Machine Translation Engines
It’s hard to imagine our globalized world without machine translation engines. The technology changed, developed and improved. MTs started as rule-based and statistical to neural engines.
In this article, we will cover some of the most well-known machine translation engines. Check out our overview and pick your preference.
Probably the most used machine translation service, Google Translate covers 103 languages. Google Translate started as a statistical machine translation service in 2006. A decade later, Google presented a neural machine translation system. This improvement is a solution for the inaccuracy Google Translate is still infamous for.
Using Google Translate for keyword translation might prove challenging. This is mainly because the localization process in the engine is not refined. These nuances are obvious for languages that are globally used but rely on local dialects. A good example is Spanish.
Despite these downsides, Google Translate is pretty simple to use and does offer more than a hundred languages. Although it’s free for the average user, getting the API will depend on your needs. The limit for free translation is at 2 million characters per month, but there is a fee for more.
Microsoft boasts its machine translation service with a support for more than 60 languages. It was based previously on statistical machine translation. Nowadays, there’s a great improvement in the direction of neural machine translation.
What do these changes mean? It means that we can expect Microsoft translations to be much closer to the natural language in the future.
Meanwhile, Microsoft Translator tech is already powering products ranging from Microsoft Office and Edge, to Bing and Skype. Since it’s an open cloud API, Microsoft Translator is easily integrated cross-platform. It’s available both free and under a subscription. However, the API for business applications requires a subscription plan.
We wrote about DeepL on LingoHub before, but we’re going to switch our focus in this article for a bit. Let’s see why DeepL is a worthy opponent to industry giants like Google Translate and Microsoft Translator.
For starters, DeepL only covers 7 languages and 42 translation combinations. This is a surprisingly small number compared to Google Translate or Microsoft Translate. However, DeepL uses AI technology, strongly focused on the deep learning systems (where the name comes from).
This is where DeepL shines. The foundation is Linguee, an online multilingual dictionary, operated by the same company. DeepL learns from the dictionary and high-quality texts. This means, you probably won’t get a word-to-word translation with DeepL. This engine catches the nuances of the language and provides a more natural translation.
DeepL is opting out of a wide scope of offered languages for translation. Instead, they are focusing on a small number of efficient translations of high quality. This strategy might prove as the winning one. More natural translations cut costs, especially for those who usually choose to invest in quality checks.
DeepL Translator is available free, while the DeepL Pro, a business-oriented version, offers the API for a fee. We’re excited to see what will be their next step of development.
This lil’ translator is mainly used to translate short text and web pages. It’s based on statistical machine translation, although a transition to neural network started last year. This change should provide users with a higher quality of translations in the future. At the same time, that is where this engine’s significant disadvantage lies. Occasional crude translations are there, and it works better for some language pairs than others.
At this moment, Yandex.Translate covers over 90 languages. Among offered languages are Tolkien’s Elvish and Emoji. That’s right – if you’re ever in a situation where you need an urgent understanding of ancient Tolkien scripture, Yandex has got you covered.
Developers can get Yandex’ API key for free here, but remember that there is a character limit without a subscription plan.
Apertium is an open-source machine translation platform, available in GitHub. This rule-based free software offers translation for 40 language pairs.
It started as a project developing in Spain, with several Universities and the government involved. This is obvious in the very translator – several specific Spanish dialects were targeted for the very beginning.
The variety of dialects and versatility of an engine like Apertium might give it a cutting edge in the future. We will make sure to follow its progress.