Low-level white noise, better transcription accuracy?

 

Yes.. you read the title correctly!

In an article posted 11/27/2007,  HealthImaging.com has noted a study at the University of Maryland Medical Center - "found that the introduction of white noise at certain levels as part of the acoustic background increased accuracy of speech recognition systems’ transcription capabilities", said Joseph Zwemmer, MD, who presented the results of the research.


This is actually pretty fascinating stuff. If white noise can be used to improve accuracy, then it's more than likely another step has been made, albeit maybe inadvertently, in the proverbial quest for the "Holy Grail" of speech recognition; to-wit software-based noise canceling.

The original article can be read here.

Hopefully, we'll be able to get some specific data soon from the University of Maryland as to what exactly comprised the white noise, what frequency ranges it included, details as to the sound itself, durations of specific frequency ranges, etc.

This has some very interesting possibilities..
Check back from time to time to see if the Medical Center has shared any data,
on just what type of sound the white noise was comprised of..