Podscope and Podzinger – An Update
I’ve had a pretty productive email exchange with a few folks from both
Podscope and Podzinger (see my previous post about the podcast search engine
tiff), and here’s the story from both sides.
First, from Chairman and Chief Executive Officer of TVEyes, Inc. (the
company behind Podscope) David Ives:
“BBN’s claims for Podzinger can only be compared to Al Gore’s
claim of inventing the Internet.
In May 2005, TVEyes launched the world’s first search engine for audio
and video podcasts. We were the first company to provide ‘spoken word
search’ for all podcasts and video blogs allowing a user to search every
word mentioned in those files. How a company can issue a press release
nearly a year later stating they are suddenly first to offer these services
can easily be dismissed. The Internet is wonderful at recording history and
podscope’s firsts are well documented. Podzinger.com was first registered as
a domain July 18, 2005…months after podscope had begun to receive
recognition for being first and best in class. TVEyes will continue to
innovate knowing that great ideas are often copied.”
Apparently, from Podzinger’s press releases, they would like to create
the impression they are alone in this space. Podscope was and always will be
first in this space and we welcome honest competition.”
And from the Vice President of Marketing over at Podzinger, Barbara Loonam:
“As for PodZinger, it is built on BBN Technologiesâ€™ speech-to-text technology which does come out of 30 years of research and development. A couple of interesting milestones, we introduced the first large vocabulary continuous speech recognition engine in 1992 from which PodZinger is built. In 1998, BBN demonstrated an â€œaudio miningâ€ application system to search for words within audio. Starting in 2002 and continuing through 2005, BBN was the leading recipient of the government funded DARPA EARS (Effective Affordable Re-useable Speech-to-Text) program where the objective was to make audio and video searchable just the same as text. Subsequently, from 2000 to the present day, BBN has been refining and highly tuning this technology and delivering it in the form of total broadcast monitoring solutions for the government.
With PodZinger, we are currently offering this service based on our underlying technology within the narrow field of use of podcasts. It came out of beta in January of this year. We just introduced video this week. As of today, we are still the only service that offers the consumer full-text search on text-based results that can be rapidly skimmed in addition to the feature that enables users to point-and-click on the text in order to listen to relevant portions of a podcast.”
I asked for some clarification from Podscope again, regarding full-text audio and video search. Here’s what they had to say:
“In addition to the falsehoods within their press release, David and I discussed the text results returned with a Podzinger search. TVEyes decided not to include the text created by their process because all forms of speech recognition have up to 20 percent or more error, depending on the quality of the file and the speakerâ€™s accent, etc. So, when you review text provided by Podzinger plenty of it is erroneous. If you read through the results you can see that it often amounts to unusable, scrambled text, surrounded by your keywords. If the idea is to help you assess relevancy of the search, itâ€™s just much more efficient to do so by listening to a short segment of the clip with your search term mentioned in context and deciding if you want to hear or see more. This is just one of the essential innovations that Podscope introduced.”
As to what differences there are in the services, Podscope and Podzinger are targeting the same market segment. Search for audio/video within podcasts. We would argue that Podscopeâ€™s proprietary combination of speech processing technologies vs. Podzingerâ€™s sole reliance on the BBN engine makes Podscope more accurate and subtle in its results.”
The bottom line of all this mess: which podcast search engine gives a better user experience? I think one could argue ad infinitum whose service came first, but in the end, it’s the searcher who needs to be focused on.