Macworld 2010: MacSpeech Scribe brings transcription to the Mac

I've been planning to write a book for a number of years and would like to do it using Dragon or MacSpeech with my newly purchased iPad 2.

I am now finding out that it is not as easy as advertised. For example, I have Dragon Dictate on my iPad but it only allows you a less-than-one-minute period of time to talk, than stops. My plan was to talk for a longer period of time (half hour or longer).

I wrote to the Nuance people for advice and have been getting VERY unclear replies. If anyone has any useful information that would help me I would really appreciate it. Such as:

- Can I use either Dragon or MacSpeech on my iPad 2 that would allow me to talk for a longer period of time?
- Could (should) I use Dragon with Windows 7 on my iPad?
- If YOU were going to write a book the way I had planned, what would you do?

Thanks.

I too work with ethnographic research and am interested either in MacSpeech Scribe or Dragon; to run in a MacBook Pro. So let ask a few things:
1- @ George and David - would it work to use, say, 6 profiles for 6 interviews, then erase them and re-set for another 6?;
2- @ Kathryn and Susan - Did you find any helpful information about which one (MacSpeech or Dragon) to go, or if either would work best for ethnographic transcriptions?;
3- @ Pegrose and Alan - Even 90% of accuracy for me would be ok, as Pegrose said, to correct punctuation and other smaller mistakes takes much less time than transcribing the whole thing typing it in. So, did any of you tested the software? Any feedbacks?

If I find anything helpful I'll post it here. Also hope to hear from you asap.

Cheers!
Eurico

... "and thoughts of sugar plums danced in their heads" comes to mind reading the "sounds like" "would be ideal for" and other dreamy comments.

Voice recognition software is in its post-infancy to early teens.

Because it is not ready for prime time -- c-o-u-r-t -- all the users with all their enthusiasm are literally fanning the fires of marketeers and hype mongers.

I've used several versions of voice recognition over the years, and had no reason to keep them. I know now that I was sold something that promised more than it delivered, and I am not even sorry to have rejected it.

I'm a regular guy, who can't forgive a machine when a human can do a better job.

For an idea of how accurate the software is, take a look at a typical YouTube video with the CC turned on.

Alan Kelly, VerbatimIT

No one has posted to this recently, so I am hoping that people have had time to consider the benefits/challenges of MacSpeech. I just purchased a MacBook Pro which has boot camp with the plan to run Dragon Naturally Speaking on Windows 7 platform rather than MacSpeech which, when I first started looking into this had no where near the approval rating of Dragon. Someone above said that he/she would run Dragon with VMWare Fusion or Parallels, but I was told, even by AppleCare folks that there was no need for VMWare or Parallels, and in fact that they might make the process difficult.

Any insight? I don't want to pay for Dragon and wish I'd gone MacSpeech, likewise I don't want to change my mind and do MacSpeech and regret not having gone with my gut....

Thanks!

Thanks for this review! You sold me as I am a grad student doing ethnographic research! This may be the key to getting hours of my life back.

As a writer, I'm interested in this product to speed up the process of transcribing long interviews. Those of you who have used Dragon and prefer it, does Dragon offer anything that recognizes multiple voices? MacSpeech Scribe sounds ideal to me and a few mistakes are not a big deal if they can be easily fixed. Thanks for your input!

thanks for the medical transcription information

I recently purchased MacSpeech Scribe, a new product from the makers of MacSpeech Dictate. The purpose of scribe is to allow you to dictate remotely (voice recorder or iPhone) and then have your voice files transcribed by the application on your Apple computer. Unfortunately, like the other MacSpeech products, it falls well short of what one would expect. Despite it's slick and minimalistic interface, short training time, and reported similar 'engine' to the much better Dragon Naturally Speaking - it has some major flaws. The recognition is no where close to the 99% that MacSpeech developers claim (90-95% is much more realistic). Also, the interface used for correcting text is awkward and slow. Finally, you cannot open multiple files to be transcribed - in other words, you must transcribe one file and then save it, then close it, then open a new file for transcription. As with all the other MacSpeech products, this one fails to produce results anywhere near the level of Dragon and it has an awkward user interface. On one final note, MacSpeech's customer service leaves much to be desired.
My recommendation for Mac-users interested in voice recognition: run Dragon Professional or Medical versions in a Windows environment through VMWare (or Parallels).

Unless I misunderstand, this application's use for transcribing things like you mention--lectures and ethnographic surveys--will be limited. Why? Its accuracy is linked to it being trained to one of six voice profiles. You would have to create a voice profile for each professor (and would only be able to use 6), and in the case of surveys there would be many more than 6 individual voices involved.