Re: Sample-rate conversion
Re: Sample-rate conversion
- Subject: Re: Sample-rate conversion
- From: David Duncan <email@hidden>
- Date: Wed, 4 Feb 2004 09:06:14 -0500
On Feb 4, 2004, at 08:39 AM, Jens Bauer wrote:
Also, to avoid aliasing artifacts in sampling, you should apply a
low-pass filter beforehand. BTW, do you know whether or not your 3.5
MHz sampled stream contains frequencies > 22050 Hz?
Uhm, I'd probably be most interested in the range (3500000 / 300) Hz
down to (3500000 / 2168) Hz, which would be 12kHz down to 1kHz.
I forgot to let you all know, that there is no noise between a
state-change on the source. The source is read from a file, which has
been calculated digitally, so I'd receive for instance 2168 hi signals
followed by 2168 lo signals (eg. the state would be low 2168 times in
a row)
If the highest frequency your interested is in 12KHz, then you only
need a 24KHz sample rate, especially since it's just a square wave.
Granted it will be less accurate for some frequencies than 3.5 Mhz, but
it will also be a LOT faster and won't alias. Personally for something
like this, I wouldn't even use a fixed sample rate, but would generate
samples on the fly at the output resolution. For example if you were
running at 44.1Khz with a command to generate 1600 hi/lo signals, 1600
* 44100 / 3500000 = 20.16 signals. You can use fractions as partial
hi/lo signals or just drop fractions altogether (although in either
case you would need to keep an accurate sample count or your timing
could drift).
--
Reality is what, when you stop believing in it, doesn't go away.
Failure is not an option. It is a privilege reserved for those who try.
David Duncan
_______________________________________________
coreaudio-api mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/coreaudio-api
Do not post admin requests to the list. They will be ignored.