This is just a PSA for those of you who–for whatever reason–may prefer dictation to typing, but have run into technical/workflow issues. Perhaps you’d like to dictate anywhere, on the go? And yet have access to all the features of your desktop computer? It’s quite possible. At any rate the following setup works for me.
Everybody’s setup is likely to be different to some degree, which makes a “how to” post a bit challenging. But OTOH there are some basic ideas that I suspect will turn out to be useful.
First off, though I strongly dislike the company, I’m going to say this: Nuance’s Dragon 15 software is the only choice I would even consider for transcription. And I mean the desktop version here–not the Dragon Anywhere app for your phone. Even more specifically, to my mind the “Professional” version is strongly preferred, as it allows for transcription of recordings. (Though of course it is possible to record text using whatever device you like, then play it using an aux cable, ideally plugged into a USB soundcard, thus achieving a very slow but workable form of delayed transcription. And saving maybe $150, too.)
You know…this could very quickly become a book. I don’t want to write that book. So if you’d like a solid introduction to Dragon in general, I recommend Scott Baker’s website, videos, and books.
So we’ll assume basic familiarity. We’ll also assume a willingness to experiment, and either a lack of fear or a strong interest in fiddling with technical details. This won’t be for the faint-hearted, but there is a pot of gold at the end of the rainbow–probably.
Let’s consider three scenarios:
- You’re sitting at your desktop (or laptop) computer and dictating into a microphone.
- You’re walking around, dictating into a recording device, planning to have software transcribe the recording later.
- You’re walking around with your phone, and you think you might like to dictate some text–but you’re not quite sure of where you were in a document, or you don’t have a lot of time and simply making a recording–which you later have to transcribe/correct–seems like more trouble than it’s worth.
This post is mostly about #3. Nuance has their Dragon Anywhere software available for both Android and iOS, which currently costs $15/month. Plus tax, and whatever. So, like, just about $200/year? You could use that. For quick notes, assuming you don’t mind the–admittedly not huge–expense, it’s pretty good.
But there are issues with DA:
- Do you pause while you record? After a non-configurable 20 seconds, DA stops listening. You can either watch the screen for that, or set the app to beep when it happens, but either way I find that extremely distracting.
- Although DA will synchronize vocabulary with your desktop version of Dragon (assuming you have that too), it does not synchronize other things–such as rules for punctuation and capitalization. For me at least, this turns into an editing hassle I’d rather do without. Who the hell decided, for instance, that an ellipsis ends a sentence? That it should be followed, always, by a space and a capital letter? Also, the vocabulary synchronization is imperfect–it includes the text form of your vocabulary, but not the actual voice training you may have performed. In other words, you end up having to redo all of that training. So what’s the synchronization for?
- Editing files in place doesn’t really work–in other words, if I have a file in Dropbox and I’d like to work with it, the best approach I’ve found is to copy/paste the contents of that file into DA, dictate there, and then copy/paste back into whatever app (I like JotterPad, because I use “markdown” to preserve things like italicization and headers in plain-text files) I use to work directly with the file in Dropbox.
- A lot of the desktop version’s editing/correction functionality is simply missing. You can do a bit of it with DA, but I find it slows me down drastically to work within the app.
So what do I do instead? I use my phone–usually with a headset plugged in–and work directly with the desktop version. Here’s how:
- I set up a VPN that allows me to connect to my home network from elsewhere on the internet.
- I have an RDP server set up that allows me to interact with Windows from a remote location–it mirrors the screen to my phone, at the proper resolution.
- I use an app (WO Mic) to send my voice from my local microphone, over the network, to my home computer.
This stuff only works if you get all three of the above steps right. And there are plenty of things that could go wrong.
For instance:
- With that VPN, aside from security concerns, you may need to either install new software on your router, buy a pre-configured router, or set up port forwarding on that router and then set up a separate VPN server in your home network. For the record, I use free open-source software called dd-wrt on my router, but set up a separate VPN server on a computer at home. I use OpenVPN, and I actually push (only!) a route to my local network’s range of IP addresses out to the VPN, which means my phone here–thus allowing all other traffic from my phone to go out over the internet normally. I could route everything to my home network, then back out, and in some scenarios I might want to do that, but most of the time I’d rather not. Also, to connect to that VPN, you’ll probably want to use something called “Dynamic DNS” or ddns so you can easily find and connect to your home network. Look around a bit, if you’re doing this–there are services for this that are actually free. Why pay if you don’t have to, right? And I do have dd-wrt handling the ddns bit.
- With that RDP server…yes, this functionality is built into Windows. However, Dragon acts badly with the Windows RDP server. You could try it if you like, but I don’t recommend doing it at all. I get around this–and many other related issues–by running Windows in a virtual machine using software called VirtualBox. VirtualBox has the capability to run an RDP server that Dragon doesn’t detect. Thus, Dragon acts normally, damnit. Note: VirtualBox 6 currently has a weird bug that may keep you from using a microphone plugged into your host computer while the RDP server is enabled. Solution: turn the RDP server off when you’re sitting at your desktop. And, a minor note for Android users: I like aRDP Pro as an RDP client. Among other nice features, it echoes sound from my computer to my phone. Handy for listening to playback of my speech and correcting Dragon’s transcription errors. (If the error was mine or because of a network connectivity issue, I’d rather not “train” the software via correction–again, Scott Baker’s website and books are a great intro to Dragon!)
- So, why the WO Mic Android app? Because VirtualBox’s RDP server does not allow for forwarding sound from your remote microphone. Which is a good thing anyway–Dragon is particular about which sound card (aka “Dictation Source” in the Dragon software) it’s using. WO Mic has to be installed under Windows too, and sets up a virtual sound card for you. Unfortunately, this app is really only designed to work semi-locally, over Bluetooth, USB, or Wi-Fi. So the workaround is to select its “Wi-Fi Direct” option, which is the app’s term–why do people make up terms for software users to figure out and translate?–for a phone-generated Wi-Fi hotspot. But connect to the VPN rather than activate an actual hotspot. A further note: Windows did not behave well when I attempted to push a route to it from dd-wrt…in other words, it could not “find” anything connected to the VPN. I had to use a “route add” command within Windows to make that work. Now…is there an equivalent app for iOS? Maybe. Couldn’t tell you. Probably? I’d like to think so, anyway.
You know, looking at this post…I could include a ton of links. But, frankly? If you need those, this post likely won’t help you anyway. There ought to be enough here to help you Google (though I prefer StartPage) your way to solutions for your particular computing environment.
It might actually be fun to write a book about using Dragon, if I ever run out of other projects. I have plenty to say about both transcription and actual-desktop use of the software. But I do have other projects, so screw it.
Incidentally? I dictated this entire post with my phone via a VPN and RDP, while pacing at my daughter’s soccer practice. And looking like a weirdo, I’m sure.
Oh yeah, one more thing. When I’m at home, but wandering around the back yard? Or pacing around the house because nobody else is home? I can use the same setup as above, but over Wi-Fi. I get better performance than over the internet–makes sense, right? The only change I make is that I don’t connect to the VPN.
Er…I can also do the same thing in my car, using a hotspot from the phone, if I take my laptop along with me–yeah, same Dragon profile I use on the desktop, stored in Dropbox to synchronize it–if I’m not sure I’ll have decent internet connectivity from my phone. Like on a road trip. Um, should I really be dictating text, which I need to stare at to edit, while driving? Maybe not. But, you know, it’s an option. Especially if you don’t mind wrecking your car or killing dogs/deer/whatever.
I hope this helps at least a few of you. And: have fun out there!
Quick note: later versions of Android do not support voice recording via a background app. Which sounds great and all, but whose phone is it? (As if we didn’t already know.) So the same phone will no longer display the Windows screen and send my voice to it via the internet. So my workaround is to have two phones. (One of which is actually quite old, but never mind.)
The first phone, the one with the actual Internet connection, is running the RDP client and creating a hotspot. The second phone connects to the hotspot, then also to the VPN I host at home, and runs WO Mic in the foreground. Possibly with a headset plugged into it, but I’ve discovered that phone microphones are actually pretty good for dictation. And it’s way easier to carry two phones in a pocket than to deal with transporting/using a headset microphone.
So far this isn’t working great for fiction. But that’s not a technical issue. It’s a problem in my brain. For other stuff, it’s pretty easy…once it’s all set up.