with another, while both pass the same scene test. Thirty days of continuous video requires 5.184 TB at the first rate and 3.888 TB at the second, using decimal units and excluding overhead. That is a calculated 25 percent difference for these assumed rates; it is not an H.265 guarantee. Apply the same reserve to both results before comparing disks. An installer should also price any client upgrade, decoder licence or support burden needed to use the second profile. If the new profile breaks reliable event playback or makes exports difficult, keep the working profile and address storage through capacity, retention policy or a better camera view. The lowest bitrate is only useful when it preserves the recording's purpose.
Quick estimate: 1 Mb/s sustained for 24 hours is about 10.8 GB. Multiply the measured average Mb/s by 10.8, then by camera count and retention days; add recorder overhead and free-space margin.
Before you sign off
- Verify codec support across the complete evidence workflow
- Measure representative day and night bitrate
- Set quality and maximum bitrate deliberately
- Calculate storage from aggregate measured Mb/s
- Test keyframe seeking and recovery after packet loss
- Inspect moving evidence—not static screenshots
- Migrate one channel before changing the estate
Can another person prove the system still works?
Record the final view, night image, bitrate, alerts, firmware and access method. A system is not finished until the owner can verify recording and export without the installer standing beside it.
Every site is different. Confirm manufacturer instructions, electrical requirements, privacy obligations and local regulations before installation.
Sources used for this guide
Primary standards and manufacturer documentation are linked so you can verify thresholds, features and current product behaviour.
Frequently asked questions
Is H.265 always better for security cameras?
No. It can reduce bitrate at comparable quality, but the recorder, viewing clients and export workflow must support it. Keep H.264 when compatibility or client capacity makes it the more dependable choice.
Will H.265 cut storage use in half?
There is no universal percentage. Match image quality and measure representative scenes on the exact camera; lighting, motion, encoder implementation and smart-codec settings affect the result.
Can I record H.265 but use H.264 for live viewing?
Some cameras support separate main and substreams with different codecs. Confirm the exact model's stream combinations and verify that the recorder saves the intended main stream.
Why will an H.265 export not play?
The player or device may lack support for the codec, profile or export container. Test an approved compatible player and retain the original recording if you create a separate viewing copy.
Does changing codec improve a blurry night image?
Not if detail was already lost through slow shutter, missed focus or inadequate light. Fix capture conditions first, then compare compression using recorded moving subjects.