Video File Formats Explained: Codecs, Containers & Export Settings

Why Video Formats Confuse Everyone

The tragedy is that this confusion is unnecessary. Video formats follow a completely logical system, and once you understand the underlying principles — the difference between a container and a codec, and how compression works — every decision on that export screen becomes obvious and intuitive. Let us fix this once and for all.

Containers vs. Codecs: The Fundamental Distinction

Every video file has two distinct layers of technology working together:

A common analogy: think of the container as a shipping box, and the codec as the packing method. The same shipping box (MP4) can contain items packed loosely (high bitrate, large file) or items vacuum-sealed and compressed (low bitrate, small file). The box is neutral — the compression method determines the density.

"The most common question I get from beginners is 'should I export as MP4 or MOV?' This is a fundamentally wrong question. The right question is 'which codec should I use?' A .MP4 file using H.264 at 20 Mbps and a .MOV file using H.264 at 20 Mbps are visually 100% identical. They will look exactly the same on every screen on earth. The only practical difference is that MP4 is accepted by every platform, device, and browser, while MOV is Apple's proprietary format. I always export MP4 for this reason."

How Video Codecs Actually Compress Data

To understand why different codecs behave so differently in terms of performance and file size, it helps to understand the fundamental compression technique that all modern video codecs use: temporal compression (also called interframe compression).

Uncompressed video at 4K resolution and 60fps generates approximately 8 to 12 gigabytes of data per second. That is physically impossible to store on any affordable drive or transmit over any network. Compression makes video practical.

Instead of storing every single frame of video independently (like a sequence of photographs), temporal compression analyzes consecutive frames and only stores the differences between them. If you record a talking head with a static background, 90% of the pixels in each frame are identical to the previous frame. The codec only needs to record what changed — the mouth moving, the eyes blinking — and can mathematically reconstruct the "unchanged" pixels from memory. This is why compressed video files are 200x smaller than uncompressed video.

The compression algorithm references three types of frames:

• I-frames (Keyframes): Complete, independently encoded frames. Every codec periodically inserts I-frames to prevent errors from accumulating.
• P-frames (Predicted frames): Frames that only store changes from the previous frame.
• B-frames (Bi-directional frames): Frames that store changes relative to both the previous AND the next frame for maximum compression efficiency.

This is why editing software struggles to play back heavily compressed codecs like H.264 and H.265 smoothly — it has to mathematically reconstruct every P-frame and B-frame in real-time during playback, which is computationally intensive.

H.264 (AVC): The Internet Standard

H.264, also known as AVC (Advanced Video Coding), was standardized in 2003 and remains the most widely used video codec in the world today. It is supported by every operating system, web browser, streaming platform, game console, smart TV, and mobile device without exception.

When to use H.264:

• Uploading to YouTube, Vimeo, TikTok, Instagram, or any online platform
• Delivering final videos to clients
• Embedding video on websites
• Any situation where compatibility is the highest priority

Recommended Bitrate Settings for H.264:

• 📱 720p (30fps): 5–8 Mbps
• 🖥️ 1080p (30fps): 12–16 Mbps
• 🎬 1080p (60fps): 16–20 Mbps
• 📺 4K (30fps): 35–45 Mbps
• 🎮 4K (60fps): 50–68 Mbps

H.265 / HEVC: Twice the Efficiency

H.265, also called HEVC (High Efficiency Video Coding), was designed as the successor to H.264. Its primary advantage is delivering the exact same visual quality as H.264 at approximately half the file size (or equivalently, twice the quality at the same file size).

The tradeoff for this compression efficiency is significantly higher computational requirements for both encoding and decoding. On older computers without dedicated hardware acceleration for HEVC, H.265 playback causes dropped frames and stuttering — the processor simply cannot reconstruct the complex frame prediction fast enough.

When to use H.265:

• Archiving finished projects where storage space matters
• Delivering 4K content to clients who have modern playback hardware
• Streaming to Apple devices (all Apple hardware since 2017 has HEVC hardware acceleration)

"I stopped using H.265 for client delivery after a client called me panicking because the video 'wouldn't open.' She was on a Windows 7 machine from 2011. Now I only use H.265 for personal archiving. H.264 is not going away anytime soon — it's too universally supported. Unless your client has explicitly confirmed they can play H.265, don't use it for delivery."

Apple ProRes & Avid DNxHD: The Editing Codecs

ProRes and DNxHD are a completely different class of codec. They are not designed for delivery or distribution — they are designed specifically for editing. They use intraframe compression (every frame is a full I-frame) rather than temporal compression.

This means every single frame is independently stored and decompressed without needing to reference previous or subsequent frames. The result is extremely fast, smooth playback in editing software — because the computer never needs to reconstruct predicted frames — at the cost of enormous file sizes.

AV1: The Rising Open-Source Standard

AV1 is a royalty-free, open-source video codec developed by a consortium of companies including Google, Netflix, Amazon, and Intel. It achieves roughly 30% better compression than H.265 at the same visual quality, making it the most efficient broadly-available codec at the time of writing.

YouTube began transcoding all uploaded videos to AV1 in 2018. Netflix now streams in AV1 on supported devices. Chrome, Firefox, and Edge all support AV1 playback natively. Hardware decoding support is available on all NVIDIA RTX 30-series and AMD RX 6000-series GPUs and newer.

However, AV1 encoding is still extremely slow on all but the latest hardware (hardware-accelerated AV1 encoding requires NVIDIA RTX 40-series or ARC Intel GPUs). For most editors in 2024-2025, exporting in AV1 is impractically slow. It is primarily relevant as a streaming delivery format that platforms handle themselves after receiving your H.264 upload.

Container Formats: MP4 vs. MOV vs. MKV

Intermediate Codecs: When to Use Proxies

When you import 4K H.264 or H.265 footage into Premiere Pro or DaVinci Resolve on a mid-range computer, the playback will stutter. This is because the temporal compression in H.264/H.265 requires the CPU to mathematically reconstruct hundreds of P-frames and B-frames every second.

The professional solution is to create proxy files — low-resolution, intraframe-compressed copies of your footage that the computer can play back smoothly. Most editors create proxies in ProRes Proxy (macOS) or DNxHR LB (Windows) at 1/4 or 1/2 of the original resolution.

The editing software works with the proxy files during editing. When you export the final video, it automatically re-links to the original full-resolution files and exports from those. You get the smooth editing experience of ProRes with the final quality of your 4K original.

The Only Export Settings You Need

"Don't fall for the 'VBR 2 Pass' trap when exporting H.264 for online delivery. Variable Bitrate 2-Pass tells Premiere to analyze the video twice — once to measure complexity, once to encode. It literally doubles your export time. On modern hardware with good CPUs and GPUs, the visual difference between VBR 1 Pass and VBR 2 Pass for a web video is indistinguishable to the human eye. Use VBR 1 Pass for 95% of exports and save those precious render hours for something that actually matters."

Frequently Asked Questions