Speex dev - May 2006 - Size of each block in a circular buffer and the samplerate

Thanks!  I hope this isn't a stupid question but if I want to play two 
or more voices at the same time, I just add the values of the samples 
right?  If I'm using an unsigned short, should I then clamp the values 
to 65535?

Jean-Marc Valin wrote:
>> Is frameSize in bytes?  If not, what unit is it in?  I need to know so
I
>> know how big to make each element in my circular buffer.
> 
> It's in samples, regardless of whether you use floats or shorts.
> 
>> Also, do I need to call
>>
>> speex_encoder_ctl( enc_state, SPEEX_SET_SAMPLING_RATE, &sampleRate
);
>>
>> Depending on sample rate I record my audio in?  If not, is there any 
>> benefit to setting the sample rate, such as reduced bandwidth?
> 
> The only thing Speex does with the sampling rate is compute the final
> bit-rate it corresponds to. If you use a "nominal" bit-rate (8
kHz for
> narrowband, 16 kHz for wideband, ...), which I strongly recommend, you
> don't even need to call this.
> 
> 	Jean-Marc
> 
>

Basically, yeah, but you want to add your samples as signed 32-bit ints, and
do the clamping (to +/- 32K) after all of the additions have taken place.

-- john
> -----Original Message-----
> From: speex-dev-bounces@xiph.org [mailto:speex-dev-bounces@xiph.org]On
> Behalf Of Kevin Jenkins
> Sent: Saturday, May 20, 2006 3:47 PM
> To: Jean-Marc Valin
> Cc: speex-dev@xiph.org
> Subject: Re: [Speex-dev] Size of each block in a circular buffer and the
> samplerate
>
>
> Thanks!  I hope this isn't a stupid question but if I want to play two
> or more voices at the same time, I just add the values of the samples
> right?  If I'm using an unsigned short, should I then clamp the values
> to 65535?
>
>

-----------

Again, once the decoder initialization is done, for every input frame:
speex_bits_read_from(&bits, input_bytes, nbBytes);
speex_decode(dec_state, &bits, output_frame);

where input_bytes is a (char *) containing the bit-stream data received 
for a frame,
nbBytes is the size (in bytes) of that bit-stream,
and output_frame is a (float *) and pints to the area where the decoded 
speech frame will be written. A NULL value as
the first argument indicates that we don't have the bits for the current 
frame.

-----------

In the snip above, the manual says to make the first argument NULL - 
however that looks like a mistake to me.  It should be the 2nd argument 
right?