¿ªÔÆÌåÓý

Date

Re: HF Digital Voice Modulation modes

 

Obrigado Rafael!

This is very interesting, and using Web standards! I'll take a look.

Abra?os,
Rafael Diniz

On 10/21/21 5:04 PM, Rafael Pinto [PU1OWL] wrote:
Rafael,
?
This Morse decoder is really good ==>

<>

and it is open source.
<>

I didn't look at the code, as I was just browsing for a?morse decoder
a few days ago, but it works really well!

73,

Rafael


Re: EI9GQ 16 Watt Linear Amp Build #homebrew #linear-amp

 

Well done! ?So the input to your 2n2219a is your homebrew? Did you have to lower the output of the 510 at that point? Or is the input to your lovely amp starting someplace else?


Re: Trouble ahead

 

On my 2018 trip to Dayton, the flight entertainment system rebooted. I could see it was booting up in linux. I pressed on random buttons and it landed me in the admin login, there seemed to be no password. I wandered around a bit on the disk system, it was remote mounted and then I went back to my "Radio Porn" (flipping old QSTs on the laptop).
- f

On Fri, Oct 22, 2021, 3:16 AM Jerry Gaffke via <jgaffke=[email protected]> wrote:
Tom has a good memory!


On Thu, Oct 21, 2021 at 08:12 AM, Tom, wb6b wrote:
I seem to remember something about a hacker (while actually traveling on the plane!) decided it would be cool to connect his laptop the the in-flight entertainment system data bus and discovered he could hack his way into the flight deck control systems.


Re: Trouble ahead

 

Tom has a good memory!


On Thu, Oct 21, 2021 at 08:12 AM, Tom, wb6b wrote:
I seem to remember something about a hacker (while actually traveling on the plane!) decided it would be cool to connect his laptop the the in-flight entertainment system data bus and discovered he could hack his way into the flight deck control systems.


Re: Trouble ahead

 

On Thu, Oct 21, 2021 at 07:05 AM, Evan Hand wrote:
You are not thinking devious enough:? IF you insert a smart chip in the I2C line inside of the chip it MIGHT be possible to use that link to take control of the rig's MPC and then use that device to link into the computer through the CAT channel.?
I seem to remember something about a hacker (while actually traveling on the plane!) decided it would be cool to connect his laptop the the in-flight entertainment system data bus and discovered he could hack his way into the flight deck control systems. Great. Glad he (and everyone else) lived to tell about it. If the story is true, hope that prompted an in-flight firewall to be added to avionics systems.?

Tom, wb6b


Re: HF Digital Voice Modulation modes

 

On Wed, Oct 20, 2021 at 07:32 PM, Tom, wb6b wrote:
However there are other prebuilt examples that can be downloaded trained and run for practice. If I find anything interesting in my notes from when I discovered this Amazon service existed I can add some links.
It is cool that we have come to the point where there a lot of plug-n-play module that can just be downloaded, and maybe modify a bit with your own training data, and do useful things. Were are entering the Arduino IDE stage of AI/ML development.

Here is a web site that was a very popular read for the folks trying to learn AI/ML in the "assembly language" days of AI.


Probably worth a cursory look at?gradient descent and?back propagation. These are the core of how these systems train. Worth it to try to get a feel for what goes on under the covers, not just the drag and drop pre-made models step we are now entering. Not that the phase we are entering is bad. Its good. I think the snake oil salesmen that tried to promote AI as magic solutions for everything have largely been exposed and gone away?

Also, in the early phase the nonlinear curve that seems to let the "neurons" produce "arbitrary" functions based on the sum of many inputs to a neuron was either an exponential function or based on calculating tangents. Seems very costly for something that is calculated over and over again while a computer preforms a big loop stepping through each "neuron" one by one recalculating the latest value. But the "sigmoid" functions were convenient for the academic early research stage of ML.

Later someone tried a nonlinear function based on a half wave rectifier. Really fast and easy to calculate and seems to work as well as the computationally complex sigmoid functions.

Anyway, I'm prompted to dig back in and start coming back up to speed so I can try some of these new models.

Tom, wb6b



Re: HF Digital Voice Modulation modes

 

On Thu, Oct 21, 2021 at 06:30 AM, Rafael Diniz wrote:
Certainly yes. An AI NN trained for such purpose could reconstruct a
good quality voice from SSB quality "Donald Duck" analog voice. Noise is
the most challenging thing here as far as I played with.
Hi Rafael,

Thanks, that is encouraging. The noise aspect is interesting. I think I ran across a paper with some experiments on AI/ML noise reduction.

Fortunately Ham Radio is only supposed to be used for speech. Since the AI voice compression is only trained on speech, wonder if it breaks down went presented with things like music. Funny, that could be a great boon to amateur radio. A transmitter that is of no use to the people that seem to think jamming and playing music on the ham bands it just a great way to spend their time.

Tom, wb6b


Re: Trouble ahead

 

You are not thinking devious enough:? IF you insert a smart chip in the I2C line inside of the chip it MIGHT be possible to use that link to take control of the rig's MPC and then use that device to link into the computer through the CAT channel.?

Not likely but it could be possible.? Similar things have been done in industrial control systems to break things.

Enough of the speculation on my part.? I will now mute the topic.
73
Evan
AC9TU


Re: HF Digital Voice Modulation modes

 

Rafael,
?
This Morse decoder is really good ==>

and it is open source.

I didn't look at the code, as I was just browsing for a?morse decoder a few days ago, but it works really well!

73,

Rafael


Re: HF Digital Voice Modulation modes

 

Hi Farhan,

Do you have any "traditional" automatic CW decoder reference for
performance comparison? I hope to get something better than what exists,
otherwise it does not make any sense.

Rafael

On 10/21/21 3:10 AM, Ashhar Farhan wrote:
Rafael,
Is a CW decoder possible with band waterfall in its input? I don't
know how to start with tensorflow. It would be dandy if someone can
write? a simple HelloWorld for us old timers to start using TensorFlow.
- f

On Thu, Oct 21, 2021, 5:09 AM Tom, wb6b <wb6b@...
<mailto:wb6b@...>> wrote:

On Wed, Oct 20, 2021 at 03:24 AM, Rafael Diniz wrote:

Also, for images, HIFIC, for example, provides significant better
compression than traditional non-AI based image encoders.

Hi Rafael,

That is really cool. I built an interface to TensorFlow many years
ago and had fun feeding images to a early neural net that was all
the rage at time. Haven't done much with AI in recent time. Looks
like some great strides have been made in real world applications.

Because you are involved currently in AI and Neural Nets, I have a
couple of random questions.

1) Regarding the AI image compression. Is it possible that there
would be adversarial patterns/images that could produce completely
unexpected results. An example was a university study where people
taped a piece of paper to their chest with a random looking
patterns printed on them. It fooled the AI into thinking the
pattern was a face and it spit out a "match" Was some time ago so
I don't remember exactly what the result was.

2) Something I've toying around with and trying to find any
possible related research. That is, would it be possible to train
an AI (possibility using the same training set as used for the
speech compression) to listen to the "Donald Duck" sounds of a SSB
transmission received by an AM receiver, to produce fully
intelligible speech? Reading about the AI speech compressors and
reading about the training data sets that are available, may make
it possible for me to give it a try.

I wonder how much it is going to drain my bank account setting up
an Amazon AWS server with GPUs and run the training?

Look forward to your paper.

Tom, wb6b



Re: HF Digital Voice Modulation modes

 

Hi Tom,

1) Regarding the AI image compression. Is it possible that there would
be adversarial patterns/images that could produce completely
unexpected results. An example was a university study where people
taped a piece of paper to their chest with a random looking patterns
printed on them. It fooled the AI into thinking the pattern was a face
and it spit out a "match" Was some time ago so I don't remember
exactly what the result was.
I never got such strange behavior with HIFIC, but it is a matter of us
trying out. Indeed, the performance is not orders of magnitude better
than VVC (H.266) - by reading the papers. But in my experiences, with
the lower bitrate HIFIC (0.160 bpp) is better than state-of-the-art
traditional image encoding (aka. VVC). Indeed, images look much better
(do not look the PSNR thou). Use the? "hific-lo" model in (I'm using the
CPU version in my old Lenovo T430, and for still images, you just need
at least 16 GB ram to play with and get the results after less than 30s!):



2) Something I've toying around with and trying to find any possible
related research. That is, would it be possible to train an AI
(possibility using the same training set as used for the speech
compression) to listen to the "Donald Duck" sounds of a SSB
transmission received by an AM receiver, to produce fully intelligible
speech? Reading about the AI speech compressors and reading about the
training data sets that are available, may make it possible for me to
give it a try.
Certainly yes. An AI NN trained for such purpose could reconstruct a
good quality voice from SSB quality "Donald Duck" analog voice. Noise is
the most challenging thing here as far as I played with.

I wonder how much it is going to drain my bank account setting up an
Amazon AWS server with GPUs and run the training?
:P
In the university we have some nice GPUs to play with. If you really
want to play with this, I bet nothing, as there will be many people
offering you GPU access to play with such computer science edge stuff.

Look forward to your paper.
Yay!

Cheers,
Rafael


Re: Trouble ahead

 

Y'know... ...it COULD happen. All it takes is one more bonehead ruling by the FCC... -- Rich WB2GXM

------ Original message------
From: Max via
Date: Thu, Oct 21, 2021 12:12 AM
Cc:
Subject:Re: [BITX20] Trouble ahead

"Every 15 minutes of having QSOs the chip will abruptly tune your rig to a radio advertising frequency and make you listen to ads."

I know you were joking, but HE11 NO !



On Wednesday, October 20, 2021, 07:21:18 PM CDT, Tom, wb6b > wrote:


On Wed, Oct 20, 2021 at 03:40 PM, Jim Sheldon wrote:
Well, the architecture of that chip only generates up to 3 simultaneous RF frequencies and nothing else.
Maybe one of the big internet companies will see a way to make a profit and produce?Si5351a like chips in volume at unbeatable low prices.?

Every 15 minutes of having QSOs the chip will abruptly tune your rig to a radio advertising frequency and make you listen to ads.?

Could be a new privatized ad supported amateur radio service.

Tom, wb6b


Re: HF Digital Voice Modulation modes

 

This maybe an interesting quick read for those interested in AI/ML in the embedded space.

Article has links to a series of training videos.?

There is also a NXP Connects coming up soon.
EMEA Nov 9-10 | Americas Nov 10-11 | Asia-Pacific Nov 16-17
https://www.nxp.com/company/about-nxp/events/nxp-connects:NXP-CONNECTS

Digikey, Arrow, Mouser are sponsors.
(in the past,? NXP sold some of the boards used in the training sessions through supplier channels at discounted rates)

Rgds,
Gary


Re: Trouble ahead

 

"Every 15 minutes of having QSOs the chip will abruptly tune your rig to a radio advertising frequency and make you listen to ads."

I know you were joking, but HE11 NO !



On Wednesday, October 20, 2021, 07:21:18 PM CDT, Tom, wb6b <wb6b@...> wrote:


On Wed, Oct 20, 2021 at 03:40 PM, Jim Sheldon wrote:
Well, the architecture of that chip only generates up to 3 simultaneous RF frequencies and nothing else.
Maybe one of the big internet companies will see a way to make a profit and produce?Si5351a like chips in volume at unbeatable low prices.?

Every 15 minutes of having QSOs the chip will abruptly tune your rig to a radio advertising frequency and make you listen to ads.?

Could be a new privatized ad supported amateur radio service.

Tom, wb6b


Re: Trouble ahead

 

I'd rather pay more then to put up with more ad's !

Max KG4PID

On Wednesday, October 20, 2021, 08:18:32 PM CDT, Ken Hansen <ken@...> wrote:


Like an Amazon "Fire Tablet" with special offers - imagine if you could get a slightly cheaper Radio if you tolerated advertising on the bandscope/display between band changes and at power-up! LOL

Ken, N2VIP

> On Oct 20, 2021, at 20:21, Tom, wb6b <wb6b@...> wrote:
>
> Every 15 minutes of having QSOs the chip will abruptly tune your rig to a radio advertising frequency and make you listen to ads.
>
> Could be a new privatized ad supported amateur radio service.







Re: Trouble ahead

 

They have only one xtal there, same as bitx IF.

At 21/10/2021, you wrote:

There are reports of bitx randomly picking up radio rwanda...

On Thu, Oct 21, 2021, 5:51 AM Tom, wb6b <wb6b@...> wrote:
On Wed, Oct 20, 2021 at 03:40 PM, Jim Sheldon wrote:
Well, the architecture of that chip only generates up to 3 simultaneous RF frequencies and nothing else.

Maybe one of the big internet companies will see a way to make a profit and produce? Si5351a like chips in volume at unbeatable low prices.?

Every 15 minutes of having QSOs the chip will abruptly tune your rig to a radio advertising frequency and make you listen to ads.?

Could be a new privatized ad supported amateur radio service.

Tom, wb6b


Re: HF Digital Voice Modulation modes

 

On Wed, Oct 20, 2021 at 05:10 PM, Ashhar Farhan wrote:
It would be dandy if someone can write? a simple HelloWorld for us old timers to start using TensorFlow.
Here is what Amazon offers for ML (AI) development.


This is Amazon's getting started page.


However there are other prebuilt examples that can be downloaded trained and run for practice. If I find anything interesting in my notes from when I discovered this Amazon service existed I can add some links. I seem to remember a basic step by step training course that let you work through examples, starting from very "simple" and let you progressively work your way up to more complex things.?

There is a big learning curve and you need to get familiar with AWS. I know AWS but only at the beginning of the learning curve of the sagemaker service.

I started to play a bit here but became involved with other priorities. The AI based speech compression that came up in these posts and the thoughts of things that might be possible by adapting that work to other uses is restoring my interest.?

Tom, wb6b


Re: Trouble ahead

 

Like an Amazon "Fire Tablet" with special offers - imagine if you could get a slightly cheaper Radio if you tolerated advertising on the bandscope/display between band changes and at power-up! LOL

Ken, N2VIP

On Oct 20, 2021, at 20:21, Tom, wb6b <wb6b@...> wrote:

Every 15 minutes of having QSOs the chip will abruptly tune your rig to a radio advertising frequency and make you listen to ads.

Could be a new privatized ad supported amateur radio service.


Re: Trouble ahead

 

There are reports of bitx randomly picking up radio rwanda...


On Thu, Oct 21, 2021, 5:51 AM Tom, wb6b <wb6b@...> wrote:
On Wed, Oct 20, 2021 at 03:40 PM, Jim Sheldon wrote:
Well, the architecture of that chip only generates up to 3 simultaneous RF frequencies and nothing else.
Maybe one of the big internet companies will see a way to make a profit and produce?Si5351a like chips in volume at unbeatable low prices.?

Every 15 minutes of having QSOs the chip will abruptly tune your rig to a radio advertising frequency and make you listen to ads.?

Could be a new privatized ad supported amateur radio service.

Tom, wb6b


Re: Trouble ahead

 

On Wed, Oct 20, 2021 at 03:40 PM, Jim Sheldon wrote:
Well, the architecture of that chip only generates up to 3 simultaneous RF frequencies and nothing else.
Maybe one of the big internet companies will see a way to make a profit and produce?Si5351a like chips in volume at unbeatable low prices.?

Every 15 minutes of having QSOs the chip will abruptly tune your rig to a radio advertising frequency and make you listen to ads.?

Could be a new privatized ad supported amateur radio service.

Tom, wb6b