When to use Machine Translation?

The quality of machine translation (MT) has advanced significantly over the past few years. The results are fluent and natural-sounding translations, but fluency is not always accuracy.

The raw MT output, without any subsequent human optimization, suits well to content that simply needs to be understood, without worrying whether it will affect your brand’s image, and in cases where speed is the most important factor. Use cases of raw machine translation are for example low-visibility non customer facing content like internal communication or user-generated content like product reviews.

Generic MT Systems vs. Customized MT Engines

Generic machine translation services such as Google Translate, DeepL, or Microsoft Bing Translator can be very good tools for instant translation and getting the gist.

Customizable MT systems are trained to improve accuracy within a specific field to make the output more reliable. Key to this is relevant and high-quality training data in the respective domain.

CBG Solutions

If you are considering using an online MT service, review the terms and conditions to ensure your data remains confidential. We provide MT systems that keep your confidential content protected and away from the public domain.

Our instant translation tool is the ideal solution when you need quick results to understand the essence of internal communication or non customer facing content.

Moreover, the tool translates all different kinds of file types while keeping the original formatting of the source file.

Light or Full Post-Editing?

In cases where reliable and accurate translation are needed, machine translation should be used with human intervention. MTPE stands for "machine translation post-editing" and is a service that combines the speed of MT, with the expertise and sensitivity of human linguists.

There are two service levels of post-editing to meet the required level of quality. The raw MT output is either “light” or “full” edited by a professional human translator, depending on the particular requirements of the project.

In light post-editing, the editor focuses only on basic essentials ensuring that the translation is understandable and factually accurate.

Full post-editing is a more detail-oriented process where the editor, besides checking the spelling, grammar, syntax and punctuation, aims to provide a completely accurate translation with consistently correct terminology and appropriate tone and style.

When to use Human Translation?

The quality of the neural machine translation output depends on the quality and domain of the source text, the language combination and the amount of translated material available as training data. Leveraging machine translations relies on these parameters and the required editing effort.

When the best translation quality is required, a human in-the-loop is essential, as machine translation is not able to render style, nuances or idiomatic expressions, marketing content and creative texts. If life or law are involved, using human translation might also be the better choice.