I’ll be honest. Before I sat down for my Asimov Hour session this morning, I had no idea what I was going to blog about for today’s process post. I realize that most people who read the blog aren’t also automatically going to watch my YouTube videos, but I felt like I’d said all I needed to say about my current state of things in the three videos I uploaded yesterday, and was a little tapped out with regard to providing updates.
At best, I thought maybe I’d address that kerfuffle from yesterday about the article that made Twitter authors lose their collective minds with the all-capped headline commanding: WRITE EVERY DAY OR STOP NOW. Maybe I’ll touch on that later since my perspective isn’t quite the apparent norm.
But that is not what I will be talking about today.
In my Asimov session this morning, as I’ve been doing for the last week-or-so, I continued working on transcribing the road-drafting sessions I recorded on my way to Pittsburgh on May 18. Now I realize I had quite a bit of trouble getting things going with Dragon in the first place, so I’ve been trying to ignore my misgivings about this method and give the software the benefit of the doubt.
Just to recap:
- I had trouble getting my recording equipment set up. Multiple purchases/returns via Amazon, multiple attempts at recording sessions that went to waste and were trashed because of poor audio, and much time wasted that I’d hoped would go toward ‘getting the hang of dictation’ but was mostly spent researching what sound equipment other dictating authors used because what the *@#!ity-@%#! was I doing wrong??
- Either from stress or seasonal allergies or overuse in all these practice sessions, I started to lose my voice before I even left for Pittsburgh!
- I finally dictated, limiting myself to about an hour to just see how it would go, then had the disappointment of seeing Dragon return total gobbledygook when I tried to transcribe those files in my hotel room.
- Extra editing time for each file was necessary because of road noise that the lavaliere mic I was using did not eliminate.
So finally, I think it was Tuesday after I wrote my weekend recap blog post, I found good instructions to clean up the road noise, and finally got somewhere with Dragon’s transcriptions. Not perfect, but far closer than I’d experienced thus far.
Encouraged, I spent all week working on those transcriptions.
Let me repeat that. I spent ALL WEEK working on TRANSCRIPTIONS of about an hour’s worth of recordings.
The reality of that didn’t strike me until this morning. I reached the end of my 9 files for transcription, and faced the possibility of walking and dictating some more (I don’t know when, it’s done nothing but rain for the last week and looks to continue that trend), and then having to return to the practice of transcribing my dictated files in future morning Asimov Hours.
So I did a little math.
I’d already been calculating the WPH from the recorded files. The output of words versus time came out to between 3700 and 4250 words per hour (calculated; the longest recording was 28 minutes). That seemed rather promising.
Except for how much time was, in reality, going into the production and transcription of those words.
Let me give you an example. And, say for the sake of simplicity, I didn’t have to deal with the anguish of the initial road-noise-impacted transcription.
That 28-minute long dictated recording.
Well, it took 28:47. That’s fairly straightforward. Hard to imagine recording in double-time or anything. So for the 1690 words that created, my pace was approximately 3496 WPH. Pretty good. Not amazing–I could easily do that hand-typing–but pretty good.
Then they took about 3 minutes to clean up in Audacity (road noise, noise levels, etc.). Not too big of an impact, you’d think. 3168 WPH, though, a noticeable drop.
Then there was the time Dragon took to run the transcription. It wasn’t fast. I don’t have the exact data on this, so it’ll have to get lumped into the sessions from that week where I also worked to correct the transcription errors. What I did, finally, was play back the audio on my Mac, following along in the DragonPad document so I could highlight sections and tell Dragon to correct them. 75% of the time I needed to hand-enter the correction because Dragon’s suggestions were way off. To transcribe this 28:47 file took me two Asimov Hours. Not the full sessions, as I also did shorter files on either end, but a fair estimate would be 35 minutes each for two days. Add 70 minutes to the time that these 1690 words took to create.
And that leaves us at 994 WPH.
If I were struggling to write one day, not free-flowing text from my fingertips like unicorn snot, I would probably get between 750 and 1000 WPH.
Which means, after the process described above, Dragon Dictation/Transcription got me no further than a BAD writing day.
Doing even rougher math than that above, I got 6400 words out of the entire experience (9 separate recordings of varying lengths), and it took me the original hour and fifteen minutes (rough estimate), plus a week’s worth of Asimov Hours (9 days, actually, let’s pretend an average of an hour each). Net WPH? 624. That’s worse than a bad day, and it cost 10 writing sessions. Even when I have a writing day, it doesn’t last more than one day. Most times I perk up before my writing session is over and make up the difference.
I do not consider the text that I created through dictation to be a usable draft. It’s a hot mess. It makes no sense. It rambles, it shambles, it walks into solid walls and bounces against them over and over, trying to figure out why it can’t move forward. Yes, a draft is meant to be shit so you can fix said shit later, but this was not writing. This was awkward fifth-grade mixer slow dancing, and there was much trodding of feet, and spinning in slow rocking circles until dizzy.
It should come as no surprise, then, to learn that I am going to stop bothering with Dragon for a while. I may return to it someday, most specifically if I am hobbled in some catastrophic way that removes my ability to type by hand. In such a desperate hour, I know I could make it work. But the effort that’s gone into it so far, the numerous frustrating audio equipment returns, lost days, and awkward drafting, has not provided the ROI I’d hoped. I have even developed wrist pain from struggling with the touch pad on the Dell laptop I was using. All of this earns Dragon a hard nope from me.
For now. Can’t return the software, so I’m not counting it out entirely. I’ll hang onto the recording equipment I bought. I might use it to dictate things later on, directly into the program, when the rest of the household has gone out and won’t be bothered by it. Might record things if this infernal rain ever stops and taking a walk is actually possible again. Recording blog posts worked pretty well, anyway.
But for right now? For a standard Asimov Hour? It’ll be finners on keys again, to get an actual draft, worth using, and an ROI in terms of WPH that won’t make me feel FUBAR.