Join 3,377 readers in helping fund MetaFilter (Hide)


What's the best, most efficient workflow for making a screencast?
January 28, 2014 2:02 PM   Subscribe

I want to make a series of instructional programming videos. I'm reasonably happy that I have all the moving parts lined up - I know what I want to cover, I have the examples and material ready, I've figured out how to make high-quality video recordings of my desktop and I've got a nice USB microphone. But if there's one thing that my brief forays into video editing have taught me, it's that it can be incredibly time-consuming. Given that I plan on making ~10 hours of content for my first batch of videos, I'd like some advice on doing this efficiently. Specific questions inside.

- Is it better/necessary to write a script from which to read? Or will that make it sound unnatural? This is going to be material that I normally deliver in the form of a lecture + slides, for which I don't use notes.

- Is it better to try and record the video + audio simultaneously? Or do the video first then dub audio on top of it? Or read the script first and then match up the video?

- Is it better to split a ~1 hour long video up into very small chunks, and record them individually then stick them together? Or is it better to try and do the whole thing in one take and edit it afterwards?

Any other advice? or if somebody can point me to a recommended workflow for this kind of thing, I'd be very grateful.
posted by primer_dimer to Computers & Internet (9 answers total) 8 users marked this as a favorite
 
Can you point out some example videos that you're seeking to emulate in terms of style and presentation? I think that would be helpful.

I think that this and this are great formats, but they also might not work for you.
posted by oceanjesse at 2:21 PM on January 28


Get right into your presentation; avoid lengthy introduction segments/pretenses.

Do it in small chunks, and try to pare it down to the bare essentials. Tangents for levity are acceptable every once in a while for audience engagement. This is why I think spontaneous Audio/Video recording is the way to go. The final product can be delivered as a whole, as long as the portions are bookended at appropriate break points.

Some summary notes per slide to keep you on track are helpful, but I wouldn't script the entire thing (personal preference as a viewer). The absolute worst thing, however, is the presenter trying to fix mistakes or humming-and-hawing over interface glitches or "it should be working" situations and if you anticipate sections having these, you might want to practice beforehand.

In my opinion, dubbing and VO work is useful on short animations or pre-recorded video segments that are timed and scripted, while slides and demonstrations are best explained on-the-fly.

Also, having a rolling record does not stop you from editing down the final product and removing/overdubbing certain sections. Perhaps most importantly, you will have a record of what you were doing / thinking at that moment, which you may not recall if you neglect to record the audio during the first session.

Hour long segments are nasty beasts to slug through in a single take. The results will also put a bigger strain on your computer system due to the filesizes and rendering required in the edit suite. Less responsive = more annoying.

Source: A/V production and editor for corporate events and their accompanying presentations.
posted by Khazk at 2:48 PM on January 28 [1 favorite]


There's no real way to do 1 hour long video in a single take. You'll make mistakes or forget or cough--and if you put the camera on a tripod, just look right at it and talk for 1 hour-- it'll be dull as hell. Write the script, rehearse it, and then figure out which parts will work best with you on-camera and which parts you'll want to show close-ups or something other than you (those clips are called b-roll.)
So, shoot your stand-ups (you on camera, talking) and then go back and shoot the other stuff--close-ups, pans, zooms, your slides, whatever you want. Then edit them together, and record a narration (voice-over track) to go over the parts where you're not on camera.
If you're used to just talking, I'd do that, rather than read a script.
posted by Ideefixe at 2:53 PM on January 28


Just from a viewers point of view, and I'm presuming that you're going to be recording applications etc, since you mention overdubbing as an option.

I much prefer it when the screen-cast is over-dubbed later. I think it gives a far more professional quality to it all. It means you can cut out video of you doing typos and backspacing, fixing a silly missed mistake etc, so everything just flows. Yes, it will be very time consuming, so you just need to weigh that against what you're happy releasing.

The other main reason is (and this is especially true of programming-related tutorials) that when you're typing code, your ability to do anything other then narrate exactly what you're typing is very reduced. I don't need to hear from you what exact words and commands you're typing, I can see that from that video and hopefully the code is available elsewhere so I don't have to copy it manually anyway. I want to know why you're writing it, the cause and effect of what you're doing, otherwise a blog would be a better outlet than a video.

Scripts don't make things seem unnatural, see movies, theatre and TV for examples. Bad delivery of the scripts does-- if you're more happy to go off the cuff, that's fine. It all depends how good you are at speaking, if you're going to umm and err, then a script could certainly help.

I was quite impressed with the quality of the videos from Team Treehouse (video picked at random, I've not actually watched that one) and I think I'd use them as a benchmark if I were to create my own.
posted by Static Vagabond at 2:55 PM on January 28


Re: pre-scripting versus improvising: either will require practice on your part if you want to do a good job. Reading from a script without sounding flat isn't easy, but neither is improvising without sounding flabby. Given that both are difficult, I'd suggest scripting, because it allows you to organize and focus your effort more carefully. Also, it's the route most people do NOT choose (which is why YouTube is rife with flabby, mumbly, cliche-ridden junk). So you can stand out more.

Beware, though, that you may write flabby, mumbly, cliche-ridden junky scripts. I don't know why, but newbies show a dismaying tendency to emulate the very worst cliched version their genre has to offer (you've surely grimaced at wannabe sportscasters and food critics). Not sure why that is. So....unless you can write tautly and clearly and free of cliches (i.e. your scripts don't sound like most improvised screencasts) then definitely improvise, so at least it won't be flabby AND flat.
posted by Quisp Lover at 4:25 PM on January 28 [1 favorite]


Audio: do both: speak as you go, then you can silence the audio track and do a voice over.

Script: Just write out a sequence and the main things you want to say. You can fix it when you dub
posted by jander03 at 7:14 PM on January 28


Many thanks for all the great responses - lots to think about here.

oceanjesse, both the examples you point out are excellent talks, but not really the kind of thing I have in mind. I'm anticipating that the video for my recordings will be 99% recordings of my terminal/editor windows/slides with maybe a tiny bit of my talking head to bookend them.

The example linked by Static Vagabond is a good example of the kind of thing I'm aiming at. I also really like the production of railscasts.com (link to random episode).

I think that the way to go is probably to make a much shorter test video and put it out there for comments. It's probably worth spending a chunk of time up-front in order to figure out a good workflow for the future.
posted by primer_dimer at 1:09 AM on January 29


One thing I do, making very similar sorts of video, is I write an outline for ideas I want to talk about. Then I sit in front of the computer and record the whole thing without looking at the script. Then, I remove that audio and overdub with a script. Sometimes I have to retake some of the video, too. It's really time-consuming, though, so I read this thread with great interest for other ideas.

Personally, I would split your video up into shorter videos unless there's a reason to keep them long format. It's not fun knowing you want 5 minutes' worth of information but have no idea where to find it in an hour long video.
posted by freezer cake at 10:01 AM on January 30


I agree with the shorter video times. I do instructional video content for a university. Try sitting through an hour long video - no way.

I think having multiple shorter videos are the way to go for people's attention spans.
posted by gregjunior at 11:12 AM on January 30


« Older Is carsort.com dead? And assum...   |  I am going to be in Paris, Bel... Newer »

You are not logged in, either login or create an account to post comments