Composite Adapter for MiSTer

[JudgeDrokk]

I'm goin to try one of these devices from Antonio Villena to try out the MiSTer with my Commodore 1701 monitor
which has composite inputs. I have also ordered a MiSter from Antonio too..

https://www.antoniovillena.es/store/pro ... o-adapter/

Has anyone else done MiSTer to composite?

The only machine I have connected to the monitor through composite was a rubber key ZX Spectrum (the C64 is via luma/chroma) and it looked pretty nice.

[akeley]

48k Spectrum's composite is actually pretty bad - the dot crawl is very heavy. I'd expect MiSTer's output to be a step up, at least that's the case with Raspberry Pi.

[JudgeDrokk]

48k Spectrum's composite is actually pretty bad - the dot crawl is very heavy. I'd expect MiSTer's output to be a step up, at least that's the case with Raspberry Pi.

Hi yeah - I tried the Speccy (with composite mod) on a flat-screen OLED connected thru the legacy composite ports on my Yamaha AV receiver
and the dot-crawn was shocking... like some flies has got behind the screen.

Then, when I connected it to the 1701 it was bright and clean.

I hope to share results soon when the device arrives.

[ash2fpga]

Some discussion on the adapter on Atari-Forum from Oct 2019:
https://www.atari-forum.com/viewtopic.php?t=37611

I do not know if this still accurate, though.

Edit: Appears to still be accurate as of earlier this month (Aug 2020), per this tweet: https://twitter.com/cunyfidefins/status ... 1904932864

[barfood]

Some discussion on the adapter on Atari-Forum from Oct 2019:
https://www.atari-forum.com/viewtopic.php?t=37611

I do not know if this still accurate, though.

Edit: Appears to still be accurate as of earlier this month (Aug 2020), per this tweet: https://twitter.com/cunyfidefins/status ... 1904932864

Ugh... I have a composite adapter from Antonio in the mail as we speak. I wish I hadn't seen this post, now, my excitement turned to anguish

[akeley]

Ouch. I was considering getting one myself...

[retrorepair]

I would like to see an example of this if someone can show me?

I have a feeling the issue has been somewhat over-hyped, I have not noticed the problem on my own adapter, though admittedly it's using a different encoder IC.

[antoniovillena]

I would like to see an example of this if someone can show me?

I have a feeling the issue has been somewhat over-hyped, I have not noticed the problem on my own adapter, though admittedly it's using a different encoder IC.

Yes. This guy is seller of MiSTer stuff and it seems he does to attack competitors. As I explain on that email, the problem is due to the core and it's present on all composite adapters on the market not only my one. I won't say me adapter it's the best, but I just receive this email from a client.

[akeley]

Yes. This guy is seller of MiSTer stuff and it seems he does to attack competitors. As I explain on that email, the problem is due to the core and it's present on all composite adapters on the market not only my one. I won't say me adapter it's the best, but I just receive this email from a client.

But the main point is that there is a flaw in the first place. Your description on the website does not mention it at all.

You also said it's possible to correct it via hardware, and will do so in the next edition. Selling the inferior unit in this situation does not seem fair. If I just bought it it'd be useless because then I'd have to buy another one instantly.

[retrorepair]

Yes. This guy is seller of MiSTer stuff and it seems he does to attack competitors. As I explain on that email, the problem is due to the core and it's present on all composite adapters on the market not only my one. I won't say me adapter it's the best, but I just receive this email from a client.

But the main point is that there is a flaw in the first place. Your description on the website does not mention it at all.

You also said it's possible to correct it via hardware, and will do so in the next edition. Selling the inferior unit in this situation does not seem fair. If I just bought it it'd be useless because then I'd have to buy another one instantly.

I still want to see evidence of this "flaw". This is composite video we are talking about here, is it really going to make that big a difference? As I mentioned, I can't see a problem with my own adapter so I can't see it being a show stopper on Antonio's either. Looks just like my consoles do.

Also would you buy a PS4 now knowing the PS5 is round the corner? A lot of people would.

People are too used to "perfection" with video quality these days, they want nostalgia but forget how it actually looked. Very far from perfect.

You are welcome to create your own adapter and suffer the same abuse when someone picks holes in it.

[akeley]

Yes, there is a big difference between normal composite video and composite video with artifacts coming from modern hardware flaws.

No, I wouldn't buy a PS4 if it had a serious flaw and I knew there is a revised PS4 without it coming out soon. And if people only learned about this flaw from some random email, then it would be a major scandal. Obviously PS4-PS5 has zilch to do with this situation. Just as some musings about nostalgia and how people should accept flawed stuff because composite wasn't perfect anyway.

The fact you don't see the problem on youradapter only means that maybe it does not exist on your adpater. Normally, yes, I would like to see evidence myself, but it's a moot point since the existence of this problem has been acknowledged by the manufacturer himself.

Trying to find out basic information about a product you want to purchase is very far from "abuse". So please think twice before you reply again with bad analogies and sulky emotional attempts at rebuttal.

[retrorepair]

Yes, there is a big difference between normal composite video and composite video with artifacts coming from modern hardware flaws.

Again, I'd like to see this.

Didn't you say you hadn't actually bought one? So you don't even know what it looks like??

Maybe keep your accusations of "serious flaws" to yourself until you know what the problem actually looks like?

I think you need to review, if anyone in this thread is coming off as "sulky", it's not me..

[akeley]

Didn't you say you hadn't actually bought one? So you don't even know what it looks like??

You mean I should buy it just so I can post in this thread, despite the fact that the guy who makes them has admitted that a problem exists? And that people who have bought it will have to buy another one soon, if they want the problem gone? That makes total sense

I'm done here, it clearly is a brick wall. And all it would take is to mention this issue on the sellers page, so then buyers can make informed decisions, instead of reading about "pure signal", "same as the original", "best in class converter", etc. Disappointing.

[antoniovillena]

But the main point is that there is a flaw in the first place. Your description on the website does not mention it at all.

You also said it's possible to correct it via hardware, and will do so in the next edition. Selling the inferior unit in this situation does not seem fair. If I just bought it it'd be useless because then I'd have to buy another one instantly.

It's not a flaw. Composite video has these artifacts and are completely normal. Fixing in hardware is not enough. It requires core modification

[antibolo]

So are those issues only specific to composite and not S-video?

In any case, count me in as another satisfied customer of your adapter, as I described here, although it's worth noting that I'm using it strictly for S-video.

BTW, to OP: I don't know the differences between a 1701 and 1702, but if your 1701 has "chroma" and "luma" RCA connectors on the back like on my 1702, you can definitely use S-video, just need the right S-video to 2xRCA splitter (bought mine a long time ago, wouldn't know where to look to get one now unfortunately).

[retrorepair]

So are those issues only specific to composite and not S-video?

Yes. The issue is related to the synchronization of luma and chroma separation. The crawl or "shimmer" is the result of the edge of either signal overlapping. Lots and lots of devices have the same issue to the point where TVs had filters integrated to try and reduce the phenomenon but it's always there to some degree.

With S-Video, they are already separated so no overlap can occur.

It would help to have a core generated colour subcarrier clock to use instead of external timing but since it would interfere with the framework, it's likely never going to be added.

Of course with perfect synchronization comes other issues such as "jailbars". You just won't win with composite video, it's the nature of the signal.

I like it's imperfections though. There's just something "warm" about seeing super metroid over RF that you can't get with RGB. I like RGB and do use it now and then, but most of my console gaming is over RF or composite.

[antoniovillena]

It would help to have a core generated colour subcarrier clock to use instead of external timing but since it would interfere with the framework, it's likely never going to be added

I have added to my hardware but the problem it's the main team of MiSTer don't want to adapt the cores to generate this signal. This solved on other FPGA products I sell (ZXDOS+ and gomaDOS+).

[akeley]

It's not a flaw. Composite video has these artifacts and are completely normal. Fixing in hardware is not enough. It requires core modification

If you talk about "fixing it" then how can it be "normal"? Your explanation in the tweet says it's normal for FPGA, but this obviously does not make it any less of a problem. Unless you are talking about fixing the flaws inherent to composite, which were present on original hardware (though then it wouldn't be "same as original", even if it would be a welcome improvement).

I really wish somebody could clarify this since it's really, really simple. If I buy this board, will I see any EXTRA artifacts caused by MiSTer cores + board itself?

Please note that I've used composite for many years with original hardware, can still do it, and I'm perfectly aware of what imperfections and artifacts it produces.

[antoniovillena]

If you talk about "fixing it" then how can it be "normal"? Your explanation in the tweet says it's normal for FPGA, but this obviously does not make it any less of a problem. Unless you are talking about fixing the flaws inherent to composite, which were present on original hardware (though then it wouldn't be "same as original", even if it would be a welcome improvement).

I really wish somebody could clarify this since it's really, really simple. If I buy this board, will I see any EXTRA artifacts caused by MiSTer cores + board itself?

Please note that I've used composite for many years with original hardware, can still do it, and I'm perfectly aware of what imperfections and artifacts it produces.

It's normal because all adapters have the same issue. You can buy, compare, make make a video and publish if you don't trust me. The encoder I use is AD723 and it's the best one I've found. It has a luminance trap (YTRAP) provides a means of reducing cross color artifacts.

The difference between FPGA generated video and original video, it's that original video uses one clock source for generate all (cpu, video signals and color subcarrier) so the artifacts are static. The FPGA video generates cpu and video signal with one clock, but the color subcarrier is generated by a different clock (crystal near the encoder). For this reason the artifacts are moving or cycling.

To solve this I have created a circuit that reads the VSYNC signal and if found a clock there overrides the crystal clock (NTSC or PAL) present on the adapter and uses the generated by the FPGA. But this is not magic. This requires changes on all cores to generate this clock on VSYNC. This will solve the problem converting moving artifacts to static artifacts like in original hardware.

[akeley]

It's normal because all adapters have the same issue. You can buy, compare, make make a video and publish if you don't trust me.

You miss the point completely. Perhaps all adapters do have the same issue, I don't question that. But that means there IS an issue. And there is no mention of that issue on your website, which instead promises the output to be pure and same as the original.

[retrorepair]

To solve this I have created a circuit that reads the VSYNC signal and if found a clock there overrides the crystal clock (NTSC or PAL) present on the adapter and uses the generated by the FPGA. But this is not magic. This requires changes on all cores to generate this clock on VSYNC. This will solve the problem converting moving artifacts to static artifacts like in original hardware.

How will this work with PAL/NTSC selection? For example NTSC (60hz) output on the core but PAL subcarrier freq on VSYNC (PAL60)? An extra core option?

To be honest, I always thought the core descriptions were misleading as really they select 50/60hz, not PAL and NTSC. In this case it would be very confusing.

[antoniovillena]

You miss the point completely. Perhaps all adapters do have the same issue, I don't question that. But that means there IS an issue. And there is no mention of that issue on your website, which instead promises the output to be pure and same as the original.

My adapter has not any issue. You really missed the point

[antoniovillena]

How will this work with PAL/NTSC selection? For example NTSC (60hz) output on the core but PAL subcarrier freq on VSYNC (PAL60)? An extra core option?

To be honest, I always thought the core descriptions were misleading as really they select 50/60hz, not PAL and NTSC. In this case it would be very confusing.

Yes. The core has an option to activate color carrier between NTSC and PAL. This option must match with the selected into the encoder.

[retrorepair]

Great, sounds like a nice solution.

How is the integration with the framework? Hopefully it's not too intrusive so may be submitted as a pull request?

[antoniovillena]

Great, sounds like a nice solution.

How is the integration with the framework? Hopefully it's not too intrusive so may be submitted as a pull request?

According with Sorgelig is not on framework
https://www.atari-forum.com/viewtopic.p ... 89#p384089

And he's not interested to support this feature. So we must do a fork

[akeley]

My adapter has not any issue.

all adapters have the same issue.

This is surreal

Unfortunately I might actually need one of these adapters soon, so still would like to find out what the score really is. It seems impossible to do so in this bizarre thread, so if anybody has one and uses it to display composite on a CRT I'd appreciate if you could drop me a PM.

[antoniovillena]

My adapter has not any issue.

all adapters have the same issue.

This is surreal

Unfortunately I might actually need one of these adapters soon, so still would like to find out what the score really is. It seems impossible to do so in this bizarre thread, so if anybody has one and uses it to display composite on a CRT I'd appreciate if you could drop me a PM.

It's easy. All adapters have the same issue except mine. My adapter has 3 options for color subcarrier.
1. External NTSC crystal
2. External PAL crystal
3. Synthetized FPGA clock

That means you can generate subcarrier from FPGA and this way avoid moving artifacts

[retrorepair]

The issue is so minor that 90% of people would be ok with it as it is. People are used to seeing it in composite video so I don't think most would see any different.

It's a well known phenomenon called "dot crawl". For a technical explanation it's where chroma and luma overlap on the edge of their waveform and the TV can't translate it so it ends up with a crawling artifact on the edges of colour changes on the image.

The fix stops these artifacts "scrolling" but they will always be there. Some TVs will handle it better than others, I'd image CRTs would be much better at hiding it than LCDs.

Some cores may not even do it. I've not done extensive testing but like I said before, I needed it pointed out to me to even see it was there.

[Dark Chili]

I was wondering why my colors were wonky and crawly on my television.

[akeley]

If I buy this board, will I see any EXTRA artifacts caused by MiSTer cores + board itself?

I've asked this really simple question some time ago, and nobody has answered it yet. "Yes" or "no" only, please.