Video Localization Guide - José Henrique Lamensdorf - translation - tradução

Go to content

Video Localization Guide



I have been translating video for dubbing since 1987, and for subtitling since 2004. In 2006 I began to subtitle DVDs. My specialty is training, technical, and institutional videos. I have translated some full-feature movies and TV series, but comparatively few.

I developed this guide to offer clients, friends, and general stakeholders an array of information on corporate video translation, as opposed to commercial film/video, i.e. TV, movies, and video rental stores.

If your company is not in the video, cinema, or TV industries, but it uses corporate video, this page was written for you. If your company is a video producer, it will be in the grey area, as it will depend on the specific project. And finally, if you are in the video, cinema or TV industry, you may read this out of curiosity, since your m.o. should be somewhat different.


Initially, a company receives a video in a foreign language, which they want to show to its clients, prospects, employees, or others.

As we’ve seen, these are usually institutional, promotional, technical, or for training purposes. Obviously, the first concern is how to get them in the local language. The second concern is how much this will cost. If this order is reversed, it won’t make much of a difference, as these two make up a closed loop.

Therefore, I’ll try to deal with both issues simultaneously. Depending on where you live, you might have seen both dubbed and subtitled films. I’ll discuss these two, plus voice-over, which is more frequent in some places than others. To cover all options, we’ll also check the hybrid dubbing+subtitling process.

Before getting started, it’s worth mentioning that translation - the first step - is different for each of these processes. Some translators work in all of them, others only in some, and there are many who specialize in just one. The key point here is that if one decides to shift, e.g. from dubbing to subtitling – or vice-versa – on the way, this will cause unnecessary rework costs. Converting a finished translation from one process to another usually requires translating it all over.



This is the most economical way. No need to describe it, if you’ve never seen it on TV, there’s plenty of it on YouTube.

The major shortcomings of subtitling are:

  • Captions draw part of the spectator’s attention. If it’s all about what some people said, it might be better to translate into written text, and send it by fax or e-mail.

  • If people talk too fast, it might be necessary to shorten phrases, losing part of the content.

  • It definitely does not work for technical instruction films. It is impossible to read a subtitle saying something like “release this lock underneath the cover to gain access to the adjusting screw underneath” and at the same time watch how it is done.

  • The original soundtrack will remain there. If the translation is bad, this may raise protests from bilingual spectators.


No need to describe it either. This entire process – at least in Brazil, and as a rule of thumb – costs about three times as much as subtitling. I charge the same for translating video for any process, but many translators who do both have different rates for dubbing and subtitling.

The major shortcomings of dubbing are:

  • It requires a more specialized translator, so that voice artists may sync their speech to the actors’ mouth movements on the screen. If it’s all off-screen narration, this problem does not exist.

  • If there are too many people talking at the camera, dubbing costs might be outrageous. According to the Brazilian dubbing artists’ union, each voice over artist may play one role, plus two extras on the same film. Maybe there are no such restrictions in your country, however if such freedom is taken to an extreme, the result may be very disappointing. In most places, any voice-over artist will charge at least for the whole first hour of work. Lip-sync dubbing requires a specialized studio to work well. As this cannot be easily done over the www, the artist will have to go to the studio, so it’s quite fair to charge at least for the displacement.

  • If there is music and sound effects (aka SFX), often called M-E, unless it has been provided by the producer on a separate soundtrack, they will have to be re-created. Depending on the content, this may come out quite expensive.


The process is similar to dubbing, however cheaper. It is found mostly in documentaries.

It means non-sync dubbing by up to three people: a narrator, a “man” and a “woman”. The (off-screen) narrator does it exactly as if it were for dubbing, replacing the whole narration. The other characters start speaking with the original audio, and its volume is immediately cut down, for the translation to be read aloud, without any synchronism, by one of the artists. The same man will read the translations for all men’s lines, while the woman will do it for all women. They’ll finish just before the original speech ends, when its volume will be restored back to normal.

It’s comparatively cheaper than dubbing, especially if the video comprises a long series of statements by different people.

The major problems of voice-over are:

  • The final result inevitably looks (or sounds) like a “cheap job”, as attention is continuously drawn to the fact that the film was translated. The impression left is that it was intended to be dubbed, but the budget was too low.

  • Under no circumstances this will be acceptable for dramatic dialogues, as the translation is read almost expressionless, like a newscast.

  • The reason it exists is that it allows to a cheaper video with the dubbing features: it is not necessary to read subtitles, having more time left to watch the film, see the images, and listen to the soundtrack.


In this process, the narrator and/or host, and/or or the leading actors, are dubbed, while occasional appearances, such as several short interviews with different people, are subtitled. Some good judgment has to be used to select what will be dubbed and what will be subtitled in a film, as this might require frequent (or not) spectator shifts from reading to listening, and back.

Dubbing, subtitling, voice-over, or hybrid is the first decision that will have to be made, and there is no use in covering it up: for corporate video, it’s just a cost/effectiveness decision. The purpose of the video should be checked, as well as its expected return, and its useful life.

It is worth reminding that if a cheaper process is initially chosen, and later, according to results, a more expensive one will be used in a remake, nothing from one may be used for the other. In other words, if a subtitled video is successful enough to justify dubbing it later, it will have to cover both the costs of the whole subtitling process, and then the whole dubbing process again, including two different translations.

It’s advisable getting estimates for the different options before choosing one of them. It should also be kept in mind that the whole process is linear; in other words, any savings at one step that result in a quality loss will have impact on all later steps. If the end result comes out unacceptable, rework will have to start at the step where quality was downgraded.


The first step is to check on what media the video is available, and what is the desired output. Nowadays any film or video can be converted to digital format, being DVD the most practical and economic one. However the original might still be on film or video tape, making a conversion necessary.  

I convert from virtually any digital video format, and even from VHS tapes, in the latter case only if the color system is US-NTSC. For all other cases, I outsource specialized services. They will mean additional costs, so it is necessary to know in advance to get an estimate.


Once the intended process has been selected, the next step is translating specifically for that output.

First, believe this: no transcribing in the original language is necessary for translation. Professional translators work directly from the original soundtrack.

There are translators who charge more when a script is not provided (and conversely others that offer discounts when it is provided), though this is not my personal practice. I’ll admit that the script helps in assigning lines to their respective characters when it’s for dubbing, but it doesn’t make so much of a difference.

Quite frankly, in a few hundred training videos, and maybe one or two dozen full-feature films I translated, only one out of every ten had a script. And from these that came with a script, only in one out of every five this script was faithful to the final edit. I had cases where a post-production transcript – made in the original film production country - was provided, but it was fraught with gross misinterpretations of what had been actually said.

There are various price structures for translating video. Many translators charge per minute of playing time, though some charge per word, or per subtitle. Watch out for those who charge “per film”, as this is a common practice for feature films in some places, but they expect playing times to be above 90 minutes, which is extremely rare in corporate video. I use an average rate per minute of total playing time, the same for dubbing, subtitling etc. If for any reason you need a transcription of the film in its original language, bear in mind that this will cost you plus as much as the translation.



In these cases, my job may be limited to translation, if you already have a dubbing studio of your choice. If you wish, I can refer to dependable dubbing studios. I deliver the dubbing script in a MS Word file, in Arial bold 14-pt type.

Delivery standards should be set by the studio that will do the job. Most often it’s just a matter of the translator setting up the required format in their word processor.

For hybrid videos, a common procedure is to assign subtitles to a “character” named “CAPTION”. Some dubbing studios also do subtitling, so it might be faster to have the whole job done at once. In digital video there is no image quality loss in successive generations, so using a copy of a copy is no longer a concern.

Dubbing will require replacing the whole soundtrack and, together with it, music and SFX (sound effects). They will have to be replaced as well. Most dubbing studios keep a stock of incidental music clips for most situations. When music and SFX are not provided separately (some zealous producers do it) and are essential, it may be necessary to re-create them, which might be an expensive process.

Usually, the translator’s job ends with the finished translation. If the film is technical, or if there is any reason for the end client to check terminology, this should be done at this time. If done later, it will involve unnecessary additional costs, or may compromise the dubbing studio’s goodwill, if done at no additional charge. All too often such changes spoil the dubbing “metrics”. For instance, the engineer may decide to replace “setting” with “whatchamacallit fine-tune adjustment”, which will not be possible to dub in the time span allowed for it on the video. Compensation for readjusting the metrics at this stage should be agreed with the translator beforehand, possibly as part of the job, but never after the fact. Otherwise the script delivered to the studio might be impossible to dub. Some studios can take care of it during the dubbing process, but it’s risky. They are usually not language experts, nor have dependable knowledge on the subject matter.

There are some rare cases when the end client wants the translator to be present at the studio during the dubbing work. While a few dubbing studios don’t like having outsiders in their working premises, others actually welcome them, as this spares them from guesswork. This is done to ensure that ad-lib changes do not compromise the intended message, as well as the correct pronunciation of foreign names. This is usually an unnecessary additional cost, but sometimes the intended purpose of the video justifies it.

It’s worth telling something about how dubbing is actually done. After the “making of” several full-feature animation pictures, dubbed by famous actors, became widely known, it’s no longer any secret. One of the earliest examples of this can be seen in the initial scenes of the movie “Mrs. Doubtfire”, where Robin Williams plays a dubber who gets fired. However it doesn’t show the backstage scenes.

The translation goes directly to the dubbing director. S/he checks it over against the video and, if necessary, makes some changes. S/he spots (or delegates to someone) the times for each “line” to be dubbed, to make it possible to easily locate them with the equipment. Next, s/he selects the cast, i.e. “who will dub who”, and prepares a schedule for the recording voices, one at a time, in the studio.

For those who imagine all dubbers together, side by side along a line of microphones, this dates back from the days of live theater on the radio. Dubbers rarely watch the whole film, unless it is one narrator from the beginning to the end. Each one enters the recording studio alone. It’s a dark, small room, with one microphone, a TV screen, earphones, and a lectern where the script is placed. Each dubber will attempt to record all their lines in one session. The operator will be in charge of finding the respective parts of the film. It’s up to the dubbing director to ensure continuity of dialogue among the different dubbers. Depending on the studio, the dubbing director may stay inside the soundproof cabin, together with the dubber, or outside, by the operator’s side. It is rare, however not impossible, to put two dubbers together in the studio, usually for heated argument or intense chit-chat scenes.

The dubbing director is responsible for ensuring scene continuity and dialog integrity. It’s up to them to avoid situations like (man:) “They call me Juniper.” being replied by a dubber who doesn't know it's a man with “How can I help you, Jennifer?”

Many dubbers do a segment in four passes. In the first time, the dubber listens to the vocal expression while reading the text to be dubbed. In the second, s/he tries to adjust the speed or rhythm s/he will say it. In the third s/he rehearses, and in the fourth s/he records it. If anything is not adequate at any time, the dubbing director will cut, and the recording will be re-taken. Sometimes it is possible to use the initial part of what was recorded, so they restart from a minor pause before the error.

Each voice is recorded on a separate audio channel. This is the “dry” dub, recorded in an echo-free acoustic chamber. If it is left as it is, it will sound artificial, as nobody lives in such a place. Furthermore, if there are various voices and/or different settings, the dub recording volumes will be out of proportion regarding the distance each voice was from the camera. So all voices are equalized, environmental sound effects are applied (e.g. inside a tunnel, in an auditorium with or without a microphone, over the telephone etc.), and they are all mixed into one (mono) or two (stereo) audio channels. On top of that, music and effects are also mixed together.

All this explanation was intended to show the devastating effect a bad translation may have on the dubbing process. First, ‘metrics’ (how long it takes to say each segment of a phrase) is essential to make dubbing or VO feasible. If the translation is careless about metrics, this will cause problems to dub.

See this example:
Original: You must  take customer needs into account as well.
Exact translation: Você precisa levar as necessidades do cliente em consideração também.
Feasible translation: Considere também as necessidades do cliente.

Of course, feasibility depends on context. The point here was to show that an exact translation wouldn’t fit.

If the translation is “not feasible”, there are a few possibilities left: Dubber and dubbing director may attempt to fix the sentence to fit. On top of wasting precious recording studio time, this will require changing the translation, which may compromise or not the intended message. The dubber may speak faster to make it fit, however this will impair lip-sync, forcing spectator awareness to the fact that the film was (poorly) dubbed.

In order to avoid such situations, upon translating I strive to offer a dubbing-feasible text. When the videos are very technical, or where terminology is critical, it’s best to have the client check the dubbing script after translation, and before dubbing. If they make changes, I – the translator - should verify it again, if the script remains compatible with the metrics, though good dubbing directors are capable of doing it on their own.

However the worst yet might be lurking. If an overly short turnaround precludes this check, the video may go through all afore described steps, and be delivered as finished. Of course, it is possible to redo just one segment if later changes come up, but the cost will be considerably higher.

In dubbing, it’s essential to ensure quality at every step because – as I said before – the process is linear. If anything calls for rework, all subsequent steps will have to be redone. And the starting point is the translation.

If a video to be dubbed includes “charts”, i.e. screens with titles, text, charts, graphs, etc. it may be convenient, at times mandatory, to replace them. When these are numerous and important, in many cases they become the actual motive for dubbing instead of subtitling, as the spectator won’t have time to watch them and read the subtitles simultaneously.

I can translate and replace these screens, as long as static elements only appear and disappear in their places on the screen. If animation is involved, i.e. the graphic elements move or change their shape on screen, this will require a cartoon animation studio.

Summarizing, the client should verify the dubbing process at specific checkpoints to ensure the desired final quality while avoiding unnecessary rework. The best checkpoints are:
1. After translation.
2. After dubbing, but before audio mixing (dry dub)
3. The screens – if any - that will replace existing ones, before they are edited into the video.
4. The final video file, before DVD authoring, if it’s the case.
5. The final DVD, before authorizing deletion of any – or all – work files.
This will minimize unnecessary rework at all stages. In case of voice-over or hybrid videos, the same process applies.


In this case, there is a first decision to make between two options: a) subtitles “burnt” onto the video; or b) subtitles overlaid on the video (DVD only). It is worth clarifying that my price will be the same for either option (assuming subtitles in one language only).

Subtitles burnt onto the video – the result is similar to that one seen on subtitled VHS tape videos. It is impossible to show the video without these subtitles, or to replace them with others (e.g. in another language).

Key features of this system are:

  • Subtitles are sharper – They are generated at the same resolution as the video, hence they look better.

  • There is no way for an operator unfamiliar with the equipment to show the video without the subtitles.

  • If, for instance, it’s a training film sold in various countries with market-specific prices, it will leave no chance for someone from its original country to buy it at a lower price elsewhere without revealing that when it's played.

  • If the video is short, it is possible to have multiple copies of it on the same DVD – which can hold up to 60 minutes (single-face, single-layer) with high quality – each one with a different set of subtitles, maybe plus one without them.

Subtitles overlaid on the video – the result is what is usually seen on commercial multilingual DVDs. One DVD can contain a video with up to 30 different, individually selectable subtitle sets, including none.

Key features of this system are:

  • It is possible to play the subtitled video in up to 30 different languages, selectable either from a menu (requires authoring) or via the “subtitle” key on the DVD player.

  • Subtitles are mounted on a “transparent video”, which the DVD player electronically overlays on the main video.

  • To save disk space, these subtitles are usually saved in lower resolution than the one used in the main video.

  • The video is always the same, hence it is not possible to replace the charts with others translated (to which language?), unless by fragmenting the main video, plus some relatively complex authoring.

  • If there is no start menu for selecting subtitles, the operator will have to go through all options with the “subtitle” key.

Closed captions – In order not to leave them unmentioned, closed captions are subtitles containing the complete script, in sync, and in the same language as the film audio. Its purpose is to fulfill the needs of hearing-impaired spectators. There is a specific way to embed them into a DVD, however for practical operation reasons, so far I’ve only had requests for closed captions as subtitles, with the complete script of a dubbed video.

As usual, the subtitling process begins with translation. From the hardware standpoint, the process is the same, but the translation technique is different. The goal is to render maximum content with minimum text, so that some time is left for the “reading spectator” to watch some of the action.

For instance:
Original: In my opinion, there is an extremely remote likelihood...
Complete translation: Na minha opinião, há uma probabilidade extremamente remota...
Subtitle: Acho bem pouco provável...

Regarding the subtitles per se, there is a whole set of both empirical and scientifically determined rules on their size, duration, spelling, conventions, and other features, that are beyond our scope here. These rules are not universal, and each subtitler has their own. Sometimes they are constrained by hardware.

One standard rule is not to have more than two lines in each subtitle, ever. The quantity of characters per line varies from one subtitler to another, each one having their standard.

When subtitling was done on video tape, a character generator (CG) was used with a device named Genlock, which overlaid the subtitles from the CG on all video frames. Depending on the CG, there could be an absolute limit for the maximum number of characters per line. This system is still being used, though computer subtitling prevails nowadays.

Anyway, when subtitling is not going to be performed by the translator him/herself, one essential piece of information the translator must have from the subtitler is that maximum number of characters allowed per line. Of course, spectator readability shall prevail as the main criterion, but this limit cannot be exceeded.

There is a very large quantity of software packages for computer-based subtitling, and each of them uses one or more different subtitle file formats. Usually, if translation and subtitling will be done by different people, it is convenient to mention the system/software that will be used for the latter, to avoid rework in reformatting the text.

Once a video has been translated for subtitling, the next step in the process is spotting. It entails specifying the exact time when each subtitle will go on- and off-screen. This may be done either by the translator or the subtitler, if they are not one and the same person. As this work has a cost attached, there is a decision to be made on who will do it. Generally, it is better to have the translator doing it, as if there is any problem in the final output, the translator will be able to modify the text accordingly.

Once the video has been spotted, it’s time to subtitle it… or not. If the subtitles are to be burnt on the video, it is possible to obtain one entire subtitled video, which later – if required (as we’ll see next, in authoring) – may be split into segments. If the subtitles are overlaid, there will be other points to consider.

One reason for overlaying subtitles is to have the same video subtitled in two or more playtime-selectable languages. In this case, it’s worthwhile to exercise some care, to recycle a set of translated subtitles and so reduce costs.

In case you don’t know, translating written text is cheaper than doing it from video. Therefore, if you need a multilingual DVD, it may be cheaper to translate and spot subtitles into one language, and later translate these subtitles into the other languages, using the same spotting. It may be necessary to make some minor adjustments in the spotting if the languages belong to the same root.

For instance, if it’s for subtitling, a video in English may be translated into Portuguese, and spotted. Next, the subtitles may be translated into Spanish, Italian, French, Romanian, and/or other Latin languages, preserving the same spotting. After that, only a revision will be needed, making occasional adjustments to the spotting.

The same could be done for German, Dutch, Swedish, and/or Norwegian, however it is not worthwhile to reuse subtitles for languages with diverse roots. For instance, if a video spoken in Portuguese is to be subtitled in English and Spanish, it’s better to have two separate translations directly from the original video.


A feature of video tape (typically our defunct VHS format) is linearity. Usually, it is played from the beginning to the end. There is no option menu, and the only way to skip some parts, or to show only certain segments, is to reset the counter in the beginning, and to rely on notes taken beforehand. VHS took long until it shifted from number counters - that varied from one player to another - to clocks that indicated relative time over the tape length.

Training videos - and 16mm films before them – often had charts saying “Stop tape/film now” for exercises, discussions, activities, etc. The instructor should then press the STOP – sometimes the PAUSE - button at the right moment, to check the participants’ response to a certain question.

One problem of such linearity in institutional videos, especially product presentation ones, was the unavailability of offering options at any time.  So, if there were a general introduction, followed by an array of different presentations, each for e.g. a specific market or product line, the solution was to make different tapes, all with the same intro, and each one with a different ending.

It’s quite possible to produce a 100% linear DVD, the cost is lower, but no benefit will be derived from its non-linearity, nor from the possibility of including menus in it, beyond those used to select soundtrack (if dubbed) and/or subtitles.

DVD authoring is the name given to the assembly of a whole interactive navigation scheme. There are two basic strategies for navigation.

The first one is simpler, and does not involve so much authoring. It uses one linear video, with various entry points, or chapters. In spite of being menu-less, it allows for jumping to the next point any time during exhibition. An opening menu allows to start playing from any chosen entry point.

However there are no exit points. Making it clear, once an entry point is selected, the video will be shown from there to its end, unless the operator presses the keys for STOP, MENU, or "skip to the next chapter". Any stop will require a manual command.

Authoring offers us many additional resources. The first step is to break the video into various self-standing segments, whose independence makes some sense. Each option in the start menu will allow to: a) go to another menu (with a different array of options); or b) play a preset sequence of segments, ending at the same or another menu.

This allows countless possibilities. The traditional “Stop tape here” in training videos may be replaced by an automatic stop at a menu, offering the presenter options to “resume”, go to some options-loaded menu, or any other. The institutional video having the same introduction and various ends for different audiences may be programmed at the outset, presenting always the same (and only) introduction, then the desired option and, if there is a common ending it may also be programmed to be shown without user intervention.

If you have a linear video, even if it’s on DVD, probably from a VHS tape (or even if it’s still on tape), and you want – while making its dubbed or subtitled Brazilian version – to develop an interactive scheme, it will be my pleasure to offer you guidance.  The only caution in DVD authoring is not to change plans after they have been drawn, since this will unnecessarily increase the overall cost.


Once translated, dubbed or subtitled, and authored, your DVD will be ready for use. If you need it mass-duplicated, it will seldom be worthwhile to do it on your computer. There are specialized companies that offer this service with both speed and quality for an attractive cost, labeling included.


I tried to give some general ideas for anyone having a corporate video, and in need of some guidance for their cost/benefit-based decision-making process. I certainly oversimplified several issues, at the risk of some technical inaccuracy. However there is no point in going too deep, as each case will be unique.

I’ll be pleased to analyze your specific case and offer options. Be welcome to contact me by clicking on E-MAIL in the menu on the left side.

Back to content