Without in-depth analysis of the source code, is it possible to make an informed decision on whether or not a codec is good? And by “good”, I mean that it would serve adequately in VOIP/Video Chat applications for the time being and one would not expect it to be outdated soon.
I’m trying to decide if a company, responsible for maintaining a particular codec, is worth their salt or if they’re on their way out thanks to skype/qik/gtalk/ichat. However, for this question, I’m trying to judge the codec itself in a vacuum, ignoring the fact that there is a lot of competition making life very difficult for the business in question.
Thanks,
~Jordan
MOS is good for evaluating audio quality, but MOS depends on many things other than “codec quality”, in particular delay, echo, packet loss — and here codec can play a part, in that different codecs have different sensitivities to packet loss – G.729 is really hurt by it, even with concealment, while iLBC at 0 loss may be a smidge behind AMR-NB and G.711, but under loss it is very resilient.
Codecs have many things to assess: no-loss quality, performance under loss and ability to conceal, ability to render non-speech (music-on-hold, etc), algorithmic delay, performance in different languages, support for variable bitrates or sample rates, etc, etc. Then there are externals: royalties/patents/patent-lawsuit-risk, quality and number of implementations, how widespread and standardized it is, etc.