.png)
Orphic Souns -Audio Engineering
1. Enhancement
Description:
The original recording had an extremely low volume, making speech nearly impossible to understand. The audio was amplified, background noise was suppressed, missing frequencies were reconstructed, and the signal-to-noise ratio was significantly improved. The result is a clear, intelligible recording that can be played at normal listening levels on any device without distracting noise or artifacts.
Credit:
Audio sample sourced from the “Diner Party Corpus.” Used under permitted academic/data-research licensing.
Examples
2. Cleanup
Description:
A set of short examples demonstrating noise reduction, speech clarification, and basic filtering across several common noise environments.
Credit:
Samples sourced from the “Noisy Speech Corpus – ITU-T P.835.”
Examples
3. Restoration
Description:
This example originates from the historic presidential recordings of Lyndon B. Johnson (LBJ). The restoration process included removal of tape hiss and noises, correction of magnetic distortions, spectral balancing, dynamic smoothing, and increasing intelligibility while preserving the authentic character of the archival material.
Credit:
Audio sourced from the “LBJ Presidential Recordings” archive.
Examples
4. Dialogue – Selective Vocal Encryption
Person 1, Person 2 → Person 3
Description:
A two-speaker dialogue. Speaker 1 (female) retains her original voice. Speaker 2 (also female) undergoes a selective vocal-encryption process that transforms her voice into a unique male voice. The transformation is non-reversible and cannot be converted back to the original speaker’s identity.
Credit:
Dialogue sourced from the “Shakespearean Dialogues Collection.”
Examples
5. Monologue – Vocal Encryption
Person 1 → Person 2
Person 1 → Person 3
Description:
A single monologue processed into two separate encrypted identities:
• Person 1 → Person 2: Original male voice transformed into a natural-sounding female voice.
• Person 1 → Person 3: Original male voice transformed into a different male voice with a new, non-reversible vocal identity.
Credit:
Audio sourced from the “Shakespearean Dialogues Collection.”
Examples
More Extreme Cases & Private Demonstrations
The examples above represent only a small portion of our daily work. Many cases handled by our lab involve severely compromised audio, including sources commonly considered a “total loss.”
Due to information-security requirements, confidentiality, and non-disclosure agreements, these extreme cases cannot be shared . However, clients are welcome to send a short excerpt from the audio they wish to improve, and we will provide a private demonstration of what can be achieved.