|
|
This example shows how to process audio data at a different sampling rate from the standard found in the Chameleon model #01, which is fixed at 48000 Hz. Since this sampling rate cannot be changed, the incoming audio stream at 48000 Hz is resampled to achieve the desired sample rate. Processing is performed at the new sample rate, and then the processed audio is resampled back to 48000 Hz to be sent to the audio output.
The process of resampling is done by filtering the input signal with a Kaiser windowed Low Pass FIR filter, set to its polyphase form properly for the resampling operation. The filters are computed in Scilab using the routines in the file 'resfilt.sci'.
The example comes with a default resampling ratio of 2/3, so the application processing is done at 32000 Hz (making it suitable and more efficient than 48000 Hz for string instruments and voice processing). The number of coefficients is set to 61 for both filters, which gives good results for audio signals with low components above 15000 Hz. The two resampling processes take about a 8% of the processing power, so a high overall performance boost is achieved (about a 25%).
E-mail: 
info@telesys.ru
