Transcriber again

[Tom Honeyman]

I believe that Transcriber does not in fact support ogg. At least, not  
on Mac OS X. Loading a test file, I can see (go to File ->  
Informations) that the file is being interpreted as: "Alaw 16kHz mono"  
not ogg vorbis, which basically means it will play back as noise.

My digging around shows inconsistent information about ogg support in  
Transcriber, which currently uses the "Snack" library for handling  
audio. However, the primary source (the home page for the Snack  
library) says it supports the following:

"Supported sound file formats: WAV, AU, AIFF, MP3, CSL, SD, SMP, and  
NIST/Sphere"

I haven't looked closely enough, but I believe this leaves you without  
the option of having a compressed format. Can anyone confirm whether  
any of the above allow for compression (excluding the fact that WAV  
files can technically wrap a compressed format)? Otherwise, I suggest  
mono, 16bit, 22.1kHz WAV files if working on speech data. I reckon  
even 11.05kHz would be acceptable if you're not transcribing close  
phonetic detail with significant information above 5000Hz.

Otherwise, you may have to wait for:

New!
May 08

Transcriber is being completely redevelopped by Bertin Technologies  
and should be released by Q2 09. That new version will be based on  
Annotation Graph, which will be the default format of the new  
annotation files. In a nutshell, an Annotation Graph represents  
linguistic annotations of time series data as a formal graph. An AG  
also abstracts away from file formats and provides a logical layer for  
annotation systems. A beta version should be available in a couple of  
month. Stay tuned.


No update so far of course...

Cheers,

Tom

On 18/01/2010, at 8:00 PM, David Nash wrote:

> At 10:42 AM +1100 4/4/09, Doug Marmion wrote:
>> He did however point me to <http://www.loria.fr/~quignard/Transcriber/ 
>> >, where you can get Transcriber 1.5.2.
>
> Further, at http://users.rcn.com/astockdale/ you can get http://users.rcn.com/astockdale/Transcription.zip 
>  which contains 'Transcriber Guide.pdf' 'A GUIDE TO USING  
> TRANSCRIBER TO TRANSCRIBE AND PREPARE DIGITAL AUDIO DATA FOR  
> QUALITATIVE ANALYSIS', where on page 3 we read:
>
> "The current version (1.5.1) doesn't correctly calculate the time  
> points within MP3 files. This appears to be a problem in the Snack  
> Sound Library used by Transcriber. (Note that Snack will be replaced  
> in the next official version.) The problem results in the timing  
> being off by slightly more than a quarter of a second every minute.  
> On very short files this is almost unnoticeable but on longer files  
> it will be very noticeable, resulting in about 17 seconds of drift  
> on a one hour interview. I would suggest using WAV or OGG files with  
> Transcriber. These are both handled correctly.'
>
> I have a 40-minute OGG file (converted from WAV) which plays fine in  
> Audacity, M Player OS X, and VLC.  However, when I open the same  
> file in Transcriber (1.5.1 or 1.5.2) it plays as noise, and the  
> waveform display is mostly black.  Any suggestions as to what is  
> going on?
>
> David

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/resource-network-linguistic-diversity/attachments/20100123/7dd01e62/attachment.htm>