Vertical
9:16Full-screen vertical feeds, where the picture fills the screen and nothing else is on it.
The frame is recomposed around whoever is speaking, and the caption block sits above the bar the platform draws along the bottom.
Service
Shapes, captions and on-screen wording
A clip is exported once for each shape you publish in, because the crop and the caption both move. Everything written on the picture comes from wording you have approved.
Shapes
Three frames, three different crops
The picture is recomposed for each shape rather than squeezed into it. Nothing is stretched, and no single export is asked to work everywhere.
Square
1:1Feeds that show a square picture with your own wording underneath it.
A tighter crop than the source, kept level with the eyeline, with the corners left clear of anything that has to be read.
Landscape
16:9Wide players, your own pages, and anywhere the clip sits inside a page rather than filling a screen.
The source framing is kept and the clip is topped and tailed only, so a landscape export stays closest to the recording as you shot it.
Why not one file for all three
A landscape file played in a vertical feed is a small picture with empty bands above and below it, and the captions are then too small to read on a handset. A vertical file in a wide player is the same problem the other way round. Each shape is its own export with its own framing and its own caption placement.
Which shapes an order covers
Named in the quotation, because each one is separate work. Adding a shape after handover is a further piece of work rather than another copy of a file, and it is quoted the same way as the first.
Captions
Five stages, and a person on the transcript
A caption that is wrong about a name or a figure is worse than no caption. That is settled at the first stage, which is not finished until a person has read the transcript against the audio.
-
Stage 01 · Transcript
A machine transcript first, because it is quick. It is then read against the audio and corrected line by line. Names, job titles, product names, figures and anything said over a laugh are where machine transcripts go wrong, and those are exactly the words a clip tends to turn on.
-
Stage 02 · Timing
Each line is timed to the speech, so it changes on the beat of the word rather than a moment behind it. Lines are broken where the sense breaks, not wherever the character count runs out.
-
Stage 03 · Setting
Your typeface and your colours where you have them; plain type with a legible weight and a background that keeps it readable where you do not. Nothing is invented to imitate a brand you have not sent us.
-
Stage 04 · Placement
Captions sit clear of the strips each platform covers with its own furniture, and clear of the frame edges. That is why the shapes are exported separately: a caption placed for a vertical frame is in the wrong place in a square one.
-
Stage 05 · Delivery
Burnt into the picture so the words survive being re-uploaded, or as a separate SRT or VTT file that can be switched off and edited, or both.
On screen
Wording you have approved, and nothing else
Everything written on a clip comes from you. Nothing is composed on your behalf and no claim is added to a frame.
Opening titles
A line of wording at the top of a clip, from text you approve. It is how the question gets supplied when the answer is the clip.
Lower thirds
Names and roles as you give them to us. We do not look people up and we do not guess at a job title.
End cards
A closing frame with your own wording and artwork. What it asks the viewer to do is your decision, and we will not write a claim into it.
Your own graphics
Logos, badges, product shots and holding frames you supply, placed consistently across the set. Nothing is downloaded from anywhere on your behalf.