AM Analog TV rendering issues (horizontal line shift)

[Vort]

I have recently found a sample of TV signal: Принимаем и декодируем аналоговое ТВ с помощью SDR и Python. And converted it to .sdriq format.

I have tried to decode it with both 4.12.3 and 4.11.6, but had no luck with getting a stable result. It may be that I am setting wrong parameters to ATV Demodulator, but, most likely, it is just glitchy. Fixing of this problem would be a good task for someone who want to improve SDRangel. Here is the file, which can be loaded with FileInput plugin: atv_sample.sdriq (941 MiB).

And here is the screenshot of current decoding with 4.12.3:

Related: #21

[Vort]

If you would like to squash some commits before you are welcome to do so.

Here is the result: https://github.com/Vort/sdrangel/commit/4afd5a7a2d9fd7cd14f1573b5ff4ff4876d14fda. It should differ from https://github.com/Vort/sdrangel/commit/ff8368470744d78b573c6977f6e94153a96b5504 only by whitespace changes.

[f4exb]

OK great! you or I can make a PR.

[Vort]

Is there a reason for it? Commit can be applied without PR. Even merge commit is not needed in this case. I am not against creation of PR, just trying to understand situation.

[f4exb]

Changes are in your fork so unknown to the f4exb repository. I don't know how this can work without a PR. I can make a diff on my clone and patch but your signature will disappear. It is as you like.

[Vort]

It is done by linking to my repository (adding remote or something). If you find how to do it, then PR will be closed automatically. If not (or do not want to do this), then just merge it: #576.

[f4exb]

Yes it seems to be working. In the same manner a diff can be done across remotes a merge can be done also and keeps the signature of the commit.

[Vort]

Can you adjust timings in modulator and fix incorrect interleaving there? (I can make separate report for it if needed)

[f4exb]

What do you mean by "incorrect interleaving" ?

[Vort]

Use this file for transmission in 625 mode:

White line


Edges of the white line should be smooth.

But instead you can see swapped scanlines:


This was not visible before because equivalent bug existed in demodulator.

[f4exb]

You mean interleaving or interlacing? The code has a m_interleaving indicator but I think this should be m_interlacing if this is interlacing then I think it is correct. The roughness observed on the edge is not due to wrong interlacing order for two reasons:

• It is observed whether image is interlaced or not (try ShortI vs ShortNI)
• It is irregular. Incorrect interlacing switches one line and the next and produces a regular pattern.

The effect on the chessboard pattern is even more dramatic

Correct interlacing


Wrong interlacing


I had the issue some time ago and I had to correct it.

Note: this is the declaration of m_interleaving:

bool     m_interleaved;      //!< true if image is interlaced (2 half frames per frame)

"interlaced" is the right term I think.

[Vort]

"interlaced" is the right term I think

Yes, I was using the term, which is used in code.

if this is interlacing then I think it is correct.

My test was not good at displaying the problem. Here is the new one:

Source:


Here is wrong encoding:


Here is good encoding:


Zoomed:


[f4exb]

OK but I think this is not an even/odd field swapped issue. The ShortI works fine and all PALxxx are affected. ShortI uses the least possible frames for vertical sync so the issue is probably in an incorrect number of frames in the vertical sync sequence or a half line mismatch.

Edit: I used your picture as an image source to test but I think this zigzag pattern is an interesting pattern to add to the list of generated patterns.

[Vort]

I was fixing line count in demodulator too. According to standard (do not remember which one, but can google it), digitized 625-lines mode should contain 576 visible lines. Maybe if you fix it, incorrect line order will go away. _ Actually, first half of the top line and second half of the last line should be black. And it is not precisely "half", but that is details._

[f4exb]

The above issue is open to specifically address the ATV Modulator issues (it now deserves its own).