Understanding a simple Linear Predictive Coder
in [DSP]
|
From: Ross Clement (Email address invalid - do not use) on 20 Feb 2006 07:05 /* cross-posted to comp.dsp and comp.speech.research */ Hi everyone. If I wanted to understand Linear Predictive Coding by writing my own programs ... my preferred method of verifying that I understand things sufficiently, then what sources would people recommend. At present I'm looking at the relevant chapters in "Digital Processing of Speech Signals" by Rabiner & Schafer. As my maths background isn't as complete as a typical DSP person, it is, on first reading, fairly heavy going. What resources would people recommend that I read to help me understand this material? I will not be offended if people recommend that I keep on reading and working on the material in R&S until it makes sense if that's the best thing to do. Cheers, Ross-c
From: Andor on 20 Feb 2006 09:01 Ross Clement wrote: .... > What resources would people recommend that > I read to help me understand this material? J Makhoul: "Linear Prediction: A Tutorial Review", Proc. of the IEEE, Vol 63, No. 4, April 1975. It only mentions adaptive methods in passing (understanding this would require further reading), but gives a very readable introduction to the different flavours of LPC, including sections on least-squares time domain modeling, all-pole modeling, frequency domain analysis, error analysis, order selection and LP coefficient quantisation. For each of these sections there exists specific literature, but the article is a good overall start into LPC. Regards, Andor
From: Ross Clement (Email address invalid - do not use) on 20 Feb 2006 10:13 Thanks. I did an author search on IEEE Explore to find it (and the correction). A number of the titles of his other papers look like high priority reading as well. Cheers, Ross-c
From: Olivier Galibert on 20 Feb 2006 11:06 > Hi everyone. If I wanted to understand Linear Predictive Coding by > writing my own programs ... I remember http://svr-www.eng.cam.ac.uk/~ajr/SA95/ to be rather good. OG.
From: Ross Clement (Email address invalid - do not use) on 20 Feb 2006 11:16
Thanks. I actually went through Tony Robinson's whole site yesterday! I didn't quite take in the parameter estimation bit, though somehow today it looks much more straightforward. I think I'll continue reading some papers already recommended but then come back to TR's pages for a first attempt at implementation. Cheers, Ross-c |