擒数网 （随信APP) | 智能眼镜功能被忽视，如今却成为必不可少的东西
文章目录[隐藏]
擒数网 （随信APP) | 智能眼镜功能被忽视，如今却成为必不可少的东西
【微信/公众号/视频号/抖音/小红书/快手/bilibili/微博/知乎/今日头条同步报道】
字幕，已经成了现代人的「外挂」。通勤嘈杂，摸鱼易露馅，学习会分心，听和看并用，才能避免进度条白白走了冤枉路。
然而，对一些人来说，字幕不只是锦上添花。
当《失控玩家》的男主角戴上眼镜，他看到了原本无法察觉的信息，不再是个被动的 NPC。
现实里也存在一种智能眼镜，可以提供「行走的弹幕」。尽管没那么科幻，但它同样通向一个信息量更大的世界。
我可以看到你说的话了
科技的意义在于，当上帝关上一道门的时候，打开一扇窗。
一位听障博主的开箱视频，在 TikTok 拿下了 80 万点赞，夙愿得偿，喜极而泣，最纯粹的情感迸发的力量，打动了素昧平生的观众。
她手里拿着一个盒子，边笑边流泪，比着手语，努力地说出完整的句子：「我等待这个盒子里的技术很多年了。」
盒子里装的是一款字幕眼镜，外形和普通眼镜没什么两样，却让她可以「看到」周围的声音，那些过去听不清的声音，变成了一行行亮绿色的、科幻感的字幕。
根据视频里的信息，这位博主用的眼镜叫作 Hearview，专为听障人士设计，今年 5 月发布，来自国内的一家科技公司，但不面向国内，市场在海外。
它可以用来娱乐，看电影、看视频、看《黑神话：悟空》的游戏直播。
从刷《老友记》的体验视频来看，文本略有些滞后但准确，配对的手机 app 还可以区分不同的说话人。
同时，它也可以用在面对面的对话，方便听障人士在餐厅点餐、在超市购物、和导购聊天、开车时导航……
HearView 的原理说起来并不复杂，和手机 app 配对，通过智能手机的麦克风捕获声音，AI 算法实时语音转文字，字幕显示在眼镜上。
如果听障人士需要回复他人，可以在 app 输入消息，将文字转换成语音。过往的字幕，也保存在 app 中。
HearView 自称，语音转文字的准确性达到 95%，最远转录 10 米外的语音。说到这里，一个问题呼之欲出：菜市场等吵闹环境里表现会变差吗？
虽然 HearView 表示手机 app 具备噪音消除功能，但我翻遍了各种体验视频，并没有找到嘈杂环境里的实测，所以很难探究效果如何。
官网只提到了一些参数信息，续航 7 小时，重 52 克，适合全天佩戴，还配备了振动和视觉警告，提醒用户注意可能的危险。
轻便，耐用，听起来让人很想剁手，但它实在太贵，很难交个朋友——1799 美元，约 12800 元人民币。
而且，Hearview 存在很大的进步空间，目前它只支持英语和西班牙语，计划支持德语和法语，还不能实现语言的互相翻译。
文本的颜色只有亮绿色，虽然确保了在各种屏幕和照明条件下都可读，然而用户没有选择，就是美中不足。
但存在，便意味着价值。字幕早已是很多人生活的一部分，这还不够，它理应以更方便、直观的形式，陪伴在最需要的人左右。
字幕眼镜虽多，但生活里的挑战更多
实时字幕眼镜，其实并非新鲜事，放在 2 年前，甚至是一个热潮，诞生过不少有趣的项目，但都很难尽善尽美。
英国 AR 初创公司 XRAI，开发了兼容多款 AR 眼镜的语音转文字应用 XRAI Glass。
然而，《连线》杂志记者戴上一款和 XRAI Glass 兼容的 AR 眼镜后发现，语音转文字固然好，但使用过程中，总有这样那样的摩擦。
先是使用体验上的，如果佩戴人工耳蜗和助听器，再戴眼镜可能会不舒服，虽然这款 AR 眼镜已经很轻，但和普通眼镜相比依然厚重。
而且，语音转文字在背景噪音大和多人讲话的时候，效果并不好，实用性大打折扣。
价格也是一个问题，300 多美元的 AR 眼镜已经让钱包隐隐作痛，XRAI Glass 还需要每月花几十美元订阅套餐，像很多语音转文字的软件那样，只能说并不意外。
国内也有类似 Hearview 的产品——亮亮视野的听语者字幕眼镜，分为助听版和翻译版，可以理解普通话、方言和外语，并且支持跨语种翻译。
它的使用方式和 Hearview 相近，下载可译 app，眼镜和手机进行蓝牙配对，用 Wi-Fi 或者热点连接网络，手机端实时语音转文本，眼镜端显示文本。
不过有用户在 App Store 反馈，听语者对方言的支持还是不够多。这是一个非常必要、却也艰难的优化方向，听力不好、习惯说方言的老年人，恰恰最需要这类产品。
也有 YouTube 博主评测发现，听语者的语音识别率高，速度也快，但口音很重的话，识别率会变低。
如果不拘泥于眼镜的形态，耶鲁大学和斯坦福大学的学生设计的一个产品原型很有意思，叫作 Transcribe Glass。
它不是一副眼镜，而是一个平视显示器，可以卡在眼镜的镜框上，开箱即用，将设备连接到 iOS 和 Android 的配套 app，让用户选择自己喜欢的语音转文本软件。
生成的字幕会通过低功耗蓝牙传输到设备，并实时叠加在用户的现实视野中。难得的是定价也格外友好，95 美元，但连测试版都还在候补，直到现在也查不到落地的信息。
除了专注语音转文字的产品，面向所有人的多功能 AR 智能眼镜，往往也包含了字幕功能。
国内价位在两三千的一体化无线智能眼镜，基本都支持语音转文字和多种语言的实时翻译，但转译效率，可能不如专为听障人士打造的 AR 眼镜，持续开着语音转文字时，续航也是一个问题。
当我咨询某个头部的智能眼镜品牌，得到的答案是「不建议听障人士使用」。
其实，这种产品最适合大厂出手，特别是 Google，在语音识别、机器翻译和 AR 方面都有深厚的技术积累，实时语音转文字和翻译的能力都遥遥领先。
Google 在 2022 年 I/O 推出了一款可以实时翻译的 AR 眼镜，支持 24 种语言，甚至支持美国手语，字幕就在对话人旁边，不过也只是原型演示，没有下文了。
科技向善是字幕眼镜的大前提，种种的「吹毛求疵」，其实是抱着殷切的心情，希望产品不满现状，不断进化，再做一次，再改进一次，变得更好一点。
字幕是一种权利，像看电影一样看世界
世界卫生组织统计，全球约有 4.66 亿人患有听力损失，其中超过 900 万人为重度耳聋。
即使使用了人工耳蜗或助听器，听障人士理解别人在说什么，仍然需要集中注意力，同时，一些发音相近的词语，光靠唇读难以区分，在理想条件下，英语中只有约 40% 的语音信息，能通过说话者的唇部动作捕捉。
这也是为什么，一个听力障碍的世界，需要文字作为补充。很早之前，就有人为之努力了。
古巴裔美国默片演员艾默生·罗梅罗，同时也是一位听障人士，在 1947 年开创了有声电影的字幕。
默片时代，观众可以通过电影画面和插入的文本了解故事情节，但当有声电影兴起，许多像他一样的听障观众反而失去了看电影的机会，因为他们无法听到台词，电影又缺乏字幕。
于是，罗梅罗决定自己手工制作字幕：当时的电影通过胶片播放，他将胶片切片，并在帧之间插入带有字幕的图像
英文版:
Subtitles have become a "cheat code" for modern people. With noisy commutes, it's easy to get caught slacking off, study distraction, listening and watching in combination is the key to avoid wasted progress bars going down the wrong path.
However, for some people, subtitles are more than just a cherry on top.
When the protagonist in "Ready Player One" puts on his glasses, he sees information that was previously imperceptible, no longer just a passive NPC.
In reality, there is also a type of smart glasses that can provide a "walking barrage". Although not as futuristic, it still leads to a world with a greater amount of information.
I Can See What You're Saying
The significance of technology lies in opening a window when God closes a door.
A unboxing video by a deaf blogger on TikTok received 800,000 likes. The long-awaited wish was fulfilled, tears of joy, the purest eruption of emotion, touching strangers.
She held a box, smiling and crying, using sign language, struggling to say a complete sentence: "I have been waiting for the technology in this box for many years."
The box contains a pair of subtitle glasses, which look just like regular glasses, but allow her to "see" the sounds around her. The sounds that were previously unclear have turned into bright green lines of subtitles with a sci-fi feel.
According to the information in the video, the glasses she used are called Hearview, designed specifically for the hearing-impaired, released in May this year by a domestic tech company, but not targeting the domestic market, focusing overseas instead.
It can be used for entertainment, watching movies, videos, and live streaming of games like "Black Myth: Wukong".
From the experience videos of watching "Friends", the text is slightly delayed but accurate, and the paired mobile app can also differentiate between different speakers.
At the same time, it can also be used in face-to-face conversations, making it convenient for the hearing-impaired to order food in restaurants, shop in supermarkets, chat with salespeople, navigate while driving...
The principle of HearView is not complicated - it pairs with the mobile app, captures sound through the smartphone's microphone, AI algorithms perform real-time speech-to-text conversion, and the subtitles are displayed on the glasses.
If the hearing-impaired person needs to reply to others, they can input messages in the app and convert the text into speech. Past subtitles are also saved in the app.
HearView claims that the accuracy of speech-to-text conversion reaches 95%, transcribing voices up to 10 meters away. At this point, a question arises: will the performance deteriorate in noisy environments like a market?
Although HearView states that the mobile app has noise cancellation, I have searched through various experience videos and have not found any real tests in noisy environments, making it difficult to determine the effectiveness.
The official website only mentions some parameter information, with a 7-hour battery life, weighing 52 grams, suitable for all-day wear, and equipped with vibration and visual alerts to remind users of potential dangers.
Lightweight, durable, sounds like a must-have, but it is just too expensive, making it difficult to purchase - $1799, or about 12800 RMB.
Furthermore, Hearview has a lot of room for improvement, currently only supporting English and Spanish, planning to support German and French, and still unable to achieve language translation.
The text color is only bright green, ensuring readability on various screens and lighting conditions, but the lack of user choice is a drawback.
But its existence signifies value. Subtitles have long been a part of many people's lives, but it should be more convenient and intuitive, to accompany those who need it the most.
While there are many subtitle glasses, there are more challenges in life
Real-time subtitle glasses are not a new phenomenon. 2 years ago, it was even a trend, giving rise to many interesting projects, but perfection has been elusive.
The UK AR startup XRAI has developed a speech-to-text app called XRAI Glass compatible with a variety of AR glasses.
However, a journalist from "Wired" magazine found that when wearing AR glasses compatible with XRAI Glass, there were always some friction in the usage process.
First is the user experience - if the wearer uses a cochlear implant or hearing aid, wearing glasses may be uncomfortable. Although these AR glasses are already lightweight, compared to regular glasses, they are still bulky.
Moreover, speech-to-text doesn't work well in loud background noise or when multiple people are talking, diminishing its practicality.
Price is also an issue; AR glasses that cost over $300 can already be painful, and XRAI Glass requires a monthly subscription package, just like many speech-to-text software, which is not surprising.
In China, there are also products similar to Hearview - Liang Liang Vision's subtitle glasses, available in hearing aid and translation versions, capable of understanding Mandarin, dialects, and foreign languages, and supporting cross-language translation.
Its usage is similar to Hearview, download the portable app, pair the glasses with the phone via Bluetooth, connect to the network via Wi-Fi or hotspot, real-time speech-to-text on the phone, and display the text on the glasses.
However, some users have provided feedback on the App Store that Liang Liang Vision's support for dialects is not enough. This is a very necessary but difficult optimization direction, as hearing-impaired elderly people who are used to speaking dialects are the ones who need such products the most.
There are also YouTube reviewers who found that Liang Liang Vision has a high speech recognition rate and speed, but if the accent is heavy, the recognition rate will be lower.
If not confined to the form of glasses, a product prototype designed by students from Yale University and Stanford University is quite interesting, called Transcribe Glass.
It is not a pair of glasses, but a heads-up display that can be clipped to the glasses' frame, ready to use, connect the device to the companion app for iOS and Android, allowing users to choose their preferred speech-to-text software.
The generated subtitles will be transmitted to the device via Bluetooth low energy and overlaid on the user's real field of view in real-time. Surprisingly, the pricing is also very friendly at $95, but even the beta version is still on a waiting list, and there is still no information on its launch.
In addition to products focusing on speech-to-text, multi-functional AR smart glasses aimed at everyone often include subtitle functionality.
Integrated wireless smart glasses priced at around two to three thousand RMB in China mostly support real-time translation of speech-to-text and multiple languages, but the translation efficiency may not be as good as AR glasses specifically designed for the hearing-impaired, and battery life is also an issue when continuously operating speech-to-text.
When I consulted a well-known smart glasses brand, the answer I received was, "It is not recommended for the hearing-impaired to use."
In fact, such products are best served by major companies, especially Google, which has profound technical expertise in speech recognition, machine translation, and AR, leading in real-time speech-to-text and translation capabilities.
Google unveiled a pair of AR glasses for real-time translation at I/O in 2022, supporting 24 languages, including American Sign Language, with subtitles placed next to the conversationalist, but it was only a conceptual demonstration, with no follow-up.
Kindness towards technology is a prerequisite for subtitle glasses, all the various criticisms are actually borne out of a keen desire for better products, for them to evolve and improve, to be refined again, and to become a little better.
Subtitles are a right, to see the world like watching a movie
The World Health Organization estimates that there are approximately 466 million people worldwide with hearing loss, with over 9 million people suffering from severe deafness.
Even with the use of cochlear implants or hearing aids, the hearing-impaired still need to concentrate to understand what others are saying. Additionally, with similar pronunciation of words, it is difficult to distinguish solely through lip-reading. Ideally, in English, only about 40% of speech information can be captured through lip movements.
This is why a world of hearing impairments requires text as a supplement. Long ago, some people worked hard for this.
Cuban-American silent film actor Emerson Romero, also a hearing-impaired individual, pioneered subtitles for sound films in 1947.
During the silent film era, audiences could understand the story through film visuals and inserted text. However, when sound films became popular, many hearing-impaired audience members like Romero lost the opportunity to watch movies because they could not hear the dialogue, and movies lacked subtitles.
Therefore, Romero decided to manually create subtitles: at that time, films were played through film reels, and he would slice the film and insert images with subtitles between frames.
Subsequent subtitle technology can be seen as a continuation of Romero's spirit, as well as more convenient designs for the hearing-impaired to watch movies, such as some American cinemas providing pre-made subtitle glasses.
In 2012, Sony introduced a pair of CC subtitle glasses for select cinemas. Unlike regular subtitles, CC subtitles not only include dialogue content but also non-verbal information like background music, sound effects, and environmental noises.
These glasses are equipped with a small projector on each side, and the green subtitles look like they are floating in front of the big screen.
Audiences sitting in any position in the cinema can clearly see the subtitles. Users can also adjust the brightness, display distance, angle, and switch between six languages.
These glasses even support 3D, eliminating the need for users to bring an extra pair of glasses to watch 3D movies. This is important, avoiding a scenario where one must wear regular glasses, 3D glasses, and CC subtitle glasses simultaneously while watching a movie.
This scene is somewhat ironic but metaphorical - to understand the world from an ordinary perspective, some people have to put in more effort.
Real-time subtitle glasses may not be perfect yet, but they show a glimpse of a promising future.
South Korean film director Bong Joon-ho said at the Oscars' award ceremony: "Once you overcome the 1-inch tall barrier of subtitles, you will be introduced to so many more amazing films."
When common subtitles in movies appear in life, they allow those in need to see more information and have the right to stay in the conversation.
Google's translation AR glasses promotional video has a line that is very touching: "I look into your eyes, and it seems like you're looking into mine."
Almost imperceptible, this is the best state of technology.
With efficient and accurate communication as the premise, interactions can return to naturalness, without occupying both hands, no need to look down at the phone, or stare at the lips continuously, but rather face-to-face, eye contact, as it has been for thousands of years.
智能眼镜这个被忽略的功能，他们已经离不开了
#智能眼镜这个被忽略的功能他们已经离不开了
关注流程：打开随信App→搜索擒数网随信号：973641 →订阅即可！
公众号：擒数网 抖音：擒数网
视频号：擒数网 快手：擒数网
小红书：擒数网 随信：擒数网
百家号：擒数网 B站：擒数网
知乎：擒数网 微博：擒数网
UC头条:擒数网 搜狐号：擒数网
趣头条：擒数网 虎嗅：擒数网
腾讯新闻：擒数网 网易号：擒数网
36氪：擒数网 钛媒体：擒数网
今日头条：擒数网 西瓜视频：擒数网
