As a slight aside, I was at an event the other night where the BBC reported they had digitised the entire World Service archive and then transcribed it automatically. They reported a 50% success rate as far as I recall, but what they were looking to do was extract metadata/tagging, not full transcriptions.
Overview of event: http://www.bbc.co.uk/academy/news/view/semantic_web
Video of presentation (see second half): http://www.youtube.com/watch?v=sCnpSsxsb5U&list=PL5906124C50F16299&index=6&feature=plpp_video
Summary of project: http://www.bbc.co.uk/blogs/researchanddevelopment/2012/03/automatically-tagging-the-worl.shtml
The person to contact is Yves Raymond (@moustaki on Twitter)
----------------------------------------------------------------------
James Morley [log in to unmask]
Website Development Manager +44 (0)20 8332 5759
Royal Botanic Gardens, Kew www.kew.org
-----------------------------------------------------------------------
> -----Original Message-----
> From: Museums Computer Group [mailto:[log in to unmask]] On Behalf Of
> Patricia Reynolds
> Sent: 11 April 2012 14:17
> To: [log in to unmask]
> Subject: Re: Dragon speech transcription
>
> Dear colleagues,
>
> I'm dictating this using Dragon professional 11.
>
> Dragon has advanced greatly over the last few years, and I expect that
> it would now be able to cope with such things as non-trilled Rs. It
> certainly manages my mangled English Rs.
>
> However, it does need to be trained. This is done in several ways,
> including an initial setup involving reading passages, providing
> program with access to one's e-mail and word documents, and by
> religiously correcting every error seen indication using the programs
> correction function rather than over typing. The person using the
> program also needs to be trained: there are lots of little tools and
> ways of using which improve the recognition of the software.
>
> This means that Dragon is not an easy way to take something such as an
> oral history interview and convert it into text. The most cost-
> effective way I know of doing this is to pay a professional audio
> typist. If that is not possible, I have known people who have played
> back the tape over headphones, and dictated what they hear into a
> microphone using either Dragon, or the Microsoft voice recognition
> system.
>
> With best wishes,
>
> Pat
>
> Ps: The errors I have corrected in making this e-mail are: capitalised
> to words where the auto capitalisation did not work, collected thrilled
> to trilled, are to Rs, and convert it to converted.
>
> Dr . Pat Reynolds
> Surrey Heritage Manager
>
> 01483 518754
> 07770 326270
>
> Surrey History Centre
> 130 Goldsworth Road
> Woking
> Surrey
> GU21 6ND
>
> Surrey Heritage services: http://www.surreycc.gov.uk/surreyheritage
>
> Archive of the Year 2012:
> http://www.exploringsurreyspast.org.uk/news/001033.html
>
>
>
>
>
> Giuseppe Sollazzo <[log in to unmask]> Sent by: Museums Computer Group
> <[log in to unmask]>
> 11/04/2012 13:40
> Please respond to
> Museums Computer Group <[log in to unmask]>
>
>
> To
> [log in to unmask]
> cc
>
> Subject
> Re: [MCG] Dragon speech transcription
>
>
>
>
>
>
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> I used it to dictate part of an essay in my native language, Italian,
> 5 years ago, and tried it with my thesis, in English.
>
> Results for me were disappointing:
> - - in English, as I expected, it was not too accurate (my accent is,
> naturally, not that of a native speaker)
> - - in Italian was more accurate but as I can't roll "Rs" it was
> missing
> all words containing any "R"... my fault, ok, but I was expecting to
> be able to train this.
>
> That said, it was 5 years ago so hopefully things have moved a bit :)
>
> Giuseppe
>
> On 11/04/12 12:25, J DAVIS wrote:
> > I was going to try using it for that a few years ago but had too
> > much problem at the time getting it to recognise my own speech.
> > Obviously the software will have improved since then. I think that
> > it could depend on the accents and could take a lot of training if
> > using with more than one voice. I would be very interested to hear
> > if anyone has done it. It would be an interesting experiment.
> >
> > Janet
> >
> > Janet E Davis
> >
> > --- On Wed, 11/4/12, Mark Osborne <[log in to unmask]>
> > wrote:
> >
> >> From: Mark Osborne <[log in to unmask]> Subject: Dragon
> >> speech transcription To: [log in to unmask] Date: Wednesday, 11
> >> April, 2012, 12:06 hi
> >>
> >> is anyone using any of Dragons software as an aid to
> >> transcription?
> >>
> >>
> >> Mark
> >>
> >>
> >> Mark Osborne Creative Technologist Carnival Archive Project
> >> UKCCA
> >>
> >>
> >> ****************************************************************
> >> website: http://museumscomputergroup.org.uk/ Twitter:
> >> http://www.twitter.com/ukmcg Facebook:
> >> http://www.facebook.com/museumscomputergroup [un]subscribe:
> >> http://museumscomputergroup.org.uk/email-list/
> >> ****************************************************************
> >>
> >
> > ****************************************************************
> > website: http://museumscomputergroup.org.uk/ Twitter:
> > http://www.twitter.com/ukmcg Facebook:
> > http://www.facebook.com/museumscomputergroup [un]subscribe:
> > http://museumscomputergroup.org.uk/email-list/
> > ****************************************************************
>
>
> - --
> ____________________________________
>
> Giuseppe Sollazzo
> Senior Systems Analyst
> Computing Services
> Information Services
> St. George's, University Of London
> Cranmer Terrace
> London SW17 0RE
>
> Email: [log in to unmask]
> Direct Dial: +44 20 8725 5160
> Fax: +44 20 8725 3583
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.10 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
>
> iQEcBAEBAgAGBQJPhXv1AAoJEAqigArPBfJXA0MH/31YfuGN30KYPPdECmAw7aff
> mwC1aYShODHx+zRikgWbEsVxQpDgl0eRVxTpuEWB2L3V8Ib+mtoXm9P7VAPylCw1
> g+0O2i0kzaRe3+Yrv3OXGIMv3clWXn8rDUiTTjw1R9kuHOW0FRVJcgRmNFImHz3+
> 1HmRCG46kUeHHK1CXroQZjjuBGmBEjN08C6MDUyc5aT3VYhd5BtRQRAMJcTrjbvj
> jGfaTIHbmVGpjRzlLFF/Ffb9lo2w+JEIYyis3IGBujs6c7GWakChWx7C+GMzWjfr
> i/EHgHGt1Qv6o/QUZEzj54p67ePYWscWnKjPvko9dRcW+Qr1APP2745Yki54DCw=
> =KvwN
> -----END PGP SIGNATURE-----
>
> ****************************************************************
> website: http://museumscomputergroup.org.uk/
> Twitter: http://www.twitter.com/ukmcg
> Facebook: http://www.facebook.com/museumscomputergroup
> [un]subscribe: http://museumscomputergroup.org.uk/email-list/
> ****************************************************************
>
>
>
> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
> *
>
> This email and any attachments with it are intended for the
> addressee only. It may be confidential and may be the subject of
> legal and/or professional privilege.
> If you have received this email in error please notify the
> sender or [log in to unmask]
> The content may be personal or contain personal opinions and
> cannot be taken as an expression of the County Council's
> position.
> Surrey County Council reserves the right to monitor all incoming
> and outgoing mail. Whilst every care has been taken to check
> this
> e-mail for viruses, it is your responsibility to carry out any
> checks upon receipt.
>
> Visit the Surrey County Council website -
> http://www.surreycc.gov.uk
>
> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
>
> ****************************************************************
> website: http://museumscomputergroup.org.uk/
> Twitter: http://www.twitter.com/ukmcg
> Facebook: http://www.facebook.com/museumscomputergroup
> [un]subscribe: http://museumscomputergroup.org.uk/email-list/
> ****************************************************************
****************************************************************
website: http://museumscomputergroup.org.uk/
Twitter: http://www.twitter.com/ukmcg
Facebook: http://www.facebook.com/museumscomputergroup
[un]subscribe: http://museumscomputergroup.org.uk/email-list/
****************************************************************
