Jin Lei from Concave Fei Temple
qubit | Official account QbitAI

The first in the worldAI streaming music app, that's how clever it is.

What are the features?

First of all, as the name suggests, the music in this app is all generated by AI.

Second, it's infinitely streamable, which means you can do it all the time you want to be generated and listened to.

Without further ado, let's get straight to the effect.

For example, our Prompt is:

Take a chilled walk with my dog.
A pleasant walk with my dog.

In the video above, we started by generating a 10-minute piece of music.

In this 10-minute period, around the feeling of "chill", the AI generated a total of four English songs with different human voices.

The atmosphere remains the same, there is no sense of disobedience in the connection, and there is no need to stop and change songs when walking the dog.

After reading the effect and the interface of the App, do you want to ask which foreign company is this AIGC product?

No, no, no.

This App is the latest overseas product created by the domestic company Kunlun Wanwei - Melodio, which also supports the web version.

So as the world's first AI streaming music platform, what other ways does Melodio play, let's continue to test it.

It's the kind that supports Chinese

Melodio is currently available in major app stores, and like any other music app, you can simply download it and start using it.

It's also extremely simple to use, and after opening and logging in (or registering), the interface looks like this:

The easiest way to do this is to type in the Prompt of the music you want to listen to and click "Let's vibe".

For example, if we enter this time:

Dynamic music, DJ, fitness, passion.

Overall, the music is in line with the dynamic rhythm that is needed when working out.

In addition to generating directly via Prompts like this, Melodio also provides a "Song Reference" mode:

Then we made it more difficult on Prompt this time, and let this overseas product generate a Chinese song:

Sad, Chinese, rainy day.

Not only does the atmosphere of the music meet the requirements of Prompt, but the AI singer's biting of Chinese is also very clear, and it is no problem not to read the lyrics.

Of course, if you want to check the lyrics, this is also available, just "right screen" (swipe right on the home screen):

If you want to save the generated music, click on the little heart at the bottom right of the home screen, and nearly 2 minutes of audio can be in your pocket~

In addition, if you don't have creative inspiration, or want to listen to what kind of music everyone has generated, Melodio's "Left Screen" also provides the function of a music square.

Here, you can listen to high-quality AI music created by netizens, including multiple styles such as light and passionate.

However, there is one thing to say, Melodio is only one of the products launched by Kunlun Wanwei in AI music this time.

There's also an AI music creation platform that beats Suno

That's right, Kunlun Wanwei has also launched an AI music creation platform - Mureka.

If Melodio is more suitable for listening to AI music, then Mureka takes AI music creation to a new level.

The barrier to entry is also very low, even if you've never created music before, it's perfectly okay to hold: just two steps.

The first step is to enter the lyrics of the song title.

In the corresponding input box on the left, we can first enter the pre-conceived song title and lyrics:

△Note: The song title and lyrics are generated by AI

The second step is to select the reference music.

In this process, we can choose the music from the platform's built-in library, or we can upload local music by ourselves:

Since we want to generate ancient Chinese music, we choose the theme song "Hongyan Tribulation" of "The Legend of Zhen Huan" this time.

Once uploaded, we can also customize the style, mood, and reference clip (30 seconds) of the music:

Finally, click "Create" and wait for about 1 minute for Mureka to provide 2 versions of the generated music:

Version 1 looks like this:

Version 2 looks like this:

The music generated by Mureka is not "disposable", but the kind that can be secondary AIGC for a certain clip:

And the music generated by Mureka is about 2 minutes long by default, if you want to continue to let the AI generate, you just need to fill in more lyrics:

The most important thing is that the music generated with Mureka is sold, and it is the kind with proof of creation!

However, one thing to say, AI music creation platforms like Mureka are not the first, after all, the previously popular Suno and Udio have already been launched.

So which is stronger or weaker than the two?

For this problem, there have been bloggers abroad who have made a nearly 20-minute experience and tutorial video, and after that, he got the summary:

Mureka, even better than Suno and Udio.

Then the next question is:

How?

For this problem, Kunlun Wanwei also gave their solution-

Sora-like architecture in the music field.

This architecture can be divided into two main parts.

The first is the Large-scale Transformer, which is responsible for composing music, learning the contextual dependencies of Music Patches, and writing and controlling songs.

The other Diffusion Transformer is responsible for the singing part, which allows Music Patches to be restored to high-quality audio through DiT, which can support the generation of 6-minute 44100Hz stereo songs with a sample rate of 44100Hz.

Therefore, Kunlun Wanwei also compared the latest achievements of the model under this architecture:

It is understood that Mureka's new model maintains a leading level in vocals, and has reached the industry's SOTA level in terms of sound quality, naturalness and pronunciation intelligibility.

At the same time, there has been a significant improvement in other more granular musicality indicators (motive quality, accompaniment texture, orchestration richness, musical structure, and mixing quality).

It's not hard to see why Melodio and Mureka are products that stand up to the test of major bloggers.

But at the end of the day, there's another question worth mentioning.

That is why Kunlun Wanwei, as a Chinese company, will pick up the track of going overseas in the field of AI music.

In fact, if you look at Kunlun Wanwei's "resume", going to sea has already become part of the company's DNA: it has 16 years of experience.

Kunlun's current business covers AGI and AIGC, information distribution, metaverse, social entertainment and games

Its markets cover China, Southeast Asia, Africa, the Middle East, North America, South America, Europe and other places. Up to now, the average monthly active users in the world are nearly 400 million, and overseas revenue accounts for 86%.

Entering the era of large models, Kunlun Wanwei is also the first group of players to comprehensively deploy here.

Its Skywork has included six AI business matrices, including AI large models, AI search, AI music, AI video, AI social networking, AI games, etc.

Its product "Tiangong" APP has also fully integrated core functions such as AI search, AI reading, and AI creation, covering multiple application scenarios such as work, study and life.

This is also the reason why Kunlun Wanwei can enter the overseas market with the most SOTA posture at the time of the rapid rise of the AI music field.

But all in all, with the official release of Melodio and Mureka, it can be said that the friends who love music have hit the threshold of creation hard.

Not only that, but even for professional music professionals, these AIGC tools provide a certain amount of motivation for their creative inspiration.

This article is from Xinzhi self-media and does not represent the views and positions of Business Xinzhi.If there is any suspicion of infringement, please contact the administrator of the Business News Platform.Contact: system@shangyexinzhi.com