作者|James
It is said that this year's May Day is the most difficult to grab tickets, but for this ticket, many music fans refer to concert tickets: Phoenix Legend, Deng Ziqi, Eason Chan...... The more special the voice, the more powerful the scene, and the more infectious, the more difficult it is to get a ticket......
At the same time, many musicians are not happy, suno, Tiangong, ACE Studio, Netease Tianyin...... The appearance of each one is like sticking a knife in the music producer, a large number of Douyin Divine Comedy studios have closed down directly, and Xiaoxu Music, like the first game music factory, has laid off half of its staff to deal with this "ChatGPT moment in the music field".
It is conceivable that the music industry is rapidly differentiating, and head, live, and performance have become the key words of the industry.
A senior performing arts agent told Entertainment Capital that even a fledgling "eighteenth-tier coffee" will try to seize the opportunity to show the audience face-to-face, and even tour 3-4 cities a week with high intensity;
On the other hand, composers and arrangers have to admit that the music created by AI has made great strides to catch up with the wisdom accumulated by humans over the years, and they have taken the initiative to cut jobs that can be replaced, and even those who are still in their jobs feel that the sense of meaning of work has been greatly diminished.
Although the positions are different, the musicians have reached a rare consensus: after the baptism of AI, the original music will increasingly become a supporting role, and the live music industry will transition to a kind of performing art, a kind of infectious economy between people.
How to evaluate the effect of AI composition
When Suno v3 and Tiangong were first launched, Liang Yi, the person in charge of the Elephant Wind performance, and his colleagues began to try to use it. By their standards, Liang Yi is not very happy if they think of AI as a human singer or composer.
"In my opinion, AI is really just a tool, at least for the time being, the songs he makes. It seems to me to be 'plasticized'. It has to be cultivated based on existing musical elements, so at the moment I think his style is very simple, and what he has made - I don't know if you feel it - I think he basically has no emotional change in the main song-chorus. Because he is a dummy, he has no emotions. ”
Lu Xiaoxu, a music producer who has been engaged in game music creation for 18 years, and CEO of Xiaoxu Music, disagrees with this. After experiencing Suno v3, he believes that AI has reached a level no less than human composition.
"We used V3 on the first day it came out, a month ahead of schedule. At the time, I thought this thing was — I felt like it was nuclear. It wasn't until March 20 that it was made public to all users, and his data and search volume began to skyrocket. We held an internal seminar and said what should we do? When this AI came out, we felt that it was not a replacement for the low-end, and many of the mid-to-high-end ones were completely at the level of a music company like ours that had been doing it for more than ten years. ”
AI Mapping by Entertainment Capital
Since the early emergence of AI content, platforms such as Douyin have urgently taken some measures to limit traffic, such as requiring all AI-generated content to be clearly labeled, and prohibiting live events without real people. Xiao Xu worries that these restrictions may actually deprive people of the opportunity to conduct unbiased "blind testing" of AI and live-action music. Some people will give a high evaluation before they know that the music is made by AI, and once they learn that it is made by AI, their attitude will take a big turn.
Some of the commercial customers displayed on the official website of Xiaoxu Music
Xiao Xu said that the level of Suno v2 has not yet reached the level of commercial application. However, the launch of V3 shocked him by the sudden disappearance of the gap they thought they thought would take a long time to catch up. This led him to believe that problems that seem intractable today could one day suddenly be solved overnight.
These unresolved issues include not being able to repeat the previous paragraph or chorus that you have just generated, sometimes working with lyrics that do not conform to the grammar, and not being able to modify the parts that have already been generated. However, Xiao Xu emphasized to Entertainment Capital and Vision Intelligence Future that after specifying in the prompt, Suno is able to make emotional progression.
"You ask why our industry has such a high evaluation of Suno v3? This generation of products has a strong melody and harmony, and in layman's terms, it sounds more pleasant. The previous product was a bit stiff, and it was very uncomfortable for us as professionals to listen to it. Secondly, its paragraph structure is extremely reasonable, for example, the arrangement of the main song and the chorus is very reasonable, and after you mark it, you will find that the chorus is handled in great detail, and some sentences will stand out, such as a five-word sentence, the last three words will be sung repeatedly. These details are exquisitely handled. ”
Xiao Xu said that the part that it is not good at is that there are not enough repertoire to train this style, such as Chinese style and folk songs, but this is not an insurmountable obstacle.
When it comes to AI music creation, it's not just Suno's all-encompassing game, and each AI tool has its own specific scope of application. In this regard, Xiao Xu, who has fully embraced AI, has his own set of experiences.
Some of the AIGC tools used in Xiaoxu Music's daily work
Xiao Xu pointed out that NetEase Tianyin's technology was actually born earlier, and it has not been reconstructed based on large models recently, so that it can instead produce some alternative advantages in the current AI composition tools based on large models, such as more accurate rhymes, and can be created in the style of famous lyricists.
Netease Cloud Music officially told Entertainment Capital that in terms of creative tools, Tianyin and X studio (developed in cooperation with Xiaoice) have been praised by many musicians. On the consumer side, the recommendation algorithm uses AI more, and there is also an intelligent song recommendation and explanation function "private DJ".
The "Divine Comedy", the algorithm of short videos, will be the first to be impacted
Although he does not believe that AI composition can replace all human music creation, Liang Yi believes that at least for some shoddy songs on the current platform, they will definitely lose their living space under the squeeze of AI. In their day-to-day work of selecting songs, the vast majority of songs written by humans are not very sophisticated, so it's hard to hear any obvious differences from AI composition.
"Some of them may not be easy to hear, but it depends on what kind of sample they are. Nowadays, we often receive a lot of samples from so-called 'copyright companies' every day, and there may be dozens or hundreds of songs for you a day. In fact, to put it bluntly, they are all following the trend, so this kind of demo is basically not more different from AI, or even inferior to AI. ”
She feels that AI music may have a greater impact on the songs produced in batches on platforms such as Douyin, and even does not rule out the independent introduction of AI composition mechanisms by video platforms, which will affect the living space of the creation of "Divine Comedy" by a third party.
At present, music dissemination is mainly carried out through the background music of short videos, which is not as good as traditional channels such as TV or albums, which can better show the value of music, and the meaning of music is weakened in this process. "Ordinary people definitely can't remember how many videos you watched today and how many BGMs you heard. But how many shows you've seen this week, do you remember? Of course you can remember. ”
Music producer Xiao Xu agrees. In response to this change, the company's current 40-odd employees in the department are expected to be reduced to 10 or 20 by the end of the year. He made it clear that all jobs that could be compressed would be cut, leaving only those who could not be replaced for the time being.
The service process displayed on the official website of Xiaoxu Music
For example, at present, AI composers can only generate complete tracks at once, and cannot make minor adjustments such as orchestration and melody style. Therefore, when they find a good AI song, they still need to manually "rip the score" to a MIDI format that can edit the audio tracks and instruments separately, and then trim them as necessary to eliminate logical problems with the song. This process is very different from the previous way humans freely created the main theme and arrangement, and they are now assisting the AI to complete the composition.
Xiao Xu mentioned that they will not secretly join AI to create content when they are not declared, so the work of composing and arranging music still needs to be done by hand, but some employees feel that their work is meaningless.
"Our current AI composition has not yet entered the commercial link, so it can only be said to be a reference song, and cannot be directly delivered to customers. But he felt meaningless. The core reason why we make these tunes for customers now is that we can't deceive customers, because we have collected money from others, and we have to do it with real people. But what doesn't make sense is that you really do it manually, you can't do it with AI. Therefore, people will feel that there is no sense of accomplishment. ”
In the past, the total supply of songs was small, and on most occasions several fixed albums would be played repeatedly, making it easy for people to remember certain popular songs and their singers. But on Douyin, even if the video marks the source of the background music, everyone quickly swipes by and quickly forgets. Therefore, the song is only a supporting role in the short video, and in case of copyright issues, it can be replaced with an AI-generated track at any time.
In January of this year, a copyright dispute broke out between Universal Music and TikTok
"This year, you will find that Douyin and Jianying have a lot of songs missing. There are a lot of people complaining. Because of the copyright issue between Douyin and record companies, a large number of them were taken off the shelves. ”
Talking about the dispute between Douyin and Universal Music, many musicians have pointed out that the Douyin platform has now stopped providing revenue sharing to short video BGM, which has had an adverse impact on the copyright owner. As for the music platform, although the share price has not changed, except for the head company, the general cooperative companies receive a low share ratio, Liang Yi bluntly said, "Even if 10,000 listens are not much." ”
Entertainment Capital has found that except for a very small number of top songs, other musicians who can generate income with one or two hits are not stable. Someone said, "It's March 2024, and the current daily broadcast volume is about 18,000, and the monthly income is about 600-800." ”
In the long run, AI may prompt music platforms to further adjust their payment models, which in turn will affect the potential development opportunities of new artists.
Liang Yi pointed out that although AI songwriting does not directly affect their company's business model, they have severely restricted the use of singers' voices for large-scale model training in their current licensing contracts. They themselves also don't plan to train the model on their own. She believes that a more direct way to increase awareness than to make artists become AI sources of music is to provide limited exemptions for new songs, which can more effectively maintain the connection with fans.
Offline performances have never been more important to artists
Liang Yi has also encountered situations similar to AI composition in the past, such as imitation shows, covers, adaptations, or using singers' sound sources for ghost animals or covers, such as the "AI Yanzi Sun" incident at the beginning of the year. These situations can also threaten the artist's earnings.
"No one's timbre is completely unique, that is, singers on the market, such as Faye Wong's timbre, there are many singers behind to imitate, so you can't say that the latter is infringement. The timbre is the same, and this thing can't be sued. ”
Every time there is a case like this, it actually points to one result: it is becoming more and more difficult to generate income for artists based on the distribution of songs and albums alone. On the contrary, through offline contact with real people, even if the number of viewers is not that large, it can achieve a good profit effect.
For an average artist to form a sustainable profit model, he only needs to maintain an audience of about 10,000 listeners. "Let's say a musician has 8 tours this year, each with 1,000 people, and the total number of people reaches 8,000, and the price of each ticket is 299 yuan. Artists can earn higher income by touring. ”
Some of the large-scale music festivals held during the May Day holiday
Liang Yi believes that AI technology and the audience of independent musicians do not overlap. Musicians need to find people who like their personality and are willing to spend money to support them. But don't you think it's very likely that people who use AI just don't want to spend money?
It is difficult for human singers to have the same influence as in the past, mainly because people's musical preferences are becoming more and more diverse. In the past, the vitality of popular songs could last for more than ten years, but now it has been reduced to a few months.
At the moment, music is heavily fragmented. Although those big stars who are household names can still attract a large audience, every musician with different styles can communicate with their fans through offline activities and earn a commensurate income.
When it comes to live performances, the factors that prompt or prevent a person from entering the venue for consumption are more complex than the average person thinks. Many people don't really pay attention to the names of the singers, but more randomly pass by, in order to feel the atmosphere of the scene.
AI Mapping by Entertainment Capital
Liang Yi has been helping a friend find independent musicians recently, and she was very impressed with one of the singers she met this month. "His live is very good, he can do guitar, synthesizer, percussion, and then edit, mix, and record himself, and he can do it all by himself, and his singing skills are also very good. ”
Before the performance began, Liang Yi was a little worried that the musicians would not be famous and the scene would be deserted. At this time, she found a couple on the sidelines who had come to Beijing for tourism.
"They just walked up there at random and were asking, 'Who is it today?' 'I don't know, I haven't heard his songs.'" As a result, when I sang the first song, and when it came time for the second loop chorus, the two of them jumped right next to me, and they were very happy. ”
Because of this experience, Liang Yi has greater confidence in the fact that newcomers with a good foundation can break through the encirclement. Even if the people who come live are not well-known, as long as there is an opportunity to perform live, most of the time it will be successful.
"Now the post-00s generation hopes to achieve the purpose of making friends through a live event. A lot of people who came to the show probably didn't even know who the band was, so they bought random tickets and came in. After that, I was infected by the music, and then I started to move there, which is a way of life for young people. ”
Therefore, in order to increase the experience of fans, music companies often provide benefits such as peripheral gifts and face-to-face signings when curating exhibitions, making it easier for fans to devote themselves to music consumption.
Artists from Liang Yi's company are performing. Photo/Elephant Wind official B station number
The importance of in-person experiences in music consumption hasn't changed even during the pandemic. With the lifting of prevention and control measures, the fees for live streaming or recording of online concerts that some copyright owners had tried earlier suddenly became unattractive.
If the audience is located in larger cities where the artist can tour, fans won't miss the event. "Therefore, in the process of doing it now, we will do it in the form of tours, that is, several major box office cities in the country, Beijing, Shanghai, Guangzhou, Shenzhen, Chengdu, Nanjing, Hangzhou, which are cities with a basic configuration for artists to do tours. ”
Artists from the same company can perform in one group, so that if the artists they like can get together, fans can stay in a city for a few more days. Some music fans will set up a schedule during the long holidays to make sure they can hear every show they want.
Musicians are selling a sense of presence and the ability to perform songs live, rather than the ability to create original tunes. Does this mean that the singer is "idolized"?
Liang Yi said that this statement is not accurate, because "in fact, the record industry was like this at the beginning, but now in my opinion, it has gone through a cycle, that is, it is back to square one." ”
Singers already have an idol or star side, especially at the turn of the century, it is very common for singers such as Stefanie Sun and Faye Wong to be not good at lyrics and composition, and people's main evaluation criteria for singers are image, voice and on-the-spot performance. With the passage of time, although singers with creative ability such as Jay Chou have appeared, even singers who can only sing can be famous.
"In fact, this kind of income offline will be more direct, and it can form a kind of stickiness. And the online thing, in my opinion, is all virtual, that is, the Internet is a very virtual thing. ”
Liang Yi believes that AI can quickly learn to create "algorithmic divine comedy", and some musicians will be greatly impacted if they rely on online data and do not do offline business or build their own fan base. Therefore, she called on existing musicians to improve their business capabilities, do more offline operations, and build personality charm and appeal to cope with the threat brought by AI.