In general, when submitting a translation project the client requests a mirror image of the source document to be delivered. This includes the formatting, the design, as well as the images containing text. Translating the text in images can be an easy or complicated task, often depending on how the PM carries out the project.
There are different ways to go about this and it will vary according to the quantity of images and the length of the project. If there are words or phrases spread across an image, the most convenient process would be to include all the images in a PDF, and request the translations to be added in small notes directly attached to each image. As such, each image would be accompanied by its translation. Considering the fact that designers are not fluent in every language we work on, this process eases their job by allowing them to easily recognize the text belonging to each image. Something important to remember is that in order to provide text consistency, when submitting the image’s text for translation we must provide the same translation memory that was used throughout the rest of the translation process.
On the other hand, if the text in the images is organized into columns or tables for example, the best way is to create an editable file where we can provide the source on one side and add an empty space on the other for the translator to fill in the proper equivalent. The positive side to having an editable file is that the use of a translation tool is still possible, thus consistency is assured.
The method that we use, whether it be through notes on a PDF or on an editable format, is not important. What does matter is the consistency we must provide to the text, regardless of how we decide to go about it.