Q: Does it cost anything to use the Lab?
A: It depends on who you are and what you want to use the Lab for. Read more about this in the Policy documents (New window).
Q: Do you give a course in Powerpoint?
A: No, the Lab does not give courses on how to use Powerpoint, and we have limited possibilities to develop such a course. We have produced a guide on how to use Powerpoint to make a poster. It is not adapted to the latest versions of Powerpoint, but the principle is the same for the later versions.
Q: I have 10 hours of recorded speech. How long does it take to transcribe it?
A: It is hard to say. If the recording is of good quality, and there is only one speaker, you can expect about one hour per recorded minute, for a transcription that can be used for linguistic coding and analysis. Then you usually get all the words, repetitions and pauses. A phonetic transcription takes considerably longer time, and if you choose to make a brief recording of the contents it may take 30 minutes instead. If the person speaks very quickly, or if there are multiple speakers, and if the speech is overlapping, the time estimation might multiply. If the transcription is later to be coded (e.g. morphological, syntactic, semantic), then expect another hour per recorded minute and coding.
Q: What is the best automatic transcription software?
A: Unfortunately there are currently no software that automatically and with certainty transcribe what someone is saying. Human speech is simply too complicated for that. This can be compared with one of the most frequently used software for speech recognition (Dragon Naturally Speaking) requires the user to train a lot with the software, and speak very clearly, to be able to use it for dictation. The website for the transcription software Transana, presents a test where they tried to automatically transcribe a recording of very good quality. The result (not very successful) can be read here: Transana website (New window).
Q: What transcription software are there?
A: In the Lab, we have primarily given courses in two software: CLAN/CHAT and ELAN. Both are free and work on several platforms (Mac, PC, Linux) but they are developed for different purposes. CLAN/CHAT is primarily developed for children learning languages, and is well suited for studies aiming to, for example, calculate the frequency of words, morphemes, syntactic structures, or lexical measurements such as lexical density and diversity. It is also relatively easy to set your own standards in terms of level of transcription (the desired orthographic/phonemic level), and add preferred coding. You can link audio and video material in CLAN, and it is suitable for many different languages.
ELAN suits multimodal data that you want to encode, for example processes happening simultaneously (such as speech, gestures, interactions, eye behavior etc.). It is less suitable for word calculations etc.
There are also many commercial software, for example Transana: Transana website (New window)
Q: Could you help with transcription?
A: No, unfortunately not. But we can often provide contact with students who have taken our transcription courses, or have followed the course Working with language corpora. Many of them have since worked extra as transcribers. Contact firstname.lastname@example.org for more information.