How to Look for Plagiarism in (AI) Music - Complete Guide

hello and welcome back to the channel so I was playing around with the beta version of yudo at yo.com and after a few iterations trying to generate an indie FK song the music generator came up with a tune I really enjoyed listening to and that I thought could have some potential as a music production or music publication in its own right then I thought why not also make a music video using the AI tools for image and video generation that are currently available to us and see what I can come up with and that's exactly what I did I basically pieced together a 100% AI generated music video now I do plan on further developing this project and I'm going to publish a separate video on the processes and workflows that got me where I'm at and also the steps I intend to take going forward but in the meantime you can actually view the whole music video and listen to the song that's called dances of Daylight by skipping to the last chapter or Tim stamp in the description below uh I may also at one point publish the finished product as a standalone Standalone music video and in that case uh I will make the link to that video available in the description as well according to udo's terms of service they do Grant full rights the music that the generator produces but they are also very clear on the fact that they won't take responsibility for any copyright claims the music that's being generated May incur and to be honest I find that uh fair enough just just as traditional musicians can fall even unwillingly in the undesirable situation of writing a tune that is maybe a bit too similar to some other song that has already been published and therefore risking to be subject to copyright or intellectual property intellectual property claims AI generated music would probably be just as prone to plagiarism even though I would think that since the developers of AI generators have access to such enormous data sets of original music the generators would be trained and capable of veering away from copying someone else's work but these are just assumptions so I'm going to try to verify if the piece of music I accomplish through udio may have any chances of bumping into intellectual property issues with existing copyrighted material to be honest this whole walkthrough can be used by anyone who wants to check if their music is plagiarizing someone else's music since these strategies would apply not just to AI generated music but they can certainly be used used to check music composed with more traditional methods as well substantially through my research I found three different types of plagiarism detectors and we'll use plagiarism detector as a general umbrella term even though in reality these tools and strategies I'm going to look at all have different purposes and produce different outcomes in general though it's safe to say that these detectors specialize either in one uh detecting plagiarism and the musical composition of the song two detecting plagiar plagiarism and the lyrics of the song or three uh checking for plagiarism by exclusion with uh song recognition applications now I only found one tool that was truly able to accomplish the task described in point1 that is basically breaking down and analyzing the musical composition or rather the vocal and instrumental components of the song and then comparing it to other songs to detect any instances of possible plagiarism unfort Ely this tool is in beta testing and although it seems to have great potential and is definitely worth keeping an eye on I do believe it is still very ripe and not ready to deliver trustworthy results when it comes to effectively detecting plagiarism in music the website I'm referring to is called uh MIP or MIP at ai.com not really sure what the acronym stands for uh it seems to be a new website that just started in January Ary of this year and it seems to be based in Korea which as you'll see later on uh may be a relevant piece of information this online app basically allows users to submit songs they want to check and in return the software uh draws up reports uh on how closely related the songs their checking are to existing published works the way their software Works uh as their website puts it is AI divides thousands of songs into four bars and and Compares their substantial similarities to fine songs with high plagiarism rates then they go on to say through AI music analysis technology it analyzes the song structure Rhythm Melody Harmony of music to detect plagiarize songs among many other songs I'll quickly walk you through the process of checking a song for plagiarism with this online app so you can see firsthand what it looks like the whole operation is pretty easy I have to say you can still sign up and use the software for free at the time of recording this video and I imagine that's because it's still in beta testing all you do once you're in is decide if you want to check only vocals only instruments or both and simply upload an MP3 file or provide a link to the song you would like to check for plagiarism once the application is done analyzing and comparing your track you'll get a full report on how similar your song is to any existing track that I imagine would be included in their database of original copyrighted works or in the pool of songs they may be scraping from the internet since every song that will be fed into this machine will most likely have some degree of similarity to other existing Publications uh at the bottom of the results page they also provide a reference scale to know the likelihood the song would have of bumping into copyright claims if you want to analyze the data even further you can then click on the check button next to each song that will load a PDF file where the entire song is broken down in four bar segments also showing which segments have the highest level of similarity uh to the reference track and would therefore be potential culprits for plagiarism as you can see the results for my AI generated song dances of daylight would be pretty reassuring since the most similar song according to the application that is Pink's Blow Me One Last Kiss has a plagiarism rate of around 22% that's well in the safe Zone according to their plagiarism rate scale I have to say the AI generated song I'm checking sounds nothing like pink song uh I can't really do a live comparison here because I don't want to risk a copyright strike from YouTube on this video but you'll have access to both my song and the other songs that were cited as possible copyright infringement so you can judge for yourself just as a benchmark I also checked Winter Winds a mford and Sun song uh that is sort of a the same Indie Fulk genre as dances of daylight my AI generated song and also pinks Blow Me One Last Kiss that was the song with the highest plagiarism rate uh compared to my AI generated song one thing that was a bit alarming for the Mumford and Sun song is that the application didn't actually find the the m and Sun's uh Winter Winds in its reference database that it seems to be pulling mainly from SoundCloud and YouTube uh I should point out that the Mumford and Sun song is fairly popular I would say so far it's collected around 23 million views on YouTube and over 96 million streams on Spotify so it's a bit worrisome that it hasn't detect it wasn't detected during the this check on the other hand though the application did find a bunch of K-pop songs if we look at the results for Winter Winds we can see it performed pretty well overall the song it was most similar to had a 26% plagiarism rate that is still in that uh safe Zone according to the website's plagiarism scale as a benchmark though I have to say I it's still performed worse than my AI generated song that got uh 22% pinks uh Blow Me One Last Kiss on the other hand did absolutely horribly uh apart from scoring 100% against itself that uh will obviously uh be disregarded as a result uh the song ranged from 37 up to 54% uh for the tracks it was compared to I think we can get a hint of what's happening and where the website may be falling short uh right down here in the small print um where uh just below the the plagiarism scale it says differences may exist depending on the genre to me uh and this is just an assumption that means that results will vary depending on the genre of the song that is being checked because the majority of the music they use to test songs up against is limited to a group of particular genres probably mostly hip-hop EDM and definitely lots of uh Pop um this uh website uh works really well technically as a web application and the idea is topnotch but I wouldn't trust the results is producing at least not yet uh once again it seems pretty ripe in my opinion and there are some warning lights uh that can't be ignored like not being able to pick up a pretty popular song with millions of views and streams like mum and Sun's Winter Winds all leads me to think that unfortunately it's not all that trustworthy nonetheless I do have to reiterate the fact that it is a brand new website that is still in beta testing uh I definitely think it's worth keeping an eye on this tool though uh for any further improvements now the second kind of detectors I found for lyrics were all text based and although not all of them are specifically designed for this purpose uh they can be used to check the lyrics of songs for signs of plagiarism I tried quite a few of these but the two that stood out are grammarly's plagiarism Checker at grammarly.com plagarism D Checker and qex Checker cex.com they're both free and neither of the two require you to sign up grammarly will also perform a test to find writing issues isues while it's checking for plagiarism and uh to its Advantage the interface uh is very stripped down and uh simple to use qex is also quite simple to use uh you just paste in the lyrics you want to check and uh it'll not only give you a plagiarism check score risk score um it will also provide a link to the documents you may be plagiarizing in my case it's actually the Bible so I don't think I would have much to worry about uh to be honest I think all these tools are most likely using the same underlying API so uh they're probably pretty much interchangeable as you can see my song's lyrics scored quite low on all of the plagiarism tools I used meaning I don't think I'll have to worry too much about the lyrics being subject to copyright finally we can move over to the third and last kind of plagiarism detection tool I listed earlier uh these are known as song recognition or song identification applications and they would be used to determine the risk of plagiarism uh by exclusion so basically if the recognition app doesn't identify your song that would be an indication of the fact that it's not included amongst the millions of tracks contained in the apps database and you could therefore conclude that it's very unlikely that the song was copied there are a few song recognition apps that are worth mentioning the most popular one would probably uh be Shazam you can download it as an application on your smartphone but you can also add it to your Chrome browser as an extension Shazam is also integrated in iOS devices and can be accessed through Siri or um their control centers the caveat with Shazam is that based on the audio fingerprint method it uh uses to detect music matches are only really possible if the song recordings that are being compared are exactly the same so there's uh really not much wiggle room for nuance one big contender in this space that we should definitely keep an eye on is Google uh they seem to be investing significant resources in developing their own AI music recognizer currently you can access their music recognition features on Android devices through Google assistant or the Google Search widget Google developers seem to be taking a slightly different approach compared to Shazam uh allowing for more versatility and recognizing musical compositions and Melodies to the point where you can simply hum a tune for the app to detect and it will return a list of possible matches also showing a percentage of how probable the match may be for each one of the results in our case it's understandable that this kind of flexibility would help increase the likelihood of finding possible cases of plagiarism I also found two web-based applications the first one is quite easy to use it's called aha music that's aha music at aha - music.com and really all you have to do is either play some music for the app to detect or you can even upload an audio file and it will analyze it and try to find possible matches the second web app uh seems to be a bit less accessible than the first one but it uh also seems to be a bit more advanced um it seems to be used uh as an API for app developers although it can also be used as a standalone web- based uh song recognizer the website is acrcloud atacr cloud.com uh as you can see they boast a database of over 100 million tracks uh accessing the app uh isn't that straightforward though you would have to click on the start testing Now button to sign up for a free 14-day trial you won't be asked for any credit card details but you will be asked to provide a few personal details including the country where you reside and your phone number then you be directed to this all products page where you can select the file scanning option bottom right as you can see it says scan and generate recognition reports for uploaded files then you'll be directed to the containers page here we'll have to create a container clicking on the blue create button top left then in the dialogue box that pops up you can give the container a name choose the audio engine I'm going to select audio fingerprinting and cover song identification then we can choose the audio Source I'm going to choose line in audio that is audio of original file or stream without noise but uh you can test out the recorded audio option as well by by default there's only the ACR Cloud bucket that I believe should contain the ACR Cloud's database of original song so we'll choose that one uh for the scanning policy I would choose Point scanning since we'll be submitting files containing a single song per file then I would leave the call back URL blank and select enable music speech recognition and then hit the confirm button once the container is created you can click on it and from this page we can either upload an audio file or a fingerprint and I'll touch on that in a second or you can link to the music online once you submit your request you can check the status that will most likely be uh processing uh to update the status hit the refresh button at the right of this uh top filter bar here once the status turns to ready you can select the tract and download uh the report by clicking on the actions tab and then selecting export from the dropdown the report is basically an Excel spreadsheet containing all the details regarding the song that's been recognized including the title uh and authors of the song and uh links to main streaming platforms like dieser Spotify and and YouTube If the status is no result that means there were no matches for your search and uh and the report you you'll get will obviously be blank pricing on this app isn't at all steep uh they use a pay as you go model with no minimum fees and the cost is basically a fraction of a cent or a penny for every valid result you get to further screen your music tracks using music recognition you may want to go as far as splitting your compositions in different music stems at least uh one for the vocals and one for the instruments and submitting them uh to the identification tool separately or you may also want to try to chop the song up in different parts or segments that could for example be uh the verse the bridge pre chorus chorus and test each segment separately to see if uh similar songs can be detected compared to the particular section of uh that track that you're submitting music recognition uh does come with some limitation those in particular in the context of using them to check for plagiarism the song that is being detected uh by the recognition app would normally have to have the same so-called audio fingerprint that's like a unique digital rendering of the track as one of the many reference songs contained in the apps database the way these apps determine their audio fingerprints and the way they match them to their POS possible counterparts uh varies from one app to another currently Shazam seems to be less flexible in determining analogies in musical composition and vocal Melodies compared to Google's most recent uh music recognizer on the other hand Google seems to be having issues extending their database of songs because of the complexity involved in their machine learning processes although each one of these different approaches reveals each apps strengths and weaknesses I'm sure that at the pace at which artificial intelligence and Computing capabilities are evolving right now uh we will surely be able to count on ever more accurate music recognition Technologies in the future one last method to detect music plagiarism I'd like to mention is also tied to the music recognition technology we just touched on and it would be to upload the music track to Media streaming platforms like YouTube or Spotify which would normally check for copyright infringement during the uploading and Publishing processes uh and the advantage of using these platforms would be the sheer volume of published music contained in their databases that is constantly being updated as user upload users upload and publish their own work so uh is this AI song I generated with udio 100% exempt from being hit by any kind of copyright or intellectual property claim uh to me the short answer is I don't think so um although I didn't really get any red flags while I was uh vetting the song for plagiarism there's definitely always a risk of creating something that is very similar to something else that already exists and as I mentioned earlier the risk of plagiarizing existing music doesn't only apply to AI generated music but it can also pertain to uh traditional music and artwork uh as well for the moment though I'm confident I can continue to work on this project without having to worry too much about uh any kind of copyright infringement if you're curious to know what the journey the outcome of this project may be uh don't forget to stay tuned And subscribe to the channel and while you're at it if you've appreciated the content so far uh why not like the video and share it with someone who may also be interested so that's all for this video uh thank you so much for watching and I'll see you in the next one take a stroll down by the Stream underneath the Willow's dream and the banjo plays a tune so sweetly oh laughter Echoes Fields of Gold stories shared and songs of old everybody's singing with their hearts so free clap your hands stomp your feet to the rhythm of heartbeats echo through the trees in this melody of living we find our way with the sun high in the sky to the close of day we come along join the song Feel The Joy it brings as we're dancing in the light that the morning s [Music] life's at canvas paint your scene in the Hues of love and Evergreen feel the pulse of the earth beneath our feet through the hush Twilight's glow we still find Our Way NE the moon soft silver light till the break of day will gather around share the sound of the night sweet [Music] refrain as the stars begin to shine music's ringing true Serena all the hearts beneath the sky so blue spin around lose your ground in the freedom felt in the twilight's magic Haze all our cares will melt and we'll find the joy in every single day from the dawn's first light to where the fireflies play [Music]

Channel: EY-EYE

Views: 4,762

Keywords: udio, udio ai music, best ai music generator, ai music generator, ai music, ai udio, ai music generator free, free ai music generator, ai music tools, free ai song generator, ai music maker, ai music platform, ai music generation, udio ai music generator, plagiarism checker, check plagiarism, online plagiarism checker, how to avoid plagiarism in music, music plagiarism, intellectual property, music copyright infringement, udio beta, udio songs, udio copyright, udio music

Id: cp0CdqAG6Og

Length: 23min 9sec (1389 seconds)

Published: Fri May 03 2024