StoryCorps is a wonderful group that works "to provide Americans of all backgrounds and beliefs with the opportunity to record, share, and preserve the stories of our lives." You have probably heard excerpts from some of their interviews on NPR. Lately they have begun working some of the interviews into short animations, of which this is my favorite:
Showing posts with label oral history. Show all posts
Showing posts with label oral history. Show all posts
Wednesday, September 8, 2010
Thursday, October 30, 2008
The Holy Grail of Audio Recognition
Big news today from the Washington State Digital Archives! (Full disclosure: I am an Assistant Digital Archivist here.) Today we put the audio files of the House of Representatives
Committee Meeting Recordings online--and they are keyword searchable.
The House of Representatives Committee Meeting Recordings cover 1973 to 2001. This is almost 6000 hours of hearings and the files take up 1 terabyte of data.
This list of house committees will help give context to some of these files. The files came from 30,000 cassette tapes.The tapes were converted to digital files and cleaned up starting in 2005. Putting them online and making them searchable is a cooperative project between the Washington State Digital Archives and the Microsoft Corporation.
The technical breakthrough is that these files are keyword searchable. Users can enter keywords or phrases and the search engine will dig through all of the files and discover when
anyone spoke those words. The search results give some details about the file but also a snippet of the text showing where on that file the words were spoken. Click on any of the strings of words between the dashes and the in-line player will take you directly to that point in the recording. Some good keyword searches are salmon and dams, "Indian gaming," "state history," and "Lewis and Clark."
This, my friends, is one of the holy grails of computing: untrained voice recognition over thousands of hours of tapes and many different voices. We rolled out this technology with the legislative hearings because we are a state archives and this gives the Washington State public unprecedented access to these public records. But think of the other uses for the keyword searching of audio files. I have never visited an archives that did not have boxes of decaying audio tapes from an oral history project that never quite got to the transcribing stage. These tapes can be digitally preserved and put online. Television and radio interviews and news and talk programs will become searchable. This is a digital history breakthrough.
Committee Meeting Recordings online--and they are keyword searchable.The House of Representatives Committee Meeting Recordings cover 1973 to 2001. This is almost 6000 hours of hearings and the files take up 1 terabyte of data.
This list of house committees will help give context to some of these files. The files came from 30,000 cassette tapes.The tapes were converted to digital files and cleaned up starting in 2005. Putting them online and making them searchable is a cooperative project between the Washington State Digital Archives and the Microsoft Corporation.
The technical breakthrough is that these files are keyword searchable. Users can enter keywords or phrases and the search engine will dig through all of the files and discover when
anyone spoke those words. The search results give some details about the file but also a snippet of the text showing where on that file the words were spoken. Click on any of the strings of words between the dashes and the in-line player will take you directly to that point in the recording. Some good keyword searches are salmon and dams, "Indian gaming," "state history," and "Lewis and Clark."This, my friends, is one of the holy grails of computing: untrained voice recognition over thousands of hours of tapes and many different voices. We rolled out this technology with the legislative hearings because we are a state archives and this gives the Washington State public unprecedented access to these public records. But think of the other uses for the keyword searching of audio files. I have never visited an archives that did not have boxes of decaying audio tapes from an oral history project that never quite got to the transcribing stage. These tapes can be digitally preserved and put online. Television and radio interviews and news and talk programs will become searchable. This is a digital history breakthrough.
Saturday, April 26, 2008
Coeur d'Alene Native Names Project
"The Native Names Project: In September of 2005 the United States Geologic Survey (USGS) funded the Coeur d'Alene Tribe’s Cooperative Agreements Program (CAP) grant to supply Native American place names to
the National Map . . . One of the most important things about this project is preserving the individual stories and pronunciations of the places. On many reservations across the United States it is the unfortunate truth that only a handful of the people know the native geographic names. The goal of this project is to collect these place names before they disappear forever." (Above is Hangman Creek, known to the Coeur d'Alenes as "Suckers in the Water.")
You can see the Coeur d'Alene names for many sites in Washington, Idaho and Montana on this flash based map. (Which actually is not loading for me this evening--check back.) Or more fun yet, download the Google Earth KMZ file. Here is what a section of the Google Earth file looks like:
Clicking on the icons brings a pop-up window with the native and Euroamerican names and a link to "site report" page with includes an MP3 of the native pronunciation and in some cases a streaming video of Couer d'Alene elder Felix Aripa
explaining the meaning of the name and the location. Here is Aripa explaining how his people called the area around Sprague Lake "Smell smell." Here is another article on the names project.
This is a wonderful combination of native tradition and modern technology. And it points the way towards other neat things one might do with oral history and Google Earth, mapping peoples memories of one room school houses, or CC camps or the like.
(This post via Metafilter.)
the National Map . . . One of the most important things about this project is preserving the individual stories and pronunciations of the places. On many reservations across the United States it is the unfortunate truth that only a handful of the people know the native geographic names. The goal of this project is to collect these place names before they disappear forever." (Above is Hangman Creek, known to the Coeur d'Alenes as "Suckers in the Water.")You can see the Coeur d'Alene names for many sites in Washington, Idaho and Montana on this flash based map. (Which actually is not loading for me this evening--check back.) Or more fun yet, download the Google Earth KMZ file. Here is what a section of the Google Earth file looks like:
Clicking on the icons brings a pop-up window with the native and Euroamerican names and a link to "site report" page with includes an MP3 of the native pronunciation and in some cases a streaming video of Couer d'Alene elder Felix Aripa
explaining the meaning of the name and the location. Here is Aripa explaining how his people called the area around Sprague Lake "Smell smell." Here is another article on the names project.This is a wonderful combination of native tradition and modern technology. And it points the way towards other neat things one might do with oral history and Google Earth, mapping peoples memories of one room school houses, or CC camps or the like.
(This post via Metafilter.)
Subscribe to:
Posts (Atom)