|
Hi Thomas You are talking about hierarchical coding and transmission, and it has been discussed here widely. We have considered it and I think it is in the informal roadmap, IE when conditions are good, we have a low priority stream that enhances the user experience. The frequency selective narrowband HF channel represents rather different character compared to the VHF narrowband channel which is primarily a flat fading condition. Frequency diversity is the most efficient system wide method of solving the problem on HF I suspect when the frequency selective performance is deemed OK by the users, and that is the frequency selective channel properties generated by say, a >600km ionospheric path, then the users will begin to push the vertical incidence ability of FreeDV, this is a high random doppler profile and will likely spur another round of development to cope with it. regards glen english VK1XX On 7/09/2015 2:01 AM, Tomas Härdin
wrote:
Hi (I tried sending this a few days ago but forgot to confirm my subscription to this list - oops! So resending)My name is Tomas and I've been following the development of codec2 on and off for a few years now, and since getting my ham license earlier this year I find myself thinking about it more. A few days ago I had what I thought was a clever idea for fixing the "digital cliff" problem Mike mentioned in a talk that's up on YouTube (I forget which). Today I see on the roadmap post[1] that this is currently being worked on using two GMSK streams, but I thought "hey, maybe someone will find it interesting". So here goes: The idea assumes that codec2 can make use of bitrate peeling. That is, that we can split the stream up into two or more streams where the first one provides a rough but usable quality, and subsequent streams improve upon this. H.264's SVC would be an example from the video world So the idea is take these streams and modulate them onto a hierarchical QAM system. The simplest would be to take the current FDMDV modem and instead of using 14x QPSK (aka 4-QAM) carriers you use 7x 16-QAM carriers. You then code the more important bitstream into the most significant bits in each 16-QAM symbol, and the improvement stream into the lower bits (assuming the two streams have identical bitrate). Since you now have half the number of carriers you can put twice the amount of power into each carrier That's about it. Lots of variants are of course possible, but this should get the point across. I may experiment with the idea once I get some suitable SSB equipment, but for now I'm interested in feedback even if it's just shooting it down :) /Tomas, SA2TMS [1] http://www.rowetel.com/blog/?p=3931 -- - Glen English RF Communications and Electronics Engineer CORTEX RF & Pacific Media Technologies Pty Ltd ABN 40 075 532 008 PO Box 5231 Lyneham ACT 2602, Australia. au mobile : +61 (0)418 975077 |
------------------------------------------------------------------------------
_______________________________________________ Freetel-codec2 mailing list Freetel-codec2@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freetel-codec2
