For anyone interested in sound and sound recordings
This group is for anyone interested in speech-to-text and voice recognition systems. Technologies that can convert the spoken word into text are growing in number, from voice-activated commands for smartphones to the conversion of digital speech archives into readable and searchable text.
The particular focus of this group is the conversion of speech archives. It is interested in the competing technologies, cost-effective models for archives and libraries, and the uses of transcribed speech collections (audio and video) for researchers.
Members: 10
Latest Activity: on Thursday
The Comment section is for general information and conversation about speech-to-text matters. The Discussion Forum will be used to list specific packages and to discuss issues relating to these.
Started by Mari King. Last reply by Mari King May 29. 2 Replies 0 Likes
Started by Mari King. Last reply by Mari King May 7. 1 Reply 0 Likes
Add a Comment
As part of the British Library's research project into speech-to-text and speech recogintion systems and their value for academic research, we would like to gather opinions from those working closely with such technologies, whether vendors, developers, R&D divisions, academic researchers or others.
We have produced a three-question survey which will only take a short while to complete, and we would be very grateful if those interested could find the time to complete it.
https://www.surveymonkey.com/s/9HXHZN7
All answers will be treated in confidence. A summary of the survey (which will not mention names) will be included in our project report.
The survey will run until July 12th.
Speech-to-text drives motorists to distraction
http://www.theregister.co.uk/2013/06/13/speechtotext_drives_drivers...
With texting so clearly dangerous while driving, users and vendors have turned to speech-to-text technologies as a safe alternative, perhaps to no avail.
According to a study published by US road safety group the AAA Foundation, speech-to-text technologies are more distracting then talking to other passengers in the car. The research backs up a simpler study carried out earlier this year in Texas.
To test cognitive distraction, the AAA conducted three experiments. In the first, volunteers performed eight tasks, and in the second, they carried out the same tasks while driving in a simulator. Finally, they drove an instrumented vehicle through a city residential area. The experimental tasks included listening to a radio; listening to an audio book; speaking with a passenger; using a hand-held mobile phone; using a speech-to-text interface; and a combination of memory and true/false maths problems.
In the laboratory baseline, the research found that compared to a single-task reaction time of about 460 milliseconds, speech-to-text operation had an impact similar to using a hand-held mobile, slowing participants' reaction times to around 570 ms.
In the driving simulator, speech-to-text operation was worse (noting the large error bars) than using a hand-held mobile phone – drivers' mean braking reaction time while using the mobile was around 950 ms, while a driver using a speech-to-text interface had a mean reaction time of about 1050 ms.
Other symptoms of distraction the researchers measured included “suppressed brain activity … missed visual cues, and reduced visual scanning of the driving environment (think tunnel vision).”
As the AAA Foundation notes: “Though shipments of these systems are expected to skyrocket in the coming years, use of speech-to-text communications presented the highest level of cognitive distraction of all the tasks we analysed.”
On the other hand – and apparently refuting an urban myth that car radios are just as great a distraction – the research found that listening to the radio or audiobooks only caused “minor increases in cognitive workload”.
The research was conducted with the University of Utah. ®
Comment by Mari King on April 4, 2013 at 14:36 Comment by Mari King on March 27, 2013 at 19:17 Relevant episode from this very popular BL affiliated series: http://www.bbc.co.uk/programmes/b01rg22g
I've written a piece on speech-to-text and search for my personal blog: http://lukemckernan.com/2013/02/05/opening-up-speech-archives/
The British Library's Opening up Speech Archives project is organising a conference on speech-to-text technologies and their use in academic research on 8 February 2013 at the BL. More information here: http://playback.ning.com/events/opening-up-speech-archives.
The event is invitee-only and we're fully booked, but if any member of the Playback speech-to-text group is interested in going (and hasn't already been invited), let me know.
The programme as it currently stands (27/1/13) is attached.
Comment by Mari King on January 18, 2013 at 12:30 User driven innovation is a key element in LARM. The infrastructure and its interface are based on user needs and are developed in close collaboration between technicians, cultural researchers and designers. The technology is tested in a series of cases where radio broadcasts are analyzed from a variety of perspectives. These research projects deliver feedback to the development of digital audio search and audio description tools in both effectual and innovative ways. Furthermore, technicians and cultural researchers work together on the development of digital tools for audio analysis and new formats for the communication of research in radiophonic cultural heritage.
"Dragon Dictate 3 for Mac gains audio file transcription, improved accuracy"
http://arstechnica.com/apple/2012/09/dragon-dictate-3-for-mac-gains...
The British Library is advertising for a 6-month contract Technical Analyst post in support of its AHRC-funded research project into speech-to-text technologies and their potential for supporting research.
The post is advertised here: http://bit.ly/ODKjVC
More information on this main Playback blog post: http://playback.ning.com/profiles/blogs/opening-up-speech-archives-...
The latest from BBC's Yves Raimond, with links to papers on Automatic Interlinking of Speech Radio and Automated Semantic Tagging of Speech Audio.
http://www.bbc.co.uk/blogs/researchanddevelopment/2012/04/notes-fro...
Diego Stocco - Duet for Leaves & Turntable from Diego Stocco on Vimeo.
June 22, 2013 from 11am to 4pm – DIVAcontemporary Studio
0 Comments 0 Likes
© 2013
Powered by
You need to be a member of Speech to Text to add comments!