Everything mentioned below is my opinion and my opinion alone. Don’t agree with me or see something I mentioned differently? Leave a response and let’s talk about it!

下面提到的所有内容都是我的观点，也是我的观点。 是不同意我的观点，还是看到我提到的不同观点？ 留下回应，让我们谈谈吧！

The music industry is one of the domains that has yet to see a full fledged restructuring as a direct cause of AI. The transition towards streaming caused a significant restructure — one that happened to be in favor of the labels and streaming services as opposed to the artists who bring the consumers in the first place. AI and ML have helped streaming services with music recommendations and targeted advertisements, but these technologies have yet to change the way we listen to/create music. This will change and it will be soon.

音乐产业是尚未将成熟的重组视为A​​I的直接原因的领域之一。 向流媒体的过渡导致了重大的重组-恰好有利于标签和流媒体服务，而不是首先将消费者带到艺术家的艺术家。 AI和ML已通过音乐推荐和有针对性的广告帮助流媒体服务，但是这些技术尚未改变我们收听/创作音乐的方式。 这将改变，并且很快。

人工智能与机器学习 (Artificial Intelligence vs Machine Learning:)

I’ll spare you the Venn diagram.

我将为您保留维恩图。

There are many differences between Artificial Intelligence and Machine Learning. There are also many similarities between the two as well given the fact that Machine Learning a subset of Artificial Intelligence. We will take and define Machine Learning at face value — the ability for a machine to learn and gain insight about a certain topic. As for Artificial Intelligence, we will define it in as the machine’s ability to take action based on the certain rules and parameters given to it. Both technologies are often intertwined in real world use cases, but in this article (and for my own sanity) we will view the two as separate tools.

人工智能和机器学习之间有很多区别。 鉴于机器学习是人工智能的一个子集，因此两者之间也有许多相似之处。 我们将按面值理解并定义机器学习，即机器学习并了解特定主题的能力。 对于人工智能，我们将其定义为机器根据赋予它的某些规则和参数采取行动的能力。 两种技术通常在现实世界的用例中交织在一起，但是在本文中(出于我个人的理性)，我们将把这两种技术视为独立的工具。

在机器学习的情况下： (In the Case of Machine Learning:)

We’re all familiar with how our data is used to personalize nearly everything, be it from advertisements to user experience. While we may not know the technicalities behind how it is these algorithms determine what is a “good fit” for us or what isn’t, we do know that the metrics that are being used are constantly compared to others to see what similar patterns/hobbies users have in common. The same idea can be implemented into artists and their music.

我们都熟悉如何使用我们的数据来个性化几乎所有内容，无论是广告还是用户体验。 虽然我们可能不知道这些算法是如何确定对我们而言是“合适”还是不合适的背后的技术细节，但我们确实知道，所使用的指标会不断与其他指标进行比较，以了解类似的模式/用户的共同爱好。 可以将相同的想法应用于艺术家及其音乐。

When using Machine Learning in music, we can replicate how we dissect and study music, be it from a lyrical or sonic perspective, at scale. This means that the idea of “similarity” can go beyond how we consider two artists within the same genre to be similar. The measures of similarity will most likely transcend genre and time period as everything about one artist serves as context for another — similar to how my context determines not only the ads that I see but the ones you do as well.

在音乐中使用机器学习时，无论是从抒情还是声音的角度，我们都可以大规模地复制和解析音乐。 这意味着“相似性”的概念可以超越我们认为同一流派中的两位艺术家相似的范围。 相似性的度量很可能会超越类型和时间段，因为关于一位艺术家的一切都将作为另一位艺术家的语境-类似于我的语境不仅决定了我看到的广告，还决定了你所做的广告。

However, if in the wrong hands, deep insights on what makes an artist who they are can be leveraged towards artificially inflating numbers to the highest bidder. This in turn becomes a new advertising market in music. The label or entity who shells out the highest payment gets put up top whereas independent artists could be left behind.

但是，如果掌握不正确，则可以利用对艺术家的深刻见解，将数字虚假地夸张给出价最高的人。 反过来，这又成为音乐领域的新广告市场。 掏出最高薪水的唱片公司或实体被置顶，而独立艺术家可能被甩在后面。

We can develop deeper insights on what makes an artist who they are musically, what makes them unique, and much more. However, if used incorrectly, these insights will only favor those who can afford it.

我们可以就什么使艺术家成为音乐人，使他们与众不同的东西等方面获得更深刻的见解。 但是，如果使用不当，这些见识只会惠及那些负担得起的人。

在人工智能的情况下： (In the Case of Artificial Intelligence:)

Do we want AI to be a tool to push humanity’s capabilities forward or do we want AI to automate tasks that we find incessant and boring? There are many cases where we can have both — having our cake and eating it too. However, in music, the answer is not as straight forward as we would like it to be.

我们是否希望AI成为推动人类能力前进的工具，还是我们希望AI自动执行我们认为无聊又无聊的任务？ 在很多情况下，我们可以两者兼得-吃蛋糕也可以吃。 但是，在音乐中，答案并不像我们希望的那样直接。

Demixing audio is a mathematical attempt at separating the various components of a song as if they were recorded separately. Spleeter by Deezer does the best job of this that I have seen so far and I am even currently building one for my undergraduate capstone project. A tool like this can drastically change how producers can sample from songs. Like the way the drums sound in a section of the song that’s alongside a few other instruments that you don’t want? Demix it and use it for your song. It may not be perfect with its current capabilities but it will get you started. Using AI in music should be no different from how VSTs work — a plug-in to make the creative process easier.

混合音频是一种数学尝试，可以将歌曲的各个组成部分分离，就好像它们是分别录制的一样。 到目前为止，Deezer的Spleeter所做的最好的工作是我迄今为止所看到的，我什至目前正在为我的大学毕业项目建造一个。 这样的工具可以极大地改变制作人从歌曲中采样的方式。 就像鼓声在乐曲中与其他一些您不想使用的乐器一起发声的方式一样吗？ 对其进行混合，然后将其用于您的歌曲。 它当前的功能可能并不完美，但是它将帮助您入门。 在音乐中使用AI应该与VST的工作原理没有什么不同-VST是使创作过程更轻松的插件。

It should be important to note that music generation is very real thing. Merge this with the top 5 companies’ computational resources and we have now shifted the music industry towards being provided thousands of hours of generated music at almost no cost to you. Now, I know what you’re thinking, who would want to listen to music made by a computer? You and I would when music generation mixes with our music taste profile and the music is now being constantly created and tuned to what we want. It’s not far from the realm of possibility and will be a reality sooner than we think it will be. The solution to this is simple, actively support human artists.

重要的是要注意音乐的产生是非常真实的事情。 将其与前5家公司的计算资源合并，我们现在已将音乐产业转向向您免费提供数千小时的生成音乐。 现在，我知道您在想什么， 谁想听计算机制作的音乐？ 当音乐的产生与我们的音乐品味相融合时，您和我会不断地创作音乐并将其调整到我们想要的音乐。 它与可能性领域相距不远，并且比我们认为的要早实现。 解决方案很简单，积极支持人类艺术家。

We can build tools to help improve the workflow of artists in creating and manipulating unimaginable songs. However, if used incorrectly, these tools could automate artists out of a livelihood.

我们可以构建工具来帮助改善艺术家创作和处理难以想象的歌曲的工作流程。 但是，如果使用不当，这些工具可能会使艺术家失去生计。

结论： (Conclusion:)

Artificial Intelligence can help usher in a new generation of creatives or replace them completely. All new and emerging technologies are similar at the core; their influence on how we operate is a direct product of how we influence their operations.

人工智能可以帮助引入新一代创意或将其完全替换。 所有的新兴技术在核心都是相似的。 他们对我们运作方式的影响是我们对他们运作方式的直接影响。