Specialized teams in international companies produce large volumes of content for distribution in various languages. Efficiency has become the mantra, with increased pressure to make content production faster and to optimize costs.
Translation technologies have actually been around for about 70 years. Since then, such tech has fundamentally changed the translation industry: enabling fast, consistently more accurate, and cost-effective translation of larger content volumes.
You will learn:
- how technological developments have changed the translation industry,
- the advantages of innovative technologies for agencies, translators, and clients,
- how automating the translation process saves time and money.
What are translation technologies?
Translation technologies include everything from:
- translation memories,
- terminology management tools,
- machine translation,
- translation management tools,
- connectors for easier system integration,
- virtual interpretation technology, and
- speech-to-text technologies.
Nimdzi has been mapping market developments for several years. In 2021, over 770 tools appeared in their extensive overview. Compared to 2020, the number increased by over 110 products.
Although the roots of some translation techniques that are still used today date back to the 9th century, the development of translation technology rode the wave of the first computers in the mid-20th century.
Until then, translators’ work had relied on manual browsing of dictionaries and striving for the best translation according to their knowledge. Without the help of technology, it was difficult to track errors, and maintain consistency of terminology, style and tone. The delivery of outputs took much longer, which ate into a major part of the budget for companies.
Translation technologies have revolutionised:
- project momentum, enabling higher translation volume faster and shortening delivery times,
- computerized quality assurance to more easily detect grammatical errors,
- translated text memories increase translation consistency,
- such memories automate work that was previously manual.
The use of CAT tools tends to give rise to the general public's false assumption that the translation is done by a computer rather than a human translator. Yet such tools actually serve to help translators.
CAT tools’ numerous useful features include:
- creation of translation memories that translators can reuse if the need arises,
- creating terminology databases and glossaries to maintain translation consistency,
- the translation is conducted in a unified environment, regardless of source document format (Word, Excel, HTML, etc.), enabling translators to concentrate on the work at hand,
- search for matches and automatic suggestions of corresponding translations,
- quality assurance,
- integration with other technologies and software.
Read also: CAT tools as excellent tools for translators
CAT tools work on the principle of creating a translation memory after each translated text. They divide the source text into individual segments = translation units. In the user interface, the translator sees individual segments in a clear grid. On the left are the source texts, and on the right are the cells where the translator inserts target texts. The CAT tool stores such translated texts in the translation memory as pairs, which can be edited again.
After storing the translated pairs, CAT tools can identify and suggest partially matching translations in the translation memory. The translator sees information about match accuracy, for example ‘100%’ if the source text has already been translated in the same wording. For lower percentage matches (99-75%), the saved text must be edited manually.
Matches and repetitions are also reflected in the client’s price offer. The more extensive the translation memory, the higher the probability of discovering matches and the lower the translation price.
Read how we achieved cost savings of 22% thanks to ZAJO translation memories.
A termbase is the set of terms that a specific brand habitually uses. Based on the created database of commonly used terms, translators receive automatic translation suggestions for individual terms. This maintains translation consistency and shortens delivery time.
Some companies have a pre-prepared list of specific or sensitive terms that need to be reflected in a translation. A glossary of terms can include common industry-related expressions, common terms, or even slang.
The creation of a glossary will ensure the correct expression and, like a termbase, contributes to maintaining translation consistency. A term used from the glossary then becomes part of the translation memory.
Machine translation enables the automated translation of content without human intervention. It works by encoding and decoding human language so that a computer can understand it and reproduce it in another language.
Until a few years ago, machine translation outputs were unreliable and were mostly used to translate individual words or short phrases. Ever more resources are being invested in the development and improvement of machine translation tools. Current translation systems perceive the whole (sentence) rather than individual words, which leads to better translation quality.
Machine translation’s development
- Rule-based machine translation(RBMT) was the first commercial machine translation system. It translated the grammatical structure from the source language to the target language based on a set of rules and extensive dictionaries. It also enabled users to suggest modifications and thereby improve translation quality.
- Statistical machine translation (SMT) worked on the principle of comparing parts of a sentence with the most frequently translated results from the corpus. Until 2016, this method was used by Google’s sentence translator. Like other machine translators, it had a problem with correct grammar and phrase context in sentences.
- Compared to its predecessors, neural machine translation (NMT) generates outputs that tend to be more grammatically accurate. AI notices a word, its position in a sentence and surrounding words, and can thereby produce better translations of longer units.
See also: The present and future of machine translation
Raw vs. edited machine translation
The quality of resulting machine translation is influenced by several factors. These include the similarity of sentence structure of the source and target languages, and the style/range of text. Although neural machine translations produce increasingly better outputs, they are still far from human-quality. For some texts – especially professional and creative, neural machine translations cannot replace traditional translators.
- Raw machine translation without the intervention of a translator can be used for quick translations of specific types of text that you need to quickly understand, or for internal purposes. (For example: recipes)
- A partially or fully edited machine translation is suitable for checking and correcting inaccuracies that may appear in a machine translation. (For example: standard order process via an e-shop)
- Choose a traditional translation for documents that require top quality and exact adherence to terminology, or creative texts where the message needs to be adapted to readers.
See also: Interest in creative translation is growing – leading to attractive specialization
Translation management systems
Translation speed is not only translators’ responsibility. A translation agency also contributes by creating optimal working conditions for the best project management, and facilitating client communication.
Translation management systems (TMS) help plan, allocate and manage translation projects from inception to completion.
Although several translation management systems are available, over time we reached the stage where it’s more beneficial to create our own solution for clients and translation suppliers.
This enables clients to get an overview of current and completed orders. In one account, the client can access the status of current requests and orders, agreed delivery dates, assigned project managers, invoices and payments.
A client can upload documents for translation at any time, we receive a notification, and work begins. So all info about projects is available in one place.
Translata Vendor Portal
This facilitates communication with translation suppliers. It reduces administrative burden and speeds up invoicing, so suppliers can better plan their new orders.
They monitor, track, and check all information related to projects and invoicing.
Connectors for system integration
Connectors enable links to be created between digital platforms and translation software. The integration replaces some manual tasks, and enables users to quickly and securely send/import translations directly from the system, which is very time-saving for multilingual content.
System integrations with translation tools deliver numerous benefits, such as:
- faster translations,
- eliminating errors when copying files,
- reducing costs by automating manual tasks,
- creation of translation memories facilitates easier overviews,
- secure encrypted data transfer (in contrast to unsecured e-mails and file forwarding).
Automation of website translations via HUB
Although some website content management platforms allow integration with CAT tools to eliminate tedious manual content copying, this is mostly not the case for e-shops. That’s why we’ve proposed our own way to help e-shop owners simplify the preparation of translations.
A client needs to invest time at the beginning of the cooperation by implementing the library in its website code and connecting to the HUB. Yet that initial implementation is balanced by time savings in subsequent cooperation.
The whole process works automatically. It recognizes that a new text has been added to a page, sends it to the CAT tool, and notifies translators about the new project. After the translation has been completed, it sends the texts back to the website.
See also: Why does a translation agency need an IT department?
Connecting CAT tool to client repository
A useful feature of CAT tools is their integration with other systems. In this way, a link can be created with various repositories where clients most often collect documents for translation.
Automating the transfer of documents eliminates the amount of e-mail communication, reduces the risk of errors that may arise during manual copying, and the process is less dependent on project manager’s responses. Its role lies pretty much in overseeing the process, hence the resulting translation price is also lower.
An advantage is the option to adapt the degree of automation to reflect clients’ need for a certain degree of project oversight.
Read: CAT tools can relieve the burden of assigning translations
Transformation is evolution
Useful technologies are a great boost for both translators and agencies. Companies that approach the transformation with an open mind have an advantage. Whether or not to use translation technologies is no longer a relevant question. The point is to understand your needs and then decide which tools will best deliver your goals.
More efficient work management raises the standard of translation services, and also strengthens cooperation with clients.
Technologies for automated translations
Current market needs are directed towards maximizing the use of technologies. Innovators respond to market demands and propose new ways to automate sub-tasks and achieve results faster. Various types of software support the automation of translations, with processes tuned to optimize translation quality and speed.
This essential aid helps translators meet the constant pressure on translation delivery and quality.
The 1990s was the key decade for computer-aided translation, during which time most commercial CAT tools were created. CAT tools fundamentally changed the translation process.
Companies and translation agencies could process large amounts of content more efficiently, which saved considerable money and time. CAT tools relieved translators of monotonous repetitive tasks, allowing them to concentrate on the translation at hand.