Codec2 Project Asks FCC To Modernize Regulations

Bruce Perens writes "The Codec2 project has developed FreeDV, a program to encode digital voice on two-way radio in only 1.125 KHz of bandwidth. But FCC regulations aren't up-to-speed with the challenges of software-defined radio and Open Source. A 24 page FCC filing created by Bruce Perens proposes that FCC allow all digital modulations and published digital codes on ham radio and switch to bandwidth-based regulation."

Comments

About Codec2

[thatkid_2002]

For those interested in knowing what Codec2 is, there's a video from Linux Conference Australia 2012 which gives a pretty good (and gentle) overview.
http://mirror.linux.org.au/linux.conf.au/2012/Codec_2_Open_Source_Speech_Coding_at_2400_bits_and_Below.ogv

good luck with that

[Anonymous Coward]

The regulations to move from 25 khz to 12.5 khz just took effect this year which forced many cash strapped agencies and municipalities to buy new radio systems. I don't think there will be much support for further narrowing bandwidth any time soon.

Re:good luck with that

[Dan Dankleton]

This will affect the amateur HF bands, not the agency bands. At the moment, US rules have separate sub-bands for voice, data and image transmissions. This does not really fit with how modern digital schemes transmit - there could be a lot of metadata carried with digital voice signals, for example. What this proposal does is do away with the rules which say where you can transmit voice and replace it with rules which say you can transmit any signal which takes less than X khz bandwidth in this segment.

Re:good luck with that

[vlm]

At the moment, US rules have separate sub-bands for voice, data and image transmissions.

Note that the FCC currently regulates by information content, not modulation. I guess a /. analogy is its like classifying networking tech first by layer 2/3 at the bottom as the "fundamental layer", then layer 1, then the upper layers, sorta. Which is obviously wrong. So if I send you a string of ones and zeros in PSK-31 modulation or whatever, that represents speech, we have to go in one subband on HF. Then if I send you ones and zeros representing text data, like this post, we have to QSY to yet another subband. Then if I send you ones and zeros representing a goatse jpeg picture (which would run afoul of the fcc reg against obscenity, but I digress), we have to QSY to yet another frequency... even if its all the same modulation technology, the same "stuff on the airwaves" for the ones and zeros. Maybe another way to put it, is our "MIME-type" is our frequency subband, not something a little more modern or realistic.

Legally/technically if I went on CW (aka morse code) and told you verbally how to draw an amplifier schematic that could be seen as illegal as its obviously image traffic, no worse than if I sent you postscript code or the ones and zeros of a .png file.

There are two killer problems which may or may not be discussed here.

First the proposal claims to promote "paperwork reduction" while installing a whole new crazy array of complicated regs ON TOP OF the existing overall rules for reasonable and prudent and good engineering practice and emergency traffic priority or WTF the exact phrases. In my opinion as a third generation ham with over three decades of experience, what works with the smart people on the VHF/UHF/microwave bands should work with the glorified CBers on the HF bands, which in summary is do whatever the heck you want as long as its good engineering practice and stays within ham band edges (note this is a simplification, but basically correct). Yes I know this is the peak of this solar cycle but when 10M is closed and dead I see no reason my buddy and I shoudn't be able to use 200 KHz of wideband FM on 10M across the city if we please, because it certainly can't hurt anyone. Or do something weird on 160M during the day time in summer, why not? So the most rational bandplan is not this proposal, but is: Do whatever the F you want between 3.5 MHz and 4 MHz as long as it stays in band edges and follows all the other numerous "content and performance based" regulations (like no intentional interference, good engineering standards, content rules wrt obscenity (which is certainly ignored on 80 and to some extent on 20 aka the high tech redneck CB bands, so why can't we accept that we'll ignore bandwidth limits too?), emergency traffic gets priority, blah blah blah)

The other thing carefully not discussed, regardless if true or not, the widely held belief was Bonnie's plan from a decade ago, mentioned in this very proposal, was just the thin edge of the wedge to fill 20M from band edge to band edge with psuedo-commercial winlink traffic. There's two problems with this. The first is it doesn't seem to modify the unattended operation rules but then again its the thin edge, the next proposal will be expanding the unattended operation subbands to 3.5 to 4 MHz for example, etc. The second is, see #1 above, why should anyone care if the vast majority of hams wanted to use winlink, if so, then let them... its not the "SSB-preservation amateur radio service" or the "AM amateur radio service" or for that matter the "CW amateur radio service".

Well, this mostly accurate history lesson outta stir the pot some.

Be careful...

[fuzzyfuzzyfungus]

"Dear economically invisible 'ham radio' users;

  After an exhaustive modernization study underwritten in part by our good friends at Verizon, we have concluded that the future of digital voice should really cost ~$100/month and rely entirely on proprietary hardware and firmware. To this end, we will be lowballing every last scrap of spectrum we can to the nation's incumbent telcos as soon as possible.

XOXOXO,

The FCC"

I applaud modernization efforts, there is no reason why 'ham radio' should be forced to stick to ancient technology for reasons of sheer regulatory inertia when it could be fertile ground for experimentation; but I worry that (given the, um, limited war chests of ham nerds vs. other spectrum users) that perfectly sensible re-examinations of legacy rules might well end up becoming an exercise in malignant entities with better lobbyists using the rexamination of legacy rules to appropriate spectrum that was protected at the cost of a certain amount of anachronism...

Re:Be careful...

[Muad'Dave]

...it could be fertile ground for experimentation...

It is a fertile ground for experimentation! You need look no farther than the recent influx of extremely spectrum-efficient modes developed by K1JT. He's developed modes tailored for most any propagation mode/band including meteor scatter, moonbounce, etc.

The newest of the lot, the JT9 modes, are capable of decoding signals as far as 42dB into the noise!. The fastest JT9 mode takes 1 minute per transmission but can decode at a S/N of -27dB - that's noise with 500x the power of the signal.

Take a look at the WSPR page - on it you can access a database of WSPR transmissions, some of them at amazingly high km/Watt ratios.

I agree with the end goal, Bruce

[Muad'Dave]

The end goal of moving toward more spectrally-efficient digital modes for all forms of communication is laudable, but I think that there still needs to be some 'semi-official' protection for the traditional SSB phone modes while they're still in widespread use. Most robust digital modulation schemes are fairly immune to interference from adjacent SSB voice transmissions; unfortunately the converse is not true - my Mark I ears are not immune to nearby digital interference. As long as we still have band plans that encourage the separation of all digital modes from the analog modes, I fully support your proposal.

A question, though: How does spread-spectrum fit into your bandwidth-based plan? Do you consider the bandwidth to be what's used by each individual chip or the SS signal over all its carriers?

How do you feel about introducing a CDMA-esque automatic listen-before-transmit rule for computer-based digital modes, particularly with the growth of unattended stations?

PS - There's a typo in item 79 in the 20m, 6kHz section of the proposed bandwidth table - you have the lower limit as 1.150 MHz instead of 14.150 MHz.

73 de K4DET

Real World Demo

[pavon]

And here is a recent demo of real world performance. Compared to SSB the encoded voice is more artificial sounding, but there is no background noise (hiss and clicks) and it uses less than half the bandwidth to transmit. There is more info and a large playlist of demo/tutorial videos on David Rowe's blog (the creator of codec2).

Not an issue in Canada

[ubergeek65536]

In Canada any codec can be used in ham radio as long as the signal fits in the allocated bandwidth for the frequency and no encryption is used. The restriction is that you must publish the method before going on the air.