Skip to main content



AV Preservation Lab participates in Video QC software focus group

By Karl Fitzke

Olin Library’s Audio-Visual Preservation Lab (AVPL) was honored with an invitation to participate in a focus group aimed at identifying potential improvements to QCTools, a well respected suite of video quality control tools developed under the umbrella of the Bay Area Video Coalition (BAVC) in San Francisco’s Mission District.   In late April, the group of approximately 15 people, all doing similar kinds of work, came together from all over the country, using grant money secured by BAVC.  I was grateful Cornell granted me the time to participate.

On day one, the group reviewed current QCTools features and operation.  On day two, we got familiar with new batch file processing capabilities and discussed what other features we’d like to see.   Batch processing promises to help us take advantage of this tool set in our regular workflow, instead of the very selective use we’ve made to date, when some video artifact in a file seems rather suspicious and/or unfamiliar to us.

For folks who are unfamiliar with our work, we use our trained eyes and ears everyday, as a form of quality assurance, in identifying potential issues with analog-to-digital transfers of the audio-visual media that we process.  But it is not practical for us to listen to and/or watch every piece of program material in its entirely.  We only routinely check beginning, middle, and end of files for obvious problems.  So software packages like QCTools, able to automate the identification of a broad range of potential issues throughout a file, are very useful.  And the fact that they quantify what can otherwise be very qualitative judgements is also useful.

So with the support of others in the AV Preservation community, we’ll be training our minds on how to interpret QCTools data, and subsequently creating some objective QC standards for our work.  The aim will be to increase confidence in our results without wasting time running after false positives that result from poorly chosen thresholds on any of the characteristics we track.

We are glad to talk about the subject further with anyone interested.  Stop by sometime!

Comments

Comments are closed.