H.264 vs H.265: Measured at Equal Quality
At the same measured SSIM of 0.992 on our stated clip, H.265 used 3,758.5 kbps and H.264 used 3,925.9 kbps. That makes this H.265 encode about 4.3% smaller, but the difference belongs to this animated clip and these two encoder configurations, not to every video. Choose H.264 when acceptance is uncertain; choose H.265 only when your documented destination and playback devices accept it.
Short answer
Choose H.264 unless you have a verified H.265 pipeline and a destination that accepts H.265. At the same measured SSIM of 0.992 on our test clip, H.265 used 3,758.5 kbps against 3,925.9 kbps for H.264, about 4.3% smaller.
Measured data
The video run used a 10 second, 1920 x 1080, 30 fps excerpt of Big Buck Bunny, an animated film rather than camera footage, licensed CC BY 3.0 by the Blender Foundation. Results describe this clip and these encoder settings, not every video or any platform pipeline.
| Codec and encoder | Measured SSIM | Interpolated bitrate | How derived |
|---|---|---|---|
| H.264, libx264 medium | 0.992 | 3,925.9 kbps | Between CRF 18 and 24 |
| H.265, libx265 medium | 0.992 | 3,758.5 kbps | Between CRF 18 and 24 |
Equal quality here means equal measured SSIM, not equal CRF. The interpolation uses the two bracketing points from each measured curve and is approximate.
Arithmetic: (1 - 3,758.5 / 3,925.9) x 100 = 4.3% smaller for this H.265 result.
When to use each option
Use H.264 for uncertain upload paths
It is the conservative handoff when a platform page names H.264 or does not publish a codec list. A container name such as MP4 does not prove H.265 acceptance.
Use H.265 for a verified HEVC path
The measured saving is modest on this clip. Confirm the exact upload surface and the devices that must decode the result before choosing it.
Keep the source for later exports
Both choices are lossy encodes. Re-encode from the original, not from a file downloaded from another platform.
Questions people ask
Is H.265 always smaller than H.264?
No. On this run it was 4.3% smaller at interpolated SSIM 0.992. Content, encoder, preset and target quality can change the result.
Can I compare the same CRF number?
No. CRF scales are encoder-specific. Compare files at a measured quality target such as SSIM, then compare bitrate or bytes.
Does MP4 mean H.265 will upload?
No. MP4 is a container. The destination must also accept the video codec inside it.
Which is safer for social uploads?
H.264 is safer when the official page names H.264 or does not name H.265. Check the owner page for the exact surface.
Measured vs sourced
Cells marked measured are Koldflux Media Lab measurements. Cells marked sourced derive from each platform / codec standards body. Cells marked guidance are expert guidance, not measured.
Sources
- FFmpeg codec documentation (read 2026-10-08)
- Blender Foundation: Big Buck Bunny and CC BY 3.0 (read 2026-10-08)
Verified at 2026-10-08.
Next steps
Inspect my file: /tools/validate
Convert / make correct version: /tools/convert
Create all target versions: /tools/pack