TechWhirl (TECHWR-L) is a resource for technical writing and technical communications professionals of all experience levels and in all industries to share their experiences and acquire information.
For two decades, technical communicators have turned to TechWhirl to ask and answer questions about the always-changing world of technical communications, such as tools, skills, career paths, methodologies, and emerging industries. The TechWhirl Archives and magazine, created for, by and about technical writers, offer a wealth of knowledge to everyone with an interest in any aspect of technical communications.
Subject:Re: Voice to text problem From:Peter <pnewman1 -at- home -dot- com> To:Elna Tymes <etymes -at- lts -dot- com> Date:Thu, 03 Feb 2000 16:59:24 -0500
Elna Tymes wrote:
> I have an interesting problem. My client has about 24 hours of videotape (which they will edit to approximately 8 hrs.) and what's said on it needs to become the source information for a course book, some marketing material, and
> the introductory sections of a number of manuals. What's the most efficient way to get this material to something a word processor can work with?
> Currently considered options are:
> 1. Using voice to text software, such as Dragon Naturally Speaking (in whatever edition seems most appropriate) and IBM's ViaVoice Gold. (After reading the reviews, I'm not inclined to recommend the Kurzweil offering in this
> Option 1 is the least expensive up front - both packages run about $150, but there's the cost of time to "train" the software to recognize the speakers' voices, and the cost of correcting misunderstood jargon.
Hi, I have extensive experience with speech to text engines. Either the
Dragon or IBM product will work however there are some things that you
must know. Do not dream of placing the mike next to the speaker and
trying to record that way.
Accurate translation from a wav file will depend upon the sampling rate.
You may want to try digitizing the tape with a digital recorder that has
not less that a sampling rate of 11x. The technical specs call for 9,
but for serious work you should have at least 11x or greater. For this
type of work, I doubt if you get away with transferring the tape, but it
is worth a try. You could get a high quality digital recorder and read
the the video tape into it, while wearing headphones, so as to minimize
background noise. Training time on any of these products is abut 5
minutes. If it is a one time job, you might want to download a thirty
day trial version of either L & H Voice Express or ViaVoice. I don't
think Dragon has a trial version. For best accuracy, the product to use
is Dragon NaturallySpeaking, Professional. It is significantly more
expensive than the Preferred edition, but the reports have been that it
is far more accurate. (its cost is about $600.00). You should have a PII
or PIII for best results, although other chips will work well if they
are at least 400-450. The key to accuracy is high memory, the more the
better. I use 256 MG and know of many who use much more.