How one app has kept me using Google Pixel for years
When people talk about locking into an ecosystem, most likely they are referring to AirDrop, iMessage and the seamless integration between iPhones and Macs in the Apple world.
Unbeknownst to a lot of them, other manufacturers also have some killer tricks up their sleeves to make sure their customers remain loyal to the brand.
On the Google side, it is the Google Recorder app that persuaded me to upgrade to another Pixel in 2021.
Initially it was the camera
Pixel has long been synonymous with its industry-leading computational mobile photography since its rebranding from Nexus.
Undoubtedly, the consistently excellent imaging results was one of the main reasons people get into Pixel, apart from the clean user interface and long software support.
I used three different Pixels from 2018 to 2021, including the Pixel 2, 2 XL and 3. They all delivered great experience and I was satisfied with how they performed.
But during same period of time, other major brands also made significant gains in the camera department by introducing larger sensors, covering wider focal range and refining the algorithm to capture better images.
Google must create new ways other than photography to make the Pixel stand out. In true Google fashion, the company focuses on the buzzword of the year - artificial intelligence.
But it is wild when CEO Sundar Pichai repeated “AI” over 100 times during the annual Google I/O event, somehow most people still are not aware of how it makes Google Recorder’s transcription feature even more powerful.
Voice transcribe… not exactly new
The technology which powers speech-to-text transcription has existed for a very long time but the results tend to be rather lackluster.
A Wired article in 2016 detailed how machine is nearly twice as bad as humans in transcribing speech. It is a difficult task as there are a lot of factors affecting the accuracy, from the quality of the microphone, the fluency and accent of the speaker to the existence of background noise.
But Google managed to solve most issues and surprised everyone with the Live Caption demo in Google I/O 2019. The company does not stop there and the tech has continued to improve.
Evolution with Tensor
When Google announced the Pixel 6 and 6 Pro with the new Tensor chip, the first processor designed in the house, engineers defininitely spent all efforts touting new capabilities unlocked by the hardware.
Thanks to years of user feedback, Recorder is now much faster in word recognition and much more accurate in dictation compared to when it was first debuted with the Pixel 4.
It also considers the context of the audio clip to suggest several keywords as title. I also like how I can backup everything and access the transcript on my computer.
As a journalism student, I cannot live without the app during interviews. It is a great tool for me to see what the interviewees say in real time and clean up my notes later on.
Sure, Google is not the only company offering audio transcribe services and there are a number of websites doing the same thing. However, in my testing they are often less accurate, more time-consuming to transcribe the entire clip and sometimes cost a bit of money.
I am not suggesting Recorder is perfect. It struggles to understand heavy regional accents like Liverpool and Yorkshire in the UK, but at least it is a better starting point than anything else and it can understand non-native speakers like me with 90% accuracy.
I also wish it could transcribe other languages with the same high level of accuracy. Probably Google can achieve this in the future when the tech gets more mature.
Still, it is the best life hack which makes my job easier and I cannot see myself spending money on another phone if it lacks the same feature or the same consistency.
I honestly have no idea why Apple and Samsung have not copied the feature yet to make their voice recorders more robust. It should be a no-brainer.
