Volume 2, Issue 10
C2PA Gets Some Updates
First up, news from the C2PA—there is a new release (2.5) of the content credential specification. The main public-facing benefits are expanded support for live video, closed captions, RSS feeds, and AI artifacts, although there are certainly other improvements. However, following on the heels of that announcement, the C2PA released an update last week that includes a guidance document for working with audio assets. I am proud to say this effort came from the Audio Task Force that we contribute to here at Catamount Music.
One of the strongest contributors to this document was Matthew Rappard, who also penned a great post on KuppingerCole last month, entitled “From Studio to Stream: Why Music Needs Content Credentials Now.” I don’t want to steal his message, so if you believe that generative tools are making it easier to replicate content in an environment with poor to no transparency, you can take some comfort that C2PA can be implemented in coordination with standards like DDEX and provide crucial (and missing) provenance information.
Some Sync Advice
Although it was a difficult decision, we decided not to attend the Taxi Road Rally this year. So we will wave from here and hopefully catch up with friends we’ve made there next year. But Taxi is more than the Road Rally. They also have Taxi TV, which covered the use of AI in music back in October. The guest speaker was Paul Croteau, who is a composer but also has a long history in technology. The video is a bit long (100 minutes or so), so I’m going to summarize a few key points here, because I think they get overlooked sometimes in the conversations:
- Rerecording an AI composition doesn’t make it yours. It just means you created a new sound recording based on an underlying composition that you don’t own.
- Generative AI can produce music that infringes unintentionally
- Don’t try to game the system by ripping streams, breaking watermarks, etc.
They go on to talk about the potential impacts of getting this wrong and how to use AI safely in the creative process. Taxi continues to promote a “zero AI in sync” approach. I don’t like the way it’s described, to be honest, because I think it allows composers to ignore AI, and you do that at your peril. The reality is that much of the sync composer business might be taken over by AI in the coming year or two. When it flips, I suspect it will flip fast to automated systems that don’t have humans at all. Which is why musicians in general should not ignore generative AI music, but rather, seek to understand it and be ready to capitalize on that knowledge as the market shifts.
MusicTectonics 2026
Honestly, this conference is a stand-out. The venues are spectacular, the attendees are all interesting, the presenters are experienced and diverse, and it’s still small enough to be intimate at under 1,000. I was there to pitch our Vervain Audio product (more on that below) and network. We also entered the Narwhal’s competition but sadly did not advance. Nevertheless, I had great experiences there and met a lot of people. What follows here is a slice of people I met with interesting technology or platforms.
Vijay Basrur / Lyrcs.ai
I saw Vijay Basrur pitch a few times at the conference, and his passion came through so clearly each time. Vijay loves South Asian music and is on a mission to build a “lyrics infrastructure for the world’s music.” But focusing on South Asian music. The site can transcribe, transliterate, and translate lyrics in 52 languages, focused specifically on South Asian countries and those close by. Now, I am an English-speaking Westerner, so as much as I appreciate what Vijay has built, I also know that there is a lot of subtlety there that I can’t fully understand, as I have almost no experience with any of the languages involved. But I have tremendous appreciation for the care that has gone into this system so that it can translate artistic and poetic concepts faithfully.
Dan Conley / Guitars for Vets
Dan has popped up on my music tech radar a few times, and we finally got a chance to meet at Music Tectonics. Dan is an impressive guy. He’s retired from a long career at Microsoft and now runs a local chapter of the Guitars for Vets organization. Guitars for Vets uses the power of guitars to bring healing to veterans who need it. They provide a 10-lesson introduction to the guitar and then give the veteran a guitar when they’ve completed the program. What a great organization! If you are looking for a new philanthropic target, please consider donating an acoustic guitar or supporting them in other ways.
Aren Barseghyan and SoundGate
I met Aren at the inventor’s expo, where I got a chance to see SoundGate in person. Before we even got to the demo, though, we started talking guitars. Aren had a beautiful new Fender Strat there for use with the demo, which led to a delightful side conversation. Aren, thanks for bringing that along. It must have been a royal pain to do, and it would freak me out to put such a beautiful new guitar on the expo floor, but it was awesome, and I think I can speak for everyone who played it—thanks!
Anyway, Aren has a suite of tools to help guitarists develop their skills. It uses visual prompting and listens to the guitar signal to provide playing feedback. There was way too much for me to explore, so I decided to try a simple scale. G major. Two things I really like about this demo:
- The display shows a guitar neck with the position of each note superimposed on the fret. These superimpositions then move along with what you played and when. This kind of feedback is comfortable and natural to me, as it is how I (and many, many guitarists) are used to visualizing exercises.
- The scales are very flexible. For example, I could select different configurations of notes per string. It defaulted to third position for G-major with two and three notes per string, but I was able to select a 3-note-per-string pattern that runs up the neck, which is what I was playing.
I have to admit, I have not used Fender Play or other apps that might be considered competitors, so this is by no means an exhaustive review. But iPhone users should be able to try this today, so here’s a chance to experience something new and innovative right from the start.
Patrick Paul / Mixfinity
I met Patrick Paul early in the conference. He’s building hardware too, so we connected on that level. And his product, Mixfinity tape, is awesome. It’s an audio player in a physical cassette-tape format with an e-ink display on the side. He’s invoking the old days of making mix tapes, but in a slick modern package. I am pretty much in awe of his engineering skills. Here’s hoping this wicked pissah playah will be a commercial success!
Stephen Zoepf / DZCO
I ran into Stephen a couple of times at Music Tectonics. We had some esoteric connections, like we both have done work in the Volpe Center in Cambridge. It is a small world, sometimes. Anyway, Stephen has founded DZCO—a service that maps concerts, venues, set lists, and other live performance information globally. So you can either look for concerts you’ve been to or plan concerts and trips in the future. The system uses a beautifully rendered globe to display all the places where it has content about shows. It also has some tie-ins, like printing a T-shirt “world-tour” of all the concerts you have attended globally.
Danny Fan and One Systemics
I met Danny at the UMG Founders Event. He’s developed a patch bay for AI models, and he’s building the idea of provenance into it. As you can imagine, we had a fabulous time talking about C2PA and ingredients and stuff like that. But his software does more than that. It allows for you to, for example, put a set of different AI models in a workflow to look for (or listen to) the differences in all the outputs without running the full model. So if you want to check denoise solutions, or simple generative models, you can select them locally and run your audio through it. Check out his paper, GroovyUI: A Patch-Bay for Local AI Audio Exploration. In one example, he builds a tool chain to stem the vocal track of a song, transcribe the lyric, translate the lyrics to another language, and speak them in that language. Pretty cool stuff, and I am very excited to see provenance represented in the work.
You can try this too! It’s on github and it is open source. At the very least, go give him a star.
Vervain Audio
Warning—this article is a total conflict of interest, because I’m writing about my own stuff. You have been warned.
We’ve been quietly building out the Vervain Audio product and story, and we started talking openly about what we are building at Music Tectonics. We don’t have a press machine or even a social media presence yet—that is all coming—because we are spending all our time trying to build an amazing product that, more importantly, serves its customers in the best way possible.
What we announced at Music Tectonics is an always-on musical instrument recorder for guitar and bass that applies Content Credentials to all the music it saves. That music is uploaded to the cloud, where it forms a personal library of everything played through the recorder, along with contemporaneous provenance data. This is a hardware device that takes a ¼” signal input and has a single buffered output so you can run the rest of your pedal chain/amplifiers as normal. We are working musicians building this company, and we care deeply about artists’ rights, provenance, and attribution. We are artists, and we are building for artists. There will obviously be much more to come on this over at Vervain Audio and its social media sites.
I’ll need a redo on that
Last month I talked about seeking out different points of view through the AI Music Creators Association. This month, however, I became aware of more provocative content of theirs that does not help advance the conversation. So I am withdrawing that advice and shifting my attention to places like the AI for Music, a set of principles espoused by UMG and Roland that lots of folks are signing on to. The group advocates for human artistry, transparency, and attribution. This is an industry org with a whole bunch of other industry orgs signing on, so let’s just be clear about that. However, I do agree with the core principles.
The Berklee AI Conference is Back in 2027
The dates and call for presenters have been issued for the 2027 Berklee AI Music Summit (AIMS). I was not able to attend this year but will plan on going next year. It is April 23rd and 24th at the Hilton Boston Back Bay.
Leave a Reply