#20 - Generating Ultra High Quality video signals with HackTV

[@ElectromagneticVideos]

Wow - thats great! I would sure like one - maybe a kickstarter project would be the way to test the market if you want to develop it further. Keep up the great work you are doing! Regards from Canada!

[@Slay3rOne]

Amazing work! Definitely would be interested in this, got a bunch of CRT TVs to mess around with here!

[@AvroVulcanXH607]

awesome. I will say SiS is a lot more fussy than I expected it to be! I have yet to manage a decent signal across any medium other than cables - no RF stuff that I have could manage it. Ah well... some day I'll get a hold of a decent microwave link in the ISM band, or perhaps get an Amateur license specifically to do experiments like this!

[@minimaxxl8]

Definitely an interesting project. i would sign-up for one if available.

[@stephenneal7373]

Hats off - amazing work. This could allow PALPlus to be software encoded potentially. That would be amazing to see. 13.5MHz sampling would probably preclude baseband MUSE, but could HD-MAC baseband squeak in? (Of course we'd need a software-based HD-MAC decoder to be implemented as I don't think any working decoders are known to exist in the wild ?) It would be amazing to see 405 NTSC generated by this route! I'd be interested in this - and the holy grail - an ADC of similar quality...

[@mattstvbarn]

Oh yes. Would love to see PALplus as an application!

[@whaka54000]

if this can interest me? absoluetly and definitely a big YES. not only (in my case) for feeding a clean and big mustached secam, but also to record it on tape. i also spent some time restoring, and trynig to fix vcr of all kind. i have a hackRF, and this would be the greatest addition to turn it into an universal generator. i would even say if s-video output is possible, it would be nice :) i'll follow this project closely, for sure. and thanks for the time you spent on that.

[@mattstvbarn]

Perfect SECAM is certainly an interesting use-case. It is quite hard to find a good SECAM source other than test pattern generators.

[@whaka54000]

@@mattstvbarn exactly. i was lucky around 2007 to buy a new philips mpeg2 DVB-T receiver, and very surprisingly it offered secam. of course mpeg2 is obsolete now. that's the last thing of a modern equipement i saw offering it. i still have it, but i've no idea how to transmit DVB with the hackRF.

[@fsphil]

@@whaka54000 The THOMSON THS804 DVB-S2 receiver (for TNTSAT) can output SECAM, though it's not the best quality.

[@kFY514]

@@mattstvbarn I don't know how "perfect" it is, but the composite encoder on Raspberry Pi 4 and below is capable of outputting SECAM that looks mostly right on a TV. I actually was the person who discovered that and first made a tool to manipulate the hardware registers to enable it (search for "tweakvec") and later integrated that in the kernel driver so now it's just supported out of the box. I know that one quirk of it is that the Pi doesn't filter the luminance information in any way, so it's very easy to trigger those dreaded smears of color by just displaying sharp text or whatever. But otherwise - if you're careful about the high frequency content of whatever you're displaying - the picture is quite nice. I would certainly love to see it examined on the oscilloscope by someone who can actually assess its true quality (or lack thereof if that's the case). Just as a side note: I can only get credit for discovering an undocumented register in the hardware that enables SECAM output, everything else just have been there in the hardware since forever. So the quality is what it is, and not much can be done about it. There are a couple of obscure registers which can manipulate some amplitudes, but that's all. And on another note: I'm not sure if the Pi's composite output actually has the right voltage. It does overdrive a consumer-grade RF modulator I have, and the image only looks right if I put a resistor between it and the modulator's input 😅 I really should finally get an oscilloscope to get to the bottom of it... And on a yet another note: The newest revision of the Broadcom SoC that's in the Pi 5 no longer contains the composite video encode that was there in all previous models. Instead, the Pi 5 includes a completely new, in-house design from the Raspberry Pi Foundation. That one has some interesting features (it could drive a VGA output if there were pins for that for example), but sadly there's no support for SECAM whatsoever. That's why I said Pi 4 and below.

[@whaka54000]

@@kFY514 thanks for your feedback. i don't have a Pi 4, and don't plan to get one soon. but "perfect" mean without the RF imperfections induced by the hackRF. luma noise, and noisy coulour transitions. by noisy transitions, i mean when you can clearly spot the sub-carrier artifacting in luma where color brutally change, really visible on philips testcard and color bars, but it's also very visible in real world use. back in the days, proper filtering was done to take care of that. and that's also true for pal anyway. this can probabely be improved in hacktv emulation, but a perfect composite signal with properly done hardware filtering would definitely be nice. other than that, the RF signal is totally usable for most cases.

[@der.Schtefan]

YES! ANOTHER VIDEO!

[@Zcooger]

I heard that current crystal in HackTV is at most average. Maybe it's worth adding TCXO module?

[@mattstvbarn]

Yeah I'm coming around to that idea. Just fix it at 13.5 MHz with a cheap SMD TCXO.

[@patrickwright8552]

I love any time heating up a chip fixes a defect. Brings back memories of people putting their 360s in the oven...

[@lachlanpatterson6390]

The rumour was it was SMD solder Turned out to be some capacitors needing to undergo a nonlinear reaction with heat because they were incorrectly made iirc

[@makomk]

Neat! I wonder how the good old Fresco Logic FL2000 would work for this (basically a USB 3 triple DAC designed for software-controlled USB to VGA, but that's been repurposed for numerous other baseband and relatively low frequency signals).

[@mattstvbarn]

The FL2000 is the chip which HackTV refers to as "FL2K". I think there is scope to improve the output of it i.e. with filtering, proper amplification, DC offset correction etc. This was something I considered however possibly even more effort than I've already expended. I'm skeptical it would get a better result however!

[@TheRealHarrypm]

@@mattstvbarn With testing it requires a distrobution amplifyer and its output resistors removed to playback CVBS/S-Video from 4fsc .tbc files properly for CRT use digital TV's dont mind much.

[@TheGmr140]

Interesting 😊

[@altebander2767]

I have found some time to look into HackTV and apparently when generating baseband video there is no filtering at all. However that can be done with software like sox. I've tried a "knife edge" 5MHz filter with it, and you get the ringing one expects. I'm currently getting an fl2k to VGA adapter and I think when combining that with sox and resampling to 140 MHz one should get a very acceptable signal with the aliases of the image being at >200 MHz where they can be filtered away easily.... This would also allow for interesting experiments. PAL is usually done over 5 MHz or more of bandwidth. How does it behave when you give it less. Theoretically it should be possible compensate the shift in hue from one sideband of the chroma carrier missing. Hoe does it behave in practice? Once I get my fl2k and have some time, I'll try it out.

[@mattstvbarn]

@@altebander2767 140 mhz. That's a lot of data! Is it reliable?

[@Capturing-Memories]

Any of this can be used to capture RF from a video tape player like in the VHS-decode project?

[@mattstvbarn]

Of course, but it'd be quite some project indeed. I believe the VHS transfer industry is already pretty well served for digitisation solutions though?

[@Capturing-Memories]

@@mattstvbarn No, not really, at least quality wise, There is the looded market of chinese gadgets that are just toys, the VHS-decode, a fan based hobby for coders, and few and far in between quality digitizers based on PCIe, USB and SDI that have their quirks, driver issues, OS issues, price and availability issues.

[@TheRealHarrypm]

@@mattstvbarn The HackRF does not have the lower frequency range for FM RF archival last I checked, this is why we use cheep CX Cards with vhs-decode, and have the MIRSC for CVBS/S-Video baseband RAW capture, and the DomesDay Duplicator optimized for LaserDisc. The issue with off-shelf digitisation hardware compared to FM RF archival is the market for Time Base Correctors is filled with inflation fruad (and info death traps like digitalFAQ) and multible AD-DA stages of lossless makes baseband capture world mess of low quality options or not worth the cost ones. Using the HackRF as a DAC to output the .tbc files from the decode projects is a neat idea also, but cheeper done with the FL2000 (FL2K) adapters and a distribution amplifyer added to them, but your hardware add on has some actually very useful playout chain value, guess I am ordering a HackRF One now.....

[@altebander2767]

Hmm,. since you are just using a tiny part of the Hack RF it might make more sense to use one of the many USB 2.0 to parallel FIFO chips. You could then just use a straight clock without having to go through complex PLLs. Oversampling would be something that could also get more quality out of the fl2k solution. After all it gets you about 150 Megasamples per second, or 30 times the maximum signal frequency of 5MHz. That's more than enough room for a simple filter as well as error diffusion for good results. In theory it should even be possible to scale the 3 channels for a higher dynamic range. I should look into it. Plus since it outputs video already, jitter likely is no problem there.

[@mattstvbarn]

Ah but the HackRF has the particular advantage that pretty well everybody who uses HackTV already has one. No need to build anything! I look forward to seeing a device which outperforms mine!

[@vincei4252]

Very cool. Can I ask that when you're talking about specific parts like the pll clock chip or the DAC's to please just say the part numbers at least once? Some of us are not familiar with this gear so don't know what parts you are referring to when pointing at them; the resolution of the video is not high enough to make out the part numbers. I think I eventually figured out Analog Devices DAC but am not familiar with the PLL chip you heated up, is it an Si part? Note: Because we now live in the KZbin dystopia where everything is disinformation or election interference placing IC component part numbers in my comments results in my comment being deleted by our friends at KZbin.

[@mattstvbarn]

Check this out: hackrf.readthedocs.io/en/latest/hardware_components.html

[@vincei4252]

@@mattstvbarn Awesome. Thank you.

[@TouYubeTom]

amateur radio televsion