Availability of texts on AU's learning and examination platforms

There are several ways to make sure that texts on AU's learning and examination platformscan be read by text-to-speech software.

• Use an original, digital and accessible copy of the text. This is the easiest and best solution, and it can be done by:
  • inserting a direct link to the text/article on Archives or other external research libraries. These will often be accessible.
  • downloading an e-book/article in PDF format from the publisher or, for example, from the Royal Danish Library.
• Use HTML or PDF format. If an original digital text is either set up in HTML format at the publishers/library or saved in PDF format, it retains its accessible digital format, and it can be read immediately by text-to-speech software. Using the original digital text also optimises the experience for the text-to-speech software user, because the text is not exposed to the mistakes that can occur during the OCR process – see below.
• Scan a printed text and convert to a text file. It is possible to scan a printed text and subsequently convert it to a text file with OCR (Optical Character Recognition) software, but this is not an optimal solution. The result will always be worse for the user than using the original text. This is because the visual conversion of text/letters in the image is heavily dependent on both the quality of the copy and the layout and content of the text.

As a general rule, if your text can be marked and copied, it can also be read aloud by text-to-speech software.

If you have scanned and subsequently converted a text to PDF format using OCR software, please check the quality of the OCR process (read more below). You can do this by copying the text from a PDF document to a Word document. In Word, you can see the text that will be read aloud. The layout of the text in Word will often differ from the layout in PDF, but the main thing is that the text itself is the same. If there are mistakes in the text, these should be corrected in the OCR programme during the OCR process or, if necessary, subsequently in the same programme (read more below).

Text layout

The OCR conversion process works a lot better on pure text, such as prose, as opposed a text with a complicated layout, for example a text that combines images, formulae, mathematical symbols and forms, etc. For some texts with a complicated layout, the OCR conversion process will be impossible and will not provide an acceptable text-to-speech solution.

In order to ensure the OCR software can recognise the text, please scan a ‘clean’ text, without highlighted passages or notes, etc. You should also avoid scanning the text in a way that leaves dark marks on the page (for example, along the spine of the book).

Programmes

Several products offer OCR software. These include:

• Adobe Acrobat Pro, which all AU employees have access to. When using pure prose text, this programme will usually provide an acceptable result. It also offers some, albeit limited, options to correct and improve the OCR process during and after scanning.
• ABBYY FineReader Pro. This is often a better alternative to Adobe Acrobat Pro, because it allows you to optimise the quality of the conversion by, for example, selecting problematic sections of texts (such as mathematical formulae) and saving them as images. These sections of text will not be read aloud, but this is preferable to them being read aloud incorrectly as a result of an inaccurate OCR conversion.

Guidance on OCR conversion

On  www.ordlab.dk, you can find short video guides on text-to-speech programmes, scanning and ABBYY FineReader Pro. All staff and students at AU have access to OrdLab via WAYF. 

Log in via the link below and read i.a. about the reading program and the OCR program that dyslexic students will be granted if they receive SPS support:

• AppWriter Windows
• FineReader Pro