Re: [ARSCLIST] voice recorders with speech-to-text software

[Francesco Martinelli <fmartinelli@xxxxxx>]

> Could you elaborate more?  Are these digital palm recorders (the digital
> equivalent of the old micro-cassette recorders) that have the Dragon
> software function built in?  Could you supply some model numbers so I can
> check them out?

I was thinking of something along the lines of SONY ICD-BM1VTP:
http://shop.voicerecognition.com/itemdesc.asp?CartId=12866499XIMD-EVEREST-J1
8&ic=SONY+ICD%2DBM1VTP&Cc=&tpc=

> 1. Their record quality generally isn't very good, and recording quality
is
> extremely important to our (Vermont Folklife Center) mission.  I argue
> everyone interested in doing this work and saving recordings should be
> concerned about this as well.  The better the recording, the greater its
> value to future researchers.

I absolutely agree. This one however seem to record at 44.1 kHz, the
recording quality should be adequate.
>
> 2. Most seem to lack a mic-level input to attach an external mic

This on has it.
>
> 3. They generally create audio files in proprietary formats and/or
> compressed formats, which flies in the face of standards-based digital
> preservation techniques

Good point. In fact they are using some sort of codec. I think they could be
converted to wave and saved on Cd, the quality will be the same but
accessibility should be granted for a long time.

> As far as the Dragon (and other voice recognition) software goes, and this
> is all based on hearsay as we are dedicated to keeping a human being
> employed as a transcriptionist, my impression is that at this point in its
> development voice recognition software still has to take a bit of time to
> "learn" a voice before it can translate it into text.  If you are
> interviewing slews of different people, the way we do, it would have to
> "learn" each new voice--and when you get into dialect and such, I imagine
> you'd have a hell of a time cleaning up the mess.  All this strikes me as
> time better spent trusting our transcriptionist's ear and typing skills.

I do agree, but I'd like to experiment and not have to rely on hearsay

> Down the road, who knows?  I imagine voice recognition software will just
> get better over time, but for now I don't think it is well suited to oral
> history/ethnogrpahic interview transcription.

It might well be true. We're doing oral history interviews for a national
jazz history project and we want to preserve them at best - the early ones
are on cassette though....
>
> If you're using it just for your own dictation, however, that's a totally
> different matter.

I think this is their best use so far.

> p.s. you might also want to check the archives of the HNET Oral History
> list on the topic of voice recognition software and the Dragon product in
> particular, I think you can access the list archives via this page:

thanks so much for pointing this resource to me and for your insightful
comments.

Francesco


---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.712 / Virus Database: 468 - Release Date: 27.06.2004