DON'S FREEWARE CORNER - JUL 2020
TRANSCRIBING AUDIO FILES
Don's Freeware Corner articles are printed in the
UTAH VALLEY TECHNOLOGY AND GENEALOGY GROUP
(UVTAGG) Newsletter TAGGology each month and are
posted on his Class Notes Page
where there may be corrections and updates.
TRANSCRIBING AUDIO FILES
©2020 Donald R. Snow - Last updated 2020-07-13
ANALOG AND DIGITAL AUDIO RECORDINGS
MOost of us have a few old audio recordings of family members, u
to-reel or cassette tape recordings. Tapes are in analog format,
phonograph records. There were home recording machines that recorded
on plastic phonograph records before there were tape recorders—so
you may have
some of those, too. Or you may have other phonograph records you want
digitize. Converting analog to digital format is called digitizing
and can be done
by playing the old media and recording it on a computer. In digital
are easy to work with, edit, copy, duplicate, and post on websites.
And digital recordings copy with exactly the same quality,
duplicating a tape recording which is like xeroxing a xerox where
lose quality with each copy.
Old recordings can be found in your home, the neighbors,
churches, schools, libraries, and various other places. For instance,
schools may have made recordings
of musical productions that you or your ancestors were in. If you
can find these
and digitize them, they are easy to edit and copy and distribute
to others. Tape
recordings need to be digitized as soon as possible because tape media
and becomes unplayable. Once digitized, copies remain the same quality as the
DIGITIZING ANALOG RECORDINGS
Digitizing analog recordings is important and sometimes complicated. The topic
of digitizing deserves an entire discussion itself and won't be discussed here.
Please review my notes and
articles posted on my website about digitizing your recordings. Here
we will assume
that your audio files are already in digital format and are ready to be
transcribed to text format.
The U.S. National Archives has a Transcription Tips website at
https://www.archives.gov/citizen-archivist/transcribe/tips . The
mostly about transcribing hand-written documents, but much of
it also applies to
audio transcriptions. The site suggests typing things exactly as
they are on the
recording, including words out of place, and so on. There are many
other sets of
transcription tips if you Google audio transcription tips. For example,
TRANSCRIBING BY SMARTPHONE
Most smartphones now have an app that will transcribe voice to text while
speaking into it. However, these apps have varying degrees of accuracy and you
may need to edit the transcribed text. My personal experience has been that it
takes more time to do it this way and edit the text than it would to directly type
the transcription text. Your experience may be different. And,
depending on your overall goal, complete accuracy may not be needed. For
example, if you want to search the transcription for certain names or places, perhaps
exact accuracy is not necessary. Audio searches must first be transcribed to text
because there are no programs that will index audio without first transcribing it.
Smartphone apps do not usually have the program grooming by "read script to
learn your voice" capabilities that computer programs have. Phone programs are
simpler and easier to use. If you do use a smartphone or tablet for transcription,
learn how to copy the text file from your phone for use elsewhere.
A smartphone app called AVA is a free app that can be downloaded
for limited use. I learned about it while attending a Hearing-
Impaired group. AVA’s icon is an ampersand "&" when you search for it and after it is installed. The app can be used for free for 5 hours per month and unlimited time for the paid subscription. I installed and tried AVA and it works fairly well. I did not check
the unlimited use price, but the free 5-hours-per-month option allows new users
to determine AVA’s usefulness. AVA seems to work better than the built-in app on
my smartphone. Hearing impaired people use AVA by holding the phone while the
other speaks, providing the hearing-impaired with a transcription. Try the free
version and see if the transcription's quality can work for you.
WINDOWS 10 BUILT-IN SPEECH-RECOGNITION
Windows 10 has speech-recognition software built-in--see
https://www.windowscentral.com/how-set-speech-recognition-windows-10 . It
requires a microphone , of course, and usually does better with a combination
earphone/boom mike. These low-cost eachphone/microphones plug into a USB jack or
directly into the sound card on your computer. Inexpensive adapters are also
available to plug round jacks into USB ports or vice versa. The sound card ports
are color coded and the mike input jack is usually colored orange. More details
are in my class notes on digitizing audio files. The article mentioned above
explains how to set up and train the WINDOWS 10 software for your voice. This
program works well, if you speak slowly and correct things as you go, but it makes
lots of errors for continuous speech. Continuous speech is the way we normally
talk to each other and the way most of your digital recordings would be. But,
since it is free, try it and see if it works for you.
DRAGON NATURALLY SPEAKING SOFTWARE
DRAGON NATURALLY SPEAKING software, considered the premier voice-
recognition software, is not free. It costs about $150, but I have seen it on sale
for much less. The company updates the program regularly, but they charge for
each new version. Even this program will not do continuous speech with 100%
accuracy. However, as mentioned above, perhaps something like 95% accuracy
will be enough for your needs. You can later go through and edit the text, but this
will take time. Your transcription goal helps determine your need for accuracy.
TRANSCRIBING BY ECHOING
I have been keeping my daily journal by recording in audio format for the past 40-
50 years, first using cassette tapes and now on a digital recorder. When I started
journal recordings in about 1975, I expected that there would be automatic
transcription programs available within a few years, but those haven't materialized yet. I thought my worst-case scenario would be to have to pay someone to
transcribe my journals later. After trying several programs, I phoned the DRAGON
NATURALLY SPEAKING tech support and they told me that DRAGON was not yet
good at continuous speech. They suggested "Echoing,” a process where you
listen through earphones and repeat what you hear. Since the voice-recognition program is trained
to recognize your voice now, it can provide a transcription. I've tried this process
which seems to work OK, but is so much extra work that I haven't done much
transcribing this way. I have not purchased the latest version of DRAGON, so I
don't really know if this works better now. The echoing method could be done
with any voice recognition program. If you find a combination that works well,
please let me know.
TRANSCRIBING USING LISTEN 'N WRITE
LISTEN 'N WRITE is a program that is free for personal use at
http://www.majorgeeks.com/files/details/listen_n_write.html. A useful 3-
minute video tutorial is viewable at
https://www.youtube.com/watch?v=WEUHZr9Bwdk (Click Skip on the ad). The
program requires the audio to be in mp3 or wav format and uses any text editor,
e.g. LibreOffice or Word, so you can use the spell checker. LISTEN N' WRITE uses
function keys to control the audio playback so you keep your hands on the keyboard
and don't need to use the mouse. F5 starts and stops audio and you can set the pause for a specified time intervals—such as 4 seconds, to
allow time to type what you have heard. F6 skips backward a specified time
interval and F7 skips forward a specified time interval. It is a helpful
program for transcribing and is free for personal use.
ADDITIONAL PROGRAMS AND METHODS FOR TRANSCRIPTION
There are many other helpful programs, most of which I have never tried. Some
of these could be used in connection with methods described above. If you find
one that works well, please let me know.
GOOGLE DOCS -- https://docs.google.com/document/u/0/?tgif=d .
Good tutorial at https://qz.com/work/1087765/how-to-transcribe-audio-fast-and-for-free-using-google-docs-voice-typing/ . Go to GOOGLE DOCS (click on the
"tic-tac-toe" icon on Google and select Docs); start a new document and click
Tools > Voice Typing. Click on the microphone symbol and start dictating. I
cannot get this to recognize my microphone, even though my mic works in the
Google searacH bpx. To hear an audible description of Voice Typing, go to Google and
click on the microphone icon at the end of the search box and say, "How to use
voice typing". From your earphones or speakers you'll hear a short description of
how to set it up.
TRANSCRIBE -- https://otranscribe.com/
A free online program from MIT to help with transcription. A tutorial and helps
are at https://transcribe.wreally.com/how-to-transcribe-audio .
REV -- https://www.rev.com/blog/resources/best-apps-to-transcribe-audio-files-to-text . This seems to have several free voice transcription programs.
EXPRESS SCRIBE (free version) -- https://www.nch.com.au/scribe/index.html . It says the free version can be used with wav, mp3, and other common file
Digitizing and then transcribing audio files is an important part of family history,
especially your own family, since you probably have some recordings of your
family. This Freeware Corner article has given you a few ideas and ways to
accomplish this. It requires hard work and stick-to-itiveness but pays off in
preserving and making old recordings available for your posterity.