Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Modeling Analog Dynamic Range Compressors using Deep Learning and State-space Models

Hanzhi Yin, Gang Cheng, Christian J. Steinmetz, Ruibin Yuan, Richard M. Stern, Roger B. Dannenberg

TL;DR

A deep learning model with S4 layers to model the Teletronix LA-2A analog dynamic range compressor is presented, which achieves roughly the same quality as previous deep-learning models but with fewer parameters.

Abstract

We describe a novel approach for developing realistic digital models of dynamic range compressors for digital audio production by analyzing their analog prototypes. While realistic digital dynamic compressors are potentially useful for many applications, the design process is challenging because the compressors operate nonlinearly over long time scales. Our approach is based on the structured state space sequence model (S4), as implementing the state-space model (SSM) has proven to be efficient at learning long-range dependencies and is promising for modeling dynamic range compressors. We present in this paper a deep learning model with S4 layers to model the Teletronix LA-2A analog dynamic range compressor. The model is causal, executes efficiently in real time, and achieves roughly the same quality as previous deep-learning models but with fewer parameters.

Modeling Analog Dynamic Range Compressors using Deep Learning and State-space Models

TL;DR

A deep learning model with S4 layers to model the Teletronix LA-2A analog dynamic range compressor is presented, which achieves roughly the same quality as previous deep-learning models but with fewer parameters.

Abstract

We describe a novel approach for developing realistic digital models of dynamic range compressors for digital audio production by analyzing their analog prototypes. While realistic digital dynamic compressors are potentially useful for many applications, the design process is challenging because the compressors operate nonlinearly over long time scales. Our approach is based on the structured state space sequence model (S4), as implementing the state-space model (SSM) has proven to be efficient at learning long-range dependencies and is promising for modeling dynamic range compressors. We present in this paper a deep learning model with S4 layers to model the Teletronix LA-2A analog dynamic range compressor. The model is causal, executes efficiently in real time, and achieves roughly the same quality as previous deep-learning models but with fewer parameters.
Paper Structure (12 sections, 1 equation, 3 figures, 1 table)

This paper contains 12 sections, 1 equation, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. Background
  3. Structured State Space Sequence Model (S4)
  4. Feature-wise Linear Modulation
  5. Methods
  6. Proposed Model
  7. Experiment
  8. Result and Analysis
  9. Objective Loss
  10. Subjective Listening Study
  11. Real-time Performance
  12. Conclusion

Figures (3)

  • Figure 1: The proposed S4 model. It mainly comprises a stack of S4 blocks, where the S4 layer models the temporal dimension, the linear layer models the channel dimension, the FiLM layer applies external controls, and PReLU layers apply non-linearities.
  • Figure 2: Subjective Evaluation Scores among all Clips
  • Figure 3: Speed Ratios in Different Buffer Lengths.