The Taylorator – All Your Frequencies Are Belong to Us
scd31.com
scd31.com
It's actually ridiculously tricky to measure transmitter power output in the terms that the FCC regs are written in, but the rule of thumb is that, given the receiver sensitivity of a typical radio, if you lose the signal when you're more than 200 feet from the transmitter, you're probably in the clear.
This means a perfectly legal power level would be more than adequate to cover your house, office, hackerspace, or whatever.
https://www.fcc.gov/media/radio/low-power-radio-general-info...
Like for example, https://www.fcc.gov/enforcement/areas/jammers
It need not saturate the bands, you technically are in violation the moment a non-licensed entity has caused interference to a licensed entity.
Willful is a pretty low bar since general intent can be derived from any related negligent acts. Its pretty safe to say transmitting on these bands without a license is illegal, and yes its a fairly big deal.
Most of the public radio spectrum has been monitored, and archived for later retrieval if circumstances dictate since the 60s. There are services that aggregate this data and make realtime info available globally from satellite systems. Trilateration of signals is fairly trivial in most cases.
You may not get caught if you always stay below the noise floor, but you can't really have any real use of the spectrum in doing so, and algorithmic scanning of structured signals is always improving.
It is really negligent to make it seem like these things are legal, when they aren't, and the author should get in trouble for that.
A simple disclaimer isn't going to cut it when then primary purpose is effectively breaking the law in many civilized nations.
Someone in a shed on a few acres playing around with an SDR? Go ahead, I don't care. Playing around in an amateur anechoic chamber, have fun if you're sure you're containing things. On a desk in a random apartment in Manhattan? What an asshole, call the FCC.
Is it potentially against the law to do this? Yeah, probably. How bad is it? Somewhere around jaywalking on an absolutely empty street and dumping waste in a park, greatly depending on multiple factors.
If amateur and citizens band radio shenanigans over the last few decades have proved anything, it's that you have to be an absolutely massive pain in the radio-wave-receiving public's ass to really get their attention.
If you had this device on 14.300, 7.200, or channel 6? Meh. No one's gonna care.
I'm pretty certain this is not legal in my locale, but I don't think the author of the article is making this out to be legal - and I don't think they should get in trouble for any of this publishing [assuming you're adressing them, not the commenter]. Perhaps the equivalent of a script kiddy could get up to no good with it, nothing new from the last 20 years of computer hacking then.
I'd far rather see enforcement and ire directed at the proliferation of poorly shielded junk that spews noise all over the spectrum, at surprisingly high power, that seems to have no trouble being sold online or imported for sale in many countries - and is often, strictly speaking, not legal with regards to necessary EMC, and probably safety, standards.
https://transition.fcc.gov/oet/info/documents/bulletins/oet6...
With the FCC maybe, but someone far worse will be interested, the RIAA. It's copyright infringement by illegal broadcast.
Do it with a private video, it won't be taken down.
I guess they've since changed their policies.
From a copyright perspective there's no issue with uploading your own personal videos with copyrighted material in it to a site like YouTube so long as they're still private. I guess YouTube just assumes if you're uploading it there, you're going to eventually want to make it public. IMO they should just show the results of their scans to clue you into potential copyright claims but not actually do enforcement if the video isn't available publicly.
Note: an unlisted video is still a public video.
And I'd bet they would make a good attempt at fining you for each channel that you modulated for.
With a cheapish Rigol spectrum analyzer and homebrew receive antennas and DIY wooden/pvc tripods you could probably do 3m open-air measurements to within 3dB of what you would measure in an anechoic chamber at a test laboratory if you were very careful. I can 100% assure you that is possible with a commercial dipole kit and horn antenna, neither of which is that expensive if you're getting paid. This in in the FM band so you could use dipoles and a low-end analyzer to the tenth harmonic.
They're written as field strength in free space (microvolts per meter), which as I understand it, either requires a calibrated field probe, or a calibrated antenna. I'm sure there's a way to perform such calibrations from first principles, but that's where I'm invoking "tricky". If you've got a trick to it, please share.
In practice, Part 15.23 is pretty generous about home-built equipment, and I think since the "200 foot" thing came from the FCC's own mouth, if you can show that you walked the perimeter with a portable radio and confirmed that you're good, that probably shows good faith on the power issue.
...
Absolutely yes a spectrum analyzer is helpful to make sure your harmonic filtering is working. I would put that forth as "good engineering practice" under 15.5, and it's far more than any hobbyist could be reasonably expected to do even just a few years ago; this equipment has gotten mindbogglingly affordable and accessible.
I wonder though if you could trust the published specs of a commercial FM transmitter to calibrate a power meter on it from a distance of a mile or so...
Even the douche canoe that intentionally interfered with firefighting efforts recently only got a hefty fine. He still has his license.
Ham radio licenses are granted by the FCC with the understanding that the the license-holders intend to experiment and explore radio technology. And further understanding that while licensees are expected to adhere to all of the rules, mistakes will sometimes be made. Hams are therefore given wide latitude in terms of self-policing and forgiveness for accidentally or unknowingly breaking them from time to time.
The primitive method for getting an antenna calibration is the three-antenna method. If you are particularly motivated, you can build three Roberts dipoles and check them. For a number of reasons, you might want to have three dipole kits anyway if you are doing a lot of these measurements.
"Tricky" in the sense that skiing is tricky. You will struggle learning from books and the internet. Starting from scratch with DIY equipment would be quite challenging. But it's all in standards and the open literature.
Solid gold search term right there, thank you!
> The primitive method for getting an antenna calibration is the three-antenna method.
More good stuff. Seriously, I hadn't gotten here and this is suddenly demystifying a lot of fundamentals.
Wave if you see me on the bunny hill! ;)
But this kind of project (and a similar I saw that broadcasts on every FRS[1] channel at the same time) is a great example of something you can do with SDR that you cannot (easily) do with transistors and boring old analog radio.[2]
It is one of the reasons I enjoy it (playing around with SDRs and DSP) so much as I keep learning new things.
[1] Family Radio Service
[2] I know, I know, building 20 modulators isn't all that tough yadda yadda.
Are there any countries where that isn't illegal?
PSK = Phase Swift Keying
TDM = Taylor Division Multiplexing
You still need to tune it but there are many "center frequencies" that can be tuned instead of just one. So, you wouldn't have to tune far but you would still have to tune :) If you had a knob like most radios in the 80's it would save you from cranking the knob several times to get the dial into the right range before fine-tuning.
This was great because it allowed for stereo. This also replaced the earlier idea of taking the speaker from the pole next to the driver's window, and then hanging from the window inside the car. This was just a mono speaker, but it was also right next to the driver's head. It reduced the drive-in's repair bills from people driving away before placing the speaker to its host stand (and other hooliganism)
*just to add some color
(And start taking requests.)
<Oops... I missed the important detail about 100 different songs... sorry. I'd suggest putting the hit "Never gonna give you up" on loop for that version of the project>
In hardware, it would be easier to feed a 200 khz square wave into a diode and pull off the first 10 mhz of harmonics, low pass filter it, and mix that with 99.1 Mhz fm audio to get 88.1 to 108.1 Mhz fm at 200 khz spacing.
In software, you should be able to precompute a complete one cycle composite of -10 to +10 Mhz sine waves spaced 200 Khz apart. You could then loop this abitrary waveform and multiply it by an FM modulated source at 99.1 Mhz to get the same effect with a lot less processing power. This is the kind of thing GNU radio[2] is perfect for.
<Correcting the above>
Precompute the FM signal for each of the songs in it's channel from -10 to +10 mhz for the length of the song. Store them as 16 bit IQ samples. Each file will be different length.
Loop though all of the songs, adding all of the components into a 32 bit I and Q sum. Round those to 16 bit values, or as appropriate for your SDR transmitter. You should have a very easy to compute 100 channel FM station with each channel separately timed.
Something that people ask occasionally is how to use their vintage TV's tuner to receive digital TV. But they don't want to tune it to channel 3 and then change the channel with a digital TV box; they actually want to change the channel with the old tuner. The current advice seems to be to spend $hundreds to get an analog CATV head-end, and a separate digital tuner for each channel, but I've never heard of anyone actually following through and doing that.
I suppose, you could call it a Taylor Series.
(I'm sorry)
I suppose if I make this comment in sotto voce it IS a taylor sum..
Like, a bunch of activists need to disable comms of some nightshift guards in an area. So... someone figures out how to put more juice into this thing to pump out music on all frequencies they could have.
Bonus points because this wouldn't just jam their communications with white noise. It would be confusing as hell. Taylor Swift on the main channel. Switch to the backup. Irish Folk. Switch to something you and Bob used some time back. Power Metal! What?
[1] 20e6*(60*3+52)*4 / 1e9
Seems like there is probably some set of clever hacks here that could get you around this (although I don't know enough about radio to propose any); I think I asked about pre-computing some state for each song on its own and he had a good response to why that either didn't work or didn't help much, but unfortunately I don't remember it!
[1] we are both currently at the Recurse Center / https://recurse.com
https://web.archive.org/web/20040923155023/http://www.emerga...
Code: https://github.com/anfractuosity/musicplayer
Based on the work of https://github.com/fulldecent/system-bus-radio
Something along the lines of taking the FM at baseband, transform to frequency domain, copy result n times and shift the coefficients to the right indices, much wider ifft to RF should work, right?
Yes, modulate then FFT for each signal, do the upconversion and mixing in frequency domain, then a single large IFFT to give the final bitstream. Roughly the inverse of this (single large FFT, filter in frequency domain, and IFFT individual signals) is how the 'wideband sdr' at utwente.nl works to support hundreds of simultaneous receivers on a single gpu.
The amateur radio folks working with these techniques on HPSDR hardware were calling the technique "Direct Fourier Conversion" but I do not know if they ever got to a releasable state. They spent an awful lot of time prematurely optimizing their stuff to run on the original Jetson board which was not really adequate to the task, so I think that was likely frustrating and killed momentum.
A power of two number (N) of equally spaced frequency channels can be efficiently combined using a short fir filter on each channel followed by an fft where one sample from each channel is input to the fft per frame. Then you get N samples of your output. There is a bit more nuance but the author's bandwidth and number of channels are trivial to handle with this method.
Since each channel is already assumed to be spaced in frequency, you are essentially already in the frequency domain and only so only one fft stage is required.
Well done :)
The idea of transmitting lots of low-bit-rate signals on different evenly spaced subcarriers is quite popular. OFDM is an example of exactly this, and the computation scales just fine. The same trick ought to work: compute in frequency space. You know what frequency-space signal you want to send on each subcarrier, so assemble the output in frequency space and iFFT it. I bet it can be done genuinely OFDM-style using transform lengths calculated to get the subcarrier spacing right with some additional care to get the boundary conditions right (conventional OFDM has a guard interval and doesn’t even try to transmit a continuous signal).
If all the songs are the same, there’s a much nicer solution. If the complex amplitude of the modulated song is A(t) (that’s just I + i⋅Q), then two copies are:
A(t) + A(t) ⋅ e^(2πit⋅200kHz)
Lots of copies makes:
A(t) ⋅ [1 + e^(2πit⋅200kHz) + … + e^(2πitn⋅200kHz)]
where n is the number of copies. The thing on the right is a geometric series and can be summed algebraically. So all that’s left is to upsample A, and it doesn’t need to be done especially precisely.
You don’t need an SDR, just put on her song and most people at traffic lights will roll their windows up!
That being said, from my experience, power will be the deciding factor. I tried both LimeSDR and BladeRF, and when maxing out the power, they overheat really fast, so they probably will only work while driving/cycling so no obstacles. And then again, most people use Spotify or similar in their cars.
If it works on AM, however, that would be great to test.
I believe the odd-numbered frequencies were chosen because if only even were allowed then there would be legal knock-down-drag-out fights over the round numbers.
Coming from a DAW background this concept was incredibly interesting. I am curious if stereo audio has ever been represented in this way, or if it is generally just used for the half sample rate?
Actually thinking on this now I guess that doesn't really make sense. But still really interested in this alternate view of sampling waveforms.
(Before someone replies about snare drums, 3:1 rule, etc. — of course if there are phase differences in the audio before you mix it, it will cause audible artifacts in the mixing process. Which is why there's an effect called a "phaser".)
Where would you source Swifty music that wasn't already at 44.1kHz or 48kHz sample rates? Curiously, what sample rates were these at, and why?
Really funny idea!