Speex dev - Aug 2004 - decoder give me values over 32000!

Hello,

    we are using  speex 1.0.3 under windows.  We code samples with 
values from  -32.000 to 32.000 (signed short), but when we decode these 
samples the values range from -45.000 to 45.000.

    What is the problem?

<p>    Thanks
--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'speex-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

I have been receiving the W32.Netsky virus from the speex email server.
Just giving you a heads up.

<p>Tony Thompson

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'speex-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

> I have been receiving the W32.Netsky virus from the speex email server.
> Just giving you a heads up.
Our mail server is running Linux, so unless Monty spent two hours to get it to
run under Wine, I think you should look elsewhere for the origin ;)

     Jean-Marc
--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'speex-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

>     we are using  speex 1.0.3 under windows.  We code samples with 
> values from  -32.000 to 32.000 (signed short), but when we decode these 
> samples the values range from -45.000 to 45.000.
> 
>     What is the problem?
I'd say the main problem is that you should have the input be that close to
the
bounds. The behaviour is probably normal (though a bit more extreme than usual)
giving that Speex is lossy and it's the perceptual quality that counts. I
suggest you just set your input level lower.

    Jean-Marc

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'speex-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

I think there is a problem.  If we are coding   PCM 16 mono, then, when  
we decode  the  samples  we should obtain PCM 16 mono again, but wave 
data over 16bits is extracted from the decoder.  If the input level is 
reduced we could loss wave information.

Thanks for your help

<p>Jean-Marc Valin wrote:
>>    we are using  speex 1.0.3 under windows.  We code samples with 
>>values from  -32.000 to 32.000 (signed short), but when we decode these 
>>samples the values range from -45.000 to 45.000.
>>
>>    What is the problem?
>>    
>>
>
>I'd say the main problem is that you should have the input be that close
to the
>bounds. The behaviour is probably normal (though a bit more extreme than
usual)
>giving that Speex is lossy and it's the perceptual quality that counts.
I
>suggest you just set your input level lower.
>
>    Jean-Marc
>
>--- >8 ----
>List archives:  http://www.xiph.org/archives/
>Ogg project homepage: http://www.xiph.org/ogg/
>To unsubscribe from this list, send a message to
'speex-dev-request@xiph.org'
>containing only the word 'unsubscribe' in the body.  No subject is
needed.
>Unsubscribe messages sent to the list will be ignored/filtered.
>
>  
>
--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'speex-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.