Creative commons : free to use for personal commercial work , free to modify 
Requirements for distribution :  attribution and/or reference link to this account

TweedyTrk1

12beat attempt aka Donald Faganesque stuff..

Did this with a copy of cakewalk 5 many years ago on a Windows 98 computer.

CakewalkIsCool

AI generated sound, created with MusicGen Private version on HuggingFace.co

Creative commons : free to use for personal commercial work , free to modify 
Requirements for distribution :  attribution and/or reference link to this account

Drowning In The City Sludge (AI generated)

This is AI generated beat boxing, created with Musicgen tool at Huggingface.co


Creative commons : free to use for personal commercial work , free to modify 
Requirements for distribution :  attribution and/or reference link to this account

Bizybee (AI generated)

AI generatedn Musicgen private mode on huggingface.co

Creative commons : free to use for personal commercial work , free to modify 
Requirements for distribution :  attribution and/or reference link to this account

Small Delta Jazz Crapification (AI generated)

AI generated using musicgen

Creative commons : free to use for personal commercial work , free to modify 
Requirements for distribution :  attribution and/or reference link to this account

Nice Beat  (AI generated)

I don’t fully recall what are used in the way of AI apps to create this but I think it wasn’t the Musicgen that I’ve used for the other tracks here￼

Creative commons : free to use for personal commercial work , free to modify 
Requirements for distribution :  attribution and/or reference link to this account

Chilli Fries Blues Demo (AI generated)

I used musicgen on huggingface.co using an A100 private duplicate of the space, for input I used about ten seconds of the flute part from howard jones hide and seek  “https://youtu.be/L2H_7ralvK0?si=TZyYB4ANF8xr8ZRO” .  The AI is not reconstructing a new song, its creating an orchestration using Howard Jones melody as a guide in the style set by the prompt, with complexity set by a temperature parameter.  A temperature of .1 will produce a simple melody, but 100% layers lots of instruments to create an orchestral sound, anything above 1 deviates from the guide and descends into chaos.. Re executions of the render produces different variations but it just tends to repeat the same pattern with very little evolution., Its fun to play with but its like raytracing programs, it takes ten times the resulting duration of the song to produce.

Creative commons : free to use for personal commercial work , free to modify 
Requirements for distribution :  attribution and/or reference link to this account

Opium Traders Indifference (AI Gen')

Repeating waltz example of MusicGen on huggingface.co

Creative commons : free to use for personal commercial work , free to modify 
Requirements for distribution :  attribution and/or reference link to this account

Bat Waltz Heavy (AI Generated)

WARNING DO NOT PLAY THIS ON HEADPHONES OR EAR BUDS, IT WILL HURT YOUR EARS

Play it on a stereo system.. 

This is the future, join the collective mind…

What you are hearing is the latest in AI generated sound, created from hand optimized prompts, specifying the instruments, the method and such, the engine at CassetteAI.com which I believe is just the commercialization of MusicGen from huggingface.co at a fraction of the cost ($4/mo) , I used to pay $50 a month for this and that was only ~16 hours of computer time. 

You can deny its existence all you want, resistance is futile, join the collective and play with the new synthesis, the collective expression of all drummers who ever lived condensed in a program that uses statistics to determine how to merge all their talent, you are not merely looking at a robot replacing people, you are looking at a reflection of the human race expressed by a different kind of database. 

Messing with samples in a thing of the past, this is the next new approach to synthesis as well as musical inspiration. Why research the past when the past is at the disposal of your children, to exploit and create. 

Technology is a heartbreak now but in the long run, you save time and explore things you could never do on your own.

Note: The big downside to this approach is that there are hiccups in the audio that I have to fix with an editor, I’m sure I didn’t get them all, but some of the composition sounds like hiccups, so its difficult to tell where they manifest, it is a new site and I’m sure its something that will eventually get fixed.

Join The Collective Mind   (Drums/Xylophone,  its Like Zappas Stuff)

Same as the other 20 minutes demo (its the first third of the hour long track of AI generated content), this was created with MusicGen available on HuggingFace.co,  CassetteAI.com is practically the same thing though you can’t use audio files to guide the melody. This was created practically from prompts even though I could have used a guidance track to control the melody. My “Opium Traders” track is an example of the guidance track feature, the melody was borrowed from a flute solo by howard jones in the song where he sings “hope you find it, hope you find it in me” which I think is a theological/creeationist reference. He’s a buddhist by the way and a vegetarian. 

Joke: I’m not a vegetarian cause I love animals, I’m a vegetarian cause I hate vegetables.

This is really just a bunch of 2 minutes renderings merged together, some redundantly, but I’ve since deleted them and recompressing these would deteriorate the quality as these were derived from a mp3 file. Thats why I call it a demo and not a song, cause it wasn’t editted..

20 minutes of hour demo DRUMS

80s Pop song, made by CassetteAI.com’s engine, editted in parts to remove hiccups (a bug in their software) and to engineer the sound.. Its named after a celebrity I am extremely attracted to.. Catherine Bell, sort of wishing I was married to her, but I am happy for her situation, I wouldn’t want to interfere.. 

She’s only 2 years old than me, and she is still fucking fine..

Happy For You Cathy Bell

An 80s style pop song no vocals created with an AI music generator from a prompt.. Understand that AI is trained on many songs in this case, and what you are hearing is the collection of choices made by humans in the depiction of a style, the AI is only approximation those choices using statistics to determine the probability of melody and orchestration.. 

Its seen as cheating by some, cause the AI does most of the work, or that it seems like its stealing from someone, but the reality is this content doesn’t exist and the content its stealing from is the pattern of everyones music.. Its just wighting its choices each time and each time it generates a composition its influenced by a section of its training to replicate the style of music. This is also why it puts in many stops in songs and is prone to feedback, 

I’m guessing, but I’ve read up on how neural nets work. 

This is the basics of it:

Neurons are arranged in layers, also called nodes, and the contents includes a summation of the inputs  to the neuron, plus a constant that is used to adjust the activation function that will be fed the input of that summation, the connections from the outputs of other nodes, the things that are summed together inside the node in question are applied to a weight, so that certain nodes have a certain contribution to the outcome of the node being input to.. 

So it looks like this 

gn = some_Node_n_layer_1’s output * weight_n;
Node_layer_2(g1,g2,g3,g4 .. gn) = activation_function(sum(g1,g2,g3,..gn) + asjustment_parameter)
the result is an input to another node on layer 3.. 

And the activation function is one of three functions, ReLU (rectified Linear Unit), Sigmoid (euler’s baby, used in modelling probabilities) or Tanh (hyperbolic tangent).. The activation function is set during training. Then as the neuraons are trained, a certain process is applied to make sure the outcome for any training data creates the best possible answers with the least error by adjusting weights and the parameter added to the sumation, which effectively adjusts the placement of the outcome on the x axis, so that the combined inputs overlap and produce whatever necessary outcomes.. Think of each neuron as being like a complex layering of if conditions in a typical programming language with some sort of inequality to determine the choice to be made. This is why its considered that machine learning codes itself, its really just creating these graphs of analog connections.. And all one needs to do is put in the stimulus and out comes the decisions based on the stimulus. 

The miracle of it all is its fairly automatic, and the training does the programming, but the training process is different for every kind of solution, and it takes a long time, but its not effective unless there is tons and tons of data  and plenty of computational resources to adjust the nets so that the neural net generates close estimates for any input data.. 

Think of it as a kind of data compression, but its data compression with a clever method of information retrieval which uses binary information, the stimuli, to locate the answers.. BEfore this we’d had put the data into a database and used something called SQL (structure query language) to query the database for information. But this way the process is easier than having to use a relational database, no coding necessary beyond the design of the neural net and the schema. 

Because this stuff is well understood, it makes it incredibly difficult to just hope it goes away, its kinda like knowing how to build an atomic bomb, its too well known how to do, so wishing it out of existence isnt going to happen. Besides that, it was inspired by human neurons.

This Is Fucked Up Oh Well

Cosmic Whimsy

Touch ext v1

Carry On Our Wayward Musician ext v1

Phillip Bailey  ext v2.2.2.1.1.2.1.1.2.1.2.2

UDIO Generated song that Harnesses what sounds like Glenn Frey's voice.. Its in good taste.. I think if Glenn had been alive he would have loved to have sung this song. Cause it has a positive message.

Riding on Faith  (Glenn Frey's Ghost)

Created with UDIO (AI Music Generation), Lyrics and music prompt written with Gemini ..

Leather Rebel - inspired by Varucca Salt

This is a song made with UDIO, an AI Music Generator, in the style of Cristopher Cross, using his lyrics, but the music is UDIO's composition selected and produced by Kiernan Holland..

What Do You Think About Love - (made with UDIO) inspired by Christopher Cross lyrics and style

Created this with UDIO, an AI Music Generator  trained on commercial content, so its legality is questionable, but its not my fault for having access to that. I'm exploiting it to show people the benefits and the ethical concern..

The Simple Things - UDIO Composition

The Sacred Thing  (to make "art") , not selling it, but making it..

The Sacred Thing - Created with UDIO, Lyrics authored by Kiernan Holland via Gemini

This Kix AZz - Made with UDIO

Created this song with UDIO about how bar room fights start and though they are enjoyed by some, the outcome is never a win for anyone, that it is a heartbreaker not a virtuous affair

Winning Ain't Winning  - Made with UDIO and Gemini (Inspired by Bob Seger and Eric Clapton)

Rock Cannibal.mp3

created in ohm studio using garageband softsynths

ScatCar

Same as my first track, only sped up and swing reduced. 

Robot "Reggae" Rock

Lowrider Bot Parade

image created with my own ipad app (unreleased).. 

Creative commons : free to use for personal commercial work , free to modify 
Requirements for distribution :  attribution and/or reference link to this account

WakingTheBot

InnerAura

Dance Of The Code Bugs

Halls Of Bot Temptation

Telephone Continuum

Dismembering Bots

The Bot War Begins

Present Focus: AI Driven Music Generation 
General Focus: Computer aided Music Composition and Sound Synthesis
Interests: Everything AI, (Art) Dreamstudio.ai, Wombo, DALLE3, (knowledge) ChatGPT, Any tools that are provided on huggingface.co and by OpenAI 
Philosophical disposition: Christian Humanist Liberal (doesn’t believe christians should vote, vote god)
Father (a christian) was a LANL Laser/Nuclear Physicist , said a good scientist is the one who realizes one does not have all the answers. We've known scientists under pressure at a mud life crisis to commit suicide, when you realize you aren't going to be Einstein, what do you do with the social pressure, where do you go with that

All my music here is
Copyright (c) 2024  
Licensing: 
Creative commons : free to use for personal commercial work , free to modify 
Requirements for distribution :  attribution and/or reference link to this account

Kiernan Holland

Pinned to spotlight

Popular tracks