Human translation VS Automatic Speech Recognition

Human translation VS Automatic Speech Recognition

Human translation VS automatic speech recognition

In this month of October, we propose you to illustrate a theme to which we attach a great importance: human translation VS automatic speech recognition. Indeed, today there are many online tools and software that can identify texts, words and translate them almost automatically. Nevertheless, despite the ultimate almost frightening power of this process, it should not be forgotten that living languages have subtleties. Especially in terms of expressions. As a result, the translation’s field always requires a human professional perspective. Thus, we are putting the expertise of translation into the abyss here, with special Halloween examples.

Translation into abyss


To begin, we will focus on human translation VS automatic speech recognition. Whereas in the past it was necessary to call human experts to translate everything from one language to another, nowadays, tools save us a lot of time. They recognize the words of a language and offer us their equivalents in a few seconds. However, can a good translation be done as « simply » as that? Is it still reliable? Who prevails when human translation is opposed to automatic speech recognition? This is what we are trying to understand with this article about translation. In fact, translating is presented as the process of moving from « one language to another, tending towards the equivalence of meaning and value of the two statements ». In this first definition, we already notice two major difficulties: meaning and semantic value.

The beginnings of automatic speech recognition

It is a reality that computer systems and software nowadays considerably improve their « understanding » of the complex nature of language. Machine translation began in the 1950s. Warren Weaver has initiated the combination of natural language processing and automated transcription. Today, speed and ease, promised by automatic and reliable speech translation, are largely made possible by many tools.

Speech recognition tools

Here are some of the automatic speech recognition tools we use today.

  • Microsoft
  • Google
  • SIRI
  • Alexa….

Some tools « recognize » our words with systems that analyze the human voice picked up by a microphone. In order to then transcribe it into a machine-readable text. Without entering into the functioning of this « artificial » technology, we understand that thanks to algorithms, some data can be quickly recognized and thus translated in a way that exceeds human understanding. This process has several advantages: it is faster than a human translator in processing requests. It’s cheaper than a human translator. It is more accessible and makes it possible to maintain a form of objectivity, neutrality with regard to translation.

A powerful technology that would almost make you feel cold! But will it replace the professional human translators?

The subtleties of language in the face of machine translation

Nevertheless, although tools are completely revolutionizing the translation industry, speech recognition and machine translation do not have a human analysis dimension. For example, some tools have limitations on the quality of translation. Indeed, if quantitatively speaking, speech recognition and automatic translation make it possible to process a large number of requests, they remain limited in their « understanding ».

As you will have understood, the tools do an excellent job of saving us time in translation. Nevertheless, they are neither complete nor infallible. For a quality translation, human intelligence is required more than artificial intelligence. In order to respond to the subtleties of language, especially when dealing with different languages of the world.

The machine can process data, but it is not currently possible for it to understand the meanings and values of a language (hence the notion of a living language…).

Linguistics and translation

Now we will talk more about linguistics and the heart of translation. That is to say the text’s adaptation, according to the geographical area, the language involved. With its expressions, its sentences constructions… For example, Quebec’s French is not quite the same as in mainland France. Just as American and British English have their differences and subtleties.

As such, translation is a true science linguistic. It includes many parameters to be taken into account: language level, semantics, cultural contextIndeed, each language reflects a whole culture, with its History, its experience, its sounds, its mode of communication and its own expressions… Expressions that cannot always be accurately translated as such.

The Expressions

Idiomatic expressions are a real puzzle in translation and our languages are full of them! While professional human translators know them, it is impossible for automatic translators to « interpret » them.

For Halloween, the famous English expression « Trick or treat » that children say when they go to their neighbours’ houses to collect candy is not well translated into French. Indeed, we literally say here « candy or a spell ». This is an interpretation that goes beyond the initial framework. Indeed, « trick » means « trap » in the sense of « joke » and « treat » is a « reward ». It is generally used for a candy but not necessarily directly for a candy.

Some frightening concrete cases in translation


Finally, for our Halloween special, let’s look at some horror movie quotes.

In the movie « That », there is a quote: « you too can float ». The « it » which means « it » has been replaced by « you ». In this illustration, there is not a huge difference in translation. Except that this type of modification compared to the original version can completely change the meaning for other scenarios.This is a first concrete case of a difficulty in translation.


Then, for the horror film « Freddy the claws of the night », we have the illustration of a second complex case for an automatic translation: song’s lyrics. Here for a nursery rhyme.


Third, for the movie « Scream » another Halloween classic we note that they take a lot of liberties to adapt the voiceover speech in the trailer. Perhaps to adapt the atmosphere to the language into which the film is translated. To respect a certain rhythm and generate certain emotions related to the words meanings, as the viewer receives them when watching a film.


Thus, to be sure to benefit from the alliance of the two, Authôt offers a proofreading service. This service saves you time and gives you a professional and accurate rendering. Thanks to the reliability of our tool, we then focus our efforts on the subtleties of the language.

We work with about a hundred professionals for the different languages offered. We used Halloween to illustrate a powerful theme: that of human translation in the face of speech recognition. With Authôt, you benefit from cutting-edge translation expertise: automatic recognition and human translation with proofreading by our professionals!

test authot application

Authôt. You speak. We write.