History of Subtitling

Subtitling used to be a complex process that often required a great deal of effort. Producing subtitles involves transcribing or translating the original dialogue, sometimes using a transcription known as a pre- or post-production script.

For celluloid film, a chemical process was used, whereby text was burned into the film material itself. For television, the "camera method" was used, where the video image was combined with the subtitles during the broadcast. While the programme was being broadcast, a camera was pointed at a card displaying a subtitle, and the two images were merged to create the subtitled broadcast. A system ensured that the subtitle cards were placed in front of the camera in the correct order. The person responsible for this was called the title director. He or she was not the translator, so practice runs were done before the broadcast to ensure the correct subtitle was displayed at the right time and for the right duration—a challenging task, especially when seeing the programme for the first time.

What made the title director’s job even more challenging was dealing with films in less common languages. It was known to happen that the title director was left with a handful of subtitles at the end of a Japanese film...

Since a feature film typically contains around 700 subtitles, one also had to be careful not to drop the stack of subtitle cards when preparing for the broadcast.

Even with modern technology, it can take up to five hours to spot (i.e., time the in and out cues) the subtitles of a one-hour talk show.

Spotting, or cueing, is the process of determining the in-cue and out-cue for each subtitle. The in-cue is the exact timestamp when the subtitle appears on the screen, and the out-cue is the exact timestamp when it disappears.

Translating a programme from one language to another takes a translator/subtitler between two to four working days of eight hours each. And that's not to mention a reference-intensive documentary, which takes much longer due to the constant need to look up terms.

It takes an experienced subtitler about five days to translate a 90-minute feature film into subtitles.

Today, modern software has transformed the subtitling process, automating timecodes and even suggesting translations. Yet, the subtitler’s craft remains essential—software can assist but cannot replace the human touch required to accurately convey meaning and tone.

Despite all technological advances, subtitling remains a craft that demands attention to detail and a deep understanding of language. The art of conveying spoken dialogue in a few words on screen will always be a unique challenge.

Most subtitlers not only translate the video; they also set the in-cue and out-cue for each subtitle. The out-cue minus the in-cue determines the duration of the subtitle—i.e., the time it remains visible on screen.

A subtitler works on a wide variety of programmes, requiring broad general knowledge.

In the 1980s, subtitles were first produced using pencil, eraser, and paper. These subtitles were then typed on an expensive character generator (costing around €35,000) by a typist, after which the translator would spot the in- and out-cues on the same device.

The process led to a qualitative improvement, as the translator/subtitler was now confronted with an overwhelming amount of text for a very limited viewing time. Reading speed became a critical factor. A subtitler must account for limitations in time and space. Since reading is slower than listening, the subtitle must match the viewer’s reading speed. The target audience’s age also influences reading speed—subtitles for older viewers are displayed longer than those for younger audiences.

A full subtitle of up to two lines is typically shown on screen for about five to seven seconds. During this time, much more can be said than what fits into a subtitle. On average, a subtitler has around 43 characters per line, including punctuation and spaces. This limit can vary depending on character width—a ‘W’ is wider than an ‘i’. It’s crucial for the subtitler to condense spoken text without losing meaning, often restructuring sentences to fit. This may confuse viewers who partially understand the source language, but it’s a necessary compromise.

A viewer without knowledge of the source language will easily understand the subtitle within the time it’s displayed.

A skilled subtitler condenses the text without losing meaning, detail, or nuance while adapting the translation to the viewer’s cultural context. It’s far from the literal translation many assume it to be.

The question every subtitler constantly asks is: How would someone of the same age and from the same social class say this in the target language, given the same circumstances as the original?

The combination of grammar and reading speed can, in extreme cases, result in a subtitle that seems to contradict the spoken word: the subtitle says "Yes," while the speaker says "No."

Subtitles appear on screen precisely when the speaker starts the sentence and disappear as soon as the words are spoken. The display time varies depending on subtitle length, always ranging from 1.5 to 7 seconds.

Because almost all subtitlers master this technique, many viewers mistakenly believe they understand the source language—especially if they claim they "never read the subtitles."

From the mid-1980s onwards, most subtitlers worked on a so-called prep unit.

This relatively expensive equipment was loaned or rented to subtitlers so they could work from home, reducing the strain on the expensive and cumbersome character generator at the subtitling company.

Many subtitlers also purchased their own prep unit. A complete setup consisted of a small character generator, a computer with its monitor, a video monitor or TV, a modified VCR, and various peripherals.

For each translation job, a subtitler received a VHS tape from the client with a transparent or black timeline on screen. This running clock, called the timecode, was not only visible on the screen but also transmitted to the computer as an audio beep or via hidden image lines. This had a purpose: it made entering the in- and out-cues easier—16/18 digits per subtitle. 

00:01:12:24 (when using frames) | 01:23:07;120 (when using milliseconds)

Using the timecode signal, the entire timecode could be entered into the timecode field with a single keystroke, similar to modern digital workflows.

This setup also enabled subtitlers to check their work by playing the VHS tape with the running timecode and ensuring the subtitles appeared at the correct time. This process, known as rehearse mode, allowed for accurate synchronization.

After completing the translation, the subtitler submitted a (floppy) disk containing a file with the translation and in- and out-cues. This file could then be further processed for subtitled video copies, TV broadcasts, or other media, such as DVDs. A DVD disc could contain up to 42 subtitle/language tracks.

Today, the prep unit has been replaced by the combination of a PC and specialized subtitle editing software. Modern subtitlers no longer receive a VHS tape but instead download the digital video file from a client's server, making the workflow faster, more efficient, and more flexible.