Columbia University

1
Columbia University

• Notable New Yorkers
• Project objective
• Digitally preserve oral history recordings on
  variety of media and paper or electronic
  transcripts of interviews

2
Columbia University

• Notables New Yorkers include
• John B. Oakes
• Bennett Cerf
• Kenneth Clark
• Ed Koch
• Mary Lasker
• Frances Perkins
• Mamie Clark

3
Columbia University

• Source files
• Analog cassettes
• 5 and 7 reels
• Lengths 7 minutes to in excess of 67 hours
• Typed transcripts
• 100 pages to maximum 5,566 pages
• Binders
• MS Word files

4
Columbia University

• Recorded media to be re-recorded to common file
  digital format
• Create preservation masters and access copies
• Preservation masters 96kHz/24bit WAVE files
• Access copies 44.1kHz/16 bit WAV files
• Web-accessible copies .mp3 format

5
Columbia University

• Transcripts
• OCLC to provide to CUL
• Archival TIFF images
• Re-keyed files
• Electronically formatted interviews

6
Columbia University

• Physical condition assessment
• Interview transcripts quality varied
• Moderate to extensive written revisions edits
• Audio files varied in quality format
• None provided for preview
• Known life-expectancies recovery issues for
  media (audio cassette, reel-to-reel, etc.)
• No evidence of sticky shed or vinegar syndrome
  identified by library

7
Columbia University

• Recommended workflow
• OCLC p/u materials at CUL
• OCLC deliver audio masters to Safe Sound Archive
• OCLC to return material to CUL

8
Columbia University

• Digitization specifications
• OCLC delivered bi-tonal text pages
• 1-bit TIFF
• Group IV TIFF compression
• 600 dpi
• XML mark-up of full-text
• OCLC facilitated
• Re-keying with 99.95 accuracy
• TEI-Lite DTD mark-up

9
Columbia University

• Audio reformatting specifications workflow
• Transmittal and trafficking policies
• Database customization
• Material log-in
• Cross-checking against packing list

10
Columbia University

• Evaluation engineers notes
• Compact audio cassettes played back on Nakamichi
  cassette decks w/mechanical electrical playback
  alignments
• Digitization w/Prism Sound analog to digital
  converters
• Output 96kHz/24bit preservation master 44.1kHz
  access copy concurrently

11
Columbia University

• Reel-to-reel tapes played back on Studer tape
  decks
• With mechanical electrical playback alignments
• Digitization w/Prism Sound analog to digital
  converters
• Output 96kHz/24bit preservation master 44.1kHz
  access copy concurrently

12
Columbia University

• CUL elected semi-monitored approach
• Up to 3 originals transferred simultaneously
• SSA guarantees 11 representation
• Monitoring alternative 100 monitored

13
Columbia University

• Quality Control Procedures
• OCLC performs 100 quality assurance of all
  original TIFF images
• Reviewed for completeness, alignment,
  illumination regularity, and detail consistency
  throughout image
• SW allows 11 viewing, zooming _at_ 100 and
  reduced full-page view

14
Columbia University

• Quality Control Procedures
• SSA expects zero returns or rework
• Achieved via database automation of file naming
  (reduces human error)
• Each file on delivery medium opened auditioned
  on separate computer to assure recoverability
• Each file checked _at_ beginning end to assure
  completeness

15
Columbia University

• Quality Control Procedures
• SSA expects zero returns or rework
• And spot-checked throughout for consistency
• File names checked by separate person for naming
  and contents against original recordings
• Person also proofs file headers

16
Columbia University

• File Delivery
• Text files XML as per CUL specifications
• Challenges encountered due to variations in the
  interview format/transcript format (some memoir
  style, some strict QA format)
• Audio files
• Portable hard drives
• ftp impractical due to file size
• Est. 270GB 1,000 hours connect time

17
Columbia University

• Digital Archive
• CUL files delivered for uploading to the OCLC
  Digital Archive repository

18
Columbia University

• Client acceptance of deliverables
• OCLC standard 30-day image file retention
• Planned 30-day customer acceptance period

19
(No Transcript)