PAL composite color video made using arb signal generator

[oz2cpu]

Finally it is made and it works,
I am super happy he was able to pull this of, it turned out to be a bit harder then first expected,

free webpage generator and full explainer here

https://blog.dzl.dk/2024/10/20/generating-loopable-pal-test-signal/

could be fun to hear if all arb signal generators you got can work ?
we only have Siglent units. 16 bit raw data, the bit rate is in the file name, so it is easy to input manually
since the raw date file is header less = loops perfectly
you need to manually input the play back speed in the generator,

[RAPo]

Nice!
Did you think of using a NTE7049, or a ZNA234E B&W only  IC, capturing the output and feed that back at the signal generator?

[oz2cpu]

what would I achive by that ?

[Benta]

Incredibly cool work. Well done!    
A hint:
investing in the book from Keith Jack, "Video Demystified" would probably have solved the issue up front. One chapter tells why you need 4 fields (2 frames) to decode PAL correctly.

Cheers.

[wasedadoc]

Incredibly cool work. Well done!    
A hint:
investing in the book from Keith Jack, "Video Demystified" would probably have solved the issue up front. One chapter tells why you need 4 fields (2 frames) to decode PAL correctly.

Cheers.

Not quite correct.  True that 625 PAL colour phase is only back to precisely the same phase after exactly 4 fields.  That does not mean that 4 fields are required to correctly decode it.  The overwhelming majority of PAL TVs ever manufactured never used more than 2 lines from a single field to decode the colour.

[wasedadoc]

free webpage generator and full explainer here

https://blog.dzl.dk/2024/10/20/generating-loopable-pal-test-signal/

I have 2 queries with the parameter table for that pattern.

1.  Black level and blanking level are given as 0.339 and 0.285 respectively.  That is incorrect.  For PAL, there is no difference between the two.  Both are 0.3 Volts.

2.  PAL burst amplitude should be 0.3 Volts peak-peak.  How does that relate to the 0.1425 in the table?

[Benta]

Not quite correct.  True that 625 PAL colour phase is only back to precisely the same phase after exactly 4 fields.  That does not mean that 4 fields are required to correctly decode it.

Right. I should have written "encode", which is what this is about.

[mr ed]

My way way back machine tells me you need 4 fields if you want smooth editing of pal  (but only 2 for us standard editing).

Anyway, very nice project. We'll done.

[TurboTom]

SUUUPER! Now it's possible to use an AWG to adjust old collectables... 

In reality, the image on the TV screen doesn't look as crappy as in the photo -- it seems the cell phone camera doesn't like the CRT's shadow mask too much -- actually, it looks like we'ld have called it 35 years ago, "acceptable" 

Thanks a lot for sharing this project! 




P.S.: In order to run the generated binary file on the DG800+, just rename the *.bin file to *.raf, then it runs without any hassle after correctly setting sample frequency and amplitude.

[oz2cpu]

wow thanks Tom for the picture and details about the filename,
I look so much forward to hear about other people who tried to see if the data file run directly on their generators of all sorts of brands,
some two channel units, can even add one channel as modulation to the other one making RF carrier,
so you can test the receiver,
I think you used the composite video input option of your TV ?

[TurboTom]

Indeed, with the SDG6000X, it's possible to modulate the signal so it can be received "via the air". Since the TV modulation scheme is inverse, i.e. white level is minimum RF amplitude, sync pulse level is maximum, the PAL converter has to be configured somewhat awkwardly, see screenshot below. To simplify this, the author of the program may possibly add an "Invert Output" switch since just negating the individual samples is a no-brainer  .

As the AM source for the RF channel (in my case CH2), the other channel, playing the arbitrary, (in my case CH1) has to be selected. For an external link, the modulation ADC isn't fast enough. Otherwise, the setup is straight forward. Negating the output of the arbitrary channel doesn't negate the modulation signal, just the "real" PAL output, hence the neccessity to "fake" the negation with the aforementioned settings. The output of the arbitrary channel has to be enabled nevertheless, otherwise there's no signal generated.

The big advantage of the SDG6000X is its wide frequency range, so the lower and upper VHF bands can be covered as well as the lower third of the UHF band.

Unfortunately, Rigol's DG800(+) locks the second channel if a custom arbitrary is played back on the first (and probably vice versa); even enabling the frequency counter disables custom arbitraries. Seems like this generator is really built to a budget and lacks in FPGA resources ... or just sloppy programming (hello, Rigol??)  But obviously, there's got to be a reason for the substantial price difference between the mentioned Rigol and Siglent AWGs.

I'm quite curious to know if this PAL modulator setup also works with the SDG2000X (obviously only lower VHF) or other, capable AWGs. To simplify this test, I uploaded the arbitrary file to a file sharing platform, here's the link. Sampling rate to be selected is 35468950 Sa/s.

Enjoy and have fun experimenting!

[oz2cpu]

Thanks for the "invert" setup values,
I will try to ask Nikolaj to add invert feature, it is very usefull and a very good suggestion