ChatGPT Explorations

[thoohil]

Today I played around with ChatGPT in regard to playlist creation and general music info. The results are quite impressive. I ended up going pro account. ChatGPT is so much better than classic web search.
I want to showcase a few examples.

1. Playlists and data sources

At first I wanted to know if the AI can access my last.fm profile. At first it worked by providing my profile link.
Then I created several playlists based on my last.fm data.

Can you create a playlist based on my last.fm profile?
Certainly! Please provide me with a link to your last.fm profile, and I will create a playlist based on your listening history.

Create a playlist with 50 electronic tracks. Here’s the link https://www.last.fm/user/xxxxx

Answer snippet:

Daft Punk - One More Time
Justice - D.A.N.C.E.
M83 - Midnight City
Boards of Canada - Dayvan Cowboy
Aphex Twin - Windowlicker
The Chemical Brothers - Block Rockin’ Beats
LCD Soundsystem - Someone Great
Bonobo - Cirrus
Radiohead - Idioteque
Massive Attack - Teardrop
Caribou - Can’t Do Without You
Four Tet - Angel Echoes
Moderat - Bad Kingdom
Underworld - Born Slippy .NUXX
The Prodigy - Breathe

That worked pretty well until I wanted to try again a few hours later. It seems last.fm blocked API access for ChatGPT.
Now I get the answer:

I’m sorry, but as an AI language model, I don’t have access to your personal information or account on Last.fm or any other platform. For privacy and security reasons, I cannot access or retrieve personal data. If you would like me to create a playlist for you, please let me know your preferred genre or style of music, and I can create a playlist based on that information.

What still works well is to access data sources like MusicBrainz, Discogs, last.fm in a non-personal fashion.

Create a playlist with 30 tracks from the genre “Berlin School” based on MusicBrainz data. Only consider tracks from the last 10 years.

Answer snippet:
Redshift - Bombers in the Desert
Johannes Schmoelling - Time and Tide (Part 3)
Beyond Berlin - Tempelhof
Tangram - Zodiak
Moonbooter - Cosmoclimax
Picture Palace Music - Vortex
Waveshape - Distant Fields
Johannes Schmoelling - The Stars and the Moon
Redshift - Nails
Pyramid Peak - Alpha III
Picture Palace Music - Remembrance
Gert Emmens - The Warlock Returns

Create a playlist with 10 tracks from the genre “Berlin School” based on last.fm data.

Answer snippet:

Tangerine Dream - Phaedra
Klaus Schulze - Crystal Lake
Ashra - Sunrain
Michael Hoenig - Departure from the Northern Wasteland
Popol Vuh - Aguirre I
Conrad Schnitzler - Ballet Statique
Cluster & Eno - Ho Renomo
Manuel Göttsching - E2-E4 (Excerpt)
Harmonia - Watussi

Getting more specific:
Create a playlist with 10 tracks from the year 1975 from the genre “Berlin School” based on discogs data.

Answer snippet:

Tangerine Dream - Rubycon Part One
Klaus Schulze - Wahnfried 1883
Ash Ra Tempel - Deep Distance
Cluster - Hollywood
Manuel Göttsching - Quasarsphere
Klaus Schulze - Picture Music
Ash Ra Tempel - Jenseits
Michael Hoenig - Hanging Garden Transfer
Tangerine Dream - Ricochet Part One
Ash Ra Tempel - Light and Darkness

Note that 1975 was a seminal year for Berlin School music, with several landmark albums and tracks released by key artists in the genre. These tracks showcase the diversity and innovation of the Berlin School sound during this period.

2. General Music Info

Show the most listened to IDM artists with data from last.fm.

Answer:
Aphex Twin
Boards of Canada
Autechre
Squarepusher
Plaid
μ-Ziq
Four Tet
Venetian Snares
Clark
The Future Sound of London

Show the most listened to IDM artists of the last 2 weeks with data from last.fm.

Answer:
Aphex Twin
Boards of Canada
Autechre
Plaid
Squarepusher
The Future Sound of London
µ-Ziq
Clark
Rival Consoles
Telefon Tel Aviv

What are the 10 most owned vinyls on discogs?

Answer
Pink Floyd - The Dark Side of the Moon
Michael Jackson - Thriller
Fleetwood Mac - Rumours
The Beatles - Sgt. Pepper’s Lonely Hearts Club Band
Nirvana - Nevermind
Joy Division - Unknown Pleasures
Radiohead - OK Computer
Led Zeppelin - IV
Prince and The Revolution - Purple Rain
Daft Punk - Random Access Memories

What are the 10 most owned hip-hop genre vinyls on discogs?

Answer:
Nas - Illmatic
Wu-Tang Clan - Enter the Wu-Tang (36 Chambers)
The Notorious B.I.G. - Ready to Die
Kendrick Lamar - To Pimp a Butterfly
A Tribe Called Quest - The Low End Theory
J Dilla - Donuts
Dr. Dre - The Chronic
OutKast - Aquemini
MF DOOM - MM…FOOD
Nas - It Was Written

This only scratches the surface what’s possible. The AI can be trained with all kinds of rules, exclusions, weighted preferences, bias etc.

[pylaczynski]

It would be cool if you could export it to Spotify/tidal etc, but whole next level of integration layer would be needed.

Generally chatGPT is outstanding tool that I am in love with and implemented it already in my work schemes already (for examples writing job offers for professionals, creating frame of project plans etc)

[thoohil]

@pylaczynski

It would be cool if you could export it to Spotify/tidal etc

That’s pretty easy. I copy and paste the result in a txt file and use it as source on https://www.tunemymusic.com/ then select my target service.
No formatting needed.

[pylaczynski]

Thanks, did not know about that tool!

[thoohil]

I also generated a Xelento 2 review with some training Fits perfectly for mainstream hifi websites.

The Beyerdynamic Xelento 2 is a premium in-ear monitor (IEM) that boasts exceptional sound quality and a sleek, ergonomic design. With its 11mm Tesla dynamic driver, the Xelento 2 is designed to deliver outstanding audio performance with high efficiency. The driver is housed in a handcrafted metal shell that is both lightweight and durable, and features a 24 ct real gold logo, gold-plated connectors, and silver-plated cables that add to its luxurious feel. The Xelento 2 comes with a comprehensive selection of earpieces in 10 sizes, providing a perfect fit for a wide range of users.

The sound signature of the Xelento 2 is characterized by its linearity and lack of coloration, with a smooth and detailed midrange that allows vocals and instruments to shine through. The treble is well-extended, providing a sense of airiness and sparkle, while the bass is precise and well-balanced, offering a solid foundation for the overall sound. The result is a soundstage that is both spacious and immersive, allowing the listener to experience their music in a truly lifelike way.

In terms of source compatibility, the Xelento 2 offers great versatility and can be paired with a wide range of devices, including smartphones, tablets, portable audio players, and high-end audio equipment such as headphone amplifiers and digital audio players. With its low impedance and high sensitivity, the Xelento 2 is easy to drive and provides exceptional sound quality even with low-power sources.

The build quality of the Xelento 2 is nothing short of exceptional, with a premium feel that is indicative of its high-end status. The metal shells are sturdy and well-constructed, offering a sense of durability that inspires confidence. The detachable cable is also of the highest quality, with gold-plated connectors that ensure a secure connection.

In conclusion, the Beyerdynamic Xelento 2 is a top-of-the-line IEM that offers exceptional audio performance and a premium design. The 11mm Tesla dynamic driver provides outstanding efficiency, while the handcrafted metal shells, gold-plated connectors, and silver-plated cables add to its luxurious feel. With its wide range of earpieces, linear sound signature, and excellent source compatibility, the Xelento 2 is a great investment for audiophiles and music enthusiasts alike.

[Rikudou_Goku]

did the same for the Shure SE215 and the Blon BL-03.

Trash reviewer, literally copy paste.

[pylaczynski]

Then add “rewrite with less frequently used words and add that the cable change would be beneficial” and you have typical Head Fi shilling review

Edit: paraphrasing is probably one of the best feature of that chat

[pylaczynski]

*Can you create a playlist based on my profile on last.fm (username: pylaczynski) that would last as long as travel from Poznań to Berlin by car?"

Amazing!

[Rikudou_Goku]

hahaha, ChatGPT > head-fiers.

[thoohil]

I’m still alternating between amazement and worry about the whole ChatGPT thing. This week I also used it for technical DevOps, IT and coding stuff at work and the results were surprisingly good. And the AI will only get better over time. Crazy times we live in.

What’s also possible.

• generate political speech with ChatGPT
• apply speech on voice synthesizer
• apply synthetic speech on politican x deep fake video
• create fake reality

[pylaczynski]

Yes, it is kind of scary that it is in unleashed phase and this good already (and dangerous from perspective of existence of many types of current jobs). But I would treat it as a tool that it is best to be learned as fast as possible for my own good and efficiency. Also having ability to interpret correctly when it is wrong and when it is not is something that would keep the best in their field professions at safe place.

At least for now…

[thoohil]

@pylaczynski

But I would treat it as a tool

Sure thing, and my work colleague is happy that I don’t bother him with annoying questions.

[NiteRiter]

I was finally exposed to chatgpt over the last week and have had a lot of fun exploring it.

I’m pretty new to the hobby and am working on building my ability and language to describe sound. I asked chatgpt what is important in the description of iems. I’m interested in what you all think about the response and what you’d add.

"In reviews of in-ear monitors (IEMs), reviewers typically evaluate several aspects of the sound signature, including:

1. Frequency Response: The range of frequencies that the IEM can produce and how well it reproduces those frequencies across the audible spectrum. This is usually represented by a graph called a frequency response curve, which shows how much sound is produced at each frequency.

2. Bass Response: The quantity and quality of the bass produced by the IEM.

3. Midrange Response: The quantity and quality of the midrange frequencies produced by the IEM, which are important for reproducing vocals and many musical instruments.

4. Treble Response: The quantity and quality of the high frequencies produced by the IEM.

5. Soundstage: The sense of space and depth created by the IEM, which can affect the listener’s perception of the music.

6. Imaging: The ability of the IEM to create a clear and accurate stereo image of the music, allowing the listener to distinguish between different instruments and voices.

7. Clarity: The level of detail and clarity in the sound produced by the IEM.

8. Dynamics: The range of volume and the ability of the IEM to handle sudden changes in volume, such as in a drum solo or a crescendo.

9. Timbre: The quality of the tone produced by the IEM and how accurately it reproduces the sound of different instruments and voices.

10. Separation: The ability of the IEM to separate different instruments and voices in a mix, allowing the listener to hear each one clearly."

[Rikudou_Goku]

Better than the majority of “reviewers” on head-fi.

[thoohil]

Great review template. Here I defined and reviewed an imaginery IEM the “El Duderino”. Make sure to

read the last review

Sounds too generic. Lets do it more pretentious.

can it even be more pretentious?

Now it’s perfect!

[NiteRiter]

Damn…that was brilliant! If the El Duderino is only 99€, I’ll take 5!!!

[Rikudou_Goku]

DanMachi - Heroic Desire "Eiyuu Ganbou" (Argonaut) [Familia Myth]

Damnnn

Although it cant look at links it seems.

[thoohil]

Cool find…every day new features to try out. I just corrected the AI about an outro sound. Can you check what you get as a reponse?
“Name the song used in the outro of the series Supernatural.”

[Rikudou_Goku]

@thoohil

“Name the song used in the outro of the series Supernatural.”

rip

[thoohil]

Well thanks, I’ll check tomorrow what the reponse is.