Transcription (service)

A transcription service is a business service that converts speech (either live or recorded) into a written or electronic text document. Transcription services are often provided for business, legal, or medical purposes. The most common type of transcription is from a spoken-language source into text such as a computer file suitable for printing as a document such as a report. Common examples are the proceedings of a court hearing such as a criminal trial (by a court reporter) or a physician's recorded voice notes (medical transcription). Some transcription businesses can send staff to events, speeches, or seminars, who then convert the spoken content into text. Some companies also accept recorded speech, either on cassette, CD, VHS, or as sound files. For a transcription service, various individuals and organisations have different rates and methods of pricing. That can be per line, per word, per minute, or per hour, which differs from individual to individual and industry to industry. Transcription companies primarily serve private law firms, local, state and federal government agencies and courts, trade associations, meeting planners, and nonprofits.

Before 1970, transcription was a difficult job, as secretaries had to write down the speech as they heard it using advanced skills, like shorthand. They also had to be at the location where the service was required. But with the introduction of tape cassettes and portable recorders in the late 1970s, the work became much easier and new possibilities emerged. Cassettes can travel through internal mail or external mail which meant for the first time, the transcribers could have the work brought to them in their own office which could be in a different location or business. For the first time, transcribers could work from home for many different businesses at their own convenience, provided they met the deadlines required by their clients.

With the birth of modern technology like speech recognition, transcription has become much easier. An MP3-based Dictaphone, for example, can be used to record the sound. Recordings for transcription can be in different media file types.[1] The recording can then be opened in a PC, uploaded to a cloud storage, or emailed within minutes to someone who could be anywhere in the world. Recordings can be transcribed manually or automatically.[2] The transcriptionist can replay the audio several times in a transcription editor and type what he or she hears to manually transcribe files, or with speech recognition technology convert audio files to the text. The manual transcription can be accelerated using different transcription hot keys. The sound can also be filtered, equalized or have the tempo adjusted when the clarity is poor. The completed document can then be emailed back and printed out or incorporated into other documents – all within just a few hours of the original recording being made.

The industry standard for transcribing an audio file takes one hour for every 15 minutes of audio. For live usage, real-time text transcription services are available for captioning purposes, including Remote CART, Captioned Telephone, and live closed captioning for live broadcasts. Live transcripts are less accurate than offline transcripts, as there is no time for corrections and refinements. However, in a multistage subtitling process with a broadcast delay and access to a live audio feed it is possible to have several correction stages and for the text to be displayed at the same time as the "live" transmission.