Request for an Article: What are the latest frontiers re:"Audio" in the current "not-yet-AGI" era?

[neRok]

As per the title, I would be very interested in any news/gossip from people "in the know" about the current what/where/why of "frontier audio" that's come about because of the current "frontier-AI"!?

To clarify; I'm not interested in stuff like "AI will EQ your music on demand" - but I am interested in anything that's along the lines of this: "An OpenAI model has disproved a central conjecture in discrete geometry"...

Now it may be that some posters here do know secrets that can't divulged publicly (those people can send me those secrets via pm though), but I also remember something posted/said by @René - Acculution.com that was along the lines of "we are simulating things that are going to make better [speakers]"....

 

[john61ct]

Master AI (Artificial Intelligence) Discussion/News Thread

I don't need to tell you that AI is upon us big time. It is important that we all keep up with its developments, issues, etc. Instead of having multiple threads on it, I thought I create this single one so that people who are not interested, can just ignore it. And others know where to find...

www.audiosciencereview.com

 

[Redacted]

Wot?
The real frontier is getting new transducer tech for speakers.

 

[Curvature]

Tech advances:
(1) Radiation pattern control through transducer, baffle and waveguide optimization.
(2) Pattern and propagation measurement and simulation.
(3) Auditory model simulation.
(4) Small room (home) and large room (concerts, etc.) correction.

3 has an immense amount of generative AI and machine learning backing new developments, mostly for medical use, more than elsewhere as far as I'm aware, but I tend to pay more attention to the perceptual side of the science.

 

[Redacted]

Bring it on with measurements* at affordable prices
*yes, I know about rooms - I have one.

 

[René - Acculution.com]

I have done quite a lot of work in topology optimization as a way of controlling sound fields, but there is a lag of probably three years between the technique being applied in products coming out in 2023 and then presenting some of this work (https://www.researchgate.net/public...mized_Tweeter_Waveguides_for_Automotive_Audio). Most companies are not keen on sharing what they/we do, so it is difficult to say much about what is going on currently, but I am looking a lot at Deep Neural Networks currently as an example. I will say that it is not the HiFi industry that is heading the way when it comes to these new frontiers. The work I do with companies making smartspeakers and conferencing equipment is more complex, with the devices having many microphones, picking up voices in difficult environments, beamforming, DGFEM/BEM/Ray acoustics simulations in rooms with the devices in it, cheap transducers that are real-time modelled within the devices to compensate for non-linear effects, and so on. But I guess it is all largely incremental. No new technique has popped up recently and changed things dramatically.

 

[EERecordist]

I would look at papers by the Audio Engineering Society, the Acoustical Society of America, and research at Dolby and Fraunhofer. Psychology papers on perception and the audio cortex would be good to monitor.

Most AI work is focused on large language models. What you are asking is in the realm of physics and medicine which are less further along.

This was a recent student paper winner: https://pubs.aip.org/asa/poma/artic...ning-distance-dependent-spatial-structure-for

 

[mhardy6647]

Wot?
The real frontier is getting new transducer tech for speakers.

Pretty much boils down to make stuff vibrate.
... until and unless neuroscience starts mainlining the stuff straight into our brains. I think I'll stick with my Altecs.

I would think the biggest frontier would be to develop tools, technology, and strategies to convince otherwise seemingly intelligent, clever, and well-educated people that wire is wire -- the only things about it that matter in a hifi context are R, C, and L, and -- at audio frequencies -- R probably matters most (and not all that much, except in extreme cases).

But maybe I'm an optimist.

 

[Salt]

Pretty much boils down to make stuff vibrate.
... until and unless neuroscience starts mainlining the stuff straight into our brains. I think I'll stick with my Altecs.

I would think the biggest frontier would be to develop tools, technology, and strategies to convince otherwise seemingly intelligent, clever, and well-educated people that wire is wire -- the only things about it that matter in a hifi context are R, C, and L, and -- at audio frequencies -- R probably matters most (and not all that much, except in extreme cases).

But maybe I'm an optimist.

Maybe R is the most impacting factor in most Minds to not accept this simple calculation ...

 

[Matias]

For every audiophile interested in new tech and frontier science it seems to be there are 3 more into vinyl, tubes and point to point soldering vintage...

 

[mhardy6647]

For every audiophile interested in new tech and frontier science it seems to be there are 3 more into vinyl, tubes and point to point soldering vintage...

... and horns.
Don't forget horns.

oh, tape, too. analog tape.

PS I am guilty as charged.