Home |
Search |
Today's Posts |
|
#1
![]() |
|||
|
|||
![]()
On 2010-03-19, Mark Conrad wrote:
medical data on my Macs and PCs by voice, commonly dictating complex 600 word medical reports with zero text errors in four minutes time, wrestling phrases such as: "perioperative transesophageal echocardiography" Using what awesome software? nb |
#2
![]() |
|||
|
|||
![]()
In article , notbob
wrote: medical data on my Macs and PCs by voice, commonly dictating complex 600 word medical reports with zero text errors in four minutes time, wrestling phrases such as: "perioperative transesophageal echocardiography" Using what awesome software? Practically any modern SR software. It is kinda like ham radio, not so much what you have, but more important is how you use it. In this case, I used "MacSpeech Dictate 1.5.8" on a 3-year old MacBook Pro. Mac has to be one of the newer Intel-based Macs, software will not run on older Macs. One can get the same results using "Dragon NaturallySpeaking" - - - the "Pro" version 10.1 - - - which I also run on my old Mac hardware, using the Vista OS from Microsoft. About the only modern speech software that is difficult to achieve such accuracy and speed is "Windows Speech Recognition", (WSR) - which comes free with both Vista and Windows-7 OS. SR is a very inexact science at the present time, best estimates are that it will take another 20 years before it is anywhere near as good as a human, when it comes to converting speech to text. http://www.kurzweilai.net/articles/a...ml?printable=1 Scroll to near the end of the above website, to this sub-section: "Building HAL 's Language Knowledge Base" Read all the way to the end, that will give you a good idea what our children will be doing with SR 20 years from now. BTW, Nuance (Dragon) recently bought MacSpeech, so they are all one company now. A typical newbie SR user will be lucky to get 70% accuracy. As he gains experience, that will edge up to about 98%. In restricted speech like medical, where the same phrases like "perioperative transesophageal echocardiography" are used over and over again, the raw accuracy will edge up to 99%. ....or in my case 100%, in 3 out of 4 tries on that 600 word example - - - the bad "4th" try is invariably my fault, for mis-pronouncing one of the 600 words. Too much time way OT, but I gave you a decent answer to your question. My post immediately following this will show the entire text of my medical dictation, with 100% raw accuracy, no correction required, total dictation time 240 seconds. Mark |
Reply |
Thread Tools | Search this Thread |
Display Modes | |
|
|
![]() |
||||
Thread | Forum | |||
WTB Yaesu FT-817ND | Equipment | |||
WTB Yaesu FT 817ND / 817 | Equipment | |||
WTB Yaesu FT-817/817ND | Swap | |||
FS: Yaesu FT-817ND HF Mobile | Swap | |||
Yaesu FT-817ND | Dx |