>Somebody posted a little program for the jack audio server to support
doing
>exactly this for multiple non sync. soundcards.
>...
>Just a pointer to existing code.

Thanks!

I guess I'll need some pointer to the underlying theory as well since code
in general is quite bad commented. Hm. :)
		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com

Please post the link when you find it.



Dan Mills wrote:
> snappy wrote:
> 
> 
>>Hello DSP:ers,
>>
>>I have a delicate problem: two audio devices, not synchronized by any
>>means, are set to the same sampling rate. In reality the sampling rates
>>will differ a little bit, i.e. 44000 and 44001 Hz.
> 
> 
> Somebody posted a little program for the jack audio server to support doing
> exactly this for multiple non sync. soundcards.
> 
> Have a poke around in the jackit-devel archive and you should find it.
> 
> I **THINK** it was Torben who wrote it, but I am not sure.
> 
> Just a pointer to existing code.
> 
> Regards, Dan.   
> 

snappy wrote:

> 
> Hello DSP:ers,
> 
> I have a delicate problem: two audio devices, not synchronized by any
> means, are set to the same sampling rate. In reality the sampling rates
> will differ a little bit, i.e. 44000 and 44001 Hz.

Somebody posted a little program for the jack audio server to support doing
exactly this for multiple non sync. soundcards.

Have a poke around in the jackit-devel archive and you should find it.

I **THINK** it was Torben who wrote it, but I am not sure.

Just a pointer to existing code.

Regards, Dan.   

Hello DSP:ers,

I have a delicate problem: two audio devices, not synchronized by any
means, are set to the same sampling rate. In reality the sampling rates
will differ a little bit, i.e. 44000 and 44001 Hz.

Now I need a way of estimating the clock drift between those two devices,
preferably in real time: One device is the playback device, and the other
one is the recording device (recording the signal that is being played
back, after it has been converted to analog and affected by an impulse
response).

At the recording device I have access to the original signal (before
affected by the impulse response) as a reference.

What I want is an estimate of the clock drift so that I can resample the
streams to the same rate.

I've seen posts about PLLs, Costas Loop, sigital servo loops, etc which
seem to be in the right direction, but I have no knowledge wbout those
things. Where is the best place to start searching for information? Or are
there any examples of this procedure to have a look at?

Best regards
		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com