Understand video bitrate and how it impacts quality, file size, and streaming performance. Learn about CBR, VBR, CRF modes, bitrate recommendations by resolution, and codec efficiency comparisons.
Bitrate is the amount of data used to represent one second of video. It is measured in bits per second, typically expressed as kilobits per second (kbps) or megabits per second (Mbps). A video with a bitrate of 5 Mbps uses 5 million bits of data for every second of footage.
Bitrate is the single most important factor in determining video quality and file size. Double the bitrate, and you roughly double the file size. But the relationship between bitrate and perceived quality is not linear: going from 1 Mbps to 2 Mbps is a dramatic improvement, while going from 10 Mbps to 20 Mbps is barely noticeable.
Understanding how bitrate works is essential for anyone who converts, compresses, streams, or uploads video. This guide explains everything you need to know.
Video codecs work by analyzing frames and discarding information that the human eye is unlikely to notice. At higher bitrates, the codec keeps more information and the result looks closer to the original. At lower bitrates, the codec must discard more, which introduces visible artifacts:
The exact bitrate where artifacts become visible depends on the resolution, codec, frame rate, and content complexity.
Video encoders offer different bitrate control modes, each with distinct advantages.
CBR maintains the same bitrate throughout the entire video. Every second uses approximately the same amount of data, regardless of scene complexity.
# CBR encoding at 5 Mbps
ffmpeg -i input.mp4 -c:v libx264 -b:v 5M -minrate 5M -maxrate 5M \
-bufsize 10M -c:a aac -b:a 192k output.mp4
Best for: Live streaming, hardware playback with limited buffers, broadcast.
VBR adjusts the bitrate dynamically based on scene complexity. Simple scenes (static backgrounds, talking heads) use less data, while complex scenes (action, fast motion) use more.
# VBR encoding with target 5 Mbps, max 8 Mbps
ffmpeg -i input.mp4 -c:v libx264 -b:v 5M -maxrate 8M \
-bufsize 10M -c:a aac -b:a 192k output.mp4
Best for: Video files for download, archival, general purpose.
CRF is a quality-targeted mode. Instead of specifying a bitrate, you specify a quality level and the encoder uses whatever bitrate is needed to achieve it. Lower CRF values mean higher quality and larger files.
# CRF encoding (quality-targeted)
ffmpeg -i input.mp4 -c:v libx264 -crf 20 -preset slow \
-c:a aac -b:a 192k output.mp4
Best for: Offline encoding where quality matters more than file size.
| Feature | CBR | VBR | CRF |
|---|---|---|---|
| Quality Consistency | Low (wastes bits on simple scenes) | High | Highest |
| File Size Predictability | Exact | Approximate | Unpredictable |
| Bandwidth Efficiency | Low | High | Highest |
| Streaming Compatibility | Best | Good | Not for streaming |
| Encoding Complexity | Low | Medium | Medium |
| Use Case | Live stream, broadcast | File delivery | Offline encode |
Pro Tip: For most video conversion tasks, CRF is the best choice. It produces the highest quality-per-byte by intelligently allocating bitrate where it is needed. Start with CRF 20 for H.264 and CRF 24 for H.265, then adjust up (lower quality, smaller file) or down (higher quality, larger file) based on your needs.
The ideal bitrate depends on the resolution, frame rate, and codec. Here are evidence-based recommendations for each combination.
| Resolution | 24fps | 30fps | 60fps |
|---|---|---|---|
| 480p (854x480) | 1-2 Mbps | 1.5-2.5 Mbps | 2.5-4 Mbps |
| 720p (1280x720) | 2.5-4 Mbps | 3-5 Mbps | 5-7.5 Mbps |
| 1080p (1920x1080) | 5-8 Mbps | 6-10 Mbps | 10-15 Mbps |
| 1440p (2560x1440) | 10-15 Mbps | 12-18 Mbps | 18-25 Mbps |
| 4K (3840x2160) | 20-35 Mbps | 25-40 Mbps | 40-60 Mbps |
H.265 is roughly 40-50% more efficient than H.264, so the recommended bitrates are correspondingly lower.
| Resolution | 24fps | 30fps | 60fps |
|---|---|---|---|
| 480p | 0.6-1.2 Mbps | 0.8-1.5 Mbps | 1.5-2.5 Mbps |
| 720p | 1.5-2.5 Mbps | 2-3 Mbps | 3-5 Mbps |
| 1080p | 3-5 Mbps | 4-6 Mbps | 6-10 Mbps |
| 1440p | 6-10 Mbps | 8-12 Mbps | 12-18 Mbps |
| 4K | 12-22 Mbps | 15-25 Mbps | 25-40 Mbps |
AV1 is approximately 30-50% more efficient than H.265, pushing bitrates even lower.
| Resolution | 24fps | 30fps | 60fps |
|---|---|---|---|
| 480p | 0.4-0.8 Mbps | 0.5-1 Mbps | 1-1.8 Mbps |
| 720p | 1-1.8 Mbps | 1.2-2 Mbps | 2-3.5 Mbps |
| 1080p | 2-3.5 Mbps | 2.5-4 Mbps | 4-7 Mbps |
| 1440p | 4-7 Mbps | 5-8 Mbps | 8-12 Mbps |
| 4K | 8-15 Mbps | 10-18 Mbps | 18-28 Mbps |
For a complete codec comparison, read our in-depth guide on H.265 vs H.264 vs AV1.
Streaming and download have fundamentally different bitrate requirements because of how data is delivered to the viewer.
Streaming video must arrive faster than it is played. If the bitrate exceeds the viewer's available bandwidth, playback stutters or drops to a lower quality level (in adaptive streaming).
Key streaming constraints:
| Service | 720p | 1080p | 4K | HDR 4K |
|---|---|---|---|---|
| YouTube | 2.5 Mbps | 4.5 Mbps | 20 Mbps | 25 Mbps |
| Netflix | 3 Mbps | 5 Mbps | 15 Mbps | 20 Mbps |
| Disney+ | 3 Mbps | 6 Mbps | 16 Mbps | 20 Mbps |
| Amazon Prime | 3 Mbps | 5 Mbps | 15 Mbps | 20 Mbps |
| Twitch | 3-6 Mbps | 6-8 Mbps | N/A | N/A |
These bitrates are lower than what you would use for a downloaded file because streaming services use highly optimized encoding (per-scene analysis, multi-pass encoding, content-adaptive settings) that extracts maximum quality from every bit.
Downloaded files do not have real-time constraints. You can use higher bitrates without worrying about buffering:
# High-quality download encode
ffmpeg -i input.mp4 -c:v libx264 -crf 18 -preset slow \
-c:a aac -b:a 256k output.mp4
For YouTube upload bitrate recommendations, see our best video settings for YouTube guide.
Different codecs extract different amounts of visual quality from the same bitrate. This efficiency is measured by encoding the same source at the same bitrate and comparing the visual quality.
| Codec | Relative Efficiency | To Match H.264 Quality | Year Introduced |
|---|---|---|---|
| H.264 | Baseline (1.0x) | 100% bitrate | 2003 |
| VP9 | ~1.4x | ~70% bitrate | 2013 |
| H.265 | ~1.5x | ~65% bitrate | 2013 |
| AV1 | ~2.0x | ~50% bitrate | 2018 |
| VVC (H.266) | ~2.5x | ~40% bitrate | 2020 |
This means that if you need 10 Mbps with H.264 to achieve a certain quality, you need only about 6.5 Mbps with H.265 or 5 Mbps with AV1 for the same visual result.
Pro Tip: The efficiency gains of newer codecs come at the cost of encoding speed. AV1 software encoding can be 10-50x slower than H.264 at similar settings. For real-time or batch use cases, hardware-accelerated H.265 often provides the best balance of quality, speed, and file size. See our 4K video conversion guide for hardware encoding details.
CRF (Constant Rate Factor) is the most popular quality control for offline encoding. Each codec has its own CRF scale.
| Quality Level | H.264 (libx264) | H.265 (libx265) | VP9 (libvpx-vp9) | AV1 (libaom-av1) |
|---|---|---|---|---|
| Visually lossless | 15-17 | 18-20 | 15-20 | 15-20 |
| High quality | 18-20 | 21-23 | 20-25 | 22-28 |
| Balanced | 21-23 | 24-26 | 25-30 | 28-32 |
| Web/streaming | 24-27 | 27-30 | 30-35 | 32-38 |
| Low quality | 28-32 | 31-35 | 36-40 | 38-45 |
The CRF scale is logarithmic. For H.264, increasing CRF by 6 roughly doubles the compression ratio (halves the file size). For H.265, the same effect occurs at a CRF increase of about 5.
# Encode a 30-second test segment at different CRF values
for crf in 18 20 22 24 26; do
ffmpeg -i input.mp4 -ss 60 -t 30 -c:v libx264 -crf $crf \
-preset medium -c:a copy "test_crf${crf}.mp4" -y
done
# Compare file sizes
ls -lh test_crf*.mp4
Then watch the test segments and pick the highest CRF value where you cannot notice quality degradation. This gives you the smallest possible file at your acceptable quality threshold.
Not all video content is created equal. The complexity of the visual content dramatically affects how much bitrate is needed.
A talking head video at CRF 23 might produce a 2 Mbps file, while an action scene at the same CRF might produce 15 Mbps. CRF automatically handles this variation, which is why it produces better results than fixed bitrate for variable-complexity content.
If your source has film grain or sensor noise, removing it before encoding can significantly reduce file size without meaningful quality loss:
# Light denoising for moderate noise
ffmpeg -i input.mp4 -vf "nlmeans=6:3:7:3" \
-c:v libx264 -crf 20 -preset slow -c:a copy output.mp4
# Aggressive denoising for very noisy footage
ffmpeg -i input.mp4 -vf "nlmeans=10:5:9:5" \
-c:v libx264 -crf 20 -preset slow -c:a copy output.mp4
For more compression techniques, see our guide on compressing video without losing quality.
The formula for estimating file size is straightforward:
File Size (MB) = (Video Bitrate + Audio Bitrate) × Duration (seconds) / 8 / 1024
Where bitrate is in kbps.
| Bitrate | 1 Minute | 5 Minutes | 10 Minutes | 1 Hour |
|---|---|---|---|---|
| 1 Mbps | 7.5 MB | 37.5 MB | 75 MB | 450 MB |
| 3 Mbps | 22.5 MB | 112 MB | 225 MB | 1.35 GB |
| 5 Mbps | 37.5 MB | 187 MB | 375 MB | 2.25 GB |
| 10 Mbps | 75 MB | 375 MB | 750 MB | 4.5 GB |
| 20 Mbps | 150 MB | 750 MB | 1.5 GB | 9 GB |
| 50 Mbps | 375 MB | 1.87 GB | 3.75 GB | 22.5 GB |
These estimates include only the video stream. Add 1-2 MB per minute for 128-192 kbps audio.
To work backwards from a desired file size:
Target Bitrate (kbps) = Target Size (MB) × 8 × 1024 / Duration (seconds) - Audio Bitrate
For example, to fit a 3-minute video into 50 MB with 192 kbps audio:
Target Bitrate = 50 × 8 × 1024 / 180 - 192 ≈ 2083 kbps (~2 Mbps)
Use the Video Compressor to handle these calculations automatically by setting a target file size.
While video typically accounts for 90-95% of file size, audio bitrate still matters.
| Audio Quality | AAC Bitrate | Opus Bitrate | Use Case |
|---|---|---|---|
| Voice only | 64-96 kbps | 32-64 kbps | Podcasts, calls, tutorials |
| Standard | 128-192 kbps | 96-128 kbps | Most video content |
| High quality | 256-320 kbps | 160-192 kbps | Music videos, concerts |
| Near-lossless | 320+ kbps | 256+ kbps | Professional audio |
For most video content, 128-192 kbps AAC is the sweet spot. Going higher wastes bytes that could be allocated to video quality. For more on audio quality, see our guide on audio bitrate and quality.
# Quick overall bitrate
ffprobe -v error -show_entries format=bit_rate \
-of csv=p=0 input.mp4
# Detailed stream-level bitrate
ffprobe -v error -show_entries stream=codec_name,bit_rate,width,height \
-of table input.mp4
# Generate per-frame bitrate data
ffprobe -v error -select_streams v:0 -show_entries \
frame=pkt_size,pts_time -of csv input.mp4 > bitrate_data.csv
This data can be graphed to visualize how bitrate fluctuates throughout the video, revealing which scenes are bitrate-heavy and might benefit from additional compression.
| Use Case | Recommended Approach | Typical Bitrate |
|---|---|---|
| YouTube upload | CRF 18 (H.264) | 8-40 Mbps (varies) |
| Social media | CRF 22-24 (H.264) | 3-10 Mbps |
| Email attachment | Two-pass target size | 0.5-2 Mbps |
| Archive/master | CRF 16-18 (H.265) | 10-50 Mbps |
| Live streaming | CBR | 3-8 Mbps |
| Web embedding | CRF 24-28 | 1-5 Mbps |
| Mobile delivery | CRF 24-26 | 1-4 Mbps |
For platform-specific bitrate settings, see our best video settings for YouTube and social media video specs guide.
No. Beyond a certain point (the "saturation point"), increasing bitrate adds file size without visible quality improvement. For 1080p H.264, this typically occurs around 15-20 Mbps. Beyond that, you are storing information the human eye cannot perceive.
YouTube re-encodes everything, so upload the highest quality source you can. Use CRF 18 with H.264 for 1080p content, or CRF 16 for 4K. The exact bitrate will depend on content complexity, but this approach ensures YouTube's encoder has plenty of data to work with.
Likely because of high bitrate, high resolution, long duration, or an inefficient codec. Use ffprobe to check the bitrate, then use the Video Compressor to reduce file size. Switching from H.264 to H.265 can cut size by 40-50% at the same quality.
Mbps (megabits per second) measures data rate. MBps (megabytes per second) measures file transfer speed. 1 MBps = 8 Mbps. A 5 Mbps video downloads at approximately 0.625 MBps.
No. Re-encoding a low-quality video at a higher bitrate simply preserves the existing artifacts at a larger file size. You cannot add detail that was already lost. Quality can only be preserved or reduced, never increased through re-encoding.
Video bitrate is the fundamental parameter controlling the balance between quality and file size. Use CRF mode for offline encoding where quality consistency matters, VBR for general-purpose files, and CBR for live streaming. Choose your codec wisely — H.265 and AV1 deliver dramatically better quality per bit than H.264.
For hands-on bitrate control, the Video Converter and Video Compressor provide intuitive interfaces for adjusting quality and file size. For detailed codec selection guidance, explore our guide on H.265 vs H.264 vs AV1 and our best video settings for YouTube.
Understanding bitrate is the key to making informed decisions about video quality, file size, and delivery. With the knowledge from this guide, you can confidently choose the right settings for any video workflow.
Marcus Rivera
Systems engineer writing about video transcoding, hardware acceleration, and large-scale media processing.