A Quick Test
Is this the first time you are recording a video? Or are you using some of your equipment for the first time?
Then do take the time to make a small test recording and play it back. This is just a sanity check whether you've set everything up correctly. Is the video quality ok? If you want to keep the audio, are you happy with the setup for that? What about the lighting you chose?
This could actually be seen as preparation, but it's the one preparation that's good to do right before recording when you will not change your setting again. And once you have a setup that works, you can skip this step or take short-cuts.
For example, I've been using the same camera and lighting for a while now. So all I do is check the live preview on the camera's small screen to check that the lighting looks ok. I also put the sheet of paper I'll be folding with on the work surface. This helps me determine whether I like the contrast of paper to backdrop or whether I need to use a differently colored paper. I also check whether I want to move the camera's position, so that the paper lies at a convenient position and is in the frame at the same time. Do remember that the more comfortably you can reach the paper, the more naturally you will be able to fold while recording. Moving the camera in the beginning is easy, but not so much later. Which brings me to my next point.
Avoid Hard Breaks
If at all possible, avoid hard breaks while recording. But what do I mean by a hard break? Basically, while you are teaching a model you are telling a story. This story is about how to fold the model, and demonstrating how you fold it yourself. The viewers will see the video in one piece, so you should try to convey that the final video was shot in one piece. Hard breaks are aspects that destroy that illusion.
Here's an example of a hard break. If you stop recording and adjust the camera position - or the model, and then continue recording it will be apparent in the final video that your scene changed. It's relatively easy to not move the camera as long as you do not do so intentionally. It may be harder with the model. So here's a tip. If you do stop recording for a bit, lay the model on the folding surface, then move your hands out of the frame. When you start recording again, move your hands into the frame and to the model and continue where you stopped. The important bit is: do not move the model while you are not recording.
An alternative is to take the model out of the frame, too, and move it back into the frame when you continue recording. However, I discourage this and would go further to say that you should avoid taking the model out of the frame once you've started recording. If you are consistently recording the model, you can be sure you recorded every single step you performed. If, on the other hand, you skip even just a small step and it's not in the final video, viewers will be confused - and presumably you will get many questions on it. Or, indeed, it will simply not be viewed by as many people, as it is perceived to be a bad explanation.
By the way, if you take a longer break, do be aware that the light situation will change, and thus a slightly less disruptive break will still be visible. This is why I try to block out most of the natural light in my recording area and use my daylight lamps instead.
And of course there are other examples of hard breaks. Some of them may be unintentional. Here's one that is hopefully not too obvious to state. Do check that the battery will last for the whole recording, and that the disk has enough space for your recording. I actually always record with my camera connected to the charger, so that - unless the electricity fails - my recording will not be disrupted by an empty battery. My camera has a 40 GB disk, but don't think that's a safe net. Every time before I record I check how much space is still free. I've had to delete old material several times, because else there would not have been enough space for the new material. It's small things you can do to avoid big annoyances, such as having to stop your recording due to such technical limitations.
Another prominent example of hard breaks is moving the model out of the frame. It may not destroy the illusion that you are folding along - but it will raise questions whether you are doing something with the model the viewer cannot see. Especially if you are indeed doing a step off-frame this destroys the illusion that you are really showing them how to fold the model - step by step. My camera has a small screen that shows a live preview - as most cameras do nowadays. When I bought my camera I did check that the screen could be tilted. So while I am recording I regularly check that I am still in the frame. With practice you also get a feeling for how much you can move your hands and the model to stay in the frame.
Adjusting the Zoom Level
When recording, try to find a zoom level that's appropriate. Usually I will set up my camera, so that the square takes up most of the frame without zooming in at all. Once I progress and the model gets smaller, I may zoom in a bit. On detail steps, I may shortly zoom in to only show a small part of the model - and zoom out again after that is completed.
There are a couple of points to consider. First, the model and all details that need to be recognizable to follow along should be clearly visible. This usually means you want the model to take up most of the frame, but you also want it to stay fully visible. Second, any change of zoom level can be seen as a hard break (see above), so only use this feature when you think it's necessary. In other words, if the value of changing the zoom level is higher than the cost of a hard break, go for it. If not, let it be. Third, always check you frame when changing the zoom level. Especially the more you zoom in, the more careful you have to be with staying in the frame. There is nothing worse than wanting to make a detail shot to later realize that you were off-frame. Also check that the recording is not blurry. Many cameras have an auto-focus feature and when you change the zoom level it may take a moment to adjust. Give it that time.
Another option is to record at one zoom level and then crop and enlarge the frame in the editing phase. There are two disadvantages here: First, it will take more time. Any editing work you can avoid while recording can save you lots of time later. If you decide to add editing time, do so consciously. Second, it is likely that when you crop and enlarge in the editing phase the video quality will suffer. This may not hold true if you are using some very fancy (i.e. specialized and expensive) equipment, but my guess is most of us don't have that. I definitely don't - and don't think it's necessary, either. Plan ahead instead, thinking about which steps need detail shots.
If you want to add verbal instructions to your videos, you have two main options: either record the audio with the visual, or add a voice over in the editing phase. In either cases, speak clearly. Check that your microphone is capturing your voice at a good level. If the microphone is too sensitive, you may have recording glitches, if it's not sensitive enough, the audio will be too quiet. The best you can do here is really doing a couple of test runs until you're happy with the audio you're getting.
When I record my videos, I add the verbal instructions right away. This does mean I sometimes splutter, but it also helps keep everything conversational. I also find it easier to explain what the viewer needs to do while I am performing the step myself. Most importantly, though, it saves a lot of time.
I do have one video from way back* when I started where I added the verbal instructions after recording. If I remember correctly it took me something like 2 hours to get a result I was happy with - for a video that was less than 10 minutes long. This did not include any other editing work, just the voice over. Granted, the quality was nicer, but it was also very stressful and hard work.
Nowadays the only voice-overs I do are on small parts of the videos. Sometimes I do mess up and say something incorrect. When I notice this only after recording, and I think it will be too confusing for the user if I keep that mistake in there, I'll do a voice over for that small section. It will be noticeable that I added a voice over for that part, as it's very hard to get the same quality of recording in a different setting and with a different microphone. But it's not about having it be perfect, it's about getting a balance that will help the viewer most.
Another aspect to consider is that when you add the audio later, you either have to overlay it with the audio of the video recording, or live with not having the sound the paper makes when being folded. I do really like hearing the paper "speak", and it can sometimes also be quite revealing. For example, if the paper clicks during a step, it speaks of the paper tension that you are working with at that point.
(*Unfortunately this video is not online, as I was not able to contact the designer for permission once I started doing that. If any of you know how to contact Kazuo Haga, do let me know! My previous attempts all failed.)
In the previous article on preparing for the recording I'd already explained that you should practice the model beforehand. Let me re-iterate here the one aspect that you should always remember when recording.
Never rush through difficult steps. If you want to get a shorter video, save time on the simple steps. But never ever dedicate less time to the steps you find difficult to explain. Always remember that the viewer will only have that video. They can play it back several times, but that's it. So if you know there's a step that's a bit tricky, give it the time it needs. Perhaps it may even be helpful to show it several times. If the step is repeated, show it from a different angle the second time round. If it doesn't get repeated, consider unfolding it and then showing it again from a different angle. Or if there are two ways of performing the step, weigh off whether you want to show both variations.
Remember that your motivation for making an instructional video is to provide instructions that others will be able to follow. Remember that you are in a teaching situation. And remember that your students won't be able to ask questions directly. Just like with diagrams.