One of my first recommendations for achieving a natural, believable voice is to employ more free narration rather than always reading a script. I recommended this because all the video tutorials on Lynda.com are narrated at the same time as they are recorded, and the less you read, the more natural your voice sounds.
However, I realize that unscripted narration, even just a few sentences, can be problematic. Eddie VanArsdall commented that using this method often results in a lot of mistakes. He says,
I always scripted my narration and sometimes improvised parts of it, but I could never record it in real time. The pressure of recording and multi-tasking seemed to guarantee that I would make mistakes.
Eddie is right. Unscripted narration can result in a lot of mistakes. Even if you're 100 percent comfortable with the app, even if you have rehearsed exactly what you'll say, even if you're fresh and alert and full of energy, if you're human, you'll make mistakes -- especially when you read a script or outline and narrate and record at the same time.
Making numerous mistakes can be frustrating. And the more frustrated and tense you are, the more mistakes you'll make. But before you smash your microphone on the floor or burst a blood vessel in your forehead, consider this comparison: when you write a help topic or article, do you write it perfect the first time? From the first word to the last, do you type out the entire concept and task details flawlessly? Of course not.
So why should we expect to do the same in speech? Speech is perhaps a trickier, more nuanced medium, since changing the tone of one word can bend the meaning in multiple ways. In addition to unintended inflections, when you string together words in real time, you're bound to have poor constructions, fumbled words, and other errors.
You can re-record the sentences that you fumble. And as you re-record the sentences, you can splice them into your original recording, similar to the way you delete some sentences and add other words in written text. The danger here is ending up with a hodgepodge of different sounds. To avoid the hodgepodge effect, consider the following tips for fixing fumbled sentences.
Set your monitor's resolution to 1024x768px and record the full screen. If you do this, the screen recording will be the same each time. If you need to re-record part of the screen, your recording will match the previous screen's position exactly -- each time.
In contrast, if you just position a capture screen at 1024x768px on a higher resolution monitor, such as 1600x1200px (my monitor's resolution), then you have to guess about where you're positioning the capture rectangle on the screen. If you need to rerecord an area, you may not reposition the capture screen in exactly the right spot. But if you change the actual resolution of your monitor and always record at full screen, you can splice in multiple recordings in a seamless way. Trust me on this one -- it works.
Wait, you say your users can't view screen dimensions that big? Okay, if you need to use 800x600, be my guest. But you'll have to implement so many pan and zoom effects to move around the capture window, it will be maddening to produce the video.
Use dual monitors when you set up to record. Put the application and capture window on the monitor with the low resolution (1024x768px). The monitor with the high resolution should have your script. I also set the high resolution monitor as my primary monitor. All my pop-ups from email and other apps appear on the high resolution monitor.
Having the ability to look at a script on my right monitor while recording the application on my left monitor reduces the number of mistakes I make. If you're trying to read a script written on paper, you'll struggle trying to figure out where to prop the paper up. If you have to look down and then up to your monitor, and then down and up, recording at the same time as you narrate won't really work.
Also, you may want to make last minute changes to your script, and if you're stuck with paper, you'll have to reprint it each time. The dual monitor provides greater ease for fine tuning your script moments before you record.
Maintaining a consistent distance between your mouth and the microphone is critical for achieving a consistent sound. If you have the mic at varying distances when you record, it will be difficult to reproduce the sentences that you fumbled and splice them in seamlessly. You'll find that some recordings sound a little louder, others have more depth, others are softer, and so on. Be consistent with your mic distance and you avoid these problems.
Although I recommend against completely reading a script from start to finish, if you write out a script, it will help prepare you for the "unscripted" delivery. Knowing what you're going to say is key to an articulate delivery. Even if you don't look at your script at all, having written it will put the words in your mind, ready to be recalled at the right time. When I give presentations at conferences, I write out my scripts via blog posts beforehand. I can talk for more than an hour without a carefully detailed list of bullet points to look at. It's the same concept with recording screencasts.
Almost every recording software has a pause and resume key. In Camtasia, it's F9. I often deliver a paragraph or concept, pause, and then rehearse the next part of the script before resuming. Pause and resume works well as a way to reduce mistakes, because you're not trying to remember a five minute script in one go. You piece together little parts to make a whole.
As soon as you finish recording, go back and listen to it. Where you make mistakes, re-record those sentences right then, while your tone and mood are still the same. If you wait until another day, chances are it will be harder to reproduce the exact sound. Your voice's sound is a combination of a lot of variables -- how you feel, what time of day it is, what you ate, what room you're in, the alignment of the stars, and so on. By fixing the fumbled sentences immediately, you increase the likelihood of a seamless patch.
Before you re-record the fumbled sentence, listen to the previous few sentences. Repeat them several times to get in key with the tone and rhythm. When you record the fumbled sentence, chances are your recording will sound much more seamless.
If you make a mistake while recording, restart the entire sentence from the beginning rather than just the word or phrase you made an error with. If you re-record mid-sentence, you'll find that some words are blended together in inseparable ways. For example, if you fumbled the sentence "I want more ice cream," saying instead "I want more axe cream," don't just stop your self and re-say "ice cream." Do you see how the word "more" blends right into the word "ice"? You can't delete a word from your recording very easily when the words aren't separate entities. But sentences usually have breaks between them, so they're much more editable. Also, shifts in tone within the same sentences are more noticeable than shifts in tone between entire sentences.
It's easy to fix fumbled sentences and recordings when you don't have any mouse movement that takes place while you're speaking. While it's not always feasible to leave the mouse stationary as you narrate, you might try to avoid speaking while dragging the mouse around where possible. Fixing those type of recordings can be more difficult, because you can't chop a second off the timeline without having the mouse jump an inch.
Get new posts delivered straight to your inbox.
I'm a technical writer based in the California San Francisco Bay area. Topics I write about on this blog include technical writing, authoring and publishing tools, API documentation, tech comm trends, visual communication, technical writing career advice, information architecture and findability, developer documentation, and more. If you're a professional or aspiring technical writer, be sure to subscribe to email updates using the form above. You can learn more about me here.