April 2025

Are We Going Too Far, Too Fast?
Knowing in advance that a particular approach is not feasible is also very valuable.
Editorial By J. Martins

  As a premise for many articles that we review, support, and approve for audioXpress, there are many that actually suggest more questions than answers and essentially make us question our acquired knowledge. That's OK, and we encourage that. We should always look at new questions in order to find solutions for problems we ignored, disregarded, or simply were not aware existed.

Please feel free to send us suggestions, abstracts, drafts, or fully expanded essays on audio-related topics. In any discipline.

I fundamentally believe we have a long way to go to fully understand sound propagation and how to optimize our own electroacoustical systems to fully leverage sound reproduction. We continuously review research work and published material going back multiple decades — something we are forced to do to be able to recognize how much of what is being written today is actually new and "advances the evolution of audio technology" as our magazine motto states. And the more we do it with today's technology and acquired knowledge, the more we identify ignored problems, oversimplifications of concepts that were not understood at the time and were not considered to be relevant, and sometimes lines of thought that lead to a conclusion that was misleading because the author didn't previously eliminate potential variables in his own research. Does that make that work invalid or less useful? Certainly not. Many such contributions have served to create foundations over which we all benefited in most cases. Truly serving the evolution of audio technology.

As with many of the article submissions we receive, trying to answer a question is always the start to finding something new. It's just that sometimes a question only brings more questions until we can identify solutions. And many solutions didn't correctly identify the problem — which is why we sometimes need to decline a submitted work.

On this topic, I cannot resist sharing this sentence that René Christensen wrote in his Simulation Techniques article included in this issue: "...knowing in advance that a particular approach is not feasible is also very valuable. I have worked with clients that had spent months taking a trial-and-error approach to solve a problem, where I could show with a combinatorics analysis that it would take literally millions of years to solve in this manner due to the number of variables involved and their interdependencies."

René Christensen is an expert in mathematical modeling, well-versed in various analytical methods, and with an admirable ability to clarify underlying problems before trying to engineer broader-level solutions. And his published materials for audioXpress reveal exactly how fundamental it is to use the correct tools to analyze even the most basic phenomena so we can build upon a solid foundation: from math to advanced numerical simulations.

In a previous editorial (audioXpress, February 2025), I alluded to the importance of artificial intelligence (AI) for the audio industry in 2025. I can tell we are at a critical stage on the evolution of technology in general — and inevitably reflecting deeply in audio.

Audio is at the core of so many application segments these days, and it's easy to see how audio is at the front of bridging AI with the physical world through sensors, transducers, human interfaces, and interactions. In this issue's Market Update, I mention examples of advanced AI-based digital signal processing. But the most exciting frontier is already becoming visible in 2025, as we witness the return of a new generation of voice assistants and voice interface applications.

I truly believe that language is the largest market opportunity for the audio industry. Because people speak — and think — in hundreds of languages, and AI technologies are finally at a state to be able to assist in bridging those "islands" that today separate people, cultures, and much of our understanding of the world.

Audio is the simplest of all signal disciplines that interface with the human senses, and sound classification through deep learning is an effective way to enhance our sense of hearing. I mentioned bridging the AI with the physical world. Hearing is the most effective human sense to approach this. Essentially because audio signals, processed through audio front ends and audio classification (trained models) plus language understanding, create our best "combo" to human extensions.

Independently of the goal being augmented reality, hearing enhancement, or hearing augmentation, AI is helping us to get there faster and more effectively. AI-based language support and translation is the most effective foundation to augment our human abilities.

J. Martins
Editor-in-Chief

Subscribe To audioXpress Magazine
Click here to subscribe to audioXpress magazine.