David Rowe wrote on 2012-06-05 22:04: > On Tue, 2012-06-05 at 17:51 +0200, Michael Hartje wrote: >> Dear list, >> >> I found it very helpful for some own live experiments to use the >> codec2-dev svn archive with the very useful pipeline calling by >> >> rec | c2enc $RATE - -| c2dec $RATE - - | play >> >> (with detailed parameters) >> export RATE=2500 >> rec -r 8000 -t raw -s -b 16 -c 1 -|./c2enc $RATE - -|./c2dec $RATE - >> -|play -r 8000 -t raw -s -b 16 -c 1 -
I'm doing the same, and it works as it should be nice in pipes. >> when I changed RATE to 1200, 1400, 1500 I did not remark any significant >> difference in quality -- this was very surprising to me. >> >> What could I expect will be the difference in audio quality? >> Can anybody explain that to me? >> Is there any simple parameter in the obtained audio at the end which >> could be heard clearly? or: is the major difference only seen by some >> measurements (e.g. noise) > > Try taking a single source file (plenty of them in the raw dir) and > processing it with the 3 rates, then listen to the output files using > the ../script/menu utility that lets you play them quickly next to each > other. David and others may correct me: What you hear is the evolution of codec2: First 2500 bit/s was a great success, and it sounded good enough for that bit rate and the target applications. Later, codec2 was improved through various clever tricks, which resulted in nearly the same quality with much lower bit rate - 1200 is less than halve of 2500. If you compare 1200 bit/sec with 2500 bit/sec you _can_ hear a clear difference, but IMO it is not big and by far not as big as the drop in necessary bitrate. There _are_ methods to measure the quality, even kind of absolute scales that make it possible to compare different systems, but so for I do not know of any value for Codec2 of any bitrate. Some of these methods are standardized (ITU), but not license free, so this would be some kind of barrier. David, Jean-Marc, would it be possible to backport the methods of the 1200 bit/sec version to 2500 bit/sec and get even better quality on 2500 bit/sec? I would say 2500 bit/sec is not yet toll quality, but maybe it could be brought there. <dream> If some kind of hierarchical modulation could be used and the codec supported both 1200 bit/sec and 2500 bit/sec, a communication system could be designed, that would support good quality with a fallback when channel properties get worse and data is not decodable. </dream> <real> Tetra, the European brother of P25, has a codec (AMR) that is grouping its bits in three classes and secures them with different levels of forward error correction (FEC). They end up at a bit below 5 kBit/sec. The documentation/specification is currently available free of charge, but not distributable. Very interesting to read! It's European Standard ETSI EN 300 395-2 See http://tetra.osmocom.org/trac/wiki/Speech_Codec </real> Regards Patrick -- Engineers motto: cheap, good, fast: choose any two Patrick Strasser <patrick at wirklich priv at> ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Freetel-codec2 mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/freetel-codec2
