OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Audio & Voice AI

Speech Emotion Analyzer

Speech Emotion Analyzer is a self-hosted machine-learning project that classifies speech audio into five emotion labels—angry, calm, fearful, happy, and sad—and male or female voice labels. Its README describes a convolutional neural network, Python, and LibROSA audio feature extraction.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Speech Emotion Analyzer — OpenViST project profile in Audio & Voice AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it as a starting point for studying or prototyping speech emotion classification when a self-hosted Python model and its five documented emotion labels fit the task.

Who it is for

  • Developers prototyping emotion-aware features for speech-based interfaces.
  • Learners and researchers exploring audio classification workflows in Python.

Documented uses

  • Extract audio features and inspect waveforms or spectrograms.
  • Train or explore a speech emotion classifier using the project’s documented datasets.
  • Review predicted emotion and voice labels for test recordings.
Before you choose

Requirements & limits

The repository is public and specifies an MIT license. It is a self-hosted project presented in Jupyter Notebook/Python.

  • The documented output covers five emotions and male/female voice labels.
  • The README reports emotion accuracy above 70% and voice-gender classification at 100% on its project tests; these are project-reported results.
  • The feature-extraction workflow times audio to three seconds, a preprocessing choice to consider when adapting it.
  • The README names RAVDESS and SAVEE as training datasets and notes that additional audio may improve emotion accuracy.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare speech-audio features for emotion modeling

Extract audio features for training using the approach described in the project README.

Before you start

  • Python audio-analysis library LibROSA
  • Speech audio files

Steps

  1. Use LibROSA to extract features from the audio files.

  2. Time each audio file to three seconds so the files yield an equal number of features.

  3. Double each file’s sampling rate while keeping the sampling frequency constant.

  4. Append labels to the resulting feature arrays.

Expected outcome

Feature arrays with labels are prepared for model learning.

Guide limitations

  • The README does not specify installation steps, code, or a complete feature-extraction configuration.
  • The documented project reports that its tuned CNN achieved a little over 70% validation accuracy; performance may vary.
  • The model’s documented emotion labels are limited to five emotions for male and female voices.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
17 October 2026
Original website title
MiteshPuthran/Speech-Emotion-Analyzer · Official homepage title · retained in its source language
How we review sourcesSuggest a correction