FilmotionCaption for the listener who is also reading
A vertical podcast clip asks the viewer to follow a face, listen to a voice and read captions in the same small space. When every word bounces, changes colour or disappears immediately, the captions become another performance to track.
I treat captions as a reading layer, not a second presenter. Their job is to preserve what the speaker means while making the conversation easy to follow with or without sound.
I start with an accurate transcript and a locked edit, then work on timing, line breaks and emphasis, in that order. Styling cannot rescue text that disappears before someone can read it.
Check reading speed before changing the font
Reading load depends on both the amount of text and how long it stays visible. A useful calculation is:
Words on screen ÷ seconds visible × 60 = words per minute.
A caption containing eight words and lasting two seconds asks the viewer to read at 240 words per minute. The same caption lasting three seconds works out to 160 words per minute. These are comparisons of timing, not targets for every viewer or clip.
I use that calculation to spot crowded blocks, then review them in context. A familiar conversational phrase and an unfamiliar job title can occupy similar space but require different amounts of attention. The viewer also needs time to glance at the guest's expression.
Consider this podcast sentence:
“When every request needs your approval, you become the bottleneck.”
Showing the whole sentence for two seconds creates a heavy reading load. Splitting it into several flashing fragments does not solve the problem. The viewer still has the same words to process, with extra visual changes added.
First, check whether the caption disappears unnecessarily early. Let it remain through the speaker's natural pause, provided it does not overlap the next spoken thought. If the delivery is consistently too fast, revise the underlying clip rather than silently removing meaningful words from the captions. Trimming a digression or choosing a less compressed passage can give the explanation room without changing the guest's point.
Break lines where the meaning bends
A caption can fit inside its box and still read badly. Automatic wrapping often divides phrases at the wrong point.
This break separates the subject's modifiers from its main noun:
> Your team's biggest
> problem is unclear ownership.
A more natural arrangement is:
> Your team's biggest problem
> is unclear ownership.
The first line establishes the subject. The second completes the thought. Each line gives the reader a useful piece of meaning.
Keep articles with nouns, names together and short verb phrases intact where possible. Avoid separating “didn't” from the action it negates. A caption that briefly shows “I would recommend that” before revealing “you don't do this” creates a misleading first impression. Keep the recommendation and its negative together when the speech timing allows.
Use one or two lines as your default. If a natural phrase cannot fit at a comfortable font size, divide it into successive caption blocks at a clause boundary. Do not shrink an entire clip's captions to accommodate one unusually long sentence.
Time blocks to phrases, not individual syllables
Word-by-word captions can follow speech closely, but they also force the viewer to keep locating new text. For an explanatory podcast clip, phrase-based blocks are often the calmer starting point.
Take this line:
“We hired another manager, but nobody knew who could approve the work.”
Try three caption blocks: “We hired another manager,” then “but nobody knew” and finally “who could approve the work.” Check the actual delivery before committing. If the guest races through the middle phrase, combining it with the last block may avoid a brief flash.
Align each block with the corresponding speech. Keep punctuation that helps explain the relationship between the phrases. The comma before “but” prepares the reader for a contrast.
Avoid exposing a punchline well before the guest says it. Equally, do not delay every caption until the phrase has finished. The text should arrive with the speech, not chase behind it.
At a speaker change, start a new block. Leaving the host's question over the guest's opening answer makes attribution harder, especially with sound off.
Emphasise the contrast, not every interesting noun
For “We needed fewer meetings, not faster meetings,” the important contrast is between “fewer” and “faster.” Highlighting “meetings” twice adds decoration without explaining the distinction.
Choose one emphasis treatment, such as a heavier weight or a restrained accent colour. Keep the rest of the caption stable. Avoid combining a colour change, a size jump and a bounce on the same word.
Do not depend on colour alone to convey meaning. The sentence should remain clear in plain text. Preserve qualifying words such as “sometimes,” “might” and “in our case.” Emphasising a claim while visually burying its qualification changes the tone of the conversation.
Review the clip at actual phone size
I finish with a phone preview, not just a large editing monitor. Captions need to clear faces, speaker labels and the destination platform's interface. A block that fits below the guest in the editor may collide with controls in playback.
My final pass checks:
- Muted comprehension without pausing.
- Caption changes against spoken phrase boundaries.
- Awkward splits or isolated words in two-line blocks.
- Names, negatives and qualifications against the audio.
- Emphasis that decorates rather than clarifies.
If a block feels rushed, I adjust its timing or phrase boundary before adding animation. Good podcast captions let the speaker remain the centre of attention. The viewer should remember the explanation, not the effort of reading it.