When I record my art classes I always provide narration, where I speak during the practical demonstrations in order to explain what I’m doing, as well as what my thought process is (showing my working), and share any other useful tidbits that might be helpful. How I record what I have to say has changed over the years as I’ve tried three different approaches, and I thought it would be interesting to document what I’ve tried, the good points and the bad for each, and my reflections on them.

Approach One: Narration recorded at the same time as the demonstration

In this approach I talk while I’m doing the demonstration. Like Bob Ross, or any live art class.

I’ve used this approach for the majority of my classes to date and it’s my second favourite way of recording the narration. The things I like best about it are:

• It’s the most natural to record as I’m literally narrating as I work and sharing whatever pops into my head.
• It’s simple to record as I can set up my mic at the same time as I set up the cameras to record, and I can use the camera audio if needed

It does have drawbacks though:

• It makes the art-making feel less pleasant for me because I can’t sink into the process and just enjoy what I’m doing. I find it strange enough to record myself making art – it’s like drawing with someone breathing down your neck as they watch you work. Then having to narrate what I’m doing while I’m doing it really takes me out of the process, unless I am only half paying attention to what I say, in which case I usually end up rambling…
• Rambling. Because what I say is unscripted it will never be as focused or precise as a scripted dialogue. Sometimes I make mistakes in what I say. All of this leads to the next point…
• Editing issues. Because of the aforementioned rambling, I always need to edit the audio. However, because this sound file will have also captured the other sounds – brushes chinking softly in the water jar, pages turning, etc – that makes editing the audio difficult. If I’ve said something that would be more helpful in a different part of the lesson, but the sound also has a water noise, I can’t put it somewhere else. If I’ve said something wrong and I’m cleaning a brush, I can’t cut if because it will remove the brush noises which will create an odd silence where the viewer would expect to hear a sound. It’s really annoying to deal with. 

If you want to do this approach, some things that can make it easier:

Have a bulleted list of things you want to make sure you cover in your chat when you’re recording your demo. Stick it up somewhere you can easily see but is out of sight of your cameras. This can help keep you on track and minimise repetition and rambling.

Do a practice run of the demonstration, talking as you go. This can help you catch anything you find tricky to explain before you’re recording yourself fumbling through an explanation so you can work out what you want to say ahead of time.

Approach Two: Narration recorded from a script, recorded after the demonstration

In this approach, which I tried in (I think) one class, I filmed the demonstrations without talking. Next I watched the demonstration back and wrote a script that covers what I’m doing in the demo and what it would be useful to explain. Then I record the narration by reading the script. 

I’ve only done this for one class, and it’s my least favourite approach. That said, it does have some good points: 

• Can get more deeply into art-making (well, as much as possible whilst also being filmed)
• No editing issues related to interfering background noise

But I feel these are more than cancelled out by the drawbacks:

• The voiceover never sounds as natural, and can easily sound dry or lifeless
• It can take a long time to write the script because it’s difficult to match the timings of what I want to say to the actual time something takes in the demo
• Can be slow to record if you need to do re-takes because you made a mistake when reading the script. And I always had to do re-takes. And sometimes I didn’t realise until I was editing and that was particularly annoying.
• Editing can also be tricky in this approach as you might want to say more in your script than the time in the demo supports, or vice versa – you might have a part of the demonstration that takes a relatively long time but that you don’t have much to say. Yes, you can add music here, but a little rambly chat might have been a more engaging experience for the viewer.

Matching the ambient sounds of the voiceover to the ambient sounds from talking heads and demos can be tricky – for example, if you record in a different room, the audio will be almost impossible to match unless you’re a sound engineer (I am not a sound engineer). After making this mistake once, I always recorded in the same room, with the set up as close to the same as possible. Even doing voiceovers at night while the talking heads were in the day makes an annoyingly large amount of difference to the ambient audio.

If you want to do this approach, some things that make it easier:

Record the narration in the same place, at around the same kind of time of day as the rest of your recordings.

Making your filming set-up quick and easy to assemble, so that when you go back to record your narration set up time is not a barrier. It also makes it less frustrating if you need to go back a third time to record more narration .

Approach Three: Narration recorded from notes, recorded after the demonstration

In this approach, which I’m trialling in the class I’m working on right now, I filmed the demonstrations without talking but with lots of lovely ASMR-esque noises, capturing the sound of opening sketchbooks, clicking pen lids shut, etc. Then I watched the demonstrations back and made bullet-point notes about what I do and what it would be useful to explain, describe, and have a little rambly chat about (including anecdotes). Then I recorded the voiceover, sitting at the same desk, in the same room where I made the demonstration and will do the talking heads. 

So far this has given what seems to be the best of both worlds and is my current favourite approach. Why do I like it? Well, I’ve found:

• I can be more in the moment while doing the demonstration / making art (well, as much as possible whilst also being filmed)
• It’s relatively quick to record because it’s naturalistic, so I’m not in any danger of mis-reading a script and thus don’t need to re-record anything (like has happened with approach 2)
• No editing issues related to interfering background noise (like I get in approach 1)

Drawbacks:

• The editing required here is slightly more than the fully scripted version (approach 2) although it is significantly less than the first approach. 
• For a few of the demos I’ve had to go back and record more audio when I’ve realised that I didn’t say enough the first time around.

Something that really helped with that last point is that for this class I’m also trialling ‘editing as I go’ which has been particularly helpful as I’ve caught anything missing right away. This allowed me to record more audio immediately while everything was still set up and so still the same day, close in time, etc. etc. This editing-as-I-go process (which I’ll write more about that later) is so much more efficient.

If you want to do this approach, some things that make it easier:

Record the narration in the same place, at around the same kind of time of day as the rest of your recordings.

Making your filming set-up quick and easy to assemble, so that when you go back to record your narration set up time is not a barrier. It also makes it less frustrating if you need to go back a third time to record more narration.

So those are the three approaches that I’ve trialled for recording narration in my art classes – live, real-time narration, scripted narration, and a blend of both.

Something worth mentioning that has made a huge difference to my editing process is that the software I use (Premiere Pro) now has in-built (and good) transcription, and you can edit by using the transcript. This has made things so, so much faster and has made all three approaches quicker than they were before. If your software has this option I highly recommend you take advantage of it. Work smarter, not harder.