Regarding text robo-coders - Why reinvent the wheel?
I think they are an *excellent* idea, but not something worth dropping voice
encoding for, and the code-base out there is already extremely well defined.
There are tons of open-source projects dedicated to solely focusing on
speech-to-text and text-to-speech.
My recommendation is look at these and find some adaptable code that could be
worked into an existing ham radio text mode, like PSK31 but with significantly
more robust error correction to ensure near-perfect text decoding.
If Codec2 ever goes "variable bit rate", it might be possible to define text
robo-coders as the ultimate bottom end of the bitrate standard. But that's
beyond the scope of where things are today...
73, KE7HQY
--- On Wed, 1/2/12, acutler22 <acutle...@yahoo.com> wrote:
From: acutler22 <acutle...@yahoo.com>
Subject: Re: [Freetel-codec2] Giving up natural sound for further bandwidth
reduction
To: freetel-codec2@lists.sourceforge.net, steve.stro...@link-comm.com
Date: Wednesday, 1 February, 2012, 9:52 PM
It seems like 1000 bits/sec or under is the *real* "sweetspot" for raw without
FEC. I remember reading some discussion between FDMDV guys that if codec2 were
in that ~1000bit/sec range it would be relatively easy to adapt it to existing
software frameworks for HF digital voice schemes.
If you get natural voice close to robovoice, it seems like keeping things
"real" is a very wonderful thing. That benefit of *voice* codecs that encode
voice vs text has already been very eloqantly pointed out, and seems the
preferred method.
Remember we aren't just working with English here! Japanese is going to have to
sound pretty good too, and that would get *real* challenging with a
text-phonetical robocoder.
73 de Anthony, KE7HQY
--- On Wed, 1/2/12, Bruce Perens <br...@perens.com> wrote:
From: Bruce Perens <br...@perens.com>
Subject: Re: [Freetel-codec2] Giving up natural sound for further bandwidth
reduction
To: freetel-codec2@lists.sourceforge.net, steve.stro...@link-comm.com
Date: Wednesday, 1 February, 2012, 9:44 PM
If I'm not mistaken, the version with statistical codebook training work by
Jean-Marc runs at 1400 without delta coding. Delta coding would probably bring
it to 1000 bits per second without diminishing quality. This is the "raw" codec
without FEC, so this number goes up depending on how much FEC you insert.
To compare this to text, this is the equivalent of 200 5-bit teletype
characters per second or 125 8-bit ASCII characters. So, we're around 5 to 8
times the rate we might achieve with "text robophone" - speech-to-text to
text-to-speech, and maybe 3 to 5 times the rate we might achieve with "phoneme
robophone" - the same but using phoneme data.
Perhaps we could use a codebook approach to further compress the robophone
data. But it hardly seems worthwhile, because we're already close to robophone
rates with a real codec.
Thanks
Bruce
-----Inline Attachment Follows-----
------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
-----Inline Attachment Follows-----
_______________________________________________
Freetel-codec2 mailing list
Freetel-codec2@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
-----Inline Attachment Follows-----
------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
-----Inline Attachment Follows-----
_______________________________________________
Freetel-codec2 mailing list
Freetel-codec2@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
_______________________________________________
Freetel-codec2 mailing list
Freetel-codec2@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
