00:00I had already been very impressed by Suno which allows you to generate music,
00:04but here, Eleven Labs Music completely blew them away.
00:07As a reminder, Suno is the artificial intelligence with this interface,
00:10where you can put a prompt, say for example that you want instrumental music,
00:14choose some settings and sliders a little to be more faithful to your prompt,
00:19to increase, if you like, a little creativity,
00:22and with one click, you will be able to generate two pieces of music each time which are pretty good.
00:27And it must be said, Suno has an incredible advantage, which is that it costs very little.
00:32With a basic Suno account, you will potentially generate 200 or even 300 songs,
00:36and this, for about ten dollars at the moment.
00:39And if you listen to the music, it's pretty good quality, and you can go from one style to another.
00:43For example, I use it to make music to illustrate my podcasts,
00:47or advertisements, or also to illustrate my anime.
00:57When you listen to it like that, it's already very good, but Eleven Labs sounds even more realistic.
01:08And Suno has another problem which is very important,
01:11is that very strongly and very potentially, she used works of artists to train,
01:17without their permission.
01:19Which in itself is quite catastrophic, because with AI there are huge problems in terms of ethics and copying artists,
01:26and you yourself, by using Suno, may end up with part of a work that has been copied.
01:32We don't create AI to copy things, we create AI to have new doors of creativity,
01:38new ideas and more performance.
01:40This is why Eleven Labs worked differently.
01:43They did something that I find extremely important,
01:45It's that they only took music where it's allowed to copy it and train on it with an AI,
01:51and above all, they have set up a remuneration program for these artists,
01:56who will therefore receive money from Eleven Labs,
01:59because their AI was trained on their music,
02:03and that these authors have given their consent.
02:05Basically, it works with production and music catalogs,
02:08and they used Cobalt and Merlin.
02:11While for the moment, Suno, but also Udio, are being sued in the UK and the United States,
02:18by what is called the RIAA, the Recording Industry Association of America,
02:23who accuse them of using works without their authorship to train their AI.
02:28It's super important that when using creative tools,
02:31that the authors give their consent and that they are also paid for it.
02:35And likewise, in its technology, Eleven Labs is committed to creating something original every time.
02:40So here we go, we're going to test all of this!
02:42As usual, you will find a direct link in the description and in the comments.
02:46This link is affiliated, meaning that if you subscribe, I will receive a percentage,
02:51and Eleven Labs did not sponsor this video in any way.
02:54I do it because I really find the result incredible,
02:57and I find it to be an incredible tool for voice, for audio and for music now.
03:02We arrive directly at the Eleven Labs interface,
03:05and on the left, we will focus on the music tab which we find here.
03:09You see here, I've already done some tests.
03:12And we're going to have fun with this part which, as usual, consists of putting a prompt
03:17and let artificial intelligence generate the result.
03:21One important thing is that on Seno, I told you,
03:23you can generate between 200 and 300 songs for 10 dollars.
03:26On Eleven Labs Music, the quality is much higher, the creativity is better,
03:30We pay the artists, but it will be expensive.
03:33That is to say, just one song, you can count it between 1 and 2 dollars.
03:38But in itself, in reality, it's nothing compared to what you would have to pay or put in as effort to generate music.
03:44So here, we will take, for example, this example here.
03:47So I gave him this prompt to create music with a bit of atmospheric emotional tension.
03:54The goal is, of course, to illustrate my anime.
03:57We end up with this interface, where we see the styles that have been included, which I can also exclude.
04:03He would then find himself in this category.
04:05And then we see that the music has been cut into several parts.
04:09The introduction, the part with the piano, the part with the strings.
04:13I find a player at the bottom, and I just have to press Play to see the work.
04:35There, we are really on something emotional and atmospheric.
04:45And from an audio point of view, it's ultra high quality.
04:47This will allow me to illustrate scenes in my anime very well.
04:51And also, there are details that are important.
04:54For example, the piano sound does not saturate.
04:56Which, conversely, happens on Sono.
04:58Now, if I want to modify these parts here a little bit, I can click on the plus to add a new section, emit a title, and the AI will come and generate the complementary music at this point.
05:09Otherwise, I can also edit the section styles.
05:13I can click here, and I can say, okay, I don't want to hear pads, or I don't want to hear kind of ghost echoes, or wind sounds.
05:23So I can delete them, or I can add more.
05:26It really allows me to move in this direction where, ultimately, all AI, all human-machine interfaces will go through conversation, through prompts.
05:34That's why it's important to learn how to prompt, and to have good diction, to have a good way of speaking to AIs.
05:39Because it allows you to maximize your results every time.
05:42So there, we saw a creation which is rather atmospheric, illustrative.
05:46But what does it sound like in another language with singing?
05:49So here, we're going to make some animated, combat-style music, with hearts behind it.
05:54And you can see that I put, right here, some Japanese.
05:58It's very clearly anime music, fighting music, and that's what I was looking for.
06:19And what's also interesting about the mix is that it's relatively very clean.
06:23Very often, on Sono, you will have sounds that will sound a little synthetic.
06:27It's going pretty well here, and I can't wait to see what happens next, and to see in a year what we can produce with Eleven Labs Music.
06:34I did encounter some flaws, though, for example, if you ask him to rap in French,
06:40You will see that the pronunciation is not yet very good, and I think they will improve that in a future version.
06:46We're going to listen to this example, and you're going to see it right away.
06:49So, it's very experimental, I'm going to ask to mix rap and dubstep.
06:53There are some misinterpretations of the sentence, but the flow is there, and the style is there.
06:59So as you heard, we have a real rap and dubstep mix.
07:19The sound is good, the voice is good, a bit robotic, but there are some sentences where it is pronounced a little badly.
07:25But anyway, Eleven Labs remains the leader in voice simulation, so I think they'll rectify that, but very quickly.
07:33What might be interesting now is to create music, perhaps pop-jazz, with a real English voice over it,
07:40and compare Suno and Eleven Labs Music.
07:43I asked ChatGPT to generate all the lyrics and instrumentals directly for me.
07:48So the style, I put it here, so it's a pop-jazz music on the theme of autumn,
07:54and I copy and paste the lyrics directly here.
07:57I click on create.
07:58In the meantime, I'm going to do the same thing on Eleven Labs Music.
08:01So I go to the prompt section, I can choose the number of variations,
08:05so the number of different music that it will generate for me, and of course, the duration automatically.
08:10Otherwise, I can also say 3 minutes, 4 minutes.
08:12We see the number of credits I have left, and I just have to click on generate.
08:16I throw in the prompt and lyrics, and click generate.
08:19So on Eleven Labs, we can see here that I have the verse that has been applied correctly,
08:24the lyrics, and surely the style,
08:26and we see that the different parts are being generated one after the other.
08:30I suggest you listen to the result on Suno.
08:34We listen to the first version.
08:35Okay, you've got this part down pat,
09:05and now we listen to the result of Eleven Labs.
09:19Suno's style is much more square, also much more expected,
09:25less creative, while this one is much more creative.
09:28It can even surprise you when you hear it for the first time.
09:30In terms of sound quality, Suno, we feel that it's a little more electronic in terms of the instrumental.
09:35Whereas on Eleven Labs, we really have a nice feeling at the level of the instruments.
09:43But hey, the result here is not particularly obvious.
09:47And you'll see it, it's when you have instrumentals that are loaded,
09:50that all the difference will be made.
09:52We're going to go into a more metal register,
09:54where you will have the sound of the guitars which will be quite obvious.
10:09If you listen carefully, it will remind you a little bit of Guitar Pro.
10:13If you're a guitarist, it really sounds like simulated instrument sound.
10:17Everything is mixed, the sound is almost muffled.
10:20Whereas, conversely, you will see that the sound of the instruments is extremely distinct on Eleven Labs.
10:42The difference is extremely impressive.
10:44If you play this for me without telling me it's AI,
10:47I could very well be wrong and tell you that this came out of a studio.
10:50While the music before generated by Suno,
10:52which is hyper generic, hyper flat and muted,
10:55where the instruments mix and merge,
10:58well here you really feel like you hear a guitar on the left,
11:01another guitar on the right,
11:02clearly the bass and the saturated vocals.
11:05So of course, if you are not a big fan of metal,
11:08Maybe it'll sound a bit like porridge.
11:10But I guarantee you that the difference you get in terms of sound cleanliness is extremely impressive.
11:16I also asked him to create some emotional music for a competition I'm running on AI.
11:21And I said to myself, hey, we're going to create music with AI too.
11:24and we will see if he manages to convey a feeling of nostalgia, sadness and emotion.
11:29And there, you will see it again, the instrumental is incredible.
11:49You have made great progress.
11:51And above all, in the sermon that I gave him,
11:53I really asked him to have progressions,
11:55to give a kind of feeling and emotion to such and such a timing,
11:59and he did it perfectly for me.
12:01So for me, Eleven Labs Music is ultra impressive.
12:04because we can hardly distinguish
12:06normal artistic creation and that of AI.
12:11Of course, there will still be small artifacts,
12:13there will still be some small sound problems,
12:16and there are improvements.
12:17But here, we really have a much higher quality result than Sono
12:20and it can be super interesting.
12:22If you want to distribute your music,
12:24I really recommend this service.
12:26I put the link in the description.
12:27You will now be able to publish your creations,
12:30your productions directly on around a hundred sites
12:33like Apple Music, Deezer, Spotify,
12:36and also start making money from your creations.
12:39And of course, Eleven Labs will then take care of
12:42to pay part of your subscription to the authors
12:45where the AI was trained.
12:47Finally, there is of course the prompt,
12:49and as you can see here,
12:50I just ask him to create an English prompt of the lyrics
12:52for Sono 4.5 because it was our base,
12:55but you can ask for Eleven Labs Music.
12:58And I ask him for pop music,
12:59so I wrote it wrong, jazz, on the theme of autumn,
13:02and directly, I had this prompt that I used for my examples
13:05and the lyrics that I copied and pasted.
13:07There's no need to rack your brains any further,
13:09the important thing is to converse with AI
13:10to tell him what you want,
13:12and it will adapt it automatically.
13:14I hope you enjoyed this video anyway,
13:16and you see new perspectives
13:18thanks to Eleven Labs Music,
13:19which I find to be of high quality,
13:21and it is also him that I will use
13:23for my next projects.
13:24In the meantime, please feel free to share with us
13:26your creations also in the comments.
13:28In the meantime, take care of yourself and your projects.
13:30Hello!
Commentaires