Utility guide
Smaller files are easy; small files that look good take method
Compression decides whether your video streams instantly or stalls — and whether it looks crisp or falls apart into blocks. This guide covers codecs, bitrate strategy, resolution ladders, and the two-pass habits that shrink files while protecting the picture.
From the Mellius product team · Updated September 28, 2026
Conspiracy theorists, you may hate this
AI Director
›
Just type what you need
- ›Remove all silent parts
- ›Add bold kinetic captions
- ›Only have Bill Ackman speaking
- ›Cut to highlights only
- ›Add cinematic transitions
- ›Match cuts to the beat
The short answer
Export H.264 or HEVC with bitrate matched to resolution, frame rate, and motion, compressing from the master rather than a re-uploaded copy. Halve bitrate expectations for talking heads versus action, keep audio near 128 to 320 kbps AAC, and verify motion scenes frame by frame before replacing the master.
Right codec
The codec matches compatibility needs — H.264 everywhere, HEVC or AV1 where support is confirmed.
Motion-aware bitrate
Bitrate follows detail and movement instead of a single number applied to every video.
Master-first workflow
Every compressed copy derives from the high-quality master, never from another compressed file.
Pick codecs and bitrates with intent
H.264 in MP4 remains the universal choice: it plays everywhere with hardware decoding on every device. HEVC halves file sizes at similar quality but needs newer devices and paid licensing in some pipelines, while AV1 compresses best of all where platforms accept it. For 1080p delivery, start near eight to twelve megabits per second for general content, dropping to five or six for locked-off talking heads and rising to sixteen or more for sports, confetti, and water.
Match frame rate and resolution to the destination before compressing: downscale 4K masters to 1080p for social delivery rather than uploading giant files platforms will crush anyway. Use two-pass or constrained-quality encoding when time allows, since single-pass constant-bitrate encodes waste data on easy scenes and starve hard ones. Keep audio at 128 to 192 kbps AAC for speech and up to 320 kbps for music-led content.
- Default to H.264 MP4 unless the destination confirms HEVC or AV1.
- Scale bitrate with motion: calm scenes need far less data.
- Downscale before compressing rather than uploading oversize files.
- Prefer two-pass or quality-based encoding for important exports.
Avoid generation loss and verify the result
Always compress from the master or timeline export, because recompressing an already-compressed download compounds artifacts with every generation. Trim, caption, and grade before encoding so the compressor sees the final pixels, and archive the master separately from platform copies. When a platform recompresses anyway — and most do — upload slightly above its recommended bitrate so the second encode has clean material to work from.
Verify on motion, not stills: scrub through the busiest five seconds and check foliage, crowd edges, gradients, and small text for blocking and banding. Watch the export on a phone over a throttled connection to feel real-world startup time. If artifacts cluster in one scene, raise bitrate selectively with segment encoding or accept a slightly larger file — viewers forgive megabytes, not mush.
Compare the workflows
| Codec | Best fit | Trade-off |
|---|---|---|
| H.264 | Universal playback across devices and platforms | Larger files than newer codecs at equal quality |
| HEVC | Small files for 4K where device support is known | Slower encodes and patchier compatibility |
| AV1 | Best efficiency for supported streaming pipelines | Heavy encoding cost and uneven upload acceptance |
A practical sequence
Put it into practice
- 01
Start from the master
Finish the edit, then export a high-quality master that all compressed copies derive from.
- 02
Set codec and bitrate
Choose H.264 for universality, set bitrate by resolution and motion, and encode two-pass.
- 03
Verify and version
Check busy scenes for artifacts, test on real connections, and keep masters archived.
Frequently asked questions
How do I compress a video without losing quality?
You cannot shrink a file with zero change, but you can make the loss invisible by encoding from the master with a modern codec and motion-matched bitrate. Talking heads compress far more cleanly than action, so set expectations by content. Verify the busiest scenes rather than judging by file size alone.
What bitrate should I use for YouTube uploads?
For 1080p SDR uploads, eight to twelve megabits per second H.264 covers most content, rising toward sixteen for high-motion scenes. Upload 4K at thirty-five to fifty megabits per second to earn better processing from the platform. YouTube recompresses everything, so clean, slightly generous uploads produce cleaner final streams.
Why does my compressed video look blocky?
Blocking usually means the bitrate was too low for the scene's detail and motion, or the file was compressed from an already-compressed copy. Raise bitrate for complex scenes, switch to two-pass encoding, and always encode from the master. Gradients and night skies may additionally need higher bit depth or dithering.
Continue reading
Primary sources
Product capabilities change. These official references support the comparisons above.
Describe the edit. Keep the timeline.
Mellius turns conversational direction into cuts, captions, graphics, motion, color, and sound you can still inspect and change.
Start editing free