KnowBrainer Speech Recognition
Decrease font size
Increase font size
Topic Title: What recording format does Dragon require?
Topic Summary: Frequency, bit rate, etc.
Created On: 04/09/2012 10:26 AM
Status: Post and Reply
Linear : Threading : Single : Branch
 What recording format does Dragon require?   - gcarson - 04/09/2012 10:26 AM  
 What recording format does Dragon require?   - Lunis Orcutt - 04/09/2012 02:05 PM  
 What recording format does Dragon require?   - gcarson - 04/09/2012 05:24 PM  
 What recording format does Dragon require?   - Lunis Orcutt - 04/09/2012 06:03 PM  
 What recording format does Dragon require?   - Chucker - 04/10/2012 08:04 AM  
Keyword
 04/09/2012 10:26 AM
User is offline View Users Profile Print this message


gcarson
Advanced Member

Posts: 189
Joined: 09/18/2007

I've been transcribing some dictation recorded with an Olympus LS-7, a Linear PCM recorder. I haven't been able to find any official description of the settings I should use for Dragon with this recorder, but I've been getting excellent accuracy (98-99%) using WAV mono, low microphone sensitivity and a Knowbrainer headset.

The specific file format with these settings is WAV 44.1 kHZ, 16 bit, Mono. This works fine, but I noticed that Olympus Sonority, the software that comes with the LS-7, is converting the WAV files before it sends them to Dragon for transcription. I couldn't tell what it was converting them to, however.

Anyway, I'm just asking about the file format Dragon requires so I can get the LS-7 set up correctly. Thanks in advance.



-------------------------

Compaq 6730b laptop Intel Core 2 Duo p8400 2.26 Ghz 1.93 G RAM (available) Intel Wireless Wifi Link 5100 adapter / Windows XP SP3 / Knowbrainer purchases: Philips 9600 DPM, Knowbrainer headset microphone, Samson Airline 77, Olympus DS5000id, DNS 10.1, Knowbrainer 2008, Parrott Half Duplex VXI USB Sound Pod

 04/09/2012 02:05 PM
User is offline View Users Profile Print this message

Author Icon
Lunis Orcutt
Top-Tier Member

Posts: 22604
Joined: 10/01/2006

The 3 Olympus LS-7 includes 3 recording formats which are PCM (WAV), MP3 and WMA. You're already using the recorder correctly for NaturallySpeaking because the Sonority application is converting your 44.1 kHz wav files to 22.5 kHz so that Dragon can handle them. You are doing everything right, down to making monophonic recordings. The only problem with WAV files is their sheer size. As long as you don't need to e-mail your sound files as attachments, require cue review, slider control or seamless punch in for editing, this recorder will work fine with NaturallySpeaking.

Keep in mind that the Olympus LS-7 is a 94 kHz 24-bit Tresmic (3 microphone) recording studio quality recorder which is really not designed for speech recognition and doesn't include basic speech recognition amenities. Because we are in the speech recognition business, we only recommend
recorders that feature Olympus 12 to 1 compressed lossless fidelity DS2 algorithms but even what we consider to be a proper digital recorder for speech recognition would likely score no higher than 98% accuracy so unless you need the other functionality, the LS-7 is performing perfectly.

Just out of curiosity, it would appear that you already own an Olympus DS-5000id which for speech recognition purposes, would be considered superior in functionality, security and recording quality. Because you already own the Olympus Cadillac, we were wondering why you downgraded to the LS-7?


-------------------------


Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 


 


 


 

 04/09/2012 05:24 PM
User is offline View Users Profile Print this message


gcarson
Advanced Member

Posts: 189
Joined: 09/18/2007

Thanks. I got the LS-7 for other uses. Just wanted to see how well it worked for dictation.

It's a fantastic recorder and surprisingly accurate with Dragon, but you're right about its limitations for dictation. Its biggest problem, like you said, is its total lack of editing functions like insert and append, a major handicap. Its onboard speaker isn't very loud, either, but it's better than other LPCM recorders I've tried. Several of them don't even have speakers.

I use recorders for several different things: dictation, stereo recording, podcasting, etc., etc. I currently use my DS5000id for dictation and the LS-7 for everything else, but I've been looking for a "universal" recorder that combines the best features of both lines of recorders into one machine. It doesn't exist, though. There's the high-end dictation market, then there's the stereo recorder market for musicians and sound professionals.

I was hoping the new Grundig Digta might do the trick since it can be used for both dictation and stereo recording. I'm assuming it's as accurate as the DS5000id for dictation, but so far I have no idea how good it is at recording stereo. You probably can't do this, but is there any chance you could post some sample stereo audio files recorded with the Digta so we could hear how it sounds?



-------------------------

Compaq 6730b laptop Intel Core 2 Duo p8400 2.26 Ghz 1.93 G RAM (available) Intel Wireless Wifi Link 5100 adapter / Windows XP SP3 / Knowbrainer purchases: Philips 9600 DPM, Knowbrainer headset microphone, Samson Airline 77, Olympus DS5000id, DNS 10.1, Knowbrainer 2008, Parrott Half Duplex VXI USB Sound Pod

 04/09/2012 06:03 PM
User is offline View Users Profile Print this message

Author Icon
Lunis Orcutt
Top-Tier Member

Posts: 22604
Joined: 10/01/2006

We're not really in a position to create high fidelity Grundig recordings of music but your assumption is correct. The Grundig Digta 703 digital recorder is really designed to be a jack of all trades with dictation in the forefront but additionally including dual high fidelity stereo microphone elements for concert recording like the LS-7. If you're interested in this recorder, contact us off-line because the unit is MAP priced but were willing to sell it at our cost.

-------------------------


Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 


 


 


 

 04/10/2012 08:04 AM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 9671
Joined: 10/10/2006

gcarson,

Just to add a smidgen to what Lunis has already pointed out, Dragon NaturallySpeaking excepts WAV, MP3, and WMA, in addition to DS2 for the Olympus and related DVR's. Dragon does not use conversion because conversion distorts the quality of the audio files. Dragon uses a unique and proprietary extraction algorithm that takes the audio file input and extract the correct sampling rate. Dragon will support any audio file that is within the range of sampling rates from 11.025 kHz-99 kHz, 16 bit Mono. The latter is the sample size that Dragon will except, the former is the sampling frequency. Together these constitute the sampling rate. All audio files that are supported within this range are automatically converted (extracted) so that they present those audio files to Dragon in the format (sampling rate) 22.05 kHz 16-bit Mono for DNS 11-11.5 and 11.025 kHz 16-bit Mono for DNS 10.1 and below. As long as your audio files all within the sampling frequency range and use a 16-bit Mono sample size, Dragon will recognize and transcribed them. The major caveat is that some users attempt to record in stereo and then try to get Dragon to recognize such audio files, which you will not. In addition, Dragon will not recognize any sampling rate below 11.025 kHz 16-bit Mono.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make Dragon NaturallySpeaking, We make it better!

Education is when you read the fine print. Experience is what you get if you don't. - Pete Seeger



-------------------------

Statistics
27366 users are registered to the KnowBrainer Speech Recognition forum.
There are currently 0 users logged in.
The most users ever online was 2028 on 04/05/2013 at 07:36 PM.
There are currently 120 guests browsing this forum, which makes a total of 120 users using this forum.

FuseTalk Standard Edition v4.0 - © 1999-2013 FuseTalk™ Inc. All rights reserved.