"Ai" Offline Radio DJ for Poweramp (feature Suggestion)

[Peyton Taylor]

An AI based radio DJ in Poweramp, like Spotify? Yes. I think it would be one of the biggest features to add to Poweramp in the form of a "feature package" I feel sometimes we miss listening to the radio because of the aesthetic of the DJ announcing or hyping up the next song. But no one misses the low quality of radio or wants to pay for the subscription of SiriusXM. The radio DJ Ai doesn't have to be GPT based, although it would be nice but there are two downfalls that come with this. Required Internet connection and the monthly subscription for the API can be expensive.

I'm thinking a large amount (for variety) of preloaded or script generated one liners in text form spoken by a high quality TTS (Text-to- speech) or multiple TTS voices to choose from. Example: "Here's Halsey (Artist) with one of their biggest hits Bad At Love (Song)." 

There could be more factors added in like (Genre), (Rating), (year), or (Album). Here's an example: previous song genre: (Rap)  "let's switch it up with a little (Rock) (Genre) here's Nightmare (Song) by Avenged Sevenfold (Artist)" or Let's say the next song is from 1997 "Let's rewind and take it back to the 1990s (Year) here's Mountain (Artist) with one of their hits Mississippi Queen (Song)." Here's an example with (rating) involved. " Here's one of your favorites queued up just for you (Rating) Let's here it for $uicideboy$ (Artist) here's Escape from Babylon (Song)

This info can be taken from the meta data or tags from a song and preloaded before the next song starts. The one liners can be shuffled and based around the songs tags and meta data. I think this option would be more suitable for shuffling songs but I think it would be good to add an option for shuffled/unshuffled playlists. Notice how the artist in one of the examples has "$" in it? give the user the option to change pronunciations of words like $ucideboy$ pronounced as Suicideboys or Roman numerals pronounced as numbers or abbreviations pronounced as full words by default ect. 

Here are some extra features that could be added: 

Custom DJ Settings Page: Provide users with options to select from a few different TTS voices (e.g., male, female, energetic, calm). Users can also adjust pitch, speed, and customize how often the DJ speaks (e.g., after every song, every few songs, or only during specific playlists).
 emphasis to make the voice more personalized and suitable for the intended DJ vibe.

Pronunciation Customization: Include a user-friendly feature within the custom DJ Settings Page that allows users to customize the pronunciation of specific words, artist names, or abbreviations. This can be a simple interface where users type in the phonetic spelling or select from common pronunciations for certain words.

Simplified Randomization Algorithm: Recommend using a basic randomization algorithm to shuffle and select DJ one-liners, so that commentary feels varied. This can be done with a lightweight script that randomly picks from a set of templates based on the song’s metadata. 

Event-Based Triggers: Propose simple event-based triggers that activate specific DJ phrases. For example, when a new playlist starts, when a user reaches a milestone (like listening to 100 songs), or when it’s a particular time of day (“Good morning! Here’s your first song of the day!”). Maybe even weather and location, an example of that would be; "it's a sunny day with a high of 86° and a low of 75° (weather) in Knoxville (location) today let's start you off with some good vibes, this is wave of you (song) by surfaces (artist) or "Brrrrrr it's cold out there (weather) in Lexington (location) stay warm with some cozy vibes from chance peña (artist) in my room (song) These are easy to implement and add a dynamic touch.

Local Text Generation: All text processing and generation can be done locally on the device. The app can generate the text using predefined templates and then send it to the TTS engine to synthesize speech in real-time.

Let me know what you think

[Phantom52347]

Oh yes

[andrewilley]

The reason I have a large collection of my own music on my own device is precisely to avoid DJs or adverts or any other interruptions to my listening...

Andre