If you’ve been reading this blog for a while you’ll know that I’m a huge fan of dictation. Even though I can type quite quickly I much preferred to write using dictation. It’s much faster and more accurate because you get fewer typos and misspellings. Lately I’ve been terribly disappointed that Nuance had decided to quit developing Dragon dictation for Mac. I’m sure this is a huge mistake, but I don’t think the company really cares. The Windows version of DragonDictate was always a couple of steps ahead than the Mac version. What they expect you to do is to buy software to run a virtual machine such as Parallels and to run Windows and the Windows version within that. It’s going to be a huge expense to put all of that in place. I really don’t like the idea of running Windows on my Mac. Even if when you are running Parallel’s in a compatibility mode it can be fairly seamless. So I was delighted to see in the WWDC Keynote 2019 mention of accurate dictation in the next version of the Mac operating system, Catalina.
Why is it in accessibility?
Because they put this dictation feature into accessibility rather than as a feature for all users I have to wonder what’s going on. It could just be a matter of emphasis. Apple likes to look after people who need accessibility features and should be commended on doing a fine job in this area. Putting the feature into accessibility doesn’t mean it can only be used by people who have a disability need. You’ll be able to turn it on and use it just because you like to dictate. Does Apple think the general populace are not interested in dictation? Has Apple created a professional dictation software which is really good but has hidden it away? I very much doubt Apple would create substandard software and put into accessibility because it wasn’t good enough for everybody to use.
How good will Siri speech recognition be?
According to the website for the updates to the operating system OS Catalina, we getting “Accurate Dictation”. The dictation feature will give us the latest advances in machine learning for audio to text transcription. Apple also said we can add custom words. This is something I like using within Dragon Dictate. There are one or two words which come up from time to time that are not recognised by the software. This could be names of things or places. If you’re a writer you might have made up a name for a character or want to use a strange spelling. It’s all part of the creative process. Trying to read into the proposed dictation system “Voice Control” improves on the existing “Enhanced Dictation”. Seems like there will be three levels of dictation.
- Standard Siri Dictation
- Enhanced Dictation
- Voice Control
Apple tells us the audio processing for Voice control happens on the device. I think that is the case for the Enhanced Dictation we already have. We have offline use and continuous dictation and you just have to turn it on in Keyboard settings. Use a shortcut to enable it.
Rich Text Editing
One of things which seems most promising with this new version of Siri dictation is the rich text editing. The webpage for Catalina tells us we can replace a phrase by saying “replace I’m almost there with I just arrived.” This looks pretty good because the software will find the words – “I’m almost there” and do the replacement. We’ll be using natural language to edit our dictated text. No need to use specific computer code words to give a command. I expect it will also be possible to tell the dictation software to select or delete a specific word. When a word is selected we’ll be able to say the word we want to use instead and it will be replaced. I’m looking forward to find out what other commands are going to be available. OS Catalina will give us both word and emoji suggestions. We are promised a new interface where we can just ask to correct a word and be presented with a list of suggested replacements. To select the word we want instead you say the number corresponding to the word in the list. Some of this I’m guessing at based on the description of the Voice Control feature on the Apple Catalina OS. It will need a proper testing. I’m hoping it will be good enough to say goodbye to Dragon Dictate for Mac.
In command of your computer with your voice
When I’m using DragonDictate in the messages app I can dictate the message and when I say the word “send” the dictated message is sent to the recipient. Catalina is promising it will have seamless transitions from dictation to commands. Other commands will allow us to open applications, search the web, use Spotlight. This is something we can already do in DragonDictate. I do sometimes use the Dragon commands to open in an application with my voice rather than resorting to the keyboard. There are other voice commands specific to people who can’t use a mouse or keyboard. These make the whole of an app clickable using numbers to make choices of sections of the screen in a grid, menus and buttons.
Research on YouTube
Since writing the above section I went into YouTube to look for videos showing Apple Voice Control. The changes to the software are only available within the developer betas. I don’t have access to those and I wouldn’t put those onto my main computers either. Apple have declared these early betas as the Wild West of operating system updates. Most definitely not something you should put onto the computer or iOS device you use on a daily basis. I’ll be waiting until the public betas are available for iOS. It’s also unlikely I’ll put a beta version of the Catalina operating system onto my Mac. Even if I am insanely curious it’s not worth messing up my iMac to have a quick look. The only way it would be possible would be to boot to an external hard drive with the new operating system. That way I could immediately go back to what I have now. I did find some interesting videos showing how Apple Voice Control works and it looks pretty good. Here is a video from someone called Soldier Knows Best on YouTube. I particularly liked the way he was able to edit the text by telling Apple voice control to change words which the dictation engine had got wrong.
I don’t think the voice recognition is as good as with Dragon dictation as demonstrated in this very short clip. I’d be interested to know if the system can learn from the user. Can the artificial intelligence learn which words I use and how I spell them and improve over time. I’m certainly looking forward to giving this voice to speech conversion software a proper try in the autumn.