6. Language Settings, Speech Recognition, AI, Grades and Reports

The four Poodll Languages activities share several features. This section explains the settings that matter most across all four, so that you can choose them confidently and interpret the results sensibly.

The target language is the language students are learning. It affects speech recognition, the voices available and the language AI uses when generating content.

Choose the variant closest to your students’ model of the language, such as English (GB) or French (Can.). More than 50 language variants are available, though not every feature is available in every language.

Your administrator sets the default for new activities, so check this setting each time you create an activity.

Figure 6.1. The target language setting in each activity.

Students often understand definitions, hints and feedback better in their first language. Each activity handles this a little differently.

A student’s language choice can apply across activities. For example, the definitions language a student chooses in WordCards is shared with MiniLesson as their native language. Whether this choice overrides your activity setting depends on options that an administrator controls.

Figure 6.2. A student choosing their definitions language in WordCards.

Translated definitions are available for words added with the Word Wizard. If you type a definition by hand or import it, students see the text you entered.

An administrator can turn on Pre-translated hints and definitions, so that hints and card translations are prepared in advance in chosen languages. AI makes these translations in the background when an item is saved, and the item form shows them so that you can review or regenerate them.

Speech recognition, also called speech-to-text or STT, turns a student’s speech into text so that it can be scored. MiniLesson, WordCards and ReadAloud let you choose between two modes.

In ReadAloud, you can make a separate choice for the Practice and Read steps. In Solo, students speak freely, so there is no expected text to guide recognition.

In some cases, speech can be recognised as the student speaks, so that results appear almost immediately. In Solo this is the Use streaming recording setting. In ReadAloud, your administrator controls which steps use realtime recognition with Open STT.

Text to speech (TTS) means you do not have to record your own model audio. Voices are listed by language. Their quality and features vary, and some have symbols in the list. In ReadAloud, for example, + marks a better-quality voice and ! indicates a voice that cannot create line breaks automatically.

Teaching tip: Listen to several voices. For beginners, a slower speed, such as Slow or Extra slow, may be more suitable.

Keep these points in mind:

• Review generated content before students see it. AI can produce errors and content at the wrong level.
• Treat automatic scores as evidence, not as the final word. Speech recognition errors affect results, especially with background noise, strong accents and names.
• Check high-stakes results. Listen to a sample of recordings, and adjust grades where you can.
• Tell students how results are produced. They should understand that a low score may sometimes reflect recognition rather than their ability.
• Follow your institution’s AI and privacy policies.

Your site may let administrators choose which AI provider handles some tasks, or restrict these features. Contact your administrator if you have questions.

All four activities can send a grade to the Moodle gradebook.

If you don’t want a gradebook entry, set Grade to None in the activity’s settings.

• What languages does Poodll support?
• Speech Transcription in Poodll ReadAloud
• Definitions in User’s Native Language
• AI Manager
• Cloud Poodll and Data: Q & A