Is it too much to ask for social media companies to come up with something new? Apparently so, since all of them are determined to copy rival TikTok’s every move. The newest copycat is Twitter. Yes, that Twitter, the last platform in your mental list of places to go to watch “cool” or “interesting” videos.
A former employee at the National Security Agency (NSA) has been charged with attempting to transmit classified defense information to a representative of a foreign government. He was arrested on Wednesday and made his initial appearance at a federal court on Thursday.
The case for heating homes with hydrogen rather than natural gas appears to be dead. In the UK, hydrogen has become an important part of the debate around decarbonizing home heating. 85 percent of all homes use natural gas to heat space and water, with the oil and gas industry pushing hydrogen as something that can leverage the existing gas pipelines. And lawmakers with close ties to the industry have claimed that hydrogen is a “silver bullet” to help the UK reach its climate targets.
According to a new study from the Regulatory Assistance Project, an NGO, such claims are a big pile of old nonsense. The project ran an extensive meta-analysis of research into hydrogen technology overall, finding that the promises of easy retrofit don’t add up. It said that it wasn’t clear if the existing infrastructure was actually suitable to take hydrogen without major adaptation. That was, after all, one of the major selling points of using hydrogen over switching to heat pumps and other low-carbon methods.
It’s something that Engadget already covered in its extensive report on the UK’s home heating situation back in 2021. The suitability of infrastructure is only one part of the problem, however, since many experts also asked where all of this hydrogen was coming from. Supplying the UK with enough hydrogen to heat 85 percent of its homes, without any work to reduce demand, would require around 10 million tons of hydrogen.
In that report, Tim Lord, who was previously responsible for the UK’s decarbonization strategy, said that to generate that much hydrogen cleanly, you would need around 75 gigawatts of offshore wind. The UK Government’s most recent figures say that the country’s total installed offshore wind capacity is just 10 gigawatts. It’s hard to see the economic case for installing seven-and-a-half times the total offshore wind capacity just to generate hydrogen.
The Regulatory Assistance Project’s report also found that trying to use hydrogen for space and hot water heating is a waste of a vital material. Green hydrogen could be put to better use in agricultural processes, like making fertilizer or in heavy industry. And we’ve already seen that green hydrogen has a part to play in decarbonizing industrial transport, like shipping, and in the railways where mass-electrification isn’t viable.
In its conclusions, the report adds that greater emphasis on hydrogen will only serve to delay the take up of better technologies, like heat pumps. There’s a political dimension to this, too, with The Guardian reporting that hydrogen lobbyists were out in force at the recent Labour Party conference, and are expected to attend next week’s Conservative Party conference as well.
Another study, from the MCS Charitable Foundation in partnership with energy analysts Cornwall Insight, found that hydrogen’s cost to consumers would be nightmarish. It found that switching from natural gas to hydrogen would likely see the cost increase by between 70 to 90 percent on average. It also warned that, unlike electricity, hydrogen would be subject to the same market volatility as other fossil fuels.
As before, this study raises the question about how much we can rely upon hydrogen given that many of its key needs are still untested. For instance, steam reformation of methane would still require carbon capture and storage at a vastly larger scale than present. (Not to mention the fact that methane is a far deadlier climate gas than carbon dioxide, so any leaks or accidents would be significantly more damaging for the planet.)
Fundamentally, on this and all of the other evidence, it would seem like legislators should avoid the expensive distraction of hydrogen in favor of full-scale electrification. That, as we’ve already covered, would provide a significant, and swift, reduction in emissions (and a timely boost to the economy).
Finally, Sonos has a subwoofer that’s more affordable and practical for smaller spaces, the Sub Mini. It only took 10 years to get here. Sonos’ original wireless Sub, which debuted in 2012, has always been targeted at its most hardcore users. With a launch price of $699, it was just as expensive as the company’s flagship Playbar, and its enormous size made it overkill for apartments. (It’s now $50 more after Sonos’ recent price hikes.)
Casual Sonos fans were basically out of luck, especially as the company released more affordable soundbars, like the Beam and Ray. You probably wouldn’t want to pair a $699 subwoofer with a speaker that costs $400 or less. Simply put, the $429 Sub Mini fills a huge gap in Sonos’s lineup. But is it actually any good?
If I could, I’d show you my cat’s shocked reaction as I blasted Baby Driver’s opening car chase on the Sonos Arc in my family room. Sonos may not be the most price-conscious company around, but it’s always made reliably great speakers. The Sub Mini is no exception.
Now I know, if you own an Arc, you’d most likely opt for the beefier Sonos Sub. But I was still impressed by how much the smaller subwoofer helped, especially for a soundbar that already delivered some fabulous low-end sound on its own. Muffled shotgun firing at the beginning of Baby Driver shook my walls (and made one cat leap into the air). I could viscerally feel the rumble of engines, the impact of car crashes and the weight shift every time Baby used the emergency brake for a sharp turn. The Sub Mini transformed the movie from something I was just watching to something I was experiencing.
I was genuinely surprised by how big the Sub Mini sounded, especially since it’s a relatively small cylinder. It weighs 14 pounds — 22 pounds lighter than the big Sonos Sub — and features dual six-inch woofers that face inward. Its sealed design means it doesn’t push out a ton of air like ported subs, but that also ensures a tighter bass response. The Sub Mini can reach down to 25Hz – more than enough to make the opening of Blade Runner 2049 hit me right in the gut.
While I wouldn’t call it portable, I appreciated how easy it was to move the Sub Mini around my home to test in different rooms. (Trust me, that was a lot less fun with the massive Sonos Sub.) You can pair the new sub with Sonos’ powered speakers, like the Beam, Ray and Play:5. Unfortunately, it doesn’t work with Sonos’s portable offerings.
That makes sense for the tiny Roam speaker, but I was genuinely disappointed to learn the Sub Mini doesn’t support the larger Move. That’s a speaker I’ve grown to love recently, since it lets me easily bring music into my backyard. It’s a shame Sonos couldn’t make the Sub Mini work while the Move was sitting on its charging base. (At that point, how is it different from a speaker with a plug?)
Despite that annoyance, the Sub Mini seems well-suited to homes with plenty of Sonos speakers. Moving it over to a Play:5 in my living room took around 30 seconds. And once it got going, it instantly added an impressive amount of depth to some of my usual test tracks. Tan Dun’s “Night Fight” from the Crouching Tiger, Hidden Dragon soundtrack sounded like I was hosting a traditional drumming concert in my home. Flying Lotus’s “Zodiac Shit,” a go-to track for low-end testing, sent another one of my cats cowering into another room.
As great as it is for music, I’d best most Sub Mini buyers would end up pairing it with one of Sonos’s soundbar for better movie and TV sound. It certainly made a huge difference on the first-gen Beam in my bedroom, which sounded twice as large during that Baby Driver chase sequence. Obviously, that’s not a room where I would want to have thumping bass all the time, but it sure is nice to have the option.
That’s my biggest takeaway. After a decade of waiting, Sonos fans finally have a viable option for beefing up their sound that isn’t obscenely expensive. And if you’re lucky enough to have multiple Sonos devices, you can easily move that bass magic all over your home. It may be called the Sub Mini, but really it’s all about maximizing sound where it matters.
The five-time Grammy winner prefers to keep the “bad names that I could have been” to herself.
The Stars Undying, the upcoming sci-fi debut from Emery Robin, introduces an interstellar princess forced into circumstances that’ll require every bit of royal strength she can muster to overcome. It’ll be released in November, but io9 is thrilled to share a generous excerpt from the book today.
Have you watched a Marvel Studios show on Disney+ and thought “I wish this was just a movie?” Well that just happened with development of the upcoming Armor Wars. The show, which stars Don Cheadle as War Machine, was being developed as a follow up from the events in next year’s Secret Invasion but now, it turns out,…
Despite claims to the contrary as recently as July, Google is shutting down its Stadia games streaming service after just three years. The company says players will still have access to their games and be able to play them until January 18, 2023. After that, Stadia will join the long, long list of products that have been killed by Google.
Those who have invested money into Stadia will be fully reimbursed. “We will be refunding all Stadia hardware purchases made through the Google Store, and all game and add-on content purchases made through the Stadia store,” Stadia vice president and general manager Phil Harrison wrote in a blog post. “We expect to have the majority of refunds completed by mid-January, 2023. We have more details for players on this process on our Help Center.”
The Stadia store is now closed, so you can’t make any new purchases. In-game transactions have been disabled too. Google says it will handle most refunds automatically, and you won’t have to return most Stadia hardware (so hey, at least some people will get a free Chromecast Ultra out of this mess). Meanwhile, Google won’t reimburse Stadia Pro subscriptions. If you have an active membership as of today, the company won’t charge for access to your library of games or other Pro features until Stadia shuts down.
The reasoning behind the “difficult decision” isn’t surprising. Google said the service “hasn’t gained the traction with users that we expected.” It’s a shame, though, since the streaming tech at the core of Stadia works very well and the service had a passionate, if ultimately small, community. The decision leaves Xbox Game Pass, NVIDIA GeForce Now and Amazon Luna as the vanguards of cloud gaming for now.
It seemed like the writing was on the wall for Stadia when Google closed its internal game development studios early last year. There have been other indications in recent months that Google was shifting resources away from Stadia to focus on licensing the underlying tech to other companies. We’ve seen that in practice over the last year or so, with AT&T offering its subscribers the chance to play Batman: Arkham Knight and Control at no extra cost. Capcom used Stadia tech for a streaming demo of Resident Evil Village as well.
Back in March, Google formally announced Immersive Stream for Games, a version of Stadia that third parties can license. It seems the Stadia tech will live on there and in other Google products. “We see clear opportunities to apply this technology across other parts of Google like YouTube, Google Play and our Augmented Reality (AR) efforts — as well as make it available to our industry partners, which aligns with where we see the future of gaming headed,” Harrison wrote. “We remain deeply committed to gaming, and we will continue to invest in new tools, technologies and platforms that power the success of developers, industry partners, cloud customers and creators.”
Harrison noted that many members of the Stadia team will continue their work in other parts of the company. It’s not clear if there will be any layoffs as a result of the Stadia closure. When asked to comment on the possibility of layoffs, Google directed Engadget to Harrison’s blog post.
They Shall Not Grow Old, a 2018 documentary about the lives and aspirations of British and New Zealand soldiers living through World War I from acclaimed Lord of the Rings director Peter Jackson, had its hundred-plus-year-old silent footage modernized through both colorization and the recording of new audio for previously non-existent dialog. To get an idea of what the folks featured in the archival footage were saying, Jackson hired a team of forensic lip readers to guesstimate their recorded utterances. Reportedly, “the lip readers were so precise they were even able to determine the dialect and accent of the people speaking.”
“These blokes did not live in a black and white, silent world, and this film is not about the war; it’s about the soldier’s experience fighting the war,” Jackson told the Daily Sentinel in 2018. “I wanted the audience to see, as close as possible, what the soldiers saw, and how they saw it, and heard it.”
That is quite the linguistic feat given that a 2009 study found that most people can only read lips with around 20 percent accuracy and the CDC’s Hearing Loss in Children Parent’s Guide estimates that, “a good speech reader might be able to see only 4 to 5 words in a 12-word sentence.” Similarly, a 2011 study out of the University of Oklahoma saw only around 10 percent accuracy in its test subjects.
“Any individual who achieved a CUNY lip-reading score of 30 percent correct is considered an outlier, giving them a T-score of nearly 80 three times the standard deviation from the mean. A lip-reading recognition accuracy score of 45 percent correct places an individual 5 standard deviations above the mean,” the 2011 study concluded. “These results quantify the inherent difficulty in visual-only sentence recognition.”
For humans, lip reading is a lot like batting in the Major Leagues — consistently get it right even just three times out of ten and you’ll be among the best to ever play the game. For modern machine learning systems, lip reading is more like playing Go — just round after round of beating up on the meatsacks that created and enslaved you — with today’s state-of-the-art systems achieving well over 95 percent sentence-level word accuracy. And as they continue to improve, we could soon see a day where tasks from silent-movie processing and silent dictation in public to biometric identification are handled by AI systems.
Now, one would think that humans would be better at lip reading by now given that we’ve been officially practicing the technique since the days of Spanish Benedictine monk, Pedro Ponce de León, who is credited with pioneering the idea in the early 16th century.
“We usually think of speech as what we hear, but the audible part of speech is only part of it,” Dr. Fabian Campbell-West, CTO of lip reading app developer, Liopa, told Engadget via email. “As we perceive it, a person’s speech can be divided into visual and auditory units. The visual units, called visemes, are seen as lip movements. The audible units, called phonemes, are heard as sound waves.”
“When we’re communicating with each other face-to-face is often preferred because we are sensitive to both visual and auditory information,” he continued. “However, there are approximately three times as many phonemes as visemes. In other words, lip movements alone do not contain as much information as the audible part of speech.”
“Most lipreading actuations, besides the lips and sometimes tongue and teeth, are latent and difficult to disambiguate without context,” then-Oxford University researcher and LipNet developer, Yannis Assael, noted in 2016, citing Fisher’s earlier studies. These homophemes are the secret to Bad Lip Reading’s success.
What’s wild is that Bad Lip Reading will generally work in any spoken language, whether it’s pitch-accent like English or tonal like Vietnamese. “Language does make a difference, especially those with unique sounds that aren’t common in other languages,” Campbell-West said. “Each language has syntax and pronunciation rules that will affect how it is interpreted. Broadly speaking, the methods for understanding are the same.”
“Tonal languages are interesting because they use the same word with different tone (like musical pitch) changes to convey meaning,” he continued. “Intuitively this would present a challenge for lip reading, however research shows that it’s still possible to interpret speech this way. Part of the reason is that changing tone requires physiological changes that can manifest visually. Lip reading is also done over time, so the context of previous visemes, words and phrases can help with understanding.”
“It matters in terms of how good your knowledge of the language is because you’re basically limiting the set of ambiguities that you can search for,” Adrian KC Lee, ScD, Professor and Chair of the Speech and Hearing Sciences Department, Speech and Hearing Sciences at University of Washington, told Engadget. “Say, ‘cold; and ‘hold,’ right? If you just sit in front of a mirror, you can’t really tell the difference. So from a physical point of view, it’s impossible, but if I’m holding something versus talking about the weather, you, by the context, already know.”
In addition to the general context of the larger conversion, much of what people convey when they speak comes across non-verbally. “Communication is usually easier when you can see the person as well as hear them,” Campbell-West said, “but the recent proliferation of video calls has shown us all that it’s not just about seeing the person there’s a lot more nuance. There is a lot more potential for building intelligent automated systems for understanding human communication than what is currently possible.”
While human and machine lip readers have the same general end goal, the aims of their individual processes differ greatly. As a team of researchers from Iran University of Science and Technology argued in 2021, “Over the past years, several methods have been proposed for a person to lip-read, but there is an important difference between these methods and the lip-reading methods suggested in AI. The purpose of the proposed methods for lip-reading by the machine is to convert visual information into words… However, the main purpose of lip-reading by humans is to understand the meaning of speech and not to understand every single word of speech.”
In short, “humans are generally lazy and rely on context because we have a lot of prior knowledge,” Lee explained. And it’s that dissonance in process — the linguistic equivalent of missing a forest for the trees — that presents such a unique challenge to the goal of automating lip reading.
“A major obstacle in the study of lipreading is the lack of a standard and practical database,” said Hao. “The size and quality of the database determine the training effect of this model, and a perfect database will also promote the discovery and solution of more and more complex and difficult problems in lipreading tasks.” Other obstacles can include environmental factors like poor lighting and shifting backgrounds which can confound machine vision systems, as can variances due the speaker’s skin tone, the rotational angle of their head (which shifts the viewed angle of the mouth) and the obscuring presence of wrinkles and beards.
As Assael notes, “Machine lipreading is difficult because it requires extracting spatiotemporal features from the video (since both position and motion are important).” However, as Mingfeng Hao of Xinjiang University explains in 2020’s A Survey on Lip Reading Technology, “action recognition, which belongs to video classification, can be classified through a single image.” So, “while lipreading often needs to extract the features related to the speech content from a single image and analyze the time relationship between the whole sequence of images to infer the content.“ It’s an obstacle that requires both natural language processing and machine vision capabilities to overcome.
Today, speech recognition comes in three flavors, depending on the input source. What we’re talking about today falls under Visual Speech Recognition (VSR) research — that is, using only visual means to understand what is being conveyed. Conversely, there’s Automated Speech Recognition (ASR) which relies entirely on audio, ie “Hey Siri,” and Audio-Visual Automated Speech Recognition (AV-ASR), which incorporates both audio and visual cues into its guesses.
“Research into automatic speech recognition (ASR) is extremely mature and the current state-of the-art is unrecognizable compared to what was possible when the research started,” Campbell-West said. “Visual speech recognition (VSR) is still at the relatively early stages of exploitation and systems will continue to mature.” Liopa’s SRAVI app, which enables hospital patients to communicate regardless of whether they can actively verbalize, relies on the latter methodology. “This can use both modes of information to help overcome the deficiencies of the other,” he said. “In future there will absolutely be systems that use additional cues to support understanding.”
“There are several differences between VSR implementations,” Campbell-West continued. “From a technical perspective the architecture of how the models are built is different … Deep-learning problems can be approached from two different angles. The first is looking for the best possible architecture, the second is using a large amount of data to cover as much variation as possible. Both approaches are important and can be combined.”
In the early days of VSR research, datasets like AVLetters had to be hand-labeled and -categorized, a labor-intensive limitation that severely restricted the amount of data available for training machine learning models. As such, initial research focused first on the absolute basics — alphabet and number-level identification — before eventually advancing to word- and phrase-level identification, with sentence-level being today’s state-of-the-art which seeks to understand human speech in more natural settings and situations.
In recent years, the rise of more advanced deep learning techniques, which train models on essentially the internet at large, along with the massive expansion of social and visual media posted online, have enabled researchers to generate far larger datasets, like the Oxford-BBC Lip Reading Sentences 2 (LRS2), which is based on thousands of spoken lines from various BBC programs. LRS3-TED gleaned 150,000 sentences from various TED programs while the LSVSR (Large-Scale Visual Speech Recognition) database, among the largest currently in existence offers 140,000 hours of audio segments with 2,934,899 speech statements and over 127,000 words.
And it’s not just English: Similar datasets exist for a number of languages such as HIT-AVDB-II, which is based on a set of Chinese poems, or IV2, a French database composed of 300 people saying the same 15 phrases. Similar sets exist too for Russian, Spanish and Czech-language applications.
VSR’s future could wind up looking a lot like ASR’s past, says Campbell-West, “There are many barriers for adoption of VSR, as there were for ASR during its development over the last few decades.” Privacy is a big one, of course. Though the younger generations are less inhibited with documenting their lives on line, Campbell-West said, “people are rightly more aware of privacy now then they were before. People may tolerate a microphone while not tolerating a camera.”
Regardless, Campbell-West remains excited about VSR’s potential future applications, such as high-fidelity automated captioning. “I envisage a real-time subtitling system so you can get live subtitles in your glasses when speaking to someone,” Campbell-West said. “For anyone hard-of-hearing this could be a life-changing application, but even for general use in noisy environments this could be useful.”
“There are circumstances where noise makes ASR very difficult but voice control is advantageous, such as in a car,” he continued. “VSR could help these systems become better and safer for the driver and passengers.”
On the other hand, Lee, whose lab at UW has researched Brain-Computer Interface technologies extensively, sees wearable text displays more as a “stopgap” measure until BCI tech further matures. “We don’t necessarily want to sell BCI to that point where, ‘Okay, we’re gonna do brain-to-brain communication without even talking out loud,’“ Lee said. “In a decade or so, you’ll find biological signals being leveraged in hearing aids, for sure. As little as [the device] seeing where your eyes glance may be able to give it a clue on where to focus listening.”
“I hesitate to really say ‘oh yeah, we’re gonna get brain-controlled hearing aids,” Lee conceded. “I think it is doable, but you know, it will take time.”
This is site is run by Sascha Endlicher, M.A., during ungodly late night hours. Wanna know more about him? Connect via Social Media by jumping to about.me/sascha.endlicher.
