克雷西 发自 凹非寺 量子位 | 公众号 QbitAI 刚刚，Gemini摇身一变，成了专业的"唱作人"。 谷歌把最新的Lyria 3模型塞进了Gemini，直接在对话框里招呼一声，Gemini就能现场给你攒个乐团。 这套玩法主打一个有手就行，给它打一段天马行空的文字，或者干脆甩过去一张刚拍的照片，它几秒钟内就能根据你的想法，吐出一首带歌 词、旋律甚至人声演唱的完整作品，整个过程快得惊人。 而且还顺便拉来了Nano Banana模型当帮手，曲子刚生成完，一张风格特搭的专辑封面也就跟着出炉了。 总之，从你想出点子到拿到带封面的专属BGM，中间的步骤简化到了极点。 网友评价，48kHz的立体声质量，加上根据照片生成音乐的功能，可见DeepMind这一波非常注重创意工作流程。 你的照片能开口唱歌了 硬指标上，Lyria 3的音频采样率来到了48KHz的高保真级别。这种规格让生成的曲子底气特别足，每一声琴鸣都显得非常扎实，有了这个音 质底座，看图唱曲的功能才更有发挥空间。 你随手上传一张在森林徒步的照片，AI就能瞬间捕捉到那种静谧感，转手给你配上一段对味的民谣，让原本静止的风景瞬间有了自己的声音。 这下，你的朋友圈也 ...

Core Insights - Google and Apple are integrating music-focused generative AI features into their consumer applications, highlighting the mainstream adoption of advanced AI tools [1] Group 1: Google Developments - Google's Gemini AI assistant can create 30-second music tracks based on user-uploaded text, photos, or videos using the Lyria 3 model, available to users over 18 in multiple languages [2] - The Gemini AI will also generate custom cover art through its Nano Banana model, enhancing the sharing experience for users [3] - Google aims to strengthen its consumer offerings amid competition with OpenAI's ChatGPT, following positive reception of its Gemini 3 AI model [4] - The AI-driven music creation feature will have usage limits, allowing free users to generate 10 tracks per day, while paying users can create between 20 to 100 tracks daily based on their subscription tier [7] - Google has implemented safeguards to prevent the AI from using specific artists' content, ensuring compliance with intellectual property and privacy regulations [9] Group 2: Apple Developments - Apple is introducing a feature called Playlist Playground in Apple Music, enabling users to create playlists from text prompts, which will include cover art, descriptions, and 25 songs [5] - This new feature is part of iOS 26.4, currently in beta, and will be more widely available in the spring [5] - Apple is working to enhance its AI capabilities across its applications, although updates to its Siri virtual assistant may face delays [6] Group 3: Industry Context - The music industry has expressed concerns regarding generative AI tools, viewing them as potential threats to business and intellectual property [8] - Major music companies have taken legal action against AI startups for copyright infringement, with some reaching settlements to ensure proper licensing and controls [8]

周三（2月18日）美股盘中，谷歌在官网博客中宣布， Gemini应用已配备公司最先进的音乐生成模型Lyria 3，用户可使用文本或图像制作30秒的音 乐片段。 新闻稿称，"只需描述一个想法或上传一张照片，Gemini就能在几秒钟内将其转化为一首高质量、朗朗上口的歌曲。为了进一步拓展创意空间，您甚 至可以让Gemini从上传的内容中汲取灵感。" 生成式AI工具在音乐行业一直受到谨慎甚至敌视的对待，许多从业者认为其可能威胁商业模式和知识产权。 谷歌在博客中强调，其系统设有保护措施，禁止AI直接挪用特定艺术家的作品。如果用户点名真实音乐人，Gemini只会将其视为"广泛的创作灵感来 源"，生成风格或氛围与之相似的曲目。 谷歌提到， Lyria的音频生成功能采用了公司最新的隐私和安全特性，所有生成的曲目都使用SynthID技术添加了不易察觉的水印，可以检测音乐是 否由人工智能创作或编辑。 消息公布后，欧美主流流媒体音乐服务平台Spotify（股票代码：SPOT）股价回吐了日内近5%的涨幅，Sirius XM（股票代码：SIRI）也一度转跌。 有分析表示，虽然谷歌的音乐生成模型预计不会成为Spotify的致命打击，但这 ...

周三(2月18日)美股盘中，谷歌在官网博客中宣布，Gemini应用已配备公司最先进的音乐生成模型Lyria 3，用户可使用文本或图像制作30秒的音乐片段。 新闻稿称，"只需描述一个想法或上传一张照片，Gemini就能在几秒钟内将其转化为一首高质量、朗朗 上口的歌曲。为了进一步拓展创意空间，您甚至可以让Gemini从上传的内容中汲取灵感。" 根据谷歌的说法，Gemini不仅可以"文字转歌曲"，还可以通过"上传照片或视频"创作一首完美契合氛围 的歌曲，歌曲长为30秒，并配有Nano Banana设计的自定义封面。 另外，创作者还可以通过YouTube的Dream Track功能探索Lyria 3，以提升Shorts短视频配乐的品质。谷 歌称，Lyria 3已在美国推出，并正在逐步推广至其他地区的创作者。 有分析表示，虽然谷歌的音乐生成模型预计不会成为Spotify的致命打击，但这些举措可能迫使后者尽快 推出AI混音功能。 另外，把音频创作工具加入移动应用，有望增强谷歌的消费者产品竞争力。对于谷歌来说，公司需要持 续向投资者证明其AI产品投入能够带来收入增长。 生成式AI工具在音乐行业一直受到谨慎甚至敌视的对待，许 ...

周三（2月18日），三大指数集体收升。 截至收盘，道琼斯指数涨0.26%，报49662.66点；标普500指数涨0.56%，报6881.31点；纳斯达克综合指数涨0.78%，报22753.63点。 | 名称 | | 電新价 | 7505 20 | 瑞珠唱 | | --- | --- | --- | --- | --- | | 道現斯指数 | 0 | 49662.66 | +129.47 | 0.26% | | 标普500指数 | | 6881.31 | +38.09 | 0.56% | | 纳斯达克综合指数 | | 22753.63 | +175.25 | 0.78% | "七巨头"集体走高，英伟达收涨1.63%，前一天与Meta宣布达成多年跨代际战略合作伙伴关系；亚马逊涨1.81%，公司获潘兴广场大幅增持，成为该 基金的第三大重仓股。 安联投资管理公司高管Charlie Ripley表示："从我们的角度来看，会议纪要支持了我们的观点，即在可预见的未来美联储是不可能降息的。" 盘中，美国方面消息称，目前没有任何迹象表明美国与伊朗的外交谈判会取得突破，与此同时，越来越多的证据表明美对伊军事行动"迫在眉睫"。 受 ...

Market Overview - The S&P 500 is attempting its first three-session win streak since late January, with markets solidly higher [1] - Sector rotation is a significant theme, with investments increasing in energy, technology, consumer discretionary, materials, and financials, while real estate, utilities, and consumer staples are being sold off [1] Economic Indicators - New orders for manufactured durable goods, excluding transportation, increased by 0.9% in December, surpassing estimates of a 0.3% increase [1] - January industrial production rose by 0.7% month over month, exceeding forecasts of a 0.4% increase [1] Company Developments - Alphabet announced the launch of Lyria 3, a generative music model available in its Gemini app, allowing users to create custom 30-second tracks [1] - The introduction of Lyria 3 is expected to enhance the quality of soundtracks for YouTube Shorts, potentially increasing creator engagement on the platform compared to competitors like TikTok and Instagram Reels [1] Upcoming Earnings Reports - Companies reporting quarterly earnings after the market close include DoorDash, Carvana, Occidental Petroleum, Figma, Blue Owl Capital, and Molson Coors Beverage [1] - Before the market opens on Thursday, Walmart, Quanta Services, Deere, and Wayfair are scheduled to report their results [1]

Group 1 - Google has launched the Gemini application featuring the advanced music generation model Lyria 3 [1] - Lyria 3 is currently available in the United States and is being promoted to YouTube creators in other countries [1]