|
Hi, I am developing code for a LPC using C6711 dsk. Could someone tell me if the raw values read from AD535 codec need to be scaled to +1 to -1 range from the 32768 to -32768 range before the autocorrelation, and reflection coefficents etc are calculated ? I have noticed that in one thesis paper they have done so, after performing pre-emphasis and hamming windowing on the raw data. I thought this scaling was only done for fixed point. Maurizio Gencarelli Electronics Engineer Royal Australian Air Force Please send any email to this newsgroup or |
|
|
Scaling Data for LPC
Started by ●August 14, 2002
Reply by ●August 15, 20022002-08-15
|
Actually it doesn't matter for the LPC analysis, because LPC
coefficients don't depend on the signal's energy You should only be careful in the further signal processing to treat the signal energy consistently Michael Shklyarman -----Original Message----- From: mgencarellli [mailto:] Sent: Wednesday, August 14, 2002 6:50 AM To: Subject: [speechcoding] Scaling Data for LPC Hi, I am developing code for a LPC using C6711 dsk. Could someone tell me if the raw values read from AD535 codec need to be scaled to +1 to -1 range from the 32768 to -32768 range before the autocorrelation, and reflection coefficents etc are calculated ? I have noticed that in one thesis paper they have done so, after performing pre-emphasis and hamming windowing on the raw data. I thought this scaling was only done for fixed point. Maurizio Gencarelli Electronics Engineer Royal Australian Air Force Please send any email to this newsgroup or _____________________________________ Note: If you do a simple "reply" with your email client, only the author of this message will receive your answer. You need to do a "reply all" if you want your answer to be distributed to the entire group. _____________________________________ About this discussion group: To Join: To Post: To Leave: Archives: http://www.yahoogroups.com/group/speechcoding Other DSP-Related Groups: http://www.dsprelated.com ">http://docs.yahoo.com/info/terms/ |
