pl | en

No. 258 November 2025

Editorial

text by WOJCIECH PACUŁA
translation by Marek Dyba
images by press releases



No 258

November 1, 2025

ABOUT FAKES
Or what would we say if tomorrow Ozzy Osbourne personally announced his new song, just recorded with the rest of the original Black Sabbath lineup?

AI is one of the hottest topics around. And no wonder, as it is one of the fastest-changing areas of technology. It also has a strong presence in music, and in recent months has become a source of controversy, intervention, musicians removing their music from streaming services, etc. And all because of one band...

ONE OF THE ADVANTAGES OF STREAMING SERVICES, apart from the super-low price for access to millions of recordings, is the opportunity to discover new music. There are various ways to do this, the most popular of which, and in fact accepted as the “standard,” is to allow the service's algorithms to suggest songs and albums. This is a good method, but it is also limiting. These scripts are not very intelligent and tend to lock us into what we already know, based on the type of music stored in the metadata, tempo, or “mood.” But, as I say, it works.

⸜ A cover of the first two The Velvet Sundown albums: Dust and Silence and Floating on Echoes • press release

It seems that “random” choices or following your favorite artist, producer, sound engineer, etc. are much more interesting. This also means sticking to certain boundaries, but if these people are interesting creators, they have crossed those boundaries in their artistic lives, and we cross them with them. So, to stick to the poetics I have chosen, we hack the algorithm and use it at the same time. I also reach for what Tidal has prepared for me from time to time. And on one hot day, it suggested the album Floating on Echos by THE VELVET SUNDOWN.

This music is classified as psych-rock, or psychedelic rock – a musical genre that emerged in the mid-1960s, characterized by influences from psychedelic culture, electronic sound effects, and innovative recording techniques. In the book Buzz Me In, which describes the ups and downs of the Power Plant recording studio, there is a long passage about one such effect, called phasing or flange, which is known from the albums of Small Faces, The Eagles, and Hendrix.

The distinctive features of psychedelic rock included improvised solos, long passages, unhurried rhythms, and a “colorful” sound. It is associated with bands such as The Beatles, The Doors, The Jimi Hendrix Experience, Pink Floyd, and Jefferson Airplane, but also The Zombies and Grateful Dead. Sounds good, right? Especially if we associate the name of the genre with the music we know. And the album Floating on Echos seemed to be the perfect reminder of this scene in a modern recording with refined details. I listened to it for a while, and then it dawned on me that it all sounded very similar – cool, it's a superbly produced and recorded album – but not very diverse.

⸜ A cover of the third The Velvet Sundown album: Paper Sun Rebellion • press release

To see how the band solved this on other albums, I turned to the internet and found that in the same year, 2025, they had released two albums so far and were preparing a third. Only then did I open the pages that had not interested me before, because that was not what I was looking for. And the title of the article on the Guardian website, An AI-generated band got 1m plays on Spotify. Now music insiders say listeners should be warned, should have made me think; this is how tunnel vision manifests itself...

Yes, exactly – the “band,” because now I can use quotation marks, does not really exist, but was entirely generated by AI. Not by AI itself, to be clear, but by people who used it. This applies to both the music and the cover, the biography, and the “promotional” photos. They feature four handsome guys in various poses, polished and retouched to the max, each one beautiful and each one from a different fairy tale, as if they had been selected in a competition for boys band members to attract different types of audiences.

As reported in the article, the “band” gained unprecedented popularity in a short period of time, as within a few weeks its songs were played over a million times on Spotify alone (today it is well over 1.5 million). And further:

Initially, the “band”, described as “a synthetic music project guided by human creative direction”, denied they were an AI creation, and released two albums in June called Floating On Echoes and Dust And Silence, which were similar to the country folk of Crosby, Stills, Nash & Young.

Things became more complicated when someone describing himself as an “adjunct” member told reporters that the Velvet Sundown had used the generative AI platform Suno in the creation of their songs, and that the project was an “art hoax”.

⸜ LANRE BAKARE, An AI-generated band got 1m plays on Spotify. Now music insiders say listeners should be warned, „Guardian”, 14th July 2025, → www.THEGUARDIAN.com, accessed on: 23.09.2025.

By the end of June, after various articles began describing this “band,” a person using the pseudonym Andrew Frelon created a fake account on X and pretended to be the spokesperson for The Velvet Sundown, which turned out to be a complicated media hoax. Frelon deceived many people, including Rolling Stone magazine. Or he was consistent in what he was doing—you can choose either option.

As we read in an article posted on the Berklee College website, following these revelations, Euronews published a statement from a Spotify spokesperson noting that the service “does not prioritize or profit from music created using AI tools. All songs are created, owned, and uploaded by licensed third parties”; more → HERE.

⸜ Photo of the band styled after the famous cover of The Beatles' album Abbey Road • press release

Actually, this should come as no surprise to us. Information that AI will soon replace musicians, Writers, journalists, and everyone who “produces” any content is presented to us on social media and even traditional communication channels with the regularity of an atomic clock. Today, we know that behind most of these revelations are companies releasing next generations of “chats”, thus generating interest in the topic and capitalizing on it.

However, the case in question is different. It was not AI that made an independent decision, but a human being—probably a producer—who used cheap software to create a fantastic fake, a brilliant dud that was difficult to distinguish from the original. Good enough that it could pass for real music, i.e., created in an act of creation, rather than a compilation of millions of the most likely paths of possibility. And, in addition, with excellent sound.

Sound

I HAVE EXCERPTED THESE PARAGRAPHS FROM THE ARTICLE because they are very interesting and not entirely “compatible” with the rest of the article. And they are interesting because it is top-shelf sound. Very distinctive, you can hear it particularly well now, but – very good.

⸜ Something interesting for audiophiles – The Velvet Sundown music is available at Tidal in 24 bits and 48 kHz resolution

When I heard the first recordings remixed using AI, I was very impressed—and I still am today. In the recording of The Beatles' Now and Then (2023), it was used to clean up John Lennon's voice from the noise and distortion of the cassette tape, but not to improve it, as Giles Martin has repeatedly stated. But on the album Revolver (2022), in additional monophonic tracks, algorithms were used to de-mix the material and remix it anew.

The same was done with the master of Ella Fitzgerald's 2020 album The Lost Tapes. Kevin Reeves took care of it using iZotope RX 8 Music Rebalance software. He used it to separate the original stereo mix into four tracks: drums, bass, piano, and vocals, which he then remixed them; more → HERE ˻PL˺. Eva Cassidy's 2023 album I Can Only Be Me was created in yet another way, with AI cutting the singer's voice out from her live recordings and the London Symphony Orchestra later adding their parts to it.

When we recall what I wrote about the sound of remasters by Mouton Records (or the studio that uses that name), we will be quite well prepared for what we get on The Velvet Sundown albums. All these recordings share a kind of warmth and lack of sharpness (and selectivity, for that matter). To a large extent, this resembles AI-generated graphics, including “photos” of The Velvet Sundown.

It's a world that's more beautiful than reality because it's pretty, warm, and super pleasant. On the one hand, this is good because it allows you to bring out a lot of details that were previously buried in distortions. But on the other hand, everything becomes similar to each other, even identical.

THE VELVET SUNDOWN was made possible by the emergence of a tool called Suno on the market. According to Wikipedia, Suno AI, or simply Suno, is a generative music program based on artificial intelligence, designed to generate realistic songs combining vocals and instrumentation or purely instrumental. Suno has been widely available since December 20th. 2023, following the launch of a web application and a partnership with Microsoft, which has integrated Suno as a plugin for Microsoft Copilot; more → HERE.

⸜ And one more photo styled after The Beatles, this time from a performance of the roof of the Abbey Road and Get Back album • press release

Suno allows user to create music in any musical style and does it very well. The free version offers a maximum of 50 credits per day, which is enough for 10 songs, after which you have to wait until the next day before you can generate more. For commercial use of the songs, you need the paid version, but it is not expensive. The Pro plan for $10 per month gives you 2,500 credits, which is enough to generate 500 songs per day (!).

However, the case of the “band” in question is much more complex. This is professional material. It must have been trained on existing music samples, and there is someone behind it who has a good sense of music. On the one hand, it is creative work, reminiscent of the imitations of the great masters. On the other hand, this “training” takes place through simple theft and is presented as a result of someone’s own creativity.

As Sophie Jones, director of strategy at the British Phonographic Industry (BPI) music trade organization, says in the Guardian article, “On the one hand, AI should help artists, but on the other, there is a real risk that large companies will cut corners.” Because cheap music is more profitable than expensive for streaming services, I would add, as that seems clear...

In this context, the opinion of musicians who play contemporary music is particularly interesting. In the article Resistance to artificial intelligence in „Dwutygodnik”, we read that they are not afraid of generative algorithms, because, first of all, there is no money in this type of art, and secondly, this type of music is not based on repetition in the strict sense of the word. These are usually completely new ideas and concepts that AI is unable to generate because its “engine” does not allow for creation, and there is nothing to train it on. This would also refer to jazz music with its idiom of improvisation:

From literature to visual arts to film, there is no shortage of enthusiasts who are convinced that if they follow their own artistic visions, their work will not be defined by a few prompts on the basis of which they can be subjected to replication. Such voices can also be heard in the world of music, but only in new music do they sound in unison.

Its composers assume that they are so skillful at evading patterns and aesthetic similarities that their work cannot be standardized, calculated, or replicated. They speak a language that processors are unable to decode, unique to each individual composition, yet stylistically consistent enough that it cannot be confused with the “language” of any other music. They are not afraid of artificial intelligence; in fact, they see it as an opportunity to overcome further limitations.

⸜ JAROSŁAW KOWAL, Odporność na sztuczną inteligencję, Dwutygodnik, August 2025, no. 419, → www.DWUTYGODNIK.com, accessed: 23.09.2025.

The author ends the article on a positive note, saying, “Stravinsky wouldn't be afraid either,” pointing to creativity as the ultimate criterion of art. But at the same time, we mustn't forget that the vast majority of works are, in a sense, derivative and imitative. And these are probably the first to be cannibalized by algorithms and their masters. Maybe that's a good thing...

⸜ The entire band – note the complete retouching of the photo • press material

From our point of view, that is, the audience's, the matter also seems manageable. Here and now, we have such incredibly rich archives at our disposal that anyone can listen to music created by people throughout their entire lives and never reach the end of this virtual playlist. Having said that, let's add that for the well-being of us, the users, albums created by AI should be labeled as such, wouldn’t you agree? I guess you would. Although the matter is not that simple.

When we think about the new remixes of The Beatles and John Lennon albums, we remember that they were remixed using... artificial intelligence. As I wrote in the introduction to the August issue of HF, entitled On Perfection. How AI does us “good”, in the contemporary ecosystem of recording, mixing, and—above all—mastering studios, algorithm-assisted programs are already a reality, slowly accepted by a growing group of sound engineers.

Should these albums also be labeled as “AI-generated”? If not, then perhaps a scale should be introduced to determine the percentage of reality that is generated and the percentage that is created, although this distinction will likely become less clear over time. But even then, we will want to know, not always, but often enough, whether the art we are experiencing is an original artifact, or perhaps a forgery. And even if the forgery is signaled by its creators, will that change our approach? It seems that the case of The Velvet Sundown is only the beginning, and with it we are entering the “uncanny valley.”

All the more so because big tech companies, including Spotify, albeit on a much smaller scale – over 50% of music distribution in Europe is generated by this Swedish streaming service – are striving to completely replace people with algorithms over time. Why? The answer to this question is clear. As Sylwia Czubkowska notes in her book Bóg techy (The Tech God), published this year, they have accumulated incredible, even “divine” power. They are focused on themselves (the owner) and are therefore narcissistic and megalomaniacal.

But above all, they exist to make money. And thanks to their hegemony, they are doing this more and more efficiently, and it is becoming increasingly difficult for us to fight it. As she writes: “Everything indicates that the deepening platformization will lead to their (big tech – ed.) domination in the public, economic, and political spheres.” We can and must try to fight back, as the author of the book encourages us to do. However, it is not clear how to do this. We can try to leave Spotify, whether as artists or users. However, this victory turns out to be a Pyrrhic one.

⸜ The same shot as above, but with different tasks assigned to the musicians • press materials

Do you remember Neil Young's action in 2022? You probably do, but just to be sure, let me remind you that the musician removed his recordings from Spotify because of the platform's exclusive distribution of Joe Rogan's podcast, which spread misinformation about COVID-19 and false information about vaccines. Young gave Spotify an ultimatum, stating that they could not have both his music and Rogan's podcast. After a two-year boycott, Young announced in March 2024 that his music would return to the service, arguing that other platforms, such as Apple and Amazon, also hosted similar misinformation podcasts. The fact that other artists did not react in the same way was probably also significant.

Similar actions took place after the disclosure of fraud in the form of The Velvet Sundown. Later, there were calls for a boycott after it was revealed that the service's owner, Daniel Ek, had invested his money in technology companies involved in the production of combat drones and military technology, such as the Prima Materia fund, which owns shares in Helsing, a company that manufactures combat drones. Artists, including Godspeed You! Black Emperor and Xiu Xiu, removed their music from the platform, highlighting the contradiction between Spotify's ethics and its CEO's investments, as well as low streaming rates.

⸜ The Velvet Sun band in the studio, with the guitarist on the right as the lead singer – none of these words are true... • press materials

And? – Nothing has changed. Like Young, they are depend on Spotify, which, like any good big tech company, occupies a strategic position and takes full advantage of its monopoly. This means that it is impossible to fight it in the traditional way by abandoning it. First and foremost, as the author of Bóg techów says, we must write, appeal, and press – in other words, take action. When we fall silent, no one will hear us. But that's only a small part of it. We need to focus more on what is worthwhile, namely musicians and the music they create. And that means – ironically – that it is worth “converting” to the faith of our fathers and invest more heavily in physical media.

Will that be enough? I doubt it, but every revolution starts with a first step. We should take action so that we don’t have to worry that next year Ozzy Osbourne personally announces his new song, just recorded with the rest of the original Black Sabbath lineup, and wait with flushed cheeks for the package containing the test press of the new Velvet Sundown album we bought, entitled And what will You do to Us? , complete with the musicians' autographs... » HF

WOJCIECH PACUŁA
Chief Editor

About Us

We cooperate

Patrons

Our reviewers regularly contribute to  “Enjoy the Music.com”, “Positive-Feedback.com”“HiFiStatement.net”  and “Hi-Fi Choice & Home Cinema. Edycja Polska” .

"High Fidelity" is a monthly magazine dedicated to high quality sound. It has been published since May 1st, 2004. Up until October 2008, the magazine was called "High Fidelity OnLine", but since November 2008 it has been registered under the new title.

"High Fidelity" is an online magazine, i.e. it is only published on the web. For the last few years it has been published both in Polish and in English. Thanks to our English section, the magazine has now a worldwide reach - statistics show that we have readers from almost every country in the world.

Once a year, we prepare a printed edition of one of reviews published online. This unique, limited collector's edition is given to the visitors of the Audio Show in Warsaw, Poland, held in November of each year.

For years, "High Fidelity" has been cooperating with other audio magazines, including “Enjoy the Music.com” and “Positive-Feedback.com” in the U.S. and “HiFiStatement.net”  in Germany. Our reviews have also been published by “6moons.com”.

You can contact any of our contributors by clicking his email address on our CONTACT  page.

AIAP
linia hifistatement linia positive-feedback


Audio Video show


linia
Vinyl Club AC Records