你是否曾在深夜辗转之际隐约听到过自己的心跳声？或在万籁俱寂的环境里，感受到一种既宁静又略带压迫的感觉？从城市的喧嚣到自然的低语，我们生活在一个被声音包裹的世界里，几乎从未有过绝对安静的体验。这间坐落在美国新泽西州默里山的消音室隶属于诺基亚贝尔实验室（Nokia Bell Labs）（下文简称默里山无回声室），建成于1947年，是目前全球最古老的、仍可正常使用的楔形无回声室之一，也是世界上最安静的房间（之一），就像一个可以吞噬绝大多数声音的“声音黑洞”。很多人可能有过这样的经历，比如在吹气球时不小心用力过猛吹破了，或是想制造一出恶作剧，在某人经过时突然踩爆一个气球，气球爆炸时，你会听到一声响亮的“砰”，声波向四周扩散、反射，仿佛充满整个空间，把自己和别人都吓一大跳。这种音色并不罕见，但如果在默里山无回声室——关上厚重的双层隔音门后——戳爆一个气球，它只会发出一声极其尖锐、短促的爆裂声，随后很快归于沉寂，声音没有任何回响或蔓延。紧接着，你会意识到，自己已然身处一个没有任何杂音、安静到近乎恐怖的世界。除了戳气球，消声室的工作人员（有时也包括艺术家）已经在这里创造出无数与我们生活息息相关的声学技术和有趣的艺术作品。比如，这里可被用于测量扬声器和麦克风的指向性和频率响应函数，模拟音乐厅声学以及阻抗面上的声音传播。贝尔实验室发明电驻极体麦克风（目前最常见的电容麦克风）的过程中，在消声室内进行的测量和验证发挥了关键作用，同时也推动了对定向麦克风系统的诸多深入研究。作为消声室实验的深度参与者，哥伦比亚大学计算机音乐中心主任塞思·克卢特（Seth Cluett）自2017年起便担任贝尔实验室的常驻艺术家，艺术家进驻是贝尔实验室对上世纪60年代“艺术与技术实验”（Experiments in Art and Technology）项目的传承与重启。其理念在于，基于媒体技术与科学研究的交叉，引入艺术家这类拥有发散性思维的“右脑思考者”，他们会在线性的工程研发环境中提出令人意想不到的问题，从而激发出创新的火花，不断拓宽听觉体验的边界。塞思介绍道，这间历史悠久的消声室名为“anechoic chamber”，字面意义就是“没有回声的房间”，它的设计初衷就是无限吸收随机的声波反射。其建筑结构为“房中房”：外部是约60厘米厚的墙体，与内部约9米见方的腔室之间由一个空气夹层隔开，以阻绝外部声波的传入。室内则铺设了大量泡沫基玻璃纤维楔形吸音体，外层包裹着金属丝网，这些吸音体密密麻麻地覆盖了四周墙壁、天花板和地面。吸音体底部截面约为0.6米×0.6米，单块楔形板长约1.37米，里宽外尖。楔形的设计实现了吸音体与周围空气的“阻抗匹配”，也可以视作一种波导，当声波射入楔形板之间的V形缝隙时，会在两个斜面之间反复反射，其声能在这个过程中被迅速吸收殆尽，最终几乎没有可以反射回房间的能量。而为了实现更均匀的角度吸收，楔形板以三个为一组，呈横竖交替的网格状排列。这一设计是默里山无回声室的首创，后来大多数现存消声室采用的都是这套方案。遍布全屋的吸音器可吸收99.995%以上大于200赫兹的入射声能，因此，默里山无回声室曾一度被《吉尼斯世界纪录》列为世界上最安静的房间。走进这里，我们并非“脚踏实地”，而是悬行在一张距离实际地面约3米高的金属网上。这是为了让实验能够在房间的几何中心进行，因为在立方体的正中心，声源到所有墙面的距离和反射路径都完全相等，是唯一能进行精确声学测量的地方。进入这个绝对寂静的空间后，人体会经历一系列奇特的感官体验。大多数人的双耳首先会有种受压感，空气仿佛变得“沉重”或“粘稠”，就像戴上降噪耳机那一刻的感受。很快，你会注意到一种低频的脉动，然后意识到这是自己的心跳——它不仅能被感觉出来，甚至能在空气中被“听”到。对于大多数人来说，他们还会听到一种持续的高频声响，起初人们可能会以为是自己耳鸣或者听力出了问题。其实这是一种名为自发性耳声发射（SOAE）的现象，在完全没有外界声音刺激的情况下，耳蜗外毛细胞会自发产生极其微弱但可被测量的听觉信号。在消声室里，通过耳蜗本身的拾音作用和颅骨的“骨传导”效应，这种伴随每个人一生的微弱声响才得以被清晰地听到。来自人体的声音一旦出现，就很难被忽视。这种感受源于“战或逃”（fight-or-flight）应激反应。在极端安静中，内耳基底膜上的静纤毛会变得异常敏感，试图捕捉一切潜在的威胁，如果你是原始人，可能须要在此时警惕“周围是否有猛虎出没”，但如果你是忙里偷闲的摸鱼打工人，就该竖起耳朵警惕领导有没有从背后偷偷靠近了。为了更直观地展示消声室的特性，塞思利用一个能够呈现360度立体声音频的高级麦克风进行了两项声音演示。首先，他以恒定的音量向正前方唱出一个单频音，同时缓慢地在原地转圈。当他正对听者时，声音最响亮、最直接；而当他背对听者、面对墙面时，声音会急剧减弱，因为绝大部分声音都被吸音体吸收了，几乎没有反射声能到达听者耳中。另一项演示则巧妙验证了声学中的“反平方定律”。该定律指出，在没有反射的自由声场中，声压级会随着与声源距离的加倍而降低6分贝，听感上的响度大约减半。塞思以正常音量说话，然后退后使距离加倍，此时他的声音响度恰好衰减为一半；当他再次将距离加倍，响度也随之再次减半。这种体验的奇特之处在于，你可以清晰地感受到，声音的响度虽已降至耳语水平，但音色和质感却毫无变化，与人们说悄悄话的气声感，或有人从很远处朝你喊话的呐喊感都完全不同。在电子技术发展史中，贝尔实验室绝对是无法忽视的存在。这里是现代多媒体技术的摇篮，从数字二进制的基本单位“比特”（bit），到数字音频录制的基础——由克劳德·香农（Claude Shannon）和约翰·皮尔斯（John Pierce）发明的脉冲编码调制（PCM），再到可以拍摄数码视频的电荷耦合器件（CCD）和核心电子元件晶体管……无数颠覆性技术都诞生于此。甚至，世界上第一部有声电影《爵士歌手》（The Jazz Singer）使用的音画同步技术，也是由贝尔实验室的前身——西部电气公司（Western Electric）开发的。自建成以来，默里山无回声室已经催生出一系列影响日常生活的重大创新。例如，很多人童年记忆中按键式电话独特的“滴滴嘟嘟”按键音，其音调和调谐并非随机，而是为了在电话网络日益庞大之时，帮助人们更容易地记忆电话号码。这项研究正是在这里通过声学与心理学（即心理声学）的交叉研究得到了优化。此外，默里山无回声室里还进行了大量关于人类语音合成与信号处理的基础研究，终极目的都是希望尽量优化信号以高效实现声音的远距离传输。值得一提的是，贝尔实验室在上世纪六七十年代开创地引入了“人因研究”，该领域的研究重点在于技术带给人的主观感受，而非纯粹的硬性技术指标。直到今天，在通讯领域，对人性化细节的考量依然具有现实意义。例如，智能手机通话中所谓“人性化”的“主动降噪”功能其实相当“非人化”，无人说话时的绝对寂静会切断通话双方微妙的共享空间感，缺少背景空间的声音暗示，人们可能会下意识地产生“你到底有没有在听”的不安。在虚拟现实（VR）和人工智能（AI）时代，人因研究再次回到业界的视野焦点。赛思指出，传统心理声学研究大多服务于电话通讯系统或军事（如声呐、雷达）领域，目标是高效传输信号或定位目标。但现在，人们在与机器交互，或隔着机器与他人交流的过程中，经常会出现“恐怖谷”效应，这是一种“几乎真实但又不完全真实”的怪异感，比如带着电流的人声听起来像诡异的“伪人”。这种感受与声音中的“软性因素”，也就是听觉上的“深度知觉”密切相关，即那些能让我们感知到情境、产生空间归属感和沉浸感的人性化细节。塞思希望弥合这“最后的10%”。通过设计新型扬声器和麦克风阵列，捕捉到那些在技术层面不够显著、却能被我们大脑解码并产生空间感的隐藏信息。有了这些信息，我们才能真正在虚拟世界里拥有“身临其境”的感受。贝尔实验室还见证了许多艺术与科技交融的趣闻。例如，很多人不知道的是，1961年，全球第一首由计算机“演唱”的歌曲就诞生于此。彼时，为了攻克数字语音合成难题，研究团队使用IBM计算机开发出一款名为“声带模型”（Vocal Tract Model）的软件，通过模拟人类声带振动、口腔共鸣和唇舌动作，从零合成音节和旋律。当时，电脑唱出的《黛西·贝尔》（Daisy Bell）听上去就像机器人在喃喃自语，低沉、缓慢，还带有明显的电子嗡鸣。1962年，科幻电影《2001太空漫游》（2001: A Space Odyssey）的编剧亚瑟·查理斯·克拉克（Arthur Charles Clark）在参观实验室时听到了这段歌声，深受启发，才有了后来影片里人工智能HAL 9000在“死亡”时唱起《黛西·贝尔》的经典一幕…… [查看全文]
Go Inside a Room That Lets You Hear Your Nervous System
Seth Cluett: This experiment just pops a balloon. And normally, when a balloon would pop, you’d hear [makes an exploding noise]—the whole room just kind of expand, right?[Pops a balloon inside a normal room, making a loud noise.]
Rachel Feltman: [Laughs.] Yeah.Feltman: Yeah, pretty loud [laughs].Cluett: But in this room there’s none of that. So you’re gonna hear it as a very sharp sound that just disappears completely.[Pops a balloon inside an anechoic chamber, making a sharp noise that dissipates immediately.]Cluett: Welcome to the anechoic chamber. Watch your step.Wow, it is already super quiet in here [laughs].Cluett: And it’s gonna get even more quiet when we close the door.Cluett: [Walks to the chamber entrance and closes the outer and inner doors.] How’s that?Feltman: It did get a lot more quiet, yeah [laughs].Welcome to Science Quickly. I’m Rachel Feltman, and today I’m here with Seth Cluett at Nokia Bell Labs. And you may notice if you’re listening to this, or if you’re watching it, that there’s some interesting stuff going on with the sound. Seth, would you tell us more about why that is?Cluett: Yeah, so we are in the historic anechoic chamber at Bell Labs. It is a room that absorbs 99.999 percent of sound-wave propagation and eliminates sound from the outside almost entirely. It is anechoic, meaning it lacks echo. So an anechoic chamber is intended to absorb as close to 100 percent of incidental reflection as you can possibly do.This room, as you kinda look around, is a [roughly] 30-by-30 cube with a wire mesh a third up from the floor. You might ask yourself, like, “Why is it a third up from the floor, not in the center?” And the answer is that you want the experiment to happen as close to the middle as possible.Cluett: Because the one place where it’s scientifically feasible to measure sound exactly is in the center of the cube because the distance to the walls and the reflections are equal in that case.The building is a shell that has [roughly] two-foot-thick walls and an air gap and then another wall, and that air gap separates the sound waves from the outside. And then inside the room there are these [roughly] four-foot wedge panels in groups of three in an offset kind of orthogonal pattern—a, a grid pattern. They capture the sound waves before they’re able to reflect back. So they, they come to a kind of inverted point in the inside of the wall, and when sound gets into that point it kind of bounces back and forth on the diagonal, and by the time it gets to the outside of the wedge there’s no more sound energy left to reflect.Feltman: Mm, so when you’re in this totally quiet space what kinds of things can you feel and perceive in your body?Cluett: Yeah, I think most people feel a kind of pressure against their ears first.Cluett: It sort of feels like the air is heavy or thick. Then, then you start to notice a kind of low pulse and realize that’s your heart beating and you can hear it in the air, in, in addition to through your body. And for most people you can even hear a high pitch, and you think, “Oh, I have tinnitus,” or “There’s something wrong with my hearing.” You’re actually hearing your nervous system ...Cluett: Bone conduction through your, through your skull, and it’s quiet enough that for the first time you can hear a part of your body that’s been along with you for the whole time you’ve been alive.Feltman: That’s very cool. Could ...Harper: Sorry—what? [Laughs.] That’s so crazy.Jeffery DelViscio: I was just like, “F—, I can hear that.”Harper: You’re just hearing your nervous system ...Feltman: Yeah, I actually ...Harper: Okay, no worries.DelViscio: I’m glad we got that. That was ...Harper: Sorry, carry on. I’m—I just—I’m so sorry [laughs].DelViscio: That was really cool.Feltman: Yeah, now I’m really distracted ’cause I can’t stop noticing it.DelViscio: Listening to your nervous system ....Cluett: And so that, that, actually, it’s starting to go away for you now ...Cluett: Because of—that’s part of the fight-or-flight response because you’re hearing it through the ...Cluett: Hypersensitive part of your stereocilia on your basilar membrane that are like, “Wake up! Are there tigers?” [Laughs.] Right? Like ...Feltman: [Laughs.] Could you show us some demonstrations to maybe help our listeners and viewers understand, like, how unique this space is? You’ve also let us borrow a very fancy mic so that we can present 360 stereo audio so that our, our listeners and viewers can, you know, experience being in the space as, as closely as we can approximate.Cluett: Yeah, absolutely. It’s subtle, and I hope that your viewers are wearing headphones.Feltman: So how does this demonstration work?Cluett: Okay, so in order to demonstrate how much of the sound the room absorbs, I’m gonna sing a sinusoid, a pure tone, straight ahead. You’re gonna hear that as the loudest, most direct sound. I am not gonna change the volume of my sound at all, but I’m gonna turn around in a circle, and so as I turn you’re gonna hear the room absorbing more and more of my sound.[Starts singing and then stops to clear his throat.] Sorry, granola bar.Feltman: [Laughs.] No worries.Cluett: [Clears his throat and then starts singing while slowly moving in a circle.]Feltman: So how does this one work?Cluett: Okay, so one of the kind of miraculous things about an anechoic chamber is it allows you to hear something that you can’t hear outside of the, of the chamber. And that’s that for every doubling of distance a sound source has to the listener there’s a cutting in half of the volume.Cluett: And so, and so I’m gonna speak to you at this level, and then I’m gonna double the distance, and you’re going to hear the same timbre, the same quality, the same emotional content as I’m talking now but just at half the volume. And then I’ll double it again, and you’ll hear even more of a falloff in the, in the volume.Cluett: Okay, so this is me talking at the normal volume. And I’m going to move back double. [Moves backward.]And this is me talking at that same volume; it’s just half as loud for you. [Moves backward.]And this is me talking at the same volume, and it’s half as loud again.Feltman: [Laughs.] Wow. Yeah, that is—it’s very weird to have it be the, you know, the, the level of a whisper but not sound like one [laughs].Feltman: So we’re obviously recording a podcast and a video right now. What kinds of technologies were created here that make this kind of multimedia possible?Cluett: It’s kind of overwhelming, honestly. The “bit” of digital binary was invented here, and as an extension Claude Shannon and John Pierce invented pulse-code modulation, which is the way we record sound digitally. The charge-coupled device, the CCD, that captures video was invented here, the transistor—you know, the list goes on and on. Even The Jazz Singer, the first [feature-length] sound film [with synchronized sound for dialogue sequences] ...Cluett: The technology for that was developed by Western Electric—Bell Labs, how Bell Labs was founded, making it possible for us to synchronize sound and image for film.Feltman: And what kind of experiments have happened in this chamber since then?Cluett: A kind of remarkable amount of things that touch our individual lives. Like, we think about the Touch-Tone phone ...Cluett: The “bop beep bop bop bop bip bop.” The individual tones and their tuning are optimized for memorizing phone numbers ...Cluett: And the acoustic research paired with psychology, or what we call psychoacoustics, was done in this room in order to, you know, help people figure out how to memorize the most phone numbers as the nation’s telephone grid got larger and larger ...Cluett: And it wasn’t just, you know, “Bronson 257,” right?So in addition to that, you know, there was a, a fundamental research into synthesizing the sound of the human voice ...Cluett: Right? We talk about microphone research and loudspeaker research, but that’s not the entirety of it. Behind that is signal-processing research, that’s about: How do you optimize a signal to go long-distance? And at first that was: How do you optimize a signal to go over a long run of copper wire, right? Reusing the telegraph system. And now in wireless: How do you, how do you encode a signal to minimize the amount of data taken up and maximize the quality of the voice?You know, we think of science as being hard measurements and, and facts, and one of the things that’s remarkable about Bell Labs in the, in the ’60s and ’70s was human-factors research, like: How does it feel to use technology?Cluett: And one of the convincing things in telephony was: How do you make sure, as you’re reducing the signal, that you’re not throwing away emotion ...Cluett: That you’re not throwing away recognition, that you can still recognize the person that you love on the other end of the line? And I think one of the things that this room actually shows us, and the audio for this podcast will actually reveal a little, is that the digital silence in cell phones is, is actually really dehumanizing ...Cluett: That, like, without the space around the signal, you, you lack context for the other person, and you’re like, “Are you there? Are you there?” And—because the absolute digital silence makes you feel, you know, quite alone.Feltman: Yeah, that’s really interesting.So I know that you are an artist-in-residence here.Feltman: How did you get involved with the lab?Cluett: Sure, so I’ve been here since 2017, so this is my eighth year in the artist-in-residence role. When I first arrived I came in as a—as part of a reboot of the historical Experiments in Art and Technology program, which was founded in the 1960s. And then in 2014 there was a reboot of, of, like, the role of artistic research at Bell Labs in, in one form or another. And because so many media technologies intersect with scientific research, the idea was: Let’s bring in artists, left-brain thinkers, to, to ask kind of divergent questions in a linear engineering space.And so when I first came I was very active, moving lab to lab, asking questions of engineers and, and technical staff of what their research was and getting them to, to enter a dialogue to try to see what sort of divergent left-brain thinking might contribute to a kind of linear engineering research and development space.Feltman: Awesome. So what kind of stuff have you been working on?Cluett: So I’ve been really interested in what I’m calling “the last 10 percent,” which is, like, in psychoacoustic research so much of the things that we know about the sound world were developed either for telephony or for the military ...Cluett: So: Can we get a signal efficiently across a line, or can we triangulate something in space—so sonar and radar and those sorts of things? In virtual reality and in artificial intelligence, now the question is: How do we pass the kind of uncanny valley of, like, this is almost real, but it’s not, right?What I’m interested in is: What are the aspects of sound that are kind of soft factors, that are the human things that ...Cluett: Make us feel like we’re situated, that make us understand the context? The auditory equivalent of depth perception ...Cluett: Was something that was never, you know, looked at seriously in the initial sort of foundations research. And so I’ve been doing experiments around loudspeaker-array design and microphone-array design to try to think about what is hiding in the signal that our brain is processing that might not appear in the physical acoustics that give us a real sense of belonging—like, a sense of space. Like: “I am in this space, and, and, and it’s me and my body, not just my mind.”Feltman: Mm, and I know that in addition to the kind of experiments that have been done here to contribute to technological advancements, there’s also been just sort of a lot of interesting art projects that have happened here ...Feltman: What are some of the, like, stranger, more interesting things that have been done in this room?Cluett: Sure. I mean, a lot of people don’t realize that the first time a computer sang was at Bell Labs, right?Cluett: So at the end of 2001: A Space Odyssey, when HAL is singing—is dying ...Cluett: And he sings, “Daisy ...”[CLIP: HAL from 2001: A Space Odyssey sings: “Daisy, Daisy / Give me your answer, do.”]Cluett: That was heard by Arthur C. Clarke when he did a tour of Bell Labs...[full transcript]
扫码关注“领研网”微信公众号
订阅最新“科学60秒”英语新闻
不再漏掉任何一次新知 plus 练耳的机会~
领研网约 19 分钟