Innovation is rampant in audio technology. New signal processing and machine learning solutions are emerging on an almost daily basis, and experimenting with audio tools frequently yields new creative approaches. However, bringing such innovation to market poses many challenges. This talk addresses these challenges while drawing on experience with several high-tech audio start-ups. It focuses on questions and dilemmas concerning, for instance, IP protection, investment, market size and potential, and early-stage growth that are specific to the audio industry. Concrete examples are given of successes and failures where audio developers have attempted to bring new technologies to market.
Link to Slides:
_
Josh Reiss
Josh Reiss is Professor of Audio Engineering with the Centre for Digital Music at Queen Mary University of London. He has published more than 200 scientific papers (including over 50 in premier journals and 6 best paper awards) and co-authored two books. His research has been featured in dozens of original articles and interviews on TV, radio, and in the press. He is a Fellow and currently President of the Audio Engineering Society (AES), and chair of their Publications Policy Committee. He co-founded the highly successful spin-out company, LandR, and recently co-founded RoEx, Tonz and Nemisindo, also based on his team’s research. He maintains a popular blog, YouTube channel, and Twitter feed for scientific education and dissemination of research activities.
_
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
Join Us For ADC24 - Bristol - 11-13 November 2024
More Info: https://audio.dev/
@audiodevcon
Three RADical Concepts in the Art of C++ Coding - Chris Nash - ADC 2023
What if MIDI was a programming language?
What if C++ had built-in audio semantics?
What if you could develop C++ plugins, live in the DAW?
This talk explores these ideas and the development of new technologies designed to blur the lines between music and code, for both artists and developers, and challenge traditional ways of thinking and working.
Drawing on concepts of flow, liveness, and rapid prototyping, the talk will present live demos, and discuss the development of:
Manhattan - a digital audio workstation and embeddable API built on a procedural music engine that integrates sequencing and programming. Used by artists, game composers, and in teaching computational thinking, example applications include crowd-driven music using machine vision, a Unity mini-game featuring a live (and somewhat mortal) orchestra, plus a growing library of famous works recomposed as code that shows the power of modelling music as both pattern and process.
Klang - an open C++ dialect (language extension) for audio, using modern language features (C++14/17) to extend the semantics of C++ to encapsulate audio, providing DSP primitives and types, and adapting the STL's concept of stream objects and operators to represent signals. Easier to read, more concise, and easily mapped to visual forms (block diagrams, Max), Klang feels like a new language (in the spirit of SOUL) but, as pure C++, retains the performance, portability, compatibility, and interoperability of the industry standard.
rapIDE - a C++ IDE inside a DAW plugin, designed for rapid audio prototyping and development of synthesisers and effects. Built on a full clang/LLVM-based toolchain, the plugin's source code can be live edited, rebuilt, reloaded and auditioned without restarting the DAW (or stopping playback). Compatible with C++ and Klang, rapIDE is designed to improve the accessibility, liveness, and immersion of audio programming, for applications in rapid prototyping and teaching, featuring realtime debugging, auto-complete, code sandboxing, and built-in audio analysis.
These technologies will support the new Music Systems Engineering (MuSE) degree programme, in development by Point Blank Music School in collaboration with industry, for launch in 2024.
_
Chris Nash
Chris Nash is a software developer, composer, educator and researcher in things that go beep in the night. Following a PhD on music software design at Cambridge, he has worked on technology and music projects across academia and industry, including for the BBC, Steinberg/Yamaha, and multiple start-ups, and independently develops and maintains several software projects, specialising in computer music and making music programming more accessible, including Manhattan (a hybrid DAW/programming language), Klang (a C++ dialect for audio), rapIDE (a plug-in based C++ IDE) and reViSiT (an award-winning plug-in based sound tracker). He is currently Senior Lecturer in Software Development for Audio, Sound, and Music at UWE Bristol, and recently founded nash.audio, a non-profit organisation supporting creativity and learning in music technology. Working with London-based Point Blank Music School, Dr Nash is the architect of the forthcoming MuSE (Music Systems Engineering) course, developed in collaboration with industry to be the world's first audio developer degree programme.
_
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
Join Us For ADC24 - Bristol - 11-13 November 2024
More Info: https://audio.dev/
@audiodevcon
Running High Channel Count Audio Applications on Linux RT - Olivier Petit - ADC 2023
Linux-based computing platforms are extremely popular to implement audio processing in embedded systems, from low power consumer devices running on ARM processors to professional multichannel solutions requiring the power of x86 based chips.
In this talk we will explore the different features that the Linux kernel offers to control real time performance and ensure glitch-free audio processing. We will study examples from a commercially available and actively maintained product, including successes and failures.
Topics that we will look at include:
• Linux kernel Real-Time patch
• Controlling thread real time priority and CPU affinity
• Measuring performance
• Common pitfalls
_
Olivier Petit
After an MSc in Integrated Circuit design, I have joined the Creative Technologies department of L-Acoustics in 2018 as a C++ software engineer. I have been taking an active part in developing innovative technologies to bring immersive audio to live performances, striving to better connect the artists with their audiences. I love the challenge of deploying audio solutions at large scales with high sound quality and (of course) reliability. Making sure our products are as intuitive and easy to use is also a priority for me. Unlike most of my peers I don’t play any instrument, I love climbing though!
_
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
Building a Music Assessment Engine: Requirements, Challenges, and Solutions - Martin Gasser - ADC23
Automatic music assessment systems have become an integral part of music (self-)education in recent years, with systems like Yousician, Smart Music, and MuseClass gaining widespread attention. In this talk, Martin Gasser will share an overview of the requirements of a music assessment system — both from a musical and a technical point of view. Technical limitations in low-resource environments like smartphones will be discussed. Solution approaches will also be described, and Martin will briefly demonstrate the performance of Muse Group's new music assessment system on real-life data.
_
Martin Gasser
Martin Gasser is a Senior Developer at Muse Group. Formerly a leading researcher in the field of Machine Learning & Music at the Austrian Research Institute for Artificial Intelligence, Martin has years of expertise in AI, musical software development, hacking C++, and more. Full bio: https://martingasser.com/bio/
_
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
NO MORE CODE: Introducing a New Graphical Programming Language for Audio Research and Developers - ChangHun Sung - ADC 2023
Audio software development is rapidly moving towards incorporating machine learning-based processing. While research scientists are continuously presenting us with inventive results in the field of AI, there is a lack of software engineering tools to utilize these results.
We are introducing a new development framework to bridge the gap between AI researchers and software engineers, allowing both fields to work together seamlessly.
By borrowing some concepts from functional programming languages, we designed a new node-based graphical programming language for audio software development, which is capable of handling a variety of data types including audio, multi-dimensional tensors and MIDI. Furthermore, the system is designed to support looping and branching logic using only nodes and node connections. This allows the implementation of highly flexible data processing algorithms.
Finally, as an AI voice technology company, we will present a demonstration of the plug-ins developed using our new programming language, showcasing its adaptability and advanced capabilities.
Link to Slides: https://data.audio.dev/talks/2023/no-more-code/slides.pdf
_
ChangHun Sung
Chang Hun is a Software Engineer at Supertone, Inc. He previously worked in the game industry and used to develop game engines. Chang Hun now develops high performance C++ frameworks to accelerate the process of productizing ML models. He is also the principal clarinetist in an amateur orchestra in Seoul.
_
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
Join Us For ADC23 - London - 13-15 November 2023
More Info: https://audio.dev/
@audiodevcon
Democratization of Audio Dev: How it Started, Where it's Going - Aurelius Prochazka
In 2018, I gave a keynote at the ADC Conference in London. In that keynote, I discussed how audio developers can benefit from opening development to all, and how to escape the errors of past and design for the future. What did I get wrong? Did I get anything right? In this talk I'll give an overview of where we are now and update where things might be going.
I'm Aurelius Prochazka, I am the founder of AudioKit, an open-source audio generation, processing, and analysis framework for the Apple ecosystem. I am a partner in AudioKitPro LLC, a company started to use the AudioKit framework to develop audio applications. Currently, AudioKit has delivered more synthesizer apps on iOS ever, including the most #1 charted music apps ever and the perennial global favorite, the free AudioKit Synth One.
In 2018 I traveled to London to give one of the keynotes for ADC in London. I have a Ph.D. from the California Institute of Technology in Computational Fluid Dynamics.
Edited by Digital Medium Ltd - online.digital-medium.co.uk
_
Organized and produced by JUCE: https://juce.com/
_
Special thanks to the ADC Team:
Sophie Carus
Derek Heimlich
Andrew Kirk
Bobby Lombardi
Tom Poole
Ralph Richbourg
Jim Roper
Jonathan Roper
Join Us For ADC23 - London - 13-15 November 2023
More Info: https://audio.dev/
@audiodevcon
Processing Audio Post Production Deliverables at Scale with Microservice Architecture - Ryan Frias - ADCx SF
With the explosion of streaming services that operate on a global footprint, the audio mastering pipeline has ballooned to include as many as 765 unique language audio packages created from original, each of which needs to be quality checked, encoded and packaged for delivery. The process can take weeks and involves creative and technical teams working to make derivative versions which must preserve the original creative intent as much as possible. The "Coda" Automated Media Ecosystem is a new extensible software platform from Skywalker Sound that automates the creation of soundtrack versions and cuts the deliverable process down from weeks to faster-than real-time. The system has already been used on premium Disney+ releases such as The Mandalorian and Moon Knight.
By automating soundtrack mastering processing from the highest original source mix format (often the Dolby Atmos mix), the automatically derived versions can be created with the same fidelity and attention to detail as the original language, improving the experience for all consumers in international markets. As the process is automated without the possibility of human error, the need to QC each pass is reduced, resulting in better results for consumers and considerable time and cost savings for content owners.
The creation of this complex media pipeline ecosystem involved the integration of multiple programming languages, frameworks, environments and hardware systems for the full stack--from low-level digital signal processing, to high level scaffolds, such as a global service mesh.
Slides: https://data.audio.dev/talks/ADCxSF/2023/audio-in-midi-out/slides.pdf
_
Ryan Frias
Ryan is a software engineer from Skywalker Sound. He has worked in the post production sound industry for over 15 years, focusing on both the creative and technical bits.
Edited by Digital Medium Ltd - online.digital-medium.co.uk
_
Organized and produced by JUCE: https://juce.com/
_
Special thanks to the ADC Team:
Sophie Carus
Derek Heimlich
Andrew Kirk
Bobby Lombardi
Tom Poole
Ralph Richbourg
Jim Roper
Jonathan Roper
Join Us For ADC23 - London - 13-15 November 2023
More Info: https://audio.dev/
@audiodevcon
Using Faust in JUCE Projects - Julius Smith - ADCx SF
Faust (Functional AUdio STream) is a domain-specific functional programming language for sound synthesis and audio processing (https://faustdoc.grame.fr). Faust development is routinely an order of magnitude faster to write and debug than C++, and the run-time performance is usually within a factor of two. Development is further facilitated by the large Faust Libraries collection of signal-processing functions (https://faustlibraries.grame.fr). Since Faust compiles to C++, it is natural to integrate C++ header files generated by the Faust compiler into a JUCE project or the like. A limitation encountered is that Faust's powerful pattern-matching facility requires its "macro arguments" to be bound at compile time, i.e., not signals or controller parameters. An example is the filter order in Faust's filter-design functions. When these arguments need to change at run time, it can make more sense to translate Faust to C++, so that the recursive pattern-matching in Faust is replaced by recursive C++ functions. Experience using ChatGPT-3.5 and ChatGPT-4 for this purpose and others will be summarized.
_
Julius Smith
Professor Emeritus of Music and by courtesy Electrical Engineering, Stanford University, based at the Center for Computer Research in Music and Acoustics (CCRMA). Activities include teaching courses in signal processing and music technology, graduate student advising, and research in signal processing applied to music and audio. https://ccrma.stanford.edu/~jos/
Edited by Digital Medium Ltd - online.digital-medium.co.uk
_
Organized and produced by JUCE: https://juce.com/
_
Special thanks to the ADC Team:
Sophie Carus
Derek Heimlich
Andrew Kirk
Bobby Lombardi
Tom Poole
Ralph Richbourg
Jim Roper
Jonathan Roper
Join Us For ADC23 - London - 13-15 November 2023
More Info: https://audio.dev/
@audiodevcon
DDSP-VST: Neural Synthesis for All - Nikhil Bhanu & Wilson Zhao - ADCx SF
We present DDSP-VST, a cross-platform audio plugin for real-time neural audio synthesis built using TensorFlow Lite and JUCE. DDSP is the underlying ML research behind this project that lets you combine the interpretable structure of classical DSP elements such as filters, oscillators, reverberation etc. with the expressivity of deep learning. This technology enables us to transform a given audio signal into any musical instrument by extracting the pitch and loudness and using those features for synthesis. In this talk we will discuss the technical and UX challenges of building an audio plugin powered by machine learning and making it accessible to musicians and artists.
_
Nikhil Bhanu
Audio software generalist with a passion for real-time systems, signal processing, musical interfaces and machine learning.
Wilson Zhao
Electronic music enthusiast + event organizer. Passionate about combining multimodal generative AI + creativity.
Edited by Digital Medium Ltd - online.digital-medium.co.uk
_
Organized and produced by JUCE: https://juce.com/
_
Special thanks to the ADC Team:
Sophie Carus
Derek Heimlich
Andrew Kirk
Bobby Lombardi
Tom Poole
Ralph Richbourg
Jim Roper
Jonathan Roper
Join Us For ADC23 - London - 13-15 November 2023
More Info: https://audio.dev/
@audiodevcon
Test Strategies for Multi-Threaded Code - Ryan Avery - ADCx SF
Multi-threaded code can be complicated. Realtime multi-threaded code even more so. And testing that code? Too often we give up faster than we can say “integration test”. But by switching around your thinking and adopting a few key strategies, you can start to tame the complexity and gain back a little bit of confidence in your critical code.
Slides: https://data.audio.dev/talks/ADCxSF/2023/test-strategies-for-multi-threaded-code/slides.pdf
_
Ryan Avery
Ryan is an audio software engineer who has worked in the industry for over 15 years. He is passionate about software quality and best practices in real-time audio software development. At Avid Technologies, he developed the Pro Limiter and Pro Multiband Dynamics plugins for Pro Tools. At Dolby Laboratories, Ryan helped develop a suite of plugins for producers to create spatial music in Atmos.
Ryan currently works at Apple, and is an active music producer under the alias of Chance's End, where he brings violin to the front of electronic production.
Edited by Digital Medium Ltd - online.digital-medium.co.uk
_
Organized and produced by JUCE: https://juce.com/
_
Special thanks to the ADC Team:
Sophie Carus
Derek Heimlich
Andrew Kirk
Bobby Lombardi
Tom Poole
Ralph Richbourg
Jim Roper
Jonathan Roper
