Sourcing the Crowd

There is a lot of buzz around crowd sourcing but most proponents of utilizing crowds for everything from opinion to transcription are missing the point. Crowds are used to replace computers that are inefficient at coping with human inflection. The quality of the crowd is key. A crappy crowd is as efficient as a crappy computer program. Some humans are as inept at understanding subtleties in speech and many are incapable of translating speech in a way traditionally referred to as phonetic transcription. Bad crowds are no better than computers at comprehending inflection or the grammatical elements required for speech segmentation.

Transcribers have struggled for years in their attempts to develop accurate transcriptions that authentically reflect the speakers stressors or inflection by developing phonetic transcriptions systems using symbols. Writing systems of the world fall into different groups. In any language, English, for example, the position of the stress in the word may need to be indicated. This is frequently achieved by placing a raised mark before the stressed syllable or by highlighting the syllable in some way. Extra length can be shown by doubling the segmental symbol, or by adding a diacritical mark. Pitch is also an essential feature of the words of ‘tone languages,’ such as Chinese, Thai, and Yoruba, and may be marked by accents over the vowel. The challenge for computer programmers is how to utilize these tools of writing so that a computer can understand these quirks of human language.

To add further complication to an already difficult task, speech in its physical form is a continuum, but transcription requires it to be split up into segments, on the basis of some kind of linguistic analysis. This is achieved with paragraphs, commas and so forth and these enable the reader to understand the inflection of the speaker. Computers have not been able to recreate this segmentation without the help of humans. To date the obstacle has proved too great and computer transcription has not been produced efficiently or accurately.

Enter the crowd. Some astute businesses have, in recent years recognized that there is no satisfactory replacement for human transcription using computers anywhere in the near future. Some have provided a service that can be described as excellent if one can afford a dedicated transcriptionist of high caliber. Those who cannot afford the high rates of personal transcribers must often suffer substandard (equal to computer quality really) services or, as is most often the case, transcribe their own recordings.

We think this has a lot to do, not with crowd sourcing per se, but with how these companies have been sourcing the crowd. Recognizing this fundamental fact places crowd sourcing in a completely new light. It is integral to the success, success being accurate transcription fast, of any crowd-sourced service.