I guess I just can not live with the "robotic" sound. This is unacceptable in this world of technology. I am trying to find away around this before I dump the whole D-Star thing. John PS: I also have a 91AD
--- In dstar_digital@yahoogroups.com, Tony Langdon <vk3...@...> wrote: > > At 05:05 AM 9/2/2010, you wrote: > > > >John, > > > >Try plugging a better quality speaker into the radio. > >I personally find this does wonders to increase fidelity. > >Icom should be ashamed of the speakers in the D-Star HT's. > > A better quality speaker certainly works, as anyone who's listened to > a DV Dongle on PC speakers can attest to. Actually, I find the > quality of the audio of my 91AD is very good (for a HT), only > complaint is there's not enough of it for some environments. > > As for the "robotic" sound, well, we're stuck with it. The whole > point of speech codecs is to literally throw away anything that's not > directly contributing to intelligibility, to reduce the > bitrate. There is a tradeoff here between bitrate and > fidelity. AMBE is designed to achieve very low bitrates, so it's not > going to sound very "natural". > > My own view of D-STAR audio is that while it doesn't sound natural, I > find it very intelligible, often more so than FM in the real world > (where there's a lot of poorly adjusted radios and radios with wonky > audio response). I'm also able to recognise who is speaking, so > enough of the voice is preserved to allow that. The AMBE vocoder > excels at what it sets out to do in my opinion - provide > communications grade speech at very low bitrates. > > > >I'm not sure, but there also may be something else going on as well. > >If I set up my 80 or 880 and connect to a reflector and do the same > >with the ID-1 using the same speaker, the fidelity improves big time > >with the 80 and 880 but the ID-1 still sounds much better (and the > >RF signal is much weaker). > > Different radios will have different audio responses. RF signal > strength is largely irrelevant in D-STAR, until the bit error rate > starts to increase significantly to the point that the FEC has > trouble correcting those errors. > > 73 de VK3JED / VK3IRL > http://vkradio.com >
