How to make royalty free AI music for your content

This content is generated from the audio voiceover so it may contain errors.

(00:01) very good to all and to all and welcome to a video more than the canal of isocial web in this case bring you a methodology with which can generate free music of rights and of automatic form thanks to the artificial Intelligence this is possible thanks to that it dip the people of Facebook has released a model with which can generate music through through the guide costs No is not so beautiful as it sounds like this Simply said Although it is quite useful and quite powerful Why is not so beautiful basically good so that you
(00:34) you do an idea is a model that has been coached with 20.000 hours of music in which they have removed him the voices this involves that you can not generate music with voices costs And apart from this as they remain 20 one thousand hours fits different types of music But there are some true musics or some cultural sounds that surely can not find or that the generation was deficient in addition to this the model has been coached with descriptions in English this involves that if you use it in another language and do the descriptions in another language does not go it to
(01:04) understand and the results go to be quite bad and now will see will see an example costs And in the last limitation that it is necessary to #take is that in function of the music that put the music does not finish of a form of increasing and goes finalising but there is many times that does the musical sound and for in the moment in which it arrives to the limitation of time Costs then those are a series of conditionings that have to #take when you go to use this Script that go you to present to continuation
(01:37) well the process is very simple have four blocs simply have to him go giving to Play Play in each one of them here as you already know in the first button and this simply goes to install the model released by Facebook that audio Craft has generated it this Script so that we can it so that we can use it I simply modified some things here only does the import of the model and of torch that is a library of learning costs afterwards do the call to the model a thing so it is interesting that is here where
(02:11) it puts large there are four different configurations of the model can use the small the medium the large and Melody costs to go trying with each one of them to me the one who improves resulted has given me has been alarge that understand that it is the one who are spending him a bigger context and can give in better results and afterwards finally have here the length of the audio that want to generate Voucher then in function of the model and in function of the size that have if you do not have an account pro and other as surely you remain you with limitation of time because the
(02:45) model the axe is quite weighed do not know are 6 gigas or something like this when it downloads Voucher then well This have to #take to see the limitations of size costs as it takes a while in Downloading the model and everything simply have generated it here with a series of examples costs here go in among comillas llas simple will have to put the type of audio or was the Chrome of audio that want to that you want to use to generate the content costs here for example have put one an alone of Spanish guitar of Flemish okay the
(03:13) I go to put so that you say it and can see pretty much How sound [Music] okay as you can see the music is quite well or was the music that generates is enough is quite currada especially if they are musics very marked very instrumental and other work very well and afterwards have done you a small test so that you see the difference this would be a classical music of piano with the description in English [Music] voucher And this would be a classical music of piano of alone of piano with the description done in Spanish
(03:55) [Music] okay as you see no whole well or was dips here a species of flute Or was does not finish to interpret very what what are asking him Then with this have done a pair of descriptions more That you can try it simply when when you open the Script have done the same with a description of a base of reggae and with an electrical guitar and the same description in Spanish and will see that in the another looks a organillo hears quite bad voucher then good simply when you use it can put
(04:25) so many lines like audios that want to generate and do it always in English that is as it works correctly so without more appreciate you the have been here and anything any doubt that have see us by the nets an embrace