Google Pixel 2 Announcement: October 4th Event Roundup
Hardware events are as a huge deal. Ever since the Pixel’s original release last year, Google take them just as seriously as their traditional software events such as I/O.
Since then, Google can now definitely be counted among serious hardware manufacturers. There are a whole host of devices bearing their name and brand lines. Pixel phones, Chromebooks, Chromecast, Google Home and other equipment… A whole Google ecosystem now exists to talk about and play with.
So lets take a look at everything Google had to say yesterday.
Google are of course, at their heart, a software a company. They’ve made their billions through search and advertising. But this is a hardware event. So why lead with software?
Well Google think that it is their software improvements which will drive home better experiences when using their hardware. This does make sense.
Like other big tech giants, Google are focusing heavily on AI-first computing. This is poised as the natural progression from mobile-first software design that has been prevalent for a while.
According to Google, the aim for AI-first computing should be to create a tailored computing experience that is:
- Able to Learn & Adapt
As such they want to focus less on raw specifications. The new ideal – to them at least – should be an intersection of AI, software and hardware.
According to Google it was Machine Learning that was the core of the first Pixel’s success in photography. They used it to create the best experience, rather than relying on promoting the best hardware.
Google’s image search / recognition techniques have apparently seen some incredible improvements over the past year. Their new AutomML (Automatic Machine Learning) algorithms apparently perform better than models designed by human data modelling experts.
Object detection is now powerful enough to recognise distinct items from similar groups. For instance Google can now accurately identify scooters, bicycles, mopeds and motorbikes from a parking rank containing all of them. Even if the picture is taken at an angle that obscures many subjects behind one another!
Google Translate is also now powerful and fast enough to enable real-time conversation between individuals. We’re getting close to a sci-fi future with some of these tools!
It is through these software tools that Google are looking to augment and improve their hardware. So when you get hold of a Pixel 2, or anything else with a Made by Google tag, the resulting experience should literally be more than the some of the component parts. This is a theme Google came back to regularly throughout the hardware announcements.
Google say that first Pixel managed to obtain 0.7% of market share. This, they admit, is a figure which could have been much higher had they manufactured and distributed more effectively. This is a challenge they’re willing to tackle this cycle.
The designs are very similar to the original. A glass shade on the back creates a two-tone effect. The body is all aluminium with a ‘hybrid coating’, for a tactile finish.
A fingerprint sensor is rear-mounted in the centre and touted as the fastest and most accurate on the market. It’s the software and hardware that enable this claim you see…
A circular polariser is integrated in the screen too. Apparently this means you should be able to view the screen properly when wearing polarised sunglasses.
There are front-facing stereo speakers but no headphone jack. This does mean that the #donglelife comes to Google. Make of that what you will. Bluetooth 5.0 is supported as part of the Android 8.0 Oreo OS.
Google have also committed to providing the same features on both phones. Essentially there are no differences between the base or XL models, aside from screen size & resolution/ratio. A subtle dig at other manufacturers there.
Following Google’s purchase of the HTC hardware group, one of many hardware rumours was also confirmed. Pixel 2 borrows the HTC U11 squeeze gimmick. Here it is used as a method to launch Google Assistant.
Getting to the Assistant quickly and intuitively is clearly very important to Google’s envisioned future. It is through the Assistant and the wider Google software/hardware ecosystem, that they see future success and customer retainment.
Cameras and Software
If the original Pixel shone in any one specific feature, it was the camera. Many outlets described it as the best mobile photography experience available in the last generation. Not willing to rest on their laurels, Google have iterated hard.
The new system still uses a single 12MP lens, as opposed to the dual systems on much of the competition. Google however insist that a dual set up is not required, due to their software enhancements.
A ‘dual-sensor’ system makes up for this. In a move very similar to traditional HDR photography, the Pixel 2 camera actually takes multiple images for every photo and combines them. This allows you to create the post-production effects offered in dual-lens devices.
OIS has been added to the Pixel 2. This complements the much-lauded software stabilisation of the first phone. As such, the Pixel 2 should provide the most accurate images available on mobile, even when shooting video.
Software also comes into play here massively. Drawing on machine-learning data from their Photos projects, the Pixel’s image processing is unique when compared to other leading photography brands.
Another benefit of using software for most of the processing? The front camera has all of the same Portrait mode and processing tricks as the rear.
Finally, Pixel 2 owners will also receive full, unlimited storage Google Photos, even if uploading 4K videos. You’ll also benefit from Photos’ automatic tagging and organisations if you provide the permissions.
Lens is integrated into the camera and can automatically recognise email addresses, phone numbers, street addresses from documents, flyers, posters etc. Rather than typing down this information or taking the picture, Lens will highlight the relevant text and offer to either store it or provide contextual actions such as sending an email or starting a Maps journey.
Lens will come to all high-end Androids eventually, debuting as a Pixel 2 exclusive feature.
Lens is also mixed with AR. Branded AR content from League of Legends, LEGO, Stranger Things, Star Wars etc. will all be available to play with. This works by dropping a ‘sticker’ into your livestream. Lens then transforms this into a 3D model that acts and reacts to other stickers, real-life lighting and obstacles. All in all it’s a good bit of fun and an example of how AR is progressing at a consumer level.
Other software features include contextual information displayed on the always-on screen. For instance the Pixel 2 can listen to ambient music, work out the artist/track details and display this to you automatically.
The new Google Home Broadcast feature (detailed below) is also available. When you’re away from home, just squeeze the phone to launch Assistant, then Broadcast a message. This will play back at home through any connected Home items you may have.
The smart home speaker is a burgeoning market and one that Google have strong competition in from Amazon. The existing Home products work as advertised, although it’s difficult to imaginge they have done particularly well in the face of Amazon’s relentless global advertising campaigns.
The thing is, Amazon Echo seems to exist primarily so Amazon can push you into buying more stuff from them. Google however want top be a bit more subtle in their approach.
At the end of the day they’re still listening to everything you say and building a profile on you. It could be argued though that Google is less interested in the fine details of what or where you actually do, go or spend your money. Where Amazon want to know what you like so they can sell it to you, Google want to point you in the right direction (and skim a little commission off the top).
So if you buy into the current Smart Home world and envision a Sci-Fi utopia/dystopia where we spend our days talking to computers, why should you be interested in the new Home products?
This is the main area where adoption is stumbling. Google say that you don’t have to ‘talk like a computer’ in order to use the new home. That’s a dig at Alexa, and the formulaic way one has to set up Skills and commands.
Of course you can’t let speak to Home in the same way Picard talks to the Enterprise Computer. That though is the goal and the new Home is a step closer to it.
Returning to the Conversational, Ambient, Contextual and Learn & Adapt approach mentioned earlier, new Home is now more able to deal with less structured commands.
Say you’ve just searched for the nearest coffee shop and got the response of “Hot Joe’s, 500 metres on your right” as a response. Now you could simply say “Plan a route”, and Home should guide you there. Until now, you might have to explicitly ask to be routed to that store i.e. “Plan a route to Hot Joe’s”. It’s a subtle difference, but creating a system that understand contextual cues is a major barrier to developing a subtle and nuanced conversational AI.
According to Google, they have used over 50 million unique voice samples to train the upgraded voice recognition. This should better help with accents and regional dialects.
A concern for some Smart Home appliances is how they can recognise individual requests from different people in the house. Voice Match is the name for the updated feature handling this.
Google Home can recognise several individual voices, matching these to separate profiles. Just like how signing in to Chrome can provide more tailored search results, different members of a household may get slightly different responses to queries.
It’s not just searches either. Let’s say you have a family Spotify account with several active profiles. Each profile has its own playlists and suggestions. This information can be tied up to the Google Home Voice Match, so asking for an open request – “Play some ambient music” – results in a more tailored response. It would also differentiate lists. So if both father and daughter have a playlist called “Rock”, Home would figure out the correct one to play from the Voice Match.
Google have also announced that hands-free calling through Home will be coming to UK devices soon. This also works with Voice Match. Just ask to call someone from your contacts and your connected Android phone will do the work.
Smart Home integration
Assistant now works with over 1000 products from 100+ brands. A big point was made yesterday that Nest products are fully integrated. This includes the thermostats, door bells with cameras, security system etc.
Smart Home routines have been extended. It requires setting up in the backend, but IFTTT style connected actions can be triggered through simple queries.
For instance, just telling Home that “it’s time to go to bed” could dim/switch off lights, change the thermostat timers, arm an alarm and just generally cycle through pre-made routines for any connected items in the house.
A new Broadcast command feature has been developed. This can be used to send a message to be played out through all synchronised Home devices in the house.
There is also a whole host of Family Link features. These are primarily aimed at children’s entertainment: both educational and play. It was mentioned on stage how the rise of screen-based entertainment has changed children’s early development in an incredibly short space of time. Google have an ideal where future games, puzzles, quizzes etc. will be more conversational and participatory, using Home as a centre.
There is a partnership with Disney which includes tons of branded content or several age groups, including movie franchises such as Marvel and Star Wars, popular Disney Channel TV shows and their more traditional characters.
Two new Home devices were announced:
Google Home Mini
This looks to compete directly with the Amazon Echo Dot. It is a very simple ‘hockey-puck’ design with a fabric top, apparently designed to sit in a home scenario where fabric is normal. The fabric top lets through some light so that you can see LED notifications.
Google Home Max
A bigger version that goes all out for sound quality. Including 4.5-inch woofers and 0.7″ tweeters, Max is 20x more powerful than the original Home. A magnetic foot allows the system to be sat in either horizontal or vertical orientation.
Machine Learning will adapt the sound output based on location. So if you put Max on a shelf or in the open in the middle of the room, it can analyse its own output and adapt accordingly.
As mentioned above Spotify is fully integrated, as well as other major streamers including Google’s Youtube Music/Red (with a free 12-month subscription to Red), Deezer etc.
And the rest
A few more bits of hardware were announced, namely the Pixelbook and Pen, Pixel Bud earphones and an updated Daydream VR headset.
Pixelbook and Pen
High end Chromebooks are an interesting proposition. The Chrome OS, while simple and very easy to use, still doesn’t offer the full desktop experience many are used to from a Windows or Mac based machine.
However Chrome OS is very popular in educational settings, due to the low cost of lower-end systems and ease of integration.
It is here that Google believe they can continue to succeed with a high end Chrome OS machine. As students progress from school, to college and university, they may opt for a Pixelbook or other high-end Chrome OS device, as they already integrated into the ecosystem.
Now that argument fails as soon as someone enters programming, graphic design or other fields that almost demand the use of Windows or Mac. Still, for a student who only needs to do research and use traditional office-based applications, Chrome OS is fine.
So the Pixelbook is a very stylish and very fast Chromebook. The aluminium frame results in the whole machine being under 10mm thick and less than a kilogram in weight. It also follows the hybrid approach of other convertible laptops, allowing you to fold the screen around fully to use in a tablet mode.
Memory and processor configuration options are available, as is an additional Wacom developed stylus with excellent latency and precision.
Overall it’s a very stylish bit of kit, although those that need something for serious computing ends will never be satisfied with Chrome OS’ restrictions.