Speaker 1: Hello, I'm Emily McKenzie from Webcertain Translates and welcome to our first tutorial in a series which examines the use of Computer Assisted Translation Tools, or CAT tools for short, in the day-to-day running of a translation company. This series aims to outline the key benefits of using CAT tools from a client perspective and will show you how you can maximise their use in your own translation workflow. This first tutorial will introduce the use of translation memory and the benefits of creating and maintaining a translation memory for each of your required languages. So what exactly is translation memory? Well, it's just one of the tools used by translation companies and freelance translators worldwide. It's basically either a bilingual or multilingual database which contains all the strings of translated text for a particular client over the longevity of their translation needs. A translation memory is created for each required language combination and all new translation requests are first analysed against the memory in order to generate leverage from only pre-existing translations which may be the same or similar to the new text. For these strings, the client often receives discounted rates from their supplier and it will also speed up turnaround times in the long run. So let's have a look at an example of a piece of text taken from a technical manual to be translated from English into German. This is a simple example which we will use to illustrate the cost and efficiency benefits of using a translation memory as part of your own translation process. The CAT tool we will be using today is STL Trados Studio 2014. Here is the main welcome page which you see when you first open the programme. So before creating the translation project, let's have a quick look at the files. For this simple example, there are three Microsoft Word documents to be translated. As we can see, they do not contain any images, tables, graphs or any other pieces of content which would need to be edited or prepared prior to translation. For this example, we are going to run the three files through an existing translation memory to illustrate where leverage can be gained from content reuse. So to begin a new project, the translation project manager will select new project from the welcome page. In the pop-up screen, the first thing we need to do is name the project and then define

Speaker 2: where the project will be saved.

Speaker 1: Next, we select the source and target languages from the available drop-down options. For this example, the source will be UK English and the target will be German for Germany. We then select the three word files for translation and add them to the project. On the next screen, we add the existing translation memory to the project by selecting the relevant language combination. Therefore, the memory we will use for this project is from UK English into German for Germany. The next couple of screens are not relevant for this particular project, so we'll skip through those for now. And let's stop at the analyse files setting screen, where we are going to activate the report internal fuzzy leverage option. This will provide us with an analysis of the word counts that also takes internal repetitions across the three selected word documents into consideration. For example, if the same or similar sentence appears in all three files, then we want to make sure that this is reflected in the analysis and charged to the client at a discounted rate. The final stage is to finish the project, which is now being created by Trados Studio. Once the project has been created, we can look at the analysis of the word counts by going to the reports area of the program, accessible from the toolbar on the left-hand side of the screen. Here we can see the breakdown of leverage generated from the translation memory. We'll look at this in more detail a little later on. If we look at the total number of new words analysed, which do not appear in the memory, we can see that there is a total of 411 new words across the three documents. The remaining 420 words have been recognised somewhere in the existing memory, with a total word count of 831 words to be translated. Let's save this analysis for comparison later on and move on to the second example, where we will analyse the same three word documents but without using a translation memory. Like before, we go back to the welcome page and create a new project. Again, we name the project and define where the project will be saved. Next, we select the same source and target languages, UK English and German for Germany.

Speaker 2: We add the same three word documents.

Speaker 1: And at this step, unlike last time, we are not going to add a translation memory to the project. This means that the files will be analysed on their own as if they are completely new content to be translated. This is the usual process when a client is translating into a particular language for the very first time and a translation memory is yet to be created. Again, we shall skip the next couple of screens and go straight to the analyse files settings as before. We mark to include the internal FuzzyMatch leverage and finish the project. Now, if we go to the reports area, we can look at the differences in the analysed word counts without using an existing memory. Here we can see that as no translation memory was available, no leverage or content reuse has been found. Instead, for the same three word documents, the number of new words calculated has increased to 816 new words out of a total of 831 compared to the total in the first example, which was 411 new words. Finally, let's save this second analysis and compare it side by side with the first one. Let's first explain what the percentages mean in the two analyses. The percentages ranging from 75 to 99% are known in the translation industry as FuzzyMatches. These can appear both in an existing memory and internally across files. A FuzzyMatch is picked up when the CAT tool finds a string of text which is similar to that previously analysed. It's not exactly the same, but there is a particular percentage of similarity, hence the range identified here. 100% matches are strings of text which are identical to a previously translated string, but the context in which they appear in the new files may be slightly different to that in the memory. Therefore, the translator still has to work their way through the file to ensure that the string of text fits within the right context and adapts as necessary to accommodate this. As previously mentioned, new words are strings of text which do not appear in the translation memory, and the CAT tool considers them as being completely new sentences to translate. An example of a new word would be the string, the cat sat on the mat, the very first time it appears in a text. A FuzzyMatch would be, the dog sat on the mat, where the sentence is very similar to the pre-existing string, but there is a slight difference. And finally, a 100% match would be, the cat sat on the mat, the second time it appears in the text. If we go back to compare the two analyses, on the left we have the analysis from the first project, which includes the translation memory, and on the right we have the second analysis without the memory. If we look at the analysis which included the translation memory, we can see that both Fuzzy and 100% matches were found within the content of the memory. The total number of new words is around 40% of the total number of words to be translated, meaning that the remaining 51% has been leveraged from the memory in some way. Here the use of a translation memory is of benefit to the client as they will only be paying the full translation rate for 411 new words, meaning they are making a considerable saving in terms of cost. In terms of productivity, by making use of a translation memory, the translator is able to process translations more efficiently and quickly as they do not have to translate everything from scratch every time new content is sent for translation. Let's compare this to the second example, where we did not include a translation memory. This time we can see that no leverage was found in terms of Fuzzy and 100% matches. We can see that by selecting to include the internal Fuzzy matches, the CAT tool has identified that across the three files to be translated, 10 words appear as internally Fuzzy. And again, the client will receive a slight discount based on this analysis. If we compare the number of new words analysed, this jumps up to around 98% of the content being considered as new. The client will therefore have to pay the full translation rate for 816 words out of the total 831. This is considerably higher than if they had an existing memory for this language combination, and a translator will have to spend more time translating this project, therefore lengthening turnaround times. Of course, this second example is often the case when a client is translating into a new language combination for the very first time. Once the initial project is complete, a translation memory can be generated and used to analyse against new translation requests sent in the future. Please be aware that it takes time to build up a substantial translation memory, and the benefits cannot always be seen across the first few projects sent. If your translation requirements for a particular language are considerable, then it's definitely in your interest to ask your supplier to create and maintain a memory on your behalf, as you're sure to reap the benefits in the long run. And with that, we've come to the end of our first tutorial on computer-assisted translation tools, and our introduction to the benefits of using translation memories. I hope you have found this incredibly insightful and provided food for thought when it comes to improving or implementing your own translation workflow. Before I go, I have one final word on translation memory. Although your translation supplier will create and maintain your translation memory, the memory itself actually belongs to you, the client. After all, it is your content, and you can request an export of your memory at any time. If you decide to change suppliers, then you can also take your memory with you, so be sure to bear this in mind. Thank you for joining me today, and stay tuned for future tutorials in this series.