Video to text transcription

All the video uploaded to Archive.org will be accompanied by a spread sheet of metadata. I was thinking to add a field containing the spoken text within each video. Unfortunately most software based - automated  audio to transcript convertors - are not of a good enough quality. This means that they are used in conjunction with hand made transcripts, such as commercial service SPeakertext.   

Other non commercial academic software convertors come with a caveat concerning there accuracy, such as transana.