Markus Wandel wrote:

> nobodyhome@mailinator.com (Markus Wandel) wrote in message news:<4d3787c3.0411061421.4bb79dd6@posting.google.com>...
> 
>>I've just bought an RCA RP-5016A voice recorder.
>>...
>>I've reverse engineered the file format enough to discover that in HQ
>>mode it consists of 80 bit blocks...
>>...
>>Samples of me saying a few words in both modes are here:
>>
>>http://wandel.dyndns.org/rp5016a_samples
>>
>>e0000001.voc     clip on HQ mode
>>e0000001.rhq     raw HQ (80 bit block) data (no encapsulation)
> 
> 
> After extensive googling, I've found only one standard speech codec
> with these characteristics, and it is G.729 CELP.  I cannot, however,
> find any software that I can just download and experiment with.  Could
> someone who does have such software run the raw data above through it,
> and tell me if it decodes?

You can download a 'free for development' G.729 codec here: 
http://www.vovida.org/applications/downloads/G729A/

> Email address in header is invalid.  I am really
> markus(AT)pinetree(DOT)org
>  
> Markus

-- 
Phil Frisbie, Jr.
Hawk Software
http://www.hawksoft.com

nobodyhome@mailinator.com (Markus Wandel) wrote in message news:<4d3787c3.0411061421.4bb79dd6@posting.google.com>...
> I've just bought an RCA RP-5016A voice recorder.
> ...
> I've reverse engineered the file format enough to discover that in HQ
> mode it consists of 80 bit blocks...
> ...
> Samples of me saying a few words in both modes are here:
> 
> http://wandel.dyndns.org/rp5016a_samples
> 
> e0000001.voc     clip on HQ mode
> e0000001.rhq     raw HQ (80 bit block) data (no encapsulation)

After extensive googling, I've found only one standard speech codec
with these characteristics, and it is G.729 CELP.  I cannot, however,
find any software that I can just download and experiment with.  Could
someone who does have such software run the raw data above through it,
and tell me if it decodes?

Email address in header is invalid.  I am really
markus(AT)pinetree(DOT)org
 
Markus

I've just bought an RCA RP-5016A voice recorder.

It is a nice gadget, and perfect in every way for what I want, except
that it uses a proprietary file format and no software is supplied to
export from this proprietary format to one understood by any other
program.

I've reverse engineered the file format enough to discover that in HQ
mode it consists of 80 bit blocks, and in SP mode it consists of 56
bit blocks, each encoding 10 milliseconds of audio.

I have verified that this is so, by reversing the order of the 80 bit
blocks in a sample HQ clip, and noting that it then plays backwards.

Does this sound familiar?  Can someone help me decode this using open
software under Linux?  Or at least point me to an algorithm?

Samples of me saying a few words in both modes are here:

http://wandel.dyndns.org/rp5016a_samples

e0000001.voc     clip on HQ mode
e0000001.rhq     raw HQ (80 bit block) data with encapsulation
stripped off
e0000002.voc     clip on SP mode
e0000002.rhq     raw SP (56 bit block) data with encapsulation
stripped off

Note that these are not Soundblaster .voc files, despite using the
same filename extension.

Can anyone help me?

Email address in header is invalid.  I am really
markus(AT)pinetree(DOT)org

Markus