Configure AVI Packet Chunk Size with FFmpeg

This article explains how to configure the packet and chunk sizes within an AVI container using FFmpeg. You will learn about the specific FFmpeg options required to adjust packet sizes, control chunk interleaving, and optimize AVI files for legacy hardware players or specific streaming requirements.

To configure the packet or chunk size in an AVI container, you must use FFmpeg’s muxer private options and general format parameters. AVI files store audio and video data in interleaved chunks; adjusting these chunks ensures better synchronization and compatibility.

Setting the Packet Size

The global -packet_size option specifies the maximum packet size in bytes for the output container. While primarily used for transport streams, it can enforce boundaries on the muxer level for AVI files.

To set the packet size to 4096 bytes:

ffmpeg -i input.mp4 -packet_size 4096 output.avi

Adjusting Chunk Interleaving

In AVI containers, “chunk size” is highly dependent on how audio and video packets are interleaved. By default, FFmpeg interleaves packets automatically. You can control this behavior using the -max_delay option, which defines the maximum muxing/demuxing delay in microseconds. This directly affects the size of the interleaved data chunks.

To configure a tighter chunk spacing (e.g., a maximum delay of 100,000 microseconds / 100 milliseconds):

ffmpeg -i input.mp4 -max_delay 100000 output.avi

For strict sector alignment (often required by older standalone DVD/CD players), you can use the -reserve_index_space option to pre-allocate index size, which helps maintain rigid chunk offsets:

ffmpeg -i input.mp4 -reserve_index_space 1024*1024 output.avi

Complete Command Example

To copy the video and audio streams without re-encoding while strictly configuring both the packet size and the interleaving delay, run the following command:

ffmpeg -i input.mp4 -c copy -packet_size 2048 -max_delay 50000 output.avi

This command keeps the original codecs (-c copy), limits the packet size to 2048 bytes, and sets a 50-millisecond maximum delay to ensure small, highly interleaved chunks for seamless playback on low-memory hardware.