Today I received the following question about creating video tutorials:
I've been asked to create a set of video tutorials for work. However, it has been difficult since I keep having to start over and re-record things whenever I make a mistake.
In my search to improve the process, I came across your article "How I create Video Tutorials." I understand that it has been a while since you wrote the post, but I was hoping you could give me some help based on your experience.
In the comments section, you mentioned that you were trying out recording the video first then adding the audio later.
Does this mean that you tape the whole audio portion in one whole take and just replace the video's audio with it? Or is it possible to record pieces of audio and then like 'copy/paste' them onto the specific portions in the video? In this way you can have several takes and then splice them together.
I'm lost here, I hope you can help me. I don't really have a budget here so I hope it is possible to use free or open source tools. So far I'm using audacity and camstudio, and a plug in mic.
Thanks in advance.
I was actually writing a post about the very topic of creating video tutorials when I received your question. The process you follow in creating a video tutorial depends on your audience and purpose. For simple how-to videos that you send to colleagues or friends, I recommend using Jing. For corporate videos that require a higher standard and more scripted delivery, I use Camtasia for Mac and follow these five steps.
Pick a short task or concept that you can describe in about 100 to 300 words. You usually speak about 100 words a minute, and videos are ideally three minutes or less (sometimes up to five minutes long, but really no longer). Write the script as conversationally as possible, and read it out loud several times as you walk through each of the steps.
I keep my scripts in a Word document and highlight in green the places in the script requiring action. Ideally, make sure you have two monitors or one large monitor so that when it comes time to record, you can open your script on one monitor while recording on another.
By simulation, I mean recording the screen actions. You need fake data and a stable test environment to set up your simulations. If you don't have fake data, you can cover it up or photoshop it out later, but it will be tedious to do so. The fake data setup can be difficult, so this step may require quite a bit of preparation.
Run through the tasks and set up any kind of information, events, or other sample data you want to show. Make sure the simulation works and is something you can demonstrate in a simple way. You may want to practice it several times to make sure your script matches all the actions you will take.
Resize the viewport on your browser to 1280 x 720. With Chrome, you can add an extension called Window Resizer to resize your browser. Open your recording software and drag the recording window to your viewport (the 1280x720 pixels size) – these dimensions will allow your video to display as HD. If your dimensions are less than this, you won't have enough pixel data to render your video in HD. This may not matter to you, though, so if you decide to record with smaller dimensions, I recommend recording at the same dimensions that you plan to publish the video. Resizing it smaller usually results in a fuzzy video.
Put your script on one monitor and record the simulation on another monitor (or another computer, so you can use two mice -- one to scroll your script, and one for the simulation).
You may find yourself deviating slightly from the script as you try to be conversational and natural – that's fine. The script usually adjusts a bit anyway as you fit it to the screen actions and the natural voice of conversation. It's important to keep your script fluid.
As you record the simulation, don't worry about your voice. Instead, focus on getting clean, smooth screen actions. Drag your mouse normally (using a regular mouse, not the jerky strides of a touchpad), and try not to move the mouse around while you talk (it makes post-production editing much harder if you do). Never wiggle your mouse to emphasize a point.
If you stutter or mispronounce a word, don't worry about it. Just make sure the timing of your actions matches the timing of your voice.
Open your recorded project up and separate out the audio. (In Camtasia for Mac, press Ctrl+Click on the recording on the timeline and choose Separate Audio from Video.). Export (or "Share" in Camtasia) the file as audio (Wav format), and then open the audio up in Audacity (or some other audio editor). Play the audio again and make sure your script matches the audio verbatim. Usually there are a few extemporaneous changes I make while recording, so I now edit my script to make sure it matches what I said.
Next, play the audio again and insert numbers where you pause. If you pause 3 seconds after an action, type 3 in that space. Format all of these numbers in red so they stand out.
Now create a new track (Command + N), put your cursor on the track, and press Shift+R to append your recording on the new track. (If you don't press Shift+R, then each time you record, Audacity will create a new track.) Read through your entire script, pausing where you noted to pause.
As far as voiceover techniques for recording, try to vary the pitch of your voice (intonation), avoid repetitive rhythms, enunciate clearly, speak confidently, and mix up your pacing to sound more natural.
After you re-record the audio, play both tracks simultaneously and match up the timing. Your pauses won't match up perfectly, but you can easily adjust the space between audio waves to get it close.
The following video shows how to sync up the new recording with the old -- dubbing over the voice. It took me about 15 minutes to sync up the dubbed voice with the original recording here.
Basically, you dub over your original voice with the new voice. Normally this wouldn't be so tricky except that you most likely have intricate timing to match. Since you can easily add or remove spaces in Audacity, it's easy to adjust the timing so that the two tracks are in sync.
Now silence the long gaps between the audio waves (Command + L) to eliminate any loud breathing noises, and you're done.
Reimport your re-recorded audio back into Camtasia and insert it on the timeline.
Listen to the audio with the simulation and remove any long, unnecessary pauses or moments of inaction. For example, if a screen takes too long to load, you can cut out some of the loading time.
Add callouts where necessary, such as when you want to call attention to some object on the screen. If you have real data to hide, add callouts that seamlessly cover those spots. You will need to zoom in a bit to make precise callouts.
Add a two-second title slide that describes the purpose of the video. Fade it out as you fade in the recording. For title slides, I often create a 1280 x 720px faded graphic of the software interface I'm demonstrating, with the main content area blanked out. I add the video title in the lower-right corner rather than the center in order to accommodate the play button, which covers up anything in the middle.
Different software programs have different options for callouts, annotations, and animation. My recommendation is to keep it simple as possible. I usually avoid pans and zooms because they create unnecessary editing. I don't add too many callouts either. Everything you add to your video will increase its file size and make it more time-consuming to produce. Keep it simple, showing the user how to do a specific task (like with the videos on lynda.com).
I usually don't integrate music tracks, but if you do, I recommend browsing istockaudio for a good "loop" or "sting." (By the way, if you find a good looping background track that isn't distracting, let me know.)
Render the video to an MP4 format and upload it to your server. At my work, we currently use Brightcove, which packages its own player over media.
I also upload videos to YouTube as well. Here's why. First, when you upload your 1280x720 MP4 video to YouTube, it renders the video automatically in HD. The HD format is critical for software screencasts, since you typically show interface with small text. Without HD, the text will look blurry. With HD, the text looks sharp.
Second, YouTube also automatically syncs your caption with the voice (through built-in voice recognition software). Take the script you used, and simply upload it in the captions section of the video. The auto-sync is pretty amazing. Not only does it help non-native speakers and people hard of hearing, it also increases your video's search engine optimization. And it helps reinforce learning. For more details on adding captions, see my post, Adding Captions to YouTube Videos.
When you embed the YouTube video, add &rel=0 after the string to avoid having YouTube pull in related videos when your video ends (if you want).
That's about it.
The time it takes me to create a video depends a lot on steps 1 and 2. If I have the knowledge to write the script myself, and the test environment is easy to set up, and there isn't a lot of private data to conceal, I can run through the whole process in about 8 hours. These are for minimalist videos, though. You can spend more time creating videos if you're trying to make them more cinematic. However, I don't believe adding fancy cinematic effects adds to the learning value of the video.
I haven't addressed the question about open source video recording tools. My recommendation: Don't waste your time trying to save money with free video recording tools. Camtasia for Mac costs only $100; the PC version is $300. Given that you might be spending 8 hours creating a video, it's not worth skimping on the software. If you're being paid $40 an hour, then each video costs more than $300 to make. With open source software, you may spend more time due to the difficulty or instability of the software. So buying good software actually reduces the costs of creating video tutorials.
If your requirements are much simpler, however, you can use Jing for free. However, you cannot edit a Jing video. What you record is what you get.
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 the following technical communication topics: Swagger, agile, trends, learning, plain language, quick reference guides, tech comm careers, and certificate programs. I'm interested in information design, API documentation, visual communication, information architecture and findability, 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. You can also contact me with questions.