Use VBR for most exported files, uploads, downloads, and video-on-demand. Use CBR when a live platform, fixed-bandwidth network, or delivery specification requires a predictable data rate.
Constant bitrate gives each part of a video roughly the same bitrate budget. Variable bitrate gives simple scenes fewer bits and complex scenes more. That usually makes VBR more efficient for finished files, while CBR remains useful when stable throughput matters more than getting the best quality per megabyte.
Quick takeaways
- Choose VBR for normal MP4 exports, website videos, social uploads, VOD, downloads, and storage.
- Choose CBR for live streaming when the platform specifies it, or for a truly fixed-bandwidth delivery path.
- At the same average bitrate, VBR usually preserves complex scenes better and wastes fewer bits on static scenes.
- CBR makes bandwidth and approximate file growth easier to predict, but it does not make every frame the same size.
- Two-pass VBR is useful when you need efficient bit allocation around a chosen average bitrate or target file size and can wait longer.
- CRF is different: it targets quality and lets bitrate—and therefore file size—vary.
- Resolution alone does not determine bitrate. Motion, noise, frame rate, codec, and acceptable quality all matter.
What is CBR?
CBR stands for constant bitrate. The encoder tries to keep the output data rate at or near a selected value over time.
If you set a video to 5 Mbps CBR, the encoder has approximately 5 megabits each second for the video stream. A static presentation slide and a fast-moving sports scene receive similar rate budgets. That predictability is useful for live transmission and constrained networks, but it is not always an efficient way to distribute quality.
Simple scenes may receive more data than they need. Complex scenes may not receive enough, causing blockiness, smearing, or lost detail.
“Constant” should not be interpreted too literally. Individual frames still vary in size, and encoder buffers smooth the rate over a window of time. Some implementations add filler data to maintain strict constant output. The practical behavior depends on the encoder and its buffer settings.
What is VBR?
VBR stands for variable bitrate. The encoder changes the data rate according to the complexity of the video.
A mostly static interview may need relatively few bits. A cut to confetti, water, foliage, camera noise, or rapid motion may need many more. VBR can save bits during easy sections and spend them where compression is more difficult.
AWS's rate-control guidance explains the trade-off directly: with VBR, an encoder allocates more bits to complex parts of the video, while CBR uses the same number of bits for all parts. AWS recommends constant rate only when the bitrate needs to remain constant, such as on a limited fixed-bandwidth network.
CBR vs. VBR comparison
| Question | CBR | VBR |
|---|---|---|
| Bitrate over time | Kept near a fixed target | Rises and falls with scene complexity |
| Quality at the same average bitrate | Usually less efficient | Usually more consistent |
| Bandwidth predictability | Higher | Lower unless peaks are capped |
| File-size predictability | Straightforward from bitrate and duration | Predictable when average bitrate is controlled; less so in quality-targeted modes |
| Encoding speed | Often one pass | One pass or slower two pass |
| Best fit | Live and fixed-bandwidth delivery | Files, uploads, downloads, VOD, archives |
| Risk | Wasted bits or poor complex scenes | Peak bitrate may exceed a delivery constraint |
The key question is not “Which mode is better?” It is which variable must remain stable:
- If the network rate must remain stable, favor CBR.
- If quality per byte matters more, favor VBR.
- If visual quality must remain stable and final size may vary, use a quality-targeted mode such as CRF.
- If final size must stay under a hard cap, control average bitrate or use a target-size workflow.
When should you use CBR?
Live streaming to a platform that asks for CBR
Live encoders cannot analyze the entire program in advance, and streaming services need manageable, predictable input. YouTube's current live encoder settings specify CBR for bitrate encoding.
Follow the platform's requirements rather than overriding them based on a generic quality rule. The platform may re-encode your stream, enforce keyframe intervals, or reject settings outside its ingest profile.
Fixed-bandwidth contribution or distribution
CBR can be appropriate when a channel has a hard, continuous capacity limit and cannot tolerate peaks. Broadcast links, managed networks, and some legacy systems fit this pattern.
Even then, leave network headroom. A 6 Mbps video stream does not mean a 6 Mbps connection is sufficient after audio, protocol overhead, retransmission, and connection variability.
A specification explicitly requires it
Some delivery contracts, hardware decoders, or production pipelines require constant-rate input. In that case, compliance is more important than the storage efficiency VBR could provide.
When should you use VBR?
Uploading a finished video
For a normal upload, the destination receives a completed file rather than a live feed. The encoder can vary bitrate without depending on a real-time connection. YouTube's recommended upload settings specify variable bitrate for uploaded H.264 video, while its live guidance specifies CBR. That contrast captures the core use cases well.
Video-on-demand and website delivery
VBR is usually the better choice for downloadable files and on-demand playback because it can produce more consistent quality for a similar average bitrate. Set a sensible maximum bitrate if the player or CDN path must avoid large peaks.
Recording and archiving
A recording may contain long static periods interrupted by bursts of motion. VBR avoids spending a constant high rate during the easy sections. For an important master, prioritize sufficient quality rather than forcing the smallest possible file.
Compressing to a known approximate size
Average-bitrate VBR is useful when duration and desired file size are known. The encoder can distribute a fixed total bit budget more intelligently than CBR, especially with two-pass encoding.
One-pass VBR vs. two-pass VBR
One-pass VBR
The encoder decides how to allocate bits as it processes the video. It is faster and is often good enough for routine exports, rapid publishing, and workflows where processing time matters.
Two-pass VBR
The first pass analyzes scene complexity. The second pass uses that information to distribute the target bit budget across the whole video. It can spend less on easy sections and reserve more for difficult sections.
Use two-pass VBR when:
- You need to fit an approximate target size.
- The average bitrate is constrained.
- Quality per byte matters more than encode time.
- The encoder's one-pass rate control struggles with your content.
Do not use two-pass merely because “two is better.” It doubles much of the work, and quality-targeted modes may be simpler when final file size is flexible.
How is CRF different from VBR?
CRF, or Constant Rate Factor, is a quality-targeted mode used by encoders such as x264 and x265. It allows bitrate to vary, so it behaves like variable-rate encoding in a broad sense, but you choose a quality target instead of an average bitrate.
That changes the promise:
- Average-bitrate VBR: “Try to average this many bits per second.”
- CRF: “Try to maintain this quality level; use the bitrate the content needs.”
CRF is excellent when quality is the priority and the final file size can float. It is a poor choice when a file absolutely must stay below 20 MB, 25 MB, or another upload limit because the size is not known until encoding finishes.
See CRF explained: what number should you use? for practical values and trade-offs.
Which mode produces a smaller file?
Neither mode automatically produces a smaller file. The result depends on the settings.
At the same average bitrate and duration, CBR and average-bitrate VBR should have broadly similar total video data. VBR's advantage is where those bits go. It can spend fewer on simple sections and more on demanding ones, often improving quality without increasing the average budget.
A quality-targeted VBR or CRF encode can be much smaller or larger because there is no fixed average bitrate target. An easy screen recording might compress exceptionally well; noisy concert footage might require far more data at the same quality setting.
How to calculate bitrate for a target file size
For a target-size encode, start with this approximation:
total bitrate in kbps ≈ target size in MB × 8,000 ÷ duration in seconds
video bitrate ≈ total bitrate − audio bitrate
Example: a 3-minute video must stay under 25 MB.
25 × 8,000 ÷ 180 ≈ 1,111 kbps total
1,111 − 128 kbps audio ≈ 983 kbps video
Do not set the video to exactly 983 kbps and assume the result will be safely below 25 MB. The container, metadata, and rate-control variance add overhead. A practical workflow reserves a safety margin and verifies the output.
At roughly 1 Mbps, detailed 1080p footage may look poor. Reducing the output to 720p can produce a cleaner image because fewer pixels share the same bit budget. This is why “keep 1080p at any cost” is often the wrong strategy for a strict file-size limit.
VidSmaller's target-size video compressor performs this calculation and adjustment for you. For automated jobs, the video compression API accepts a target_size_mb value.
CBR or VBR for shrinking 1080p video?
For a finished 1080p file, start with VBR or CRF, not CBR.
- Use CRF if you want a chosen quality level and can accept an unpredictable size.
- Use average-bitrate VBR, preferably two pass, if you need an approximate size.
- Use a target-size workflow if the result must remain below a hard upload limit.
- Use CBR only when the delivery system requires stable bitrate.
1080p is only the frame size. A talking head against a plain wall may look good at a much lower bitrate than handheld footage of trees, water, crowds, film grain, or gameplay. Frame rate matters too: 1080p60 normally needs more data than 1080p30 at similar quality.
If the calculated bitrate is too low for the content, reducing resolution to 720p often looks better than preserving 1080p with severe artifacts.
CBR or VBR for 720p H.264?
The same rule applies: choose based on delivery, not on resolution.
- 720p H.264 live stream: use CBR if the streaming platform calls for it.
- 720p H.264 file or upload: use VBR for better bit allocation.
- 720p file with a strict size limit: use average-bitrate VBR or a target-size encoder.
- 720p archive where quality matters most: use CRF or another quality-defined mode at a sensible level.
YouTube's current upload guidance lists reference bitrates by resolution and frame rate, but those values are upload recommendations—not universal minimums or ideal compression settings for every destination. A messaging attachment, website preview, editing master, and live stream have different requirements.
What should you set for maximum bitrate?
A maximum bitrate limits how high VBR may spike. It can protect a constrained decoder or delivery path, but setting it too close to the average removes much of VBR's flexibility.
There is no universal ratio. Check the destination specification first. If it gives an average and maximum, use those numbers. If it gives only a connection requirement, remember that audio and protocol overhead also consume capacity.
For files played locally or uploaded for another platform to transcode, a restrictive peak cap may be unnecessary. For direct web delivery, a reasonable cap can make buffering behavior more predictable.
Common mistakes
Choosing CBR because the file size must be exact
CBR makes the size easy to estimate, but it may allocate bits inefficiently. Two-pass average-bitrate VBR is often a better way to use a known total bit budget. Always leave a margin when the limit is strict.
Choosing VBR without setting a peak for a constrained network
An acceptable average does not guarantee acceptable short-term peaks. Set a maximum bitrate and buffer appropriate to the delivery system.
Treating bitrate recommendations as universal quality laws
A recommended 1080p bitrate is a starting point for a particular platform, codec, frame rate, and workflow. It is not proof that every 1080p video needs the same amount of data.
Ignoring audio
Audio counts toward file size and transmission rate. A 320 kbps track consumes 24 MB over ten minutes before counting any video data. For speech and ordinary web video, a lower AAC bitrate may leave more budget for the picture.
Re-encoding an already damaged source
Rate control cannot restore lost detail. Start from the best source available and avoid repeated generations. See the full guide to compressing video without losing visible quality.
Frequently asked questions
Is CBR better quality than VBR?
Usually not at the same average bitrate. VBR can allocate more data to difficult scenes and less to simple ones, which generally improves quality consistency. CBR's advantage is predictable rate, not maximum quality per byte.
Is VBR good for streaming?
It depends on what “streaming” means. VBR is common for encoded video-on-demand files and adaptive streaming renditions. Live ingest platforms may require or recommend CBR because real-time bandwidth needs to remain predictable. Follow the destination's specification.
Does VBR take longer to encode?
One-pass VBR does not necessarily add much time. Two-pass VBR takes longer because the encoder analyzes the video before the final encode.
Is CBR exactly constant?
Not always at every instant. Encoders use buffers and rate-control windows, and frame sizes vary. Strict CBR implementations may use filler data. Treat CBR as a constrained delivery rate, not identical bytes for every frame.
Which is better for YouTube: CBR or VBR?
For uploaded files, YouTube's recommended settings specify variable bitrate. For live encoder input, YouTube specifies CBR. Check current platform guidance because requirements can change.
Which is better for Discord or email attachments?
Use a target-size or average-bitrate VBR workflow. The hard constraint is the final file size, and VBR can distribute that limited bit budget more efficiently. For email specifically, remember that attachment encoding can make a file larger in transit; see why a 25 MB email limit may mean a smaller practical video file.
Sources and further reading
- AWS Elemental MediaConvert: Comparison of bitrate control modes
- YouTube: Recommended upload encoding settings
- YouTube: Live encoder settings, bitrates, and resolutions
- Microsoft Media Foundation: Encoding methods
Technical facts and linked platform guidance were reviewed on October 10, 2026.