EchoRate
Bilingual Audio Evaluation Portal
Listen • Compare • Evaluate
Sign in
Access the evaluation portal
Bilingual Audio Evaluation Portal
Access the evaluation portal
Your account registration request has been submitted to the administrator.
You will be able to log in normally as soon as an administrator approves your registration.
Live overview of your bilingual speech evaluation experiment
Calculated in real time across all submitted evaluations in SQLite
Latest scores submitted by human evaluators
| # | Evaluator Name | Email Address | Sample # | Scores (N, I, E, O) | Timestamp |
|---|
Upload a folder or files containing audio pairs, CSV metadata, or document transcripts
Choose the folder or files that contain audio pairs and metadata (e.g., CSV, DOCX/TXT transcripts, JSON)
.wav or .mp3 files for both Model A and Model B in both target languages.metadata.csv) with columns: sample_id, audio_a_path, audio_b_path, language, transcript_a, transcript_b..docx, .pdf, or .txt formatted with sentence pairs and domain tags.| Dataset Name | Upload Type | Samples | File Size | Upload Date | Status | Action |
|---|
View, search and monitor all evaluation records
| # | Name | Sample ID | N | I | E | O | Time |
|---|
Consolidated single-row user activity, evaluation metrics, and registration approvals
| Name | Email ID | Registered Date | Action |
|---|
| # | Name | Evaluated | Overall Score | Download | Action |
|---|
Download complete research evaluation results and statistical summaries
All 82,431 submitted evaluation ratings with evaluator IDs, sample IDs, timestamps, and metric scores.
Raw JSON structured data with audio waveform seeds, duration metadata, and comparative scores.
Single row per evaluator containing Audio used, Text used, and Overall Score (0.3×AD + 0.4×I + 0.3×NATu).
Create immediate SQLite snapshots, download database archive, or restore evaluations from a previous backup.
Configure research scoring metrics, admin security, and portal parameters
Mean Opinion Score (MOS) weighting for benchmark ranking:
Type your score directly for each metric (0 – 100 scale).