On Wed, Apr 17, 2013 at 2:05 PM, David Rowe <[email protected]> wrote:
> Yes I think something like 800 bit/s is possible at roughly the same
> quality using Vector Quantisation of the LSPs. However this wouldn't be
> robust to channel errors.
I am not sure what you mean by "robust to channel errors". It seems to me
that the job of the codec itself should be to compress the audio and the
job of the FEC should be to make the communication channel robust. Is it a
design goal for codec 2 itself (absent FEC) to minimize the loss
in intelligibility due to bit errors? If so, it seems that it may lead to
a design that is sub-optimal for communication channels like IP in which
FEC (at the IP level) doesn't work (because packets with errors get dropped
before the application has a chance to apply FEC).
I would suggest that the design goals for the codec itself should be to use
minimum bandwidth and add minimum latency, with some consideration for
making the job of the FEC easier by minimizing how long an uncorrected bit
error would affect the audio (as Bruce referred to). Then the FEC should
use the remaining available bandwidth to make the channel as robust as
possible. I suppose the counter-argument might be that giving up some
compression in the codec in order to minimize the effect of uncorrected bit
errors would be more effective (in the RF case) than doing more compression
and more FEC.
I guess this touches on the issue where "digital" radio protocols in
general tend to maintain their audio quality as the signal gets weak until
they suddenly "fall off the cliff". As long as the FEC is able to correct
all of the errors, there is no loss of quality. When it can't, we get to
hear how "robust" the underlying codec is when dealing with uncorrected
errors. So there is a tradeoff between pushing the cliff out as far as
possible (more compression and more FEC) versus giving it a soft knee
(making the codec fall apart less with uncorrected errors). Which is more
effective at making communication possible at a given bit error rate is
probably going to take a lot more experimentation to figure out, and even
then be subjective.
Steve
--
Steve Strobel
Link Communications, Inc.
1035 Cerise Rd
Billings, MT 59101-7378
(406) 245-5002 ext 102
(406) 245-4889 (fax)
WWW: http://www.link-comm.com
MailTo:[email protected]
------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Freetel-codec2 mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
