opensmile

Status: Current Last updated: 2026-05-02 07:30 EDT

Extract acoustic features from audio files using openSMILE. Produces .opensmile.csv output, not CHAT. This is the only processing command that does not produce .cha output.

Uses positional INPUT_DIR OUTPUT_DIR arguments (not the shared PATHS... -o OUTPUT form used by align, morphotag, etc.).

Quick start

# Extract default eGeMAPSv02 features from all audio in a directory
batchalign3 opensmile input_dir/ output_dir/

# Use a different feature set
batchalign3 opensmile input_dir/ output_dir/ --feature-set ComParE_2016

# Use the remote server
batchalign3 --server http://your-server:8001 opensmile input_dir/ output_dir/

Pipeline

flowchart TD
    start([opensmile invoked]) --> resolve[Resolve audio files]
    resolve --> prep[Rust audio prep\nprepare mono PCM artifact]
    prep --> feature_check{--feature-set?}
    feature_check -->|eGeMAPSv02| egemaps[eGeMAPSv02 features\n88 acoustic descriptors]
    feature_check -->|ComParE_2016| compare[ComParE_2016 features\n6,373 acoustic descriptors]
    feature_check -->|Custom| custom[Custom feature set name]

    egemaps --> worker
    compare --> worker
    custom --> worker

    worker["execute_v2(task='opensmile') → Python worker\nExtracts acoustic features from prepared audio"]
    worker --> output[Write CSV output\nContent-type: csv]
    output --> done([Output .opensmile.csv files])

Options

Positional arguments

Argument	Meaning
`INPUT_DIR`	Directory containing audio files
`OUTPUT_DIR`	Directory for output `.opensmile.csv` files

opensmile options

Option	Default	Meaning
`--feature-set SET`	`eGeMAPSv02`	Feature set: `eGeMAPSv02`, `eGeMAPSv01b`, `GeMAPSv01b`, or `ComParE_2016`
`--lang CODE`	`eng`	3-letter ISO language code
`--bank NAME`	:	Server media bank name from `server.yaml` `media_mappings` (server-backed runs only)
`--subdir PATH`	:	Subdirectory under the selected `--bank` to scope the run

Each audio file produces FILE.opensmile.csv with feature names as column headers and one row per file (or one row per frame for frame-level sets). BA3 uses a row-oriented CSV (feature names as columns), which differs from BA2’s transposed feature-per-row export.

Command I/O: opensmile, I/O patterns
Command Flowcharts: opensmile, full architecture flowchart

This page last changed: 2026-07-29 (commit f4f12680). The whole book last changed: 2026-09-16 (commit 34d249d8).

TalkBank Toolchain