Tag: real-time

Leveraging Pruning & Quantization for Real-Time Audio Applications – Dharanipathi Rathna Kumar

  • Lobby
  • Tag Archives: real-time

Join Us For ADC24 - Bristol - 11-13 November 2024
More Info: https://audio.dev/
@audiodevcon​

Leveraging Pruning and Quantization for Efficient Real-Time Audio Applications - Dharanipathi Rathna Kumar - ADCx India 2024

In the constantly changing world of audio processing, the demand for real-time response and high-quality output is a relentless challenge, especially in the presence of computational constraints. The rapid growth in model complexity, especially in deep learning, has made it challenging to deploy complex models involving architectures such as TCN, LSTM, etc. on resource-constrained devices and/or to achieve fast inference speeds for tasks such as real-time audio effects, audio style transfer, and source separation. Model compression is vital to address these challenges, making it possible to retain high performance while using fewer computational resources. This talk delves deep into two paramount model compression techniques, namely pruning, and quantization, and explores their applicability in the context of audio applications.

Pruning is a method of eliminating redundant or less contributive weights from the model to reduce the computation resources required to run the network. We'll explore its variants, methodologies, and outcomes, and how it can drastically reduce computational complexity without significantly undermining model performance. Quantization is the process of reducing the precision of the weights, biases, and activations such that they consume less memory. By reducing the bit-width of model parameters, we can achieve sizeable savings in memory and computational power, making it indispensable for on-device audio applications for real-time audio contexts.

In this presentation, I elucidate that by adopting strategic weight pruning and parameter quantization, it is feasible to significantly enhance the efficiency of sophisticated audio models, paving the way for robust, real-time audio processing even in resource-constrained environments.

Link to Slides: https://data.audio.dev/talks/ADCxIndia/2024/pruning-and-quantization-for-efficient-real-time-audio-applications.pdf
_

Edited by Digital Medium Ltd - online.digital-medium.co.uk
_

Organized and produced by JUCE: https://juce.com/
_

Special thanks to the ADC24 Team:

Sophie Carus
Derek Heimlich
Andrew Kirk
Bobby Lombardi
Tom Poole
Ralph Richbourg
Prashant Mishra

#adc #audiodev #dsp #audio #audio

Filed under: UncategorizedTagged with: , , ,

Real-time Confessions in C++ – Fabian Renn-Giles – ADC23

  • Lobby
  • Tag Archives: real-time

https://audio.dev/ -- @audiodevcon​
_

Real-time Confessions in C++: The Most Common “Sins” in Real-Time Code - Fabian Renn-Giles - ADC 2023

This talk examines the most prevailing misconceptions and frequent errors encountered when audio developers handle real-time code in C++. With my background as a contractor in the audio industry, I’m often called in to help fix subtle bugs in, or review code with real-time constraints. Yet, I see (and have myself made) the same type of mistakes over and over again resulting from a few common misconceptions of real-time C++ code.This talk offers an in-depth analysis of each of these misconceptions, debunking them with compelling examples from the audio industry.

Ranging from the ignorance of C++'s data-safety rules altogether, to the overuse of std::atomic and misbeliefs about the forbidden usage of locks and exceptions in real-time code, this presentation navigates the landscape between the theoretical rules of the C++ standard and real-world practical realities. This talk is an essential guide for developers seeking to avoid common pitfalls and write more efficient, reliable real-time code.
_

Link to Slides: https://docs.google.com/presentation/d/1rJNebpRg3xJO4AiE5AntrkEKIXRta25DCOxHxDMzC0c
_

Fabian Renn-Giles

Fabian is a freelance C++ programmer, entrepreneur and consultant in the audio software industry. Before this, he was staff engineer at ROLI Ltd. and the lead maintainer/developer of the JUCE C++ framework (www.juce.com) - an audio framework used by thousands of commercial audio software companies. Before joining ROLI, he completed his PhD at Imperial College London, developing a numerical quantum optics solver with modern digital signal processing techniques and C++/MPI/OpenCL. Fabian is now a regular consultant specializing on low-level real-time C++, embedded audio, time sensitive networks, audio over WiFi and audio plug-in formats. His clients range from early startups to FAANG companies. Additionally, he is a regular speaker at the audio developer conference ADC and other C++ conferences.
_

Streamed & Edited by Digital Medium Ltd: https://online.digital-medium.co.uk
_

Organized and produced by JUCE: https://juce.com/
_

Special thanks to the ADC23 Team:

Sophie Carus
Derek Heimlich
Andrew Kirk
Bobby Lombardi
Tom Poole
Ralph Richbourg
Jim Roper
Jonathan Roper
Prashant Mishra

#adc #audiodev #cppprogramming #audio

Filed under: UncategorizedTagged with: , , , , ,