Let’s start with the most obvious question – What is PropTech? The name PropTech is an amalgamation of the words “property” and “technology”, and is also commonly referred to […]
The direct reason why voice technology has been improving through the last 10 years is the high demand for comfort, efficiency, and increased speed of processing. Users’ expectations of digital products have changed. Today, the ability to perform multiple operations at once is what creates the hierarchy in the device category. Most of us just want to accomplish things faster and voice recognition development makes it almost possible. Almost.
We’ve experienced an explosion of voice technology. It’s incorporated into our smartphones but also into our home devices. There are tasks that we can accomplish just by saying commands to our phones and gadgets. It’s really awesome! However, we still need to learn how to overcome a set of challenges that make the overall usage less convenient. Before we discuss them, let’s quickly remind ourselves of the moment when it all started.
Today words such as text-to-speech, speech recognition, machine learning, and adaptive VUIs are commonly known. But where did it all begin?
2011 – The launch of Apple’s Siri – the first voice-enabled assistant on a smartphone! ?
2014 – Alexa enters the market! Amazon introduces its player in the voice recognition category ?
2016 – Google Assistant meets the world! The company celebrates its first voice-enabled device ?
The rest of the events escalated pretty quickly. Only in 2018, there were 9,5mln people in the UK using smart speakers. The demand for voice recognition software was rapidly increasing and the competitors to this day have much work to do to meet the final users’ expectations. Although the last years were spent on minimizing errors in voice recognition development, the main problems are yet to be solved.
Here are some challenges we still have to face when implementing any voice features.
Many have already noticed how inaccurate speech-to-text technology can be. Voice recognition software won’t always put your words exactly how you said on the screen. It’s usually something close to it but since software can’t understand the context of language, it often leads to misinterpretation.
Voice recognition still has problems with slang, technicalities, acronyms, and distinguishing words that sound familiar but mean something completely different.
The famous “duck” instead of “f***” ??
Same as with understanding the context, voice recognition software can have problems with different accents and local dialects.
Even though some devices may learn to decode your speech, it takes time. You’ll have to talk slowly and clearly to minimize the errors and help the software learn how your pronunciation sounds. What’s more, the sound and quality of your voice may also affect the speed and accuracy of processing. Sometimes a throat problem is enough to make the technology work less efficient.
For sure, we aren’t ready to make the voice technology compatible with all languages yet.
If we have even minor problems with the most popular languages, think about how far away we must be with the quality in the case of national minorities. Probably there’s a long way before the technology will be available to all no matter the language.
Reach and loud environment is unfortunately still a problem. Using a voice-enabled device in a quiet place is considered the best practice to get the best out of it. Having a lot of background noise, the system can’t distinguish your speech and gets crazy. One that may be helpful here is wearing close-talking microphones or noise-canceling headsets.
You have to actually spend some more time to review and edit a message to correct. errors. Also, you have to learn how to properly use it. Sometimes when we talk too fast we only increase spelling and grammar errors but with the right tone and pace, we can be more effective. Oh, and remember that you have to say the punctuation out loud:
[…] question mark [..] dot […] exclamation mark.
It can be annoying sometimes and for sure affects the flow of our conversation.
Voice recognition is not a technology that will stop right where it is. As long as we see that people use it and enjoy the comfort (even partially) of accomplishing things just by using voice commands – we will work on the devices to be only more effective. After all, what matters is the quality of user experience.
Recently we finished a great project for a Swedish company. Check how we powered an eLearning platform with voice recognition technology!