Mark you said..."The question I have been struggling with is how much is enough/too much. I guess what I am looking for is a curve showing bandwidth vs. throughput for parallel tone modems, or maybe more precisely where is the point of diminishing returns? "
Yes...perhaps you are right about seeing curve...but I'm not sure that there is enough data (on/off the air or simulator) available to make such a curve. Rick...you said..."The maximum accepted bandwidth for most modes is the width of an SSB transmitter since you can not go wider than that and communicate with the typical rigs of the day." But is this any reason to not use wider modes? One luxury that we (amateur radio) don't have is all the frequencies we need to use modes that are frequency/band dependent. If you are going to work short distances, you have 160M, 80M, 40M and maybe 30M. If you are going to work longer distances (DX), you will probably need to use 30M, 20M and above. Thus is a parallel tone modem works as you describe (and I believe that you are correct), then we need to focus on parallel tone modems. And Paul raise the question of needs..."There seems to be an idea that there's a 'one size fits all' solution. That's probably the END of a (failed) search for a solution, rather than the beginning." If we can agree that the parallel tone modems are the modem of preference, then we need to look at how to optimize them for varying bands and band conditions as well as operating modes. Right now I see MFSK-16 as one of the best "chat" modes and MT63-2K are the robust high throughput mode. MultiPSk and the Linux equal have both MFSK16 and MT63 at various configurations. What we aren't defining is actually what features do we want/need modes to fill. As Rick said..."If it wasn't for the potential for supporting emergency communications, none of this would really matter very much since we rarely would need to send large amounts of data for typical radio amateur chats and the superficial hello/goodbye type contacts that most radio amateurs prefer." I work with/in disaster relief communications and high throughput and robustness is greatly needed...if MT63-2K is 200 wpm with a -5 dB SNR, I can see the need for 600-800 WPM throughput at a -10- to -15 dB SNR. But this is rare. Even if I send a picture of my QSL card or shack, rig, etc, I don't need this throughput but would like the robustness so I could work at 5 watts just like I would at 100 watts (SSB PEP). If I use a chant mode, (and my XYL, WB5WXY) is typing, I want a mode that will operate and 60+ WPM and again work down in the -5 to -15 dB range so my 5 watt QRP rig will give me the same performance as my 100 watt rig. I might also send a small picture of my camping site I made with my 3 MPix digital camera. Others may want to send high resolution photos or digital maps or schematics, etc. I don't know what mode is or will be the most robust and what throughput can be obtained. I do think we need to keep the bandwidth as small as reasonable to accomplish the desired results. If I only need a robust, low throughput chat mode, then something under 100 KHz would seem reasonable. Medium throughput and robust modes may be able to use 500 KHz up to 2-3 KHz, But high throughput, robust modes, as needed for disaster communications, may require bandwidths more than are currently available using current hardware except perhaps the SDR-1000 and like SDR transceivers. But again, that will just create a new batch of hardware to buy. I would like to hear your thoughts. 73, Walt/K5YFW -----Original Message----- Rick, KV9U wrote: The maximum accepted bandwidth for most modes is the width of an SSB transmitter since you can not go wider than that and communicate with the typical rigs of the day. We already have the basic modes to work high speeds with good conditions and slower speeds under difficult conditions. What we don't have are the types of ARQ modes that are adaptive to those conditions. It seems that the serial tone modems might work close to the MUF, but parallel tone modems are typically superior for a wider array of conditions. So my thinking is that we would want an adaptive parallel tone modem that can change the number of tones, spacing of the tones, length of tones, and modulation of the tones, to dynamically meet the conditions. When conditions get down below zero db S/N, the mode may have to change to the ones that can handle working deeper into the noise. And another factor would be to have the ability for the operator to select the maximum bandwidth they wish to use. Maybe that might be only 500 Hz, but maybe a lot more for certain purposes such as sending larger amounts of data, images/documents, emergency communications, etc. If it wasn't for the potential for supporting emergency communications, none of this would really matter very much since we rarely would need to send large amounts of data for typical radio amateur chats and the superficial hello/goodbye type contacts that most radio amateurs prefer. 73, Rick, KV9U -----Original Message----- Paul Schmidt, K9PS wrote: There's the billion [insert local currency here] question. Or actually two questions: what's the maximum, and what's appropriate? There seems to be an idea that there's a 'one size fits all' solution. That's probably the END of a (failed) search for a solution, rather than the beginning. Each of the bands has its unique characteristics which vary depending on time of day, time of the sunspot cycle, etc. The maximum needs to be set high enough to handle anything that is reasonably expected to be appropriate - much like the speed limit on the roads we use. But that doesn't mean that the maximum is appropriate all the time. Most states have a speed limit on the interstates... but if there's snow and slush on the road, anyone running maximum legal speed should be removed from the highway system. The speed limit signs on the interstates of any major city during rush hour certainly make travelling the roads *legal* at speeds much greater than the 0-5 MPH the traffic often allows. When the regulation by bandwidth issue came up in the U.S. it got combined with issues of automatic control All of the sudden, instead of the regulatory change being centered on "what should we allow as options?", it became focused on "what might some self-centered, egotistical bozo with a big amplifier, big antenna, and small brain do?"... and there went our regulatory flexibility. Somehow, if we're going to update our antique regulations, we're going to have to get past the idea that "legal" should be the most restrictive definition, thereby forcing things to work -- and replace it with the idea that "legal" should be the least restrictive definition which will allow things to work. -----Original Message----- Mark Miller, N5RFX wrote: I think there is no doubt that this is true. The question I have been struggling with is how much is enough/too much. I guess what I am looking for is a curve showing bandwidth vs. throughput for parallel tone modems, or maybe more precisely where is the point of diminishing returns? I am sure there are many factors that would affect the curves. I know from experience that MT63 is a great mode when making very long and many hop contacts. I have watched the fading move across the waterfall, and my text be 100% correct. I am sure that this is because of the redundancy of the code spread out over many frequencies. MFSK16 sometimes performs better under certain conditions with a quarter of the bandwidth. What my question boils down to is generally, what is the accepted maximum bandwidth of any signal in the Amateur HF bands, given the finite spectrum and many interests? 73, Mark N5RFX
