Tech Tuesday: Mac Dictate

Welcome to the first edition of Tech Tuesday.  Tech Tuesday is where we review a piece of technology – gear, software, gadgetry in general – and see how it fits into our lives as photographers.  Being a web designer for years before I got serious about photography, and before that, a *gulp* tech guy, technology is generally somewhere mixed in my thoughts when it comes to photography and how we can more easily do what we do.

So thats tech tuesday in a nutshell.

This week we have Matt Brandon giving us a quick review of a piece of software he uses quite regularly called Mac Dictate. Mac Dictate is produced by Mac Speech and the purpose of the software is to, as they say “…forget about what you are typing, and focus on what you have to say.”  I’ve always been skeptical of speech recognition software, mostly because I find that it turns out to be more hassle than it was worth and usually not free on top of that.  Nonetheless, I’m incredibly impressed by Mac Dictate and equivalently intrigued by how integrated it is into Matt’s workflow.

Matt wanted me to point out that he used Mac Dictate to answer these questions – He made zero corrections and to be honest, he didn’t even proofread it when he sent it to me. Matt’s by no means lazy and I’m by no means that bad of an editor, he just wanted you guys to get a feel for how accurate and inteligent the software actually is and what better way than to put your money where your mouth is (I’m avoiding an obvious pun here).  There are mistakes but for the most part its way more impressive than I would have imagined.  Check out some example videos here and here.

Undoubtedly a piece of technolgy that could prove useful for photographers.  Check out the interview after the break.

Describe this product/technology for us and where does it fit into your life as a photographer?

The product were talking about today is Mac dictate. It’s dictation software produced by MacSpeech. I’ve waited for a good version of dictation software to come around for years. The closest thing was Dragon NaturallySpeaking for the Windows platform.

What inspired you to use Mac Speech? How’d you hear about it.

I’ve been aware of MacSpeech for years. I’ve threatened to get Mac dictate many times, but the reviews never seemed to give me much hope that it would actually work well. But I knew I needed something. Then, when my daughter entered seventh-grade and her writing workload increased I realized we needed help. You see, my daughter, like me is dyslexic. You might not know that dyslexia very rarely shows up as an isolated learning disability.  It is usually paired with other disabilities  like dysgraphia, a learning disability that hampers the ability to write. With her increased workload we knew we had to get some help. So I went ahead and ordered Matt Dictate. It’s amazingly accurate. Though it can be very frustrating at times. I’m dictating this whole tech review using Mac dictate. I will say this, there is no way I could do a complete report without using the keyboard. Part of that is because I don’t fully know all the commands. The program is very powerful. You can tell it to jump from one place to another, you can select words and replace them. The accuracy of the spelling has been a huge help for me. The spelling is pretty contextual. Meaning it knows the difference between the use of the word there are or their. I’m not sure how it does that, but that’s a little concern to me. I’d rather just know it works, and it sure seems to.

How accurate is it, are there necessary “work-around’s” and does it have to learn your voice?

MacSpeech claims it’s 99% accurate. There’s no way that can be true. I would say it’s closer to 80% accurate. But then, to be honest some of the errors might be operator error. Certainly, if I mumble it messes up and spells a word incorrectly or uses a wrong word. If I take my time, and enunciate it’s amazingly accurate. The one downside of this, is it doesn’t understand inflection, pauses and punctuations. Meaning it doesn’t intuitively know when a sentence is done and you need a punctuation like a. Or! See what I mean? It couldn’t tell the difference between the command and the word. So when I write a sentence and I need  insert a comma I have to say the word  “COMMA”.

As far as training it to learn your voice, MacSpeech gives you text to read and train the program to your voice. It’s about 15 minutes worth of reading. But once it’s done, as you can see it’s fairly accurate. Its vocabulary is pretty extensive. But there are certain words it doesn’t know. For instance, I had to train it to spell Lumen Dei, duChemin and other names and places it just would not be in its vocabulary. Some words that never learns to make its pretty frustrating. One such word is my wife’s name, Alou. To be able to use her name in that sentence, I had to change to spell mode. If I just would’ve dictated it would come out, glue, a Lou, Alou. It looks like it’s spelled it right when I put the emphasis A. The one thing they tell you that you need, and I think they’re right is a good microphone with noise canceling abilities. If I drop something on the desk or on the floor, sometimes they will try to spell the sound. Kind of weird, as if it’s spelling an onomatopoeia. The other thing you need to watch out for is someone walking in the room and talking to. It might start spelling their words and sentences in the middle of your document or if you forget it’s on it will start dictating your conversation to them. This has happened to me on many occasion.

How does this help you as a photographer?  Workflow?  Communication?

I never would have thought it would’ve helped as a photographer outside of blogging. But the other day when I was uploading photos to my photo shelter account (you see it doesn’t know the name photo shelter is one word) I realized I needed to add many more keywords and captions to the images. My typing is atrocious and my spelling is worse. So I thought I would give MacSpeech a try. It really speeded things up and for most of the, it was pretty accurate. I was surprised at some of the words it was able to spell and recognize. Like the words, masjid, Benares. But there certainly were plenty words I didn’t know, but I didn’t have time to train and it was easier just to type it in. Words like Varanasi, it wanted to spell “Berenice” or “there are Nazis”. It can’t do contextual spelling if you’re just giving a list of words. So when I needed to spell the word “Sikh”  it’s spelled “seek”. But, nevertheless it was a huge timesaver overall. And it did very well in captioning.

I use it when I need to do long e-mails or interviews. Again, I don’t rely on it totally. I don’t have the patience to learn all the commands. In thing still slip through to you really need to watch her punctuations and still proofread it. For me I tried to get other people to proofread if it’s a long letter were blog posts.

On a scale of 1-10, how cool is this and why?

I’d say it’s an 8 to 9 on the cool factor. I am impressed how well it understands and can translate what I say into words and sentences. I’ve only scratched the surface of its powers and abilities (sounds like I’m talking about a superhero). Apparently it can do global commands like, open documents key commands open and close applications and more. I simply use it to dictate but found it to be well worth the money spent.

By the way, remember I told you we bought it originally for my daughter? Well, it didn’t work for her because apparently her voice is too high-pitched. It seems to like more mature voices than that of a 13-year-old girl. She tries to use it every once in a while, but without much luck. At one point she started lowering her voice like she was imitating a monster or something and it did seem to help. But she felt too silly to continue dictating that way. I can understand.