Questions & Answers

02.09.2010

How long a transcription takes depends on your typing speed and the complexity of the transcription rules you are following. It is therefore difficult to say exactly how long it takes. The top speed we have observed was a type-up three times as long as the recording being transcribed. This record was achieved under the following conditions: a very rapid typing speed of over 250 strokes per minutes using the ten-finger system, USB footswitch support, a good quality recording with clear sound, simple transcription rules.
On average the transcription duration for a text with simple transcriptions rules is around five or ten times the length of the recording.

How can I transcribe as quickly as possible?
Typing up verbal data is a longwinded process. You will usually require between five and ten times the amount of time that the recording lasts.

You can speed up the process by up to 30% by using a footswitch. The slower your typing speed, the more the footswitch can help. Even if you type very efficiently (with a typing speed of over 220 strokes per minute) and have until now always used key combinations to operate your transcription software, a footswitch can still make the transcription process 10-20% quicker.

Do take breaks. This may sound trivial, but it is not possible to transcribe for eight hours at a time without requiring medical assistance! We recommend one five to ten minute break per one hour of computer work, which is a recognized standard. It is extremely taxing to transcribe for more than six hours a day.

 

Does automatic speech recognition yield useful results?

Unfortunately, it doesn’t… (Not yet!)

Unfortunately, speech recognition software is still absolutely useless for transcribing interview recordings. The only situation where it can be useful is where single speakers, who have trained the program to recognize their voices, read out a letter, for example. (It takes around three hours to convert ten minutes of speech into text). It is impossible for a speech recognition program to automatically recognize the speech of a person it doesn’t know or to make head or tail of two people speaking at the same time. The software currently available is also incapable of deciphering group conversations. Its task is made yet more difficult in that people often speak with an accent or use slang, which the program cannot identify. It also does not register pauses and intonation.   

The only way to use speech recognition for the transcription of interviews is to train the program to recognize your own voice and then to play your interviews and repeat what the interviewees say. That would mean sitting in front of the computer with headphones and a microphone and parroting the entire interview. This procedure is only worthwhile for people who type very slowly with the “three-finger system” – those who have a typing speed of under 100 strokes per minute. Our tests showed that transcribing text using this speech recognition method was hopelessly inferior to the standard transcription approach.   

However, anyone wanting to try out this software for repeating an interview, for example, should use Dragon Naturally Speaking (~100-900 Euro). Before you can install the current version you will need to download a 50MB update from the internet. 

 

Can I also transcribe videos with the support of a footswitch?

Yes. The Transana software allows you to transcribe videos using hotkeys. Our newly developed footswitch can be set to correspond to any hotkey you choose and can therefore be used in conjunction with Transana. For further information or to place an order please send us an e-mail or visit our online shop. (Price: €119)

 

Does "digital transcription" mean that the computer automatically types up the recording as a text?

No, and this is highly unlikely to change over the next few years. Transcription software only assists you to write the transcription yourself, by performing certain tasks which conventional, analogue transcription equipment (an analogue cassette recorder plus footswitch) also fulfils. These functions are adjustable playback speed, automatic brief interval rewind and footswitch support. Automatic recognition of group conversation recordings is absolutely impossible with the software that is currently available (e.g. Dragon Naturally Speaking).

 

Can I record in the same way with a digital recorder as with an analogue voice recorder?

No. The equipment we recommend works similarly to an analogue voice recorder. The only limitation of all digital voice recorders is that it is not possible to rewind and write over an existing text. Every time the recording is stopped, a physically new file is created, which you can jump to by pressing a key. Recordings can also be stopped using the pause button. It is not possible to jump, rewind and fast forward during recording.   

A digital device creates a new file every time playback is stopped. This function has the disadvantage of not allowing you to write over mistakes, for example, but the advantage is that it is simpler to navigate around the dictated/recorded text, since every section can be directly accessed as an individual file (you don’t have to note down the time code on the tape deck).   

The Olympus DS-2300 digital voice recorder, among others, has an editor function which allows some existing recordings to be overwritten.    (Since this device is primarily intended for dictation, and not for interview situations, we have not included it in our standard assortment. If you would like to purchase one, however, we can deliver you this item for €230. Simply send us your order or query via e-mail.)

 

What is the difference between analogue and digital recording?

Alongside the technical differences in the way audio material is saved, there is a practical difference between the two different recording approaches. Digital recorders allow you to record higher quality at a more affordable price. A cassette recorder plus footswitch and microphone is much more expensive than a digital recorder plus software and USB footswitch.   

If you record digitally, you also have the option of processing and archiving your audio recordings on your computer.   

Please note! MiniDisc records in digital format but only the latest, and thus most expensive, versions – so-called HI-MD – allow you to transfer data quickly and unproblematically onto a computer (one hour’s worth of interview in under two minutes. The old, standard MiniDisc recordings can be transferred to your PC but only at real-time speed (meaning that to transfer ten hours’ worth of interviews you will need about eleven hours). 

 

When choosing a digital recorder, do I need to worry about what format it records in?

Most digital voice recorders record in dss format. We believe that this format is ideal for dictations but is rarely appropriate for interview recordings. We therefore recommend you use equipment which records in wav, wma or mp3 format (e.g. the Olympus WS range of products).    V

ersion 2 of our transcription software supports all the formats mentioned above, apart from dss. Since this format requires a licence, we could not incorporate it. However, the free software Switch enables you to convert files from dss format. It takes a maximum of two minutes to transfer a one-hour interview in dss format from a voice recorder onto a PC in, for example, mp3 format.   

The Express Scribe transcription software can read dss files without any need for conversion.

 

Why do I have to convert digital recordings?

In order to save memory, audio data is usually downsized or compressed by various algorithms called codecs. One of the best-known codecs is the mp3 format. In order for a program to play a compressed file, the program must contain the codec. Our f4 software can read the codecs wma, wav, mp3 and ogg. Therefore, if your data is in another format, it must be converted. To do this, we recommend the program SWITCH, which can convert an enormous number of diverse and rare formats into better-known formats such as mp3.

 

What is the difference between converting and transferring data?

If you have recordings that you have made on a cassette recorder, and would like to play them on your PC, they must be in a format which a computer can read. You must therefore “transfer” the audio data onto your computer, like when you record the contents of one cassette onto another.   

Audio data which already exists in one digital format is “converted” into another digital format. 

  

What is the difference between a MiniDisc recorder and a digital voice recorder ?

The main difference is in how the devices store recordings, Whereas MiniDisc recorders save recordings onto a small CD, digital voice recorders usually store data on a built-in memory. The two devices also employ different formats for saving audio data, although little difference in quality can be detected between them. However, there is a slight tendency for MiniDisc recordings to be of better quality. But digital voice recorders are quickly catching up and are sure to overtake the MiniDisc in the near future.

 

What do the abbreviations mp3, ogg, wav, aif, dss, etc. mean?

Audio data can be stored and played in various formats. Here we provide you with an overview of some of the most common formats.   

wav        

Sound travels in waves. wav format saves these sound waves directly, without the need for any complex conversions. The disadvantage of this format is that the data uses up a lot of memory.   

aiff        

This is an audio data format for MACs. As with wav, this format does not compress data, i.e. aif files use up a lot of memory but provide good sound quality.

In order to get around the problem of memory consumption, various methods were developed to compress data so that it  

1. retains as much of its quality as possible and  

2. uses up as little memory as possible.   

mp3        

mp3 is a well-known format which has achieved great popularity through its use on internet music-sharing sites. The advantage of this popularity is that many devices easily support this format (mp3 players, CD players with mp3 function, etc.) The Fraunhofer Institut owns the rights to the mp3 format. Therefore, many applications which use mp3 require licences. WinLAME has developed an open-source rival to the mp3 codec   

ogg

As with mp3, ogg saves memory capacity as it filters out non-audible sections of the audio data. However, unlike mp3, it is a non-patented and copyright-free format. In some audio tests, ogg turned out to yield better results than mp3.   

dss       

dss was developed by the companies Olympus, Grundig and Philips for their digital voice recorders. The use of this format requires a licence and it is recognized by few PC programs.