| Author |
Message |
Tom
Guest
|
 Posted:
Thu Jan 06, 2005 7:27 pm Post subject:
AEC on two different devices |
|
|
Hi,
I'm working on an acoustic echo canceller but I
can't get it to work when two different devices
are used for recording and playback.
I believe that this is because the near end signal
is recorded at a slightly different samplerate than
the samplerate by which the far end signal is played.
If someone has any idea on how to solve this problem,
please let me know.
Any help/pointers would be appreciated a lot.
Tom. |
|
| Back to top |
|
 |
Jon Harris
Guest
|
| Posted:
Thu Jan 06, 2005 11:13 pm Post subject:
Re: AEC on two different devices |
|
|
I've heard this can be a significant problem. The quick answer is to
synchronize them, but I don't know exactly how. The ideal would be to use
hardware where synchronization was guaranteed. If this is not feasible, some
type of sample rate conversion will probably be necessary.
"Tom" <ck_1979@hotmail.com> wrote in message
news:1105021622.208841.193960@f14g2000cwb.googlegroups.com...
| Quote: |
Hi,
I'm working on an acoustic echo canceller but I
can't get it to work when two different devices
are used for recording and playback.
I believe that this is because the near end signal
is recorded at a slightly different samplerate than
the samplerate by which the far end signal is played.
If someone has any idea on how to solve this problem,
please let me know.
Any help/pointers would be appreciated a lot. |
|
|
| Back to top |
|
|
Tom
Guest
|
| Posted:
Fri Jan 07, 2005 1:53 pm Post subject:
Re: AEC on two different devices |
|
|
Thanks for your answer.
Unfortunately, I can't use specific hardware because the AEC
runs on a regular PC.
I agree that I need to resample to a matching samplerate, but
my problem is that I do not know what samplerate I should resample
to. Do you know some how I can detect this? |
|
| Back to top |
|
|
Richard Owlett
Guest
|
| Posted:
Mon Jan 10, 2005 1:16 am Post subject:
[Naive(?) Idea] Re: AEC on two different devices |
|
|
Tom wrote:
| Quote: | Thanks for your answer.
Unfortunately, I can't use specific hardware because the AEC
runs on a regular PC.
I agree that I need to resample to a matching samplerate, but
my problem is that I do not know what samplerate I should resample
to. Do you know some how I can detect this?
|
My DSP skills being weak, I was looking for another solution path.
My underlying assumption are that:
1. although the sample rates are different, they do *NOT*
drift *significantly*.
2. the input of the recording device can [ at least temporarily ]
be connected to playback device output.
Arbitrarily define sampling frequency of playback device to be reference
standard F1.
Connect playback device output directly to recording device input.
Write a program to:
1. produce square wave at playback device output at FF=(F1)/(2^n) .
2. take absolute value of input to recording device
3. Subtract average value leaving a signal related to "beat frequency"
of F1 and sample rate of recording device F2. FB = |F1-F2|
4. Iterate to minimize FB.
( I suspect a first pass minimization may be minimizing *amplitude*
of the beat signal. )
PS to experts
Please verbosely tell me if I'm off in left field. Thanks. |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in