The public sources name these chunks by GUID, listed on Chunks. Only the summary list has a published layout. Readers skip chunks they do not understand, by size.
What is known of each
| Chunk | What is known | Layout |
|---|---|---|
levl | a peak-level chunk; the name matches WAV’s levl peak envelope | not confirmed |
list | a list chunk; the name matches WAV’s LIST | not confirmed |
junk | filler, as in WAV’s JUNK; ignored by readers | any |
bext | broadcast metadata; the name matches WAV’s bext | not confirmed |
| acid | a chunk libsndfile names ACID; the name matches WAV’s acid loop chunk | not confirmed |
| marker | a marker chunk; only the GUID is published, and it is not read by FFmpeg | not confirmed |
| summary list | tag text | below |
“Not confirmed” means no public description was found; they may follow the WAV chunks of the same name, but this page does not assume so.
Summary list
As FFmpeg reads it. Offsets are from the start of the chunk’s data.
| Offset | Size | Type | Name | Description |
|---|---|---|---|---|
| 0 | 4 | u32 | count | the number of entries |
| 4 | … | entries | the entries, below |
| Offset | Size | Type | Name | Description |
|---|---|---|---|---|
| 0 | 4 | ID | key | a four-character key |
| 4 | 4 | u32 | size | the size of the value in bytes |
| 8 | size | UTF-16LE | value | the tag text |
The meaning of each key is not documented; FFmpeg passes them through as metadata.