Machine translation ≠ creative localisation
A translation performed by a machine is not comparable with the creative localisation of texts. The machine is not able to translate creative slogans and advertising copy, since word-for-word translations are not usually suitable in these cases. In addition, when the aim is to create content that speaks to your target group and takes cultural and regional linguistic features into account – in a process known as localisation – then a native translator is practically essential.
However, if texts do not need to be adapted to specific markets or if a large volume of words needs to be translated in a short time, then machine translation is a good option. This can include help documents, forums and live chats, internal documents and communications, notes and drafts, manuals etc.
Post-editing is key
Post-editing – reading, correcting and optimising the machine translation – carried out by a specialist translator who is a native speaker of the target language ensures the best quality standards for your machine translation projects. However, it is not always easy to identify right level of optimisations during this process: How much manual work is appropriate?
This depends largely on what the text is intended for. If the text is purely for informational purposes, then the facts must be 100% correct and the translation must not contain any misunderstandings. This is the only level at which the translator intervenes when post-editing this kind of translation, as the style and flow of the text are not a priority.
The situation is different when texts need to have a particularly high-end style – such as company publications. In this case, not only does the content need to be accurate, the style is also important. A corporate language must often be observed. This can mean that sentences and full paragraphs often need to be completely re-translated.
If the time required for the post-editing process increases significantly, it makes sense to weigh up whether a specialist translation by our translation experts would be a better alternative in terms of the cost-to-benefit ratio. Even though machine translations have greatly improved with the use of neural software, they’re still not perfect. Although grammatical mistakes occur less frequently these days, it is still not uncommon for the meaning or context of the source text to be misinterpreted in machine translations. Therefore, it is important not to lose sight of the cost-to-benefit ratio mentioned previously.
Here is what a professional post-editing stage should cover:
- Correcting terminology
Termbases ensure consistent corporate language. If this is not programmed into the machine translation algorithm, then the use of corporate language must be checked during the post-editing stage.
- Adapting the context
When the context of the source language is not recognised by the machine, a greater amount of work is required during the post-editing process.
- Harmonising formulations
Expressions and linguistic conventions are not uniform: The form of address changes from formal to informal or expressions are not translated the same way throughout. These are good improvements that can be made during post-editing./li>
The bottom line?
For our machine translation projects, we work with professional, adaptive machine translation engines that can be trained. With good planning, consistent processes and texts, as well as for very large volumes, machine translation is an effective tool for translating content into different languages.
Get advice from our experts.