A neophyte's introduction to video

Photography helps us preserve precious memories, document the world around us, and convey ideas in an unambiguous way. Because of its immense utility, almost everyone knows how to take passable photos. With a bit of effort, it’s also possible to get a hang of the lighting and composition techniques that set remarkable pictures apart from mediocre ones.

In a way, videos are to photos what images are to text. When done right, a video can effortlessly break down a complex process or capture the atmosphere of a memorable day. Alas, producing a competent video appears far harder than taking a good photograph.

The crux of the issue is that film adds a whole new dimension to your work — time — and uses it in a nonlinear way. In most long-form videos, frequent cuts are necessary to maintain good pacing, to switch between vantage points or filming locations, and to allow for retakes and coffee breaks. At the same time, as viewers, we’re easily thrown off by gaps in spatial or temporal continuity. To solve this problem, professional filmmakers rely on a variety of tricks — including visual and audio cues — to foreshadow transitions and guide us along the way. A video assembled with no regard for these techniques is bound to be hard to watch.

I’m not at a point where I could authoritatively teach the principles of filmmaking, but I’m over the first couple of humps — so I figured it might be useful to write down some practical tips.

The filming process

My life as a hobbyist filmmaker revolves around two types of videos. The first type are DIY projects. These videos are usually shot in my garage with a stationary camera mounted on a tripod. The setup is minimal; the only other piece of equipment is a Westcott Solix lamp providing diffuse light.

Videos shot from a tripod tend to lack dynamism; to deal with this issue, I stick to short takes, often just 10-20 seconds long and highlighting a single action; I also move the camera around in between shots. The approach gives me the freedom to use a 50mm lens and zoom in on a slice of the work area while keeping clutter out view.

My other cinematic endeavors are family videos. These do not have complex storylines; instead, I try to capture the zeitgeist of everyday moments, such as day hikes or trips to the beach. I talk about the purpose of these films in an earlier article, “Digital detritus”.

For such projects, I opt for less aggressive pacing; I stick to longer takes, often 20-40 seconds each, and I move around to follow the action. When recording handheld videos, I find it best to employ wide-angle lens, somewhere around 24mm. Longer focal lengths produce too much camera shake and make it difficult to keep up with children and pets.

For outdoor takes with a prosumer camera, I recommend enabling manual exposure or AE lock; otherwise, panning the camera can cause jarring brightness shifts in the middle of a clip. It’s also good to switch to a logarithmic tone curve, such as Canon Log 3; this gives the device a blessing to cram a wider range of sensor readings into a standard video file, allowing blown-out highlights or harsh shadows to be recovered later on. The trade-off is that the video won’t look right until the appropriate inverse operation is applied in post.

As for other video parameters, I go for the highest resolution my camera and laptop can handle (UHD 4K, 3840 x 2160) at roughly 24 frames per second. This framerate yields a distinctive look associated with cinematic works. Smartphones typically record at about 30 fps, giving the videos more of a TV vibe; it’s also possible to film at approximately 50 or 60 fps for butter-smooth motion at the expense of file size. The exact choice is less important than knowing where you stand: mixing incompatible framerates and resolutions in a single video project is a recipe for pain.

If you plan to record audio in the field, I strongly recommend getting a good on-camera microphone. I’m using Rode Stereo VideoMic adorned with a “dead kitten” windscreen. With +20 dB amplification enabled on the mic and automatic gain turned off in the camera, I get brilliant directional audio and zero issues in noisy and windy environments.

Pre-editing #1: The storyboard

The initial stages of the editing process can be accomplished without specialized tools. After copying the videos to my laptop, I manually sort through the recordings, move obvious rejects out of the way, give the remaining clips descriptive names, and sort them in the order I want them to appear in the film (e.g., 010 - hike - dogs running around.mp4).

I almost never follow strict chronological ordering; instead, I try to keep context switching in check. For example, on a recent camping trip, I organized the recordings into three seamless chapters — “activities on the trail”, “daytime at the camp”, and “nighttime stories” — and I opted for that structure instead of cycling through individual days and nights. Similarly, when filming a DIY project that involves several parallel tasks, I reorder the clips to show each activity start to finish before moving on to the next one.

Pre-editing #2: Soundtrack and voiceovers

The second step in my workflow is picking a soundtrack. Not all videos need one — a heavily-narrated clip or a lecture might be better off without any background music — but a well-chosen track helps in several ways. It sets the mood, helps mask audio edits, and drastically improves the appearance of video cuts that are synchronized to the beat of the song.

If you want a soundtrack, resist the temptation to consult your Spotify playlist; copyright issues aside, pop songs are ill-suited for background use. There is a separate genre of ambient cinematic music meant to convey moods without getting in the way. A hike featuring breathtaking mountain vistas might benefit from an epic orchestral tune; a lake on an overcast day calls for a nostalgic and ethereal track.

For all this and more, I recommend Soundstripe. It’s an endless source of high quality royalty-free tracks (and audio effects, if you need some). The service costs about $70 a year and is well worth the price.

For certain types of projects, this is also the right time to record voiceover narration. I have an accent, so I don’t narrate my DIY projects, but I had positive experiences with murf.ai — an text-to-speech platform that produces incredibly natural and expressive results. Alas, they discontinued their pay-as-you-go pricing, so if you want to use them, you need to shell out $20 for a full month.

Editing #1: The timeline

There are quire a few video editors out there, but it’s hard to beat DaVinci Resolve. The basic version is available for free; the pro edition is $300 for life. Unlike most alternatives, the app is yours to own: it doesn’t require a subscription or an internet connection to work.

Whichever option you choose, be aware that video editors have rather hermetic UIs that take some getting used to. I suggest setting time aside for reading tutorials and messing around with the app before doing any serious work. With that out of the way, the “real” editing process begins with importing the clips, trimming away the awkward parts, placing the videos on the timeline, and then dragging the music onto a separate audio track.

The next challenge is sorting out video transitions. The default is a hard cut: the last frame of one clip is followed by the first frame of another. Hard cuts are a good, unobtrusive choice when the video maintains spatial and temporal continuity — but problems can creep up with ease. A classic blunder is filming a person walking down the street and switching between their left and right profiles. Most viewers would interpret this progression of clips as an inexplicable 180° turn. Non-overlapping backgrounds in the shots might hint at teleportation, too.

Because hard cuts can cause problems of this sort, novice filmmakers sometimes fall back to exaggerated transition effects such as VHS-era wipes. These techniques work in the sense that they adequately hint at the passage of time; but they are distracting and don’t age well. A less obnoxious visual cue is a fade-in (the old clip cuts to black and then a new scene gradually fades into view) or a fade-out (slow fade to black and then a sudden beginning of a new take). A longer dwell on a black screen is also an effective way of separating chapters of a video or building suspense.

As noted before, all transitions appear more natural if they are synchronized to the beat of a background audio track; we instinctively expect something to happen when the beat drops, so a well-timed cut becomes less of a “huh” and more of an “a-ha”.

Editing #2: Mixing audio

With the video timeline sorted out, it’s time to fine-tune the audio track. This may involve adjusting clip or soundtrack volumes to mute random noises or to accentuate the important bits. But above all, it’s crucial to fix any abrupt sound level or background noise differences across adjacent takes.

The canonical audio transition is a cross-fade: the old clip gradually fades out in sync with the new one fading in. Cross-fades can be combined with l-cuts and j-cuts: an l-cut lets the audio from a preceding clip play over the beginning of a subsequent video; a j-cut is the opposite, foreshadowing video transitions by switching audio ahead of the time.

J-cuts and l-cuts, with letters illustrating the origin of their names.

The benefit of l-cuts and j-cuts is that they provide additional hints of temporal continuity, even if the visuals change. The other practical consideration is that they are relatively unobtrusive, so you can rely on them to make up for unusable audio at the beginning or at the end of a clip.

Editing #3: Color and tone

The final stage of my editing workflow is to correct exposure, color, and tone. For videos recorded using logarithmic color curves, this should begin with applying the correct inverse transformation, known in the filmmaking community as a lookup table (LUT). After that, the process is not different from what’s used in photography: exposure, white balance, and shadows + highlights are the most important adjustments for each clip.

With the per-clip tweaks made, it’s also important to play the entire movie start to finish and check for consistent tone; for example, you might discover distracting shifts in white balance depending on the filming location or the time of the day.

Thematic color grading is also possible: some filmmakers may want to reproduce that orange-tinted Wes Anderson look — or go full Chernobyl and give everything a sickly shade of green. I don’t do this often, but if you’re interested, make friends with the color balance tool.

Rendering, hosting, and backing up films

Having made painful mistakes in the past, I recommend rendering your films at the highest effective resolution and bitrate you can handle, and keeping at least two well-labeled backups, one of them offline. You will want to keep some of your films for decades — and on these timeframes, quality standards change, computers break down, and online services go defunct.

I render my videos using the H.265 codec at a QHD 2K resolution (2560x1440). I make one nearly-lossless archival copy by configuring 4:2:2 chroma subsampling and using a constant-quality encoder with a very low CQ value (2-5). The resulting bit rate is about 20-40 Mbit/s — perhaps too much for online uses, but not an issue for local storage. If necessary, I also render an “online-friendly” copy at a CQ value of 15-20 — resulting in a 2-3x reduction in file size.

As with photography, video color management on the internet remains a chore. There is no guarantee that the people viewing your videos have properly-calibrated monitors, or that their software will respect the metadata embedded in your files. For maximum compatibility, most sources recommend using the Rec.709 color space with 2.2 gamma, but YouTube doesn’t render such clips quite right. For this particular platform, switching to 2.4 gamma can help.

As far as hosting goes, YouTube is the only game in town if you’re aiming for on-platform engagement; that said, the ad-filled and clickbait-riddled interface, along with minimal customer support and aggressive content policing, make it a poor choice for private videos or for anything you’re planning to embed somewhere else on the internet. For such purposes, Vimeo is a more dependable pick; they have a reasonable free tier for up to 25 videos, and a paid plan that’s about $90 a year with a discount code.

Closing words

There's plenty I didn’t cover: from simple video stabilization and noise removal techniques, to motion tracking and compositing tricks essential for modern visual effects, to storytelling approaches such as cutaway shots, jump cuts, or cross cuts.

This is on purpose; I think it’s easier to learn a skill if we’re not getting pulled in too many directions at once, and learn the fundamentals without the temptation to cover up issues with glitzy special effects.