You Say Distortion, I Say Nasturtium
Sometimes it seems like speech recognition software will never fully flower.
Here’s EveryZing’s rendering of the audio track of a video recently posted at Boston.com about one of the annual rituals at the Isabella Stewart Gardner Museum in Boston:
“When we display the distortions of the beginning of spring frosts.”
Here’s what was actually said:
“When we display the nasturtiums it’s the beginning of spring for us.”
Cambridge-based EveryZing has a great system for indexing multimedia content so that Web surfers, search engines, and ad-placement software can understand it—I’ve written about the company a couple of times for Xconomy. But it’s far from perfect, as this little example shows. Maybe it’s not fair to ask speech-recognition software to understand an obscure word like “nasturtiums.” But in this case it’s the most important term in the whole transcript. Better luck next time, EveryZing.