A rule of thumb for transcription work is that if you are a professional transcriber (or transcriptionist as they are known in the US!) then it is likely to take four times as long as the length of the recording to turn an audio recording into a written document.
So the quick answer to how long does it take to transcribe about 20 minutes of audio? 80 minutes of transcriber time would be the rough rule of thumb.
Assumption of a clear recording
This is based on the assumption that the recording is one or two people speaking, and the audio is clear, no-one is speaking quickly, and the accents are not too strong. This rule of thumb applies to any language around the world where the transcription is being done by a professional transcriber who is transcribing in their native language.
Hard to hear and other issues
Where this rule of thumb does not apply is if the recording is hard to hear, has a lot of background noise, the accents are strong, the recording device is placed quite a distance from the speakers, the type of file is not of sufficient quality in order to render accurate transcription (some types of audio file are better than others), the speakers are talking over the top of each other, the speakers are talking extremely quickly or somebody is mumbling.
Most professional transcribers will tell you that the vast majority of work is never straightforward!
Issues are common
The reason that transcription companies exist in the first place is because most recordings have issues like those above, which means that in order to get an accurate written version of an audio file, the actual time will very much depend on the specific recording. Furthermore, quite a few recordings will have subject-specific terminology that in some cases a transcriber may have been working on hundreds of similar recordings for the same client, and so transcribing is a lot easier, but similarly this may be the first time they have come across some of the terminology, which will slow things down.
Most transcription companies pay for work on a set hourly rate, which is not the hourly rate for the time taken to do the work, but the hourly rate based on the length of audio. This gives both the transcribers and the transcription company certainty in terms of the cost, and enables accurate quotes to be given to clients.
Some transcription companies will charge their clients extra for any recordings that are hard to hear, or they had problems with, or take longer to complete, because of some of the factors affecting the recording as above. Other companies, including ourselves, do not charge any extra for hard-to-hear recordings, or any issues that mean it takes the transcribers longer to complete the recordings. This is because we want to give our clients a set price for all work, so that our prices are set at the start of a project, and they do not change just because some of the recordings are hard to hear. Furthermore we would spend most of our working week informing clients of problems arising and asking for more money, which would of course drive us and all our clients nuts!