LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Programming (https://www.linuxquestions.org/questions/programming-9/)
-   -   audio sampling Skews! (https://www.linuxquestions.org/questions/programming-9/audio-sampling-skews-66768/)

microtim 06-19-2003 11:41 PM

audio sampling Skews!
 
hi, i am working on the Zaurus handheld and am trying to have a bunch of these devices simultaneously record audio with a microphone. ideally, given that all Zauri "hear" the same sound, the resulting audio files would be almost exactly the same. for example, a clicking sound at regular 1 second intervals results in audio files that produce clicking sounds at regular 1 second intervals.

however, the problem is that the resulting audio files are NOT the exactly the same. after an hour or so, the audio files are off by up to 2 seconds which is unacceptable. doh!

i am looking for suggestions. is it hardware, software, or other related? how can i test and experiment to hopefully identify the problem? thanks.

-Tim

nowonmai 06-23-2003 04:50 AM

Re: audio sampling Skews!
 
Quote:

Originally posted by microtim
hi, i am working on the Zaurus handheld and am trying to have a bunch of these devices simultaneously record audio with a microphone. ideally, given that all Zauri "hear" the same sound, the resulting audio files would be almost exactly the same. for example, a clicking sound at regular 1 second intervals results in audio files that produce clicking sounds at regular 1 second intervals.

however, the problem is that the resulting audio files are NOT the exactly the same. after an hour or so, the audio files are off by up to 2 seconds which is unacceptable. doh!

i am looking for suggestions. is it hardware, software, or other related? how can i test and experiment to hopefully identify the problem? thanks.

-Tim

slight variations in sample rate, perhaps... as to why this is happening.. issues with multitasking, maybe? or varying processor/RTC speeds... 2 seconds in an hour is less than 0.03% variance

microtim 06-23-2003 06:45 PM

i've determined the problem/solution
 
google lead me to the answer: UCLA published a paper about their solution to synchronizing audio-recording on several devices (iPAQs). apparently, the problem is, essentially, that different devices have different clock sampling rates. Also, audio hardware has its own clock which causes further variation in audio sampling.

http://lecs.cs.ucla.edu/Publications...ots-sensor.pdf


All times are GMT -5. The time now is 11:14 PM.