Why is the audio out of sync after converting?

Common causes: (1) variable framerate source rendered as constant (use -vsync vfr to preserve VFR); (2) different audio sample rates not resampled (add -ar 48000 ); (3) container limitations (MP4 with variable framerate misbehaves — prefer MKV during editing, encode to MP4 only at the end). Always run ffprobe on both source and output to compare timing.

Why is the audio out of sync after converting?

Common causes: (1) variable framerate source rendered as constant (use -vsync vfr to preserve VFR); (2) different audio sample rates not resampled (add -ar 48000 ); (3) container limitations (MP4 with variable framerate misbehaves — prefer MKV during editing, encode to MP4 only at the end). Always run ffprobe on both source and output to compare timing.

Guide

H.264/AVC Video Codec: Complete Technical Guide

PC By Pablo Cirre • Updated May 7, 2026

What Is H.264?

H.264, also known as AVC (Advanced Video Coding) and MPEG-4 Part 10, is the most widely deployed video compression standard in history. Developed jointly by the ITU-T Video Coding Experts Group (VCEG) and ISO/IEC Moving Picture Experts Group (MPEG) under the Joint Video Team (JVT), H.264 was first standardized in 2003 and achieved mainstream adoption by 2005–2008.

H.264 dramatically improved compression efficiency over its predecessor MPEG-2 (used on DVDs): H.264 can store video at half the bitrate of MPEG-2 at equal quality, enabling HD video streaming over broadband connections, Blu-ray disc formats, web video (YouTube, Netflix), video calling, and virtually every modern video application.

H.264 Encoding Architecture

H.264 encodes video by predicting how each frame can be derived from previously decoded frames, then encoding only the difference (residual) between the prediction and the actual frame.

Frame Types

Frame Type	Description
I-frame (Intra)	Compressed without reference to other frames; standalone. Larger than P or B frames.
P-frame (Predictive)	Predicted from one previous reference frame. Uses motion vectors to describe how blocks moved.
B-frame (Bidirectional)	Predicted from both past and future frames. Smallest; highest compression but requires decoder buffering.

A Group of Pictures (GOP) is the sequence of frames between two I-frames. Common GOP structures:

IBBBPBBBP... — typical film encoding (high compression)
IPPPPP... — low-latency encoding (live streaming)
Keyframe-only (all I-frames) — editing-optimized codecs

Motion Estimation and Compensation

Motion compensation is H.264's primary compression tool. For each block in a P or B frame, the encoder finds the best matching block in reference frames:

Divides each frame into macroblocks (16×16 pixels)
Searches reference frames for similar blocks (motion search)
Records the displacement vector (motion vector)
Computes residual = actual block - predicted block
Applies DCT transform to the residual
Quantizes DCT coefficients (discards high-frequency detail)
Entropy codes the quantized coefficients

Sub-pixel motion compensation: H.264 allows motion vectors with 1/4-pixel precision (quarter-pixel), achieved through fractional-pixel interpolation filters. This provides much better prediction accuracy than integer-pixel motion vectors.

Intra Prediction

Even within an I-frame, H.264 uses spatial prediction: each 4×4 or 8×8 luma block is predicted from already-decoded neighboring blocks. H.264 defines 9 intra prediction modes for 4×4 blocks:

Mode 0: Vertical (propagate top edge downward)
Mode 1: Horizontal (propagate left edge rightward)
Mode 2: DC (use mean of top and left neighbors)
Modes 3-8: Diagonal and off-diagonal directions

The encoder chooses the mode that minimizes the residual energy (Rate-Distortion optimization).

Transform and Quantization

The prediction residual undergoes:

4×4 DCT — converts spatial residuals to frequency domain
Quantization — divides DCT coefficients by a Quantization Parameter (QP), rounding to integers. Higher QP = more compression, lower quality.
Inverse quantization + IDCT on decoder side to reconstruct residual

CRF (Constant Rate Factor): x264 and x265 use CRF instead of fixed QP. CRF targets a consistent perceptual quality across the video, automatically adjusting QP per frame based on complexity. CRF 18 ≈ transparent; CRF 23 = default; CRF 28 = lower quality.

CABAC Entropy Coding

H.264 High Profile uses CABAC (Context-Adaptive Binary Arithmetic Coding), a powerful entropy coder that models the probability of each symbol based on the context of neighboring symbols. CABAC typically reduces bitrate by 10–15% vs. CAVLC (used in Baseline/Main Profile) for the same quality.

H.264 Profiles

Profile	Features	Typical Use
Baseline	No B-frames, no CABAC, limited to 2 reference frames	Old mobile devices, video calls
Main	B-frames, CABAC, up to 16 reference frames	Standard broadcasting
High	All Main features + 8×8 intra DCT, custom quantization matrices	High-quality video, Blu-ray, web
High 10	10-bit color depth	Professional production, HDR
High 4:4:4	Full chroma sampling, lossless mode	Professional production

Most consumer content (YouTube, Netflix 1080p, Blu-ray) uses High Profile.

H.264 Levels

Levels define maximum decoder complexity (resolution × frame rate × bitrate):

Level	Max resolution	Max bitrate	Example
3.0	720×480 @ 30fps	10 Mbps	SD video
3.1	1280×720 @ 30fps	14 Mbps	720p HD
4.0	1920×1080 @ 30fps	20 Mbps	1080p HD
4.1	1920×1080 @ 30fps	50 Mbps	Blu-ray
5.0	1920×1080 @ 60fps	168 Mbps	High-frame-rate HD
5.1	4096×2048 @ 30fps	300 Mbps	4K
5.2	4096×2160 @ 60fps	480 Mbps	4K/60fps

H.264 Encoding with x264

x264 is the gold-standard open-source H.264 encoder. Key FFmpeg commands:

# CRF 23 (default quality) — balanced quality/size
ffmpeg -i input.mp4 -c:v libx264 -crf 23 -preset medium output.mp4

# High quality (CRF 18, slower encoding)
ffmpeg -i input.mp4 -c:v libx264 -crf 18 -preset slow output.mp4

# Specific bitrate (2-pass encoding for accurate file size)
ffmpeg -i input.mp4 -c:v libx264 -b:v 2000k -pass 1 -f null /dev/null
ffmpeg -i input.mp4 -c:v libx264 -b:v 2000k -pass 2 output.mp4

# Web-optimized (faststart moves moov atom to front)
ffmpeg -i input.mp4 -c:v libx264 -crf 23 -preset medium -movflags +faststart output.mp4

# 1080p @ 60fps for social media
ffmpeg -i input.mp4 -c:v libx264 -crf 22 -preset slow -vf scale=1920:1080 -r 60 -c:a aac -b:a 192k output.mp4

x264 Presets

The preset controls encoding speed vs. compression efficiency trade-off:

Preset	Encoding speed	File size at same quality
ultrafast	Very fast	Largest
superfast	Fast	Large
veryfast	Fast	Medium-large
faster	Medium	Medium
fast	Medium	Medium
medium	Default	Default
slow	Slow	Smaller
slower	Very slow	Smaller
veryslow	Extremely slow	Smallest

For most use cases, medium or slow is recommended. veryslow takes much longer for marginal quality improvement.

H.264 Compatibility

H.264 High Profile is supported by:

All modern web browsers (Chrome, Firefox, Safari, Edge)
All smartphones (iOS, Android)
Smart TVs, gaming consoles (PS4/5, Xbox)
Blu-ray players
Streaming services (Netflix, YouTube, Twitch, etc.)
Video editing software (Premiere, DaVinci Resolve, Final Cut Pro)

This universal support makes H.264 the safest choice for video distribution even in 2024, despite H.265 and AV1 offering better compression.

H.264 Containers

H.264 video is almost always stored in one of these containers:

Container	Extension	Common use
MPEG-4 Part 14	.mp4 / .m4v	Universal — web, streaming, mobile
Matroska	.mkv	Open container, PC media players
QuickTime	.mov	Apple ecosystem, production
Flash Video	.flv	Legacy — Adobe Flash (obsolete)
AVI	.avi	Legacy — Windows Media Player
MPEG-2 Transport Stream	.ts	Broadcasting, DVB, HLS segments

Related conversions

Common video conversions that pair well with this guide:

Frequently Asked Questions

CRF (Constant Rate Factor) controls quality in x264. The scale runs from 0 (lossless) to 51 (worst quality), with 23 as the default. For archival or high-quality masters, CRF 18 is considered "visually lossless" — most viewers cannot distinguish it from the original. CRF 20-23 is excellent for most content. CRF 24-28 is acceptable for casual viewing or streaming where file size matters. For web upload to platforms like YouTube (which re-encodes your video anyway), CRF 20 with the "slow" preset gives a clean source that survives re-encoding well. Never go above CRF 28 for video that will be distributed.

CRF (Constant Rate Factor) controls quality in x264. The scale runs de 0 (lossless) to 51 (worst quality), com 23 como o default. para archival ou alta-quality masters, CRF 18 is considered "visually sem perdas" — most viewers cannot distinguish it de the original. CRF 20-23 is excellent para most content. CRF 24-28 is acceptable para casual viewing ou streaming where tamanho do arquivo matters. para web enviar to plataformas like YouTube (which re-encodes your video anyway), CRF 20 com the "slow" preset gives a clean source that survives re-encoding well. Never go above CRF 28 para video that will be distributed.

CRF (Constant Rate Factor) controls quality in x264. The scale runs von 0 (lossless) to 51 (worst quality), mit 23 als das default. für archival oder hoch-quality masters, CRF 18 is considered "visually verlustfrei" — most viewers cannot distinguish it von the original. CRF 20-23 is excellent für most content. CRF 24-28 is acceptable für casual viewing oder streaming where Dateigröße matters. für web hochladen to Plattformen like YouTube (which re-encodes your video anyway), CRF 20 mit the "slow" preset gives a clean source that survives re-encoding well. Never go above CRF 28 für video that will be distributed.

CRF (Constant Rate Factor) controls quality in x264. The scale runs de 0 (lossless) to 51 (worst quality), con 23 como el default. para archival o alta-quality masters, CRF 18 is considered "visually sin pérdidas" — most viewers cannot distinguish it de the original. CRF 20-23 is excellent para most content. CRF 24-28 is acceptable para casual viewing o streaming where tamaño de archivo matters. para web subir to plataformas like YouTube (which re-encodes your video anyway), CRF 20 con the "slow" preset gives a clean source that survives re-encoding well. Never go above CRF 28 para video that will be distributed.

AV1 is the most efficient (royalty-free, ~30% smaller than H.265) but encoding is slow. H.265 (HEVC) saves ~30–50% over H.264 and is supported by every modern phone and desktop. H.264 remains the safest baseline for legacy compatibility. Rule of thumb: archives → AV1, daily use → H.265, broadest reach → H.264.

H.264 profiles define which encoding features are used. Baseline Profile uses the simplest tools: no B-frames, no CABAC entropy coding, maximum 2 reference frames. It was designed for low-power devices like older phones and hardware video calling. Main Profile adds B-frames and CABAC, improving compression 10-15%. High Profile adds 8×8 intra DCT transforms and custom quantization matrices for another 5-10% efficiency gain. All modern devices support High Profile. When encoding for distribution, always use High Profile unless you specifically need to target very old hardware (pre-2008 phones or early set-top boxes).

H.264 profiles define which codificação features are used. Baseline Profile uses the simplesst ferramentas: no B-frames, no CABAC entropy coding, máximo 2 reference frames. It was designed para baixa-power devices like older phones e hardware video calling. Main Profile adds B-frames e CABAC, improving compressão 10-15%. alta Profile adds 8×8 intra DCT transforms e custom quantization matrices para another 5-10% efficiency gain. All dispositivos modernoos support alta Profile. When codificação para distribution, always usar alta Profile unless you specifically precisar target very old hardware (pre-2008 phones ou early set-top boxes).

H.264 profiles define which Codierung features are used. Baseline Profile uses the einfachst Werkzeuge: no B-frames, no CABAC entropy coding, maximal 2 reference frames. It was designed für niedrig-power devices like older phones und hardware video calling. Main Profile adds B-frames und CABAC, improving Komprimierung 10-15%. hoch Profile adds 8×8 intra DCT transforms und custom quantization matrices für another 5-10% efficiency gain. All moderne Geräte support hoch Profile. When Codierung für distribution, always verwenden hoch Profile unless you specifically need to target very old hardware (pre-2008 phones oder early set-top boxes).

H.264 profiles define which codificación features are used. Baseline Profile uses the simplest herramientas: no B-frames, no CABAC entropy coding, máximo 2 reference frames. It was designed para baja-power devices like older phones y hardware video calling. Main Profile adds B-frames y CABAC, improving compresión 10-15%. alta Profile adds 8×8 intra DCT transforms y custom quantization matrices para another 5-10% efficiency gain. All dispositivos modernoos support alta Profile. When codificación para distribution, always usar alta Profile unless you specifically need to target very old hardware (pre-2008 phones o early set-top boxes).

CRF (Constant Rate Factor) is the best default for offline files: ffmpeg picks the bitrate frame-by-frame to maintain perceived quality. Two-pass is only better when you must hit an exact final size (DVD targets). Constant bitrate is for streaming with a fixed channel. For "smallest at quality X" always use CRF.

This is usually because the MP4 moov atom (the metadata header that tells players how to decode the file) is at the end of the file instead of the beginning. Web players need the moov atom before they can start playing, so they buffer the entire file first. Fix this by adding the -movflags +faststart option to your FFmpeg command: ffmpeg -i input.mp4 -c:v libx264 -crf 23 -movflags +faststart output.mp4. This moves the moov atom to the beginning of the file, enabling progressive streaming without buffering the whole file.

This is Geralmente because the MP4 moov atom (the metadata header that tells players how to decode o arquivo) is at the end of o arquivo em vez de the beginning. Web players need the moov atom antes they can start playing, so they buffer the entire arquivo first. Fix this by adding the -movflags +faststart option to your FFmpeg command: ffmpeg -i input.mp4 -c:v libx264 -crf 23 -movflags +faststart output.mp4. This moves the moov atom para o beginning of o arquivo, enabling progressive streaming sem buffering the whole file.

This is Normalerweise because the MP4 moov atom (the metadata header that tells players how to decode die Datei) is at the end von die Datei anstatt the beginning. Web players need the moov atom vor they can start playing, so they buffer the entire Datei first. Fix this by adding the -movflags +faststart option to your FFmpeg command: ffmpeg -i input.mp4 -c:v libx264 -crf 23 -movflags +faststart output.mp4. This moves the moov atom zum beginning von die Datei, enabling progressive streaming ohne buffering the whole file.

This is Normalmente because the MP4 moov atom (the metadata header that tells players how to decode el archivo) is at the end de el archivo en vez de the beginning. Web players need the moov atom antes they can start playing, so they buffer the entire archivo first. Fix this by adding the -movflags +faststart option to your FFmpeg command: ffmpeg -i input.mp4 -c:v libx264 -crf 23 -movflags +faststart output.mp4. This moves the moov atom al beginning de el archivo, enabling progressive streaming sin buffering the whole file.

Common causes: (1) variable framerate source rendered as constant (use <code>-vsync vfr</code> to preserve VFR); (2) different audio sample rates not resampled (add <code>-ar 48000</code>); (3) container limitations (MP4 with variable framerate misbehaves — prefer MKV during editing, encode to MP4 only at the end). Always run <code>ffprobe</code> on both source and output to compare timing.

For maximum compatibility, use H.264. It plays on every device and every browser without exception. For best quality-per-file-size ratio, use H.265 (HEVC) — it achieves the same quality as H.264 at roughly half the bitrate, ideal for storage or high-bandwidth content. For web delivery, consider AV1 (supported by Chrome, Firefox, Edge) or H.265 for browsers that support it. The practical rule: H.264 if your primary concern is compatibility; H.265 if your primary concern is file size while maintaining high quality; AV1 if you target modern browsers and want the best compression. Many workflows use H.265 for storage and H.264 for delivery.

For máximo compatibilidade, usar H.264. It plays on every device e every browser sem exception. para melhor qualidade-per-file-size ratio, usar H.265 (HEVC) — it achieves the same quality as H.264 at roughly half the bitrate, ideal para storage ou alta-bandwidth content. para web delivery, consider AV1 (supported by Chrome, Firefox, Edge) ou H.265 para browsers that support it. The practical rule: H.264 if your primary concern is compatibilidade; H.265 if your primary concern is tamanho do arquivo while maintaining alta qualidade; AV1 if you target moderno browsers e want the best compressão. Many workflows usar H.265 para storage e H.264 para delivery.

For maximal Kompatibilität, verwenden H.264. It plays on every device und every browser ohne exception. für beste Qualität-per-file-size ratio, verwenden H.265 (HEVC) — it achieves the same quality as H.264 at roughly half the Bitrate, ideal für storage oder hoch-bandwidth content. für web delivery, consider AV1 (supported by Chrome, Firefox, Edge) oder H.265 für browsers that support it. The practical rule: H.264 if your primary concern is Kompatibilität; H.265 if your primary concern is Dateigröße while maintaining hohe Qualität; AV1 if you target modern browsers und want the best Komprimierung. Many workflows verwenden H.265 für storage und H.264 für delivery.

For máximo compatibilidad, usar H.264. It plays on every device y every browser sin exception. para mejor calidad-per-file-size ratio, usar H.265 (HEVC) — it achieves the same quality as H.264 at roughly half the bitrate, ideal para storage o alta-bandwidth content. para web delivery, consider AV1 (supported by Chrome, Firefox, Edge) o H.265 para browsers that support it. The practical rule: H.264 if your primary concern is compatibilidad; H.265 if your primary concern is tamaño de archivo while maintaining alta calidad; AV1 if you target moderno browsers y want the best compresión. Many workflows usar H.265 para storage y H.264 para delivery.

Yes if you only change the container: <code>ffmpeg -i in.mkv -c copy out.mp4</code>. This remuxes the stream without re-encoding, takes seconds even for hours of footage. Limitations: codec must be supported by the target container (e.g. you cannot put H.264 in WebM, only VP8/VP9/AV1). To shrink size you must re-encode.