Depending on the signal format used in the corpus you will need a simple tool to replay parts of the signal, optimally in conjunction with a rough signal display where the transcriber may mark parts of the signal and listen to them via headphones. If you are working on multi-channel transcripts (for instance the dialogue between two persons recorded with two separate close microphones), it will help to identify overlapping speech by displaying both signals synchronously on the screen.
To enter the text use a simple text editor that does not perform any automatic formatting functions. As mentioned above, in complex transcription schemes you may consider using an editor with built-in parsing and/or high-lighting capabilities e.g. xemacs). You may speed up the transcription process by using hot-key functions to insert marker strings, to automatically transform digits and numbers into strings etc. To reduce the number of typos in the transcript the editor might use a built-in spelling test that keeps an adaptable list of words.
Please also refer to section at the end of this chapter for the WWWTranscribe tool that might be useful if you intend to work with a distributed transcriber group over the Internet.