So as every followup seems to have detailed, there is an increase in desired bandwidth with a direct need in required spectrum. If we can reduce spectrum, we increase distance the signal can transit. If we increase bandwidth for a particular size spectrum, we improve the amount of information we send.
The problems with current voice compression being understood have to do with remedial compression techniques based on available compute power. I suggested FPGA because of exactly this issue. Sure, people pick the easy route because they can buy those solutions and get into the marketplace faster. What needs to happen is the "Apple" thing. We need a company that actually cares enough about the quality of what it can ship, worries about power requirements and optimizes performance to create a truly awesome voice CODEC standard.
The cell phone market keeps trying to optimize the bandwidth needs to increase their spectrum's available capacity.
We are frustrated by the attributes of AM-VSB television characteristics vs ATSC coded VSB television. Because, the minimal available information transitions to no available information in a very short distance and signal level change. Thus we can't hear the TV at least. Either we get everything, or we get nothing.
This is where we are at with digital emission standards at this point. It's not the perfect solution because we are not sending enough information to recreate a perfect version of the original audio sample, for audio stuff. But, we are able to use the complete 12.5khz that D-Star is using (down from 20khz wide band FM is at now, and less than half of the old 30khz stuff that the old mobile phone radios were using). That 12.5khz has 2 channels in it. One for voice an done for data. So more information is bandwidth is available.
This is one of those experimentation moments. Not everyone is happy with where it is at, but without some more participation, those experimenting now will be the ones setting the standards, and if you are not happy with those results, it will be your fault not theirs, because you chose not to participate.
Gregg Wonderly W5GGW
On 4/25/2011 6:10 AM, Ben Jackson wrote:
On 4/23/2011 2:42 PM, Gordon JC Pearce wrote:
On Sat, 2011-04-23 at 10:42 -0500, Gregg Wonderly wrote:
In the end, digital compression of spectrum space is going to happen more and more. AM style broadcast is hugely inefficient even though it is painfully
Okay, but *why*? Why are we so obsessed with squeezing bandwidth down and down, at the expense of intelligibility?
You unfortunately provided data on why we should get ahead of crunching down bandwidth: Because sooner or later, we're going to get squeezed for bandwidth due to our spectrum being fairly empty and everyone and their brother wanting to push IP to their new wireless toaster service.
I'm not a fan of proprietary codecs but our lack of an alternative back in the 2000s caused D-STAR to be used with AMBE. Too bad, so sad. Don't support it, probably not going to use it. My worry is that even though we provided a alternative with Codec2, what cutting edge technology that will be here five years from now are we not developing because we were playing catch up?