Digital Humanities Workbench |
Home page > From source to data > Transcription > Speech Transcription of speechIf you have recorded spoken language, it is usually necessary to transcribe it first before subjecting it to further analysis. This means that whatever is said is represented in text as accurately as possible. Unfortunately, it is usually not possible to automate this process with acceptable results using speech recognition, so this must typically be done by hand. This process has two main activities: listening and typing. Assuming the recording is digital, there are various tools available to support this process:
Note: for some types of linguistic research, transcription does not only involve the representation of speech in written form, but also the notation of various phonological and phonetic characteristics, such as pauses, stress and pitch. The International Phonetic Alphabet (IPA) and the ASCII equivalent SAMPA are common systems for this type of transcription. There are various useful tools for inserting IPA symbols (phonetic symbols) into a document, such as IPAKLICK (only works well in Internet Explorer) and the IPA Console. |
Other topics in this section: Text |