RGB image issue with Super Famicom :(

[Alcahest]

Hey All,
I have an issue with Super Famicom outputing RGB 60hz to HDTV.
Here's the trouble:

http://videoff7.free.fr/IMG_0743.jpg
http://videoff7.free.fr/IMG_0746.jpg

http://videoff7.free.fr/snes_240p_signal_flaw.jpg

As you can see, the upper part of the screen seem badly deinterlaced, the bottom part is OK!
I thought it was the cable being dodgy at first, as extremely old, so bought an official Gamecube RGB cable, received it today, but problem is still there 

I have a 2nd Super Famicom which exhibits the same exact visual flaw so I think it's safe to rule out a problem on the Super Famicom side (?)
As well, the cable is brand new and official.
Regarding the TV,  no such effect whatseover when feeding it Playstation 480i RGB 60Hz or any other 480i sources via RGB.

So I'm beginning to wonder where it could be coming from.
Could the Super Famicom(s) be at fault (a capacitor responsible for the video output that badly aged)? (but on 2 famicoms, that would be surprising!)
Could it come from the power supply?

Both SFC are totally unmodded, untouched.

It seems as if the signal the TV is receiving is not perfect enough for the tv to deinterlace it properly, but then why only the upper part of the picture is affected?
My TV has 2 scart inputs, both accepting RGB and both have the issue.

I have another SFC cable which outputs SVHS, no problem whatsoever with it, the image is sharp from bottom to top, but colors are not as good as RGB.

Have you guys already experienced this? Anything I could try to definitely pinpoint the problem?
Thanks, again, a bunch ^^
Later,

Alcahest

edit: just tried to connect to old SDTV set, it looks asbolutely awesome but i guess the scanlines are just masking the problem :/

[vkp_ashley]

Few things, could you give me the make and model of your SCART tv/monitor? Just curious cause the SNES is outputing 240 lines progressive and not 480 lines interlaced. So it makes sense for your TV/monitor to display ps2 games fine cause RGB is 480i which is what your monitor is expecting. For a quick test, play a PSone game on your PS2 and give me back the results cause PSone games output at 240p. Thanks.

[NFG]

This doesn't seem to be an interlaced issue so much as a sync problem.  The horizontal sync is not working properly at the start of the frame, but gets its act together by about halfway.

vkp_ashley: what the hell is 240p?  Back in my day we called it 'low res', or non-interlaced.  I don't think a signal is 'progressive' just because it's not interlaced.  

[Alcahest]

> Few things, could you give me the make and model of your SCART tv/monitor?

It's a Sony 40W2000 HDTV, it is 1 year old.

> Just curious cause the SNES is outputing 240 lines progressive and not 480 lines interlaced.

Hmm, are you sure, the signal coming from Super Famicom RGB cable isn't 480i then? o_O

> So it makes sense for your TV/monitor to display ps2 games fine cause RGB is 480i which is what your monitor is
> expecting. For a quick test, play a PSone game on your PS2 and give me back the results cause PSone games output at
> 240p. Thanks.

I tried Vagrant (^^;) PS1 USA on PS2 USA connected to tv with RGB cable.
no problem whatsoever when it comes to the image, nothing such as what i get with the SFC.

> Lawrence: This doesn't seem to be an interlaced issue so much as a sync problem.  The horizontal sync is not working
> properly at the start of the frame, but gets its act together by about halfway.

Yes, it makes sense!
But since both Super Famicoms i have behave the same when it comes to this problem... it must be cable related?
I'm using a EURO Gamecube RGB cable, so it's wired expecting a EURO NGC, not a Jpn Super Famicom, could it be the issue..
I opened the official NGC RGB cable on both ends, on the Super famicom multiAVout side, only pins 7 & 8 are not connected to anything (Y(S-Video) &  C(S-Video) respectively), at the end of the scart cable it looks like this (unmodified, received a couple hours ago).

On which pin would "horizontal sync" data be passed from SFC through to the TV?
On JPN SFC, pin 3 is CSYNC, it is currently connected to scart pin 8.. (as a GC outputs 12v on pin 3 (pin 8 on scart) to autoswitch to the right tv channel)
I can't seem to find the SFC pin responsible for Horizontal Sync ( http://dempa.jp/rgb/heaven/g_sfc.html & http://gamesx.com/wiki/doku.php?id=av:nintendomultiav )
Arg, the more i'm looking into it, the more it seem hard to understand 

Alcahest

[albino_vulpix]

On which pin would "horizontal sync" data be passed from SFC through to the TV?
On JPN SFC, pin 3 is CSYNC, it is currently connected to scart pin 8.. (as a GC outputs 12v on pin 3 (pin 8 on scart) to autoswitch to the right tv channel)
I can't seem to find the SFC pin responsible for Horizontal Sync (

The SFC feeds it sync through the composite video signal.

[Alcahest]

vkp_ashley, do you have any specific PS1 titles that are confirmed to be output as 240 progressive?
I want to make more tests on the tv side.
If PS1 shows OK on tv, i'll try to join the SFC CSYNC pin to VIDEO pin 20 of scart...

Albino, On SFC, the picture is OK in composite if i disable the Select pin (scart pin8) to force composite display instead of RGB, no strange artefacts on the upper part.
Maybe the CSYNC / CVBS output of NTSC SNES needs to be filtered/enhanced to get a proper horizontal sync signal...?

I would very much like to see how the official Super Famicom Japanese RGB cable is built...
See you,

Alcahest
Edit: vkp would you know if Bloody Roar 1 PSX is a "240p" game, here's the output when played on ps2 with RGB/Scart to TV:
http://videoff7.free.fr/IMG_0759.jpg
Raystorm: http://videoff7.free.fr/IMG_0761.jpg
it looks pretty flawless to me. So at least i can rule out a problem on the TV side.

[marqs]

Notice that there's a resistor parallel to ground (and perhaps a serial capasitor, see http://members.optusnet.com.au/eviltim/gamescart/gamescart.htm#gamecube) if you're using gamecube rgb cable. I've also had sync problems with snes and removing those components fixed it for me. I suggest you open the cable and:
1. Disconnect the csync ilne going to mode select pin 8
2. Take off any extra components from cvbs line (the resistor probably lies within the small block which goes to multiav)
3. Use either cvbs or csync for sync (pin 20)

[Alcahest]

marqs, awesome recommendations, will do and report ASAP.
Are you the author of this great page? http://members.optusnet.com.au/eviltim/gamescart/gamescart.htm#gamecube
I've been meaning to contact you, my official nintendo Gamecube RGB cable (DOL A RGB) is different from your schematics.
that's what the end of the plug looks like, unmodded.
http://videoff7.free.fr/IMG_0758.jpg
As you can see, i have a pin on scart 17 instead of 18, i checked and they're both grounds so it shouldn't matter right?
Do you own a NTSC Snes or PAL? And connected to TV via RGB too?
Thanks for those great tips,

Alcahest

[marqs]

Alcahest: I'm not really the author of that page, I guess it's Viletim from this forum. 17&18 are both grounds as you stated so it doesn't matter. The diagram isn't 1 to 1 with the official cable, if I recall correctly the resistor value for rgb status pin (16) was around 100ohms in the official cable, not 180. Anyway, that site should point out the differences between gc and ntsc snes rgb cables.

My snes is usa version (early revision) and it's connected to my projector via rgb. The projector is quite sensitive with sync since on normal crt tv:s there were no problems with the unmodified gc rgb cable. In fact, I've tried a newer revision snes (not snes jr) with it and the sync wasn't right even with the modified cable! Perhaps the earlier us models had better sync, but I don't know about jap revisions.

[acem77]

I had a problem with a rgb scart cable back in the day for my snes when using my Xrgb2 plus.(well plus some other mod work that had nothing for the problem)
They used composite sync and i would get this sub blocky pixel crawl  effect in the picture.
to fix the problem i cut open the snes end of the cable and used pin 8  S-Video C (chroma) as sync.
This trick work for a few other systems also.

[Alcahest]

My snes is usa version (early revision) and it's connected to my projector via rgb. The projector is quite sensitive with sync since on normal crt tv:s there were no problems with the unmodified gc rgb cable. In fact, I've tried a newer revision snes (not snes jr) with it and the sync wasn't right even with the modified cable! Perhaps the earlier us models had better sync, but I don't know about jap revisions.

Thanks for these extra precisions. I now think my problem has nothing to do with the cable though..

I had a problem with a rgb scart cable back in the day for my snes when using my Xrgb2 plus.(well plus some other mod work that had nothing for the problem)
They used composite sync and i would get this sub blocky pixel crawl  effect in the picture.
to fix the problem i cut open the snes end of the cable and used pin 8  S-Video C (chroma) as sync.
This trick work for a few other systems also.

Thanks for your message acem77, I ended up trying this too, much unfortunately it didn't correct this issue.

Here's what I tried:
- connect pin 3 (CSYNC Snes) to Scart Sync (pin20) -> problem is there
- connect pin 9 (Composite NTSC Snes) to Scart Sync (pin20) -> problem is there
- connect pin 8 (S-video chroma Snes) to Scart Sync (pin20) -> problem is there
I tried connecting pin 3, 9 & 8 all at once to Scart Sync (pin20) -> problem is there.

However I made a major discovery, I ran the "SNES SERVICE TEST MODE" rom on the SFC (i own a backup device) and tested the system...
It turns out when the SNES is outputting Hi-Res, the picture is perfect from bottom to top!!

Low-res: http://videoff7.free.fr/IMG_0763.jpg
Hi-res : http://videoff7.free.fr/IMG_0764.jpg

Which immediately rules out a cable problem, doesn't it, since at least Hi-Res displays perfect? 
(unless hsync is ok for hires, but not for low-res, could come from a cable problem as well?)

So it seems that:

1. Either my TV doesn't support "240p" properly (oh noes  )
-> therefore I would really need the title of a confirmed PS1 "240p" game to see if the TV exhibits the same flaw with PSX.

2. Or the SNES PPU has a problem (on both units I have, that would be surprising; both early JPN SFC... )
-> Dead/dying capacitor somewhere in the PPU?

3. Or the SNES was never intended to be played on anything other than sdtv with scanlines and this issue is inherent to the old technology. (which is pretty much ruled out instantly as you guys have no problems with it!)

From a wiki article:
"Older video game consoles such as the Nintendo Entertainment System generated a non-standard version of NTSC or PAL in which the two fields did not interlace, and instead were displayed directly on top of each other, keeping the orientation of the scanlines constant. This would be 240p and 288p respectively"

It seem that the 2 fields created by the SFC aren't exactly "aligned" to start with, therefore when the image is reconstructed by the TV, the problem shows. It's is exacerbated by the pure quality of RGB. Since composite & SVHS are 480i signals, the issue doesn't show up as the SFC is making the conversion to 480i internally (?).

For now i'll run some tests on the tv side, but it's making more and more sense that it could be a SFC internal issue. (aging PPU..?)
If you guys have any more infos/suggestions you're more than welcome
See you,

Alcahest

Edit: Something to back up the "incorrect fields generated by SFC" theory:
http://videoff7.free.fr/sfc.avi
When I force the TV to display the SFC output as "PAL", in addition to the image being squeezed of course, it is perfect, but slightly bouncing from left to right, and this more on the top of the picture than on the bottom...
When set to Auto =NTSC, the image is stable but has the damn issue, pq is too bad to see it on my video but trust me the issue is there :/
I don't know exactly what all that means, but i think i'm onto something

[Silver]

What type of TV is it? CRT/LCD/Plasma etc?

have you tried your snes + cable on another TV? (do this first).

Whatever the snes is outputting, remember your TV is having to to internally process & scale this image to fit. Quite a lot of scaling to HD.


NTSC/PAL Colour conversion video - eh? If you are connecting via RGB, PAL/NTSC should be irrelevant - they are standards which include colour coding, which are not used (obviously) with RGB.

What is relevant is the timings - 50hz  vs 60Hz, number of lines etc...

From a wiki article:
"Older video game consoles such as the Nintendo Entertainment System generated a non-standard version of NTSC or PAL in which the two fields did not interlace, and instead were displayed directly on top of each other, keeping the orientation of the scanlines constant. This would be 240p and 288p respectively"

= very badly written. Basically they are saying the snes usually outputted progressive (240p) not interlace (480i) - but they have the same scan rate.



A snes outputting 240p basically only writes to the even scanlines, all the odd scanlines will be permanently black (not as noticable as you think on an old CRT). It could be that your problem lies with how your TV is processing this image.

The strange thing is that it only affects part of the picture - makes it hard to pin down.

[Alcahest]

Hey Silver,

> What type of TV is it? CRT/LCD/Plasma etc?

It's a Sony 40W2000 LCD HDTV.
After much testing, I can rule out a problem on the TV side totally (it's a relief!) as:

PS1 games output at 240p display perfectly via RGB (tested on a PS2 with RGB scart cable)
http://videoff7.free.fr/psx_240pRGB_1.jpg
http://videoff7.free.fr/psx_240pRGB_2.jpg

Also I can rule out the responsibility of the SFC RGB-SCART cable as the problem only occurs when the SFC is outputting Low-res (240p). When the SFC is outputting Hi-res (480i) the display is perfect!
(if sync or cable itself was bad, the flaw would show both for low-res & hi-res output wouldn't it?)

Low-res: http://videoff7.free.fr/snes_240p_RGB.jpg
Hi-res : http://videoff7.free.fr/snes_480i_RGB.jpg

The only thing that remains is.. the SFC itself!!
For some reason the signal it is sending in low-res 240p isn't exactly correct and when the TV reconstructs the image, the flaw shows up.
This same SFC with same RGB cable connected to a SDTV looks wonderful, but I do see some very slight flickering towards the top of the screen...
Since my 2nd SFC displays the exact same flaw, and since they're both very early models
Serial : S16296637 & S21890081, I can only believe that those early jpn models either aged badly or were never outputting a perfect 240p signal in the first place.

2 questions:
- Could the AC Adapter be responsible for the flaw? I'm sending the SFC 9 volts when it is expecting 10v per specs.. dumb question i guess, but we never know..

- Could it be that the official Nintendo SHVC-010 RGB cable corrects this "flaw" by modifying the low-res 240p signal before it reaches the TV?

Later,

Alcahest

[Silver]

Hmm.

In producing 240p old systems "cheat" slightly. In normal NTSC/PAL output, each interlaced field starts/ends on a "half scanline". When going progressive 60hz (by skipping the odd scanline field) it must refresh either half a scanline early or late. (I know you are RGB, but the sync output will be similar).

The fault in your pics is that alternate scanlines are shifted slightly, with the fault reducing over the top half of the image. I was going to suggest that processing in your TV was getting confused by this (i.e. programmed to expect a field start/end of half a scanline). However, you say it works fine on a PS1 RGB output at 240p. I can only assume the PS1 does the same trick, so I'm barking up the wrong tree.

The fault looks like your horizontal blanks are of different timings from line to line (they should not be - should be fixed) at the start of each frame and converge gradually.

[Alcahest]

Hmm.

In producing 240p old systems "cheat" slightly. In normal NTSC/PAL output, each interlaced field starts/ends on a "half scanline". When going progressive 60hz (by skipping the odd scanline field) it must refresh either half a scanline early or late. (I know you are RGB, but the sync output will be similar).

Ok, I understand.

The fault in your pics is that alternate scanlines are shifted slightly, with the fault reducing over the top half of the image. I was going to suggest that processing in your TV was getting confused by this (i.e. programmed to expect a field start/end of half a scanline). However, you say it works fine on a PS1 RGB output at 240p. I can only assume the PS1 does the same trick, so I'm barking up the wrong tree.

Yes, I'm puzzled as well. TV works just fine with PS1/240p. So it seems that indeed the Super Famicom isn't generating either the odd field or the even field properly? That one field is "shifted", a little bent compared to the other and so it produces the flaw, yes?

So let's consider the TV to be OK.
About the RGB cable, am I right to suppose that since Snes 480i output is OK and the problem only occurs with 240p, it cannot be a cable issue?

The fault looks like your horizontal blanks are of different timings from line to line (they should not be - should be fixed) at the start of each frame and converge gradually.

So would you say that the Super Famicom itself could be responsible for this?
2 systems with the same issue, would be very bad luck, but if it's a generalized problem i'm very surprised the issue hasn't been raised by other people with JPN SFC connected to their display via RGB 
Would an oscilloscope be able to detect this flaw somehow (connected to the CSYNC or CVBS pin at the SFC AVMulti-out plug)?
This would be definite proof.

Btw Silver, did you register just to help on my issue, i'm very grateful 
See you,

Alcahest
Edit: Also could the AC Adapter i'm using be somehow responsible for this..  if the SFC is a bit underpowered (9v instead of 10), has it any influence on video output?

[NFG]

I'm pretty convinced that you shouldn't be calling it 'progressive' when it's just a 240 vertical pixel display.  No one in the history of gaming called it that until they started using the term to impress soccer moms and football dads who couldn't be expected to learn something technical like 'non interlaced'.

Progressive means that it's filling all the lines of a screen in one pass, and in the case of low-res consoles, it's not: it's only filling half the screen, so interlaced or not, it's not a progressive signal.

[kyuusaku]

Silver, what is this half scanline thing? TVs toggle a flag every line, and because there are an odd number of lines, at retrace this will indicate whether the last field was even or odd. TVs are tricked into progressive mode by simply using a consistent number of lines every field (262/263 depending on whether you want to display on odd/even lines).

I think it's proper to say the SNES generally outputs 240p(60) video since if you looked at the waveform, that's exactly what it is, it doesn't blank a field of 480i video to achieve half the resolution (not that that's possible since the phosphors don't glow that long). For all the SNES knows it is driving a fixed 240 line display like LCD (and then a 480i display when it switches to a "high resolution" mode).

[Alcahest]

Lawrence & kyuusaku, what are your takes on my problem given the new infos I gathered in Reply #14?
Any ideas at all for the source of the issue, bad batch of SFCs? 
Thanks,

Alcahest

[RGB32E]

I think it's proper to say the SNES generally outputs 240p(60) video since if you looked at the waveform.....

No... the snes and other low rez consoles typically output in "double strike" mode (i.e. where each field is treated as a full frame with the same scanline row position).  Hence, thats why low rez games don't exhibit the same level of "flicker" as higher resolution games (over ~255 rows of pixels) on standard def CRTs.  So, to discuss 240p or 262.5P is a misnomer. 

[kyuusaku]

Lawrence & kyuusaku, what are you takes on my problem given the new infos I gathered in Reply #14?
Any ideas at all for the source of the issue, bad batch of SFCs? 
Thanks,

Alcahest

I think it's your TV's DSP, but probably the console's fault. Have you tried using a NES via composite? I wonder if it has the same problem. It appears the TV isn't locking fast enough on every other line for the first half of the screen.

Edit: looking closer, the problem goes all the way down to the bottom... Does the cable have a resistor and/or cap on the sync line? I guess it could be a RC constant though logically, all the lines would gradually shift, not every other line. Every other line = the TV is doing something with the signal, perhaps "fixing" it.

I think it's proper to say the SNES generally outputs 240p(60) video since if you looked at the waveform.....

No... the snes and other low rez consoles typically output in "double strike" mode (i.e. where each field is treated as a full frame with the same scanline row position).  Hence, thats why low rez games don't exhibit the same level of "flicker" as higher resolution games (over ~255 rows of pixels) on standard def CRTs.  So, to discuss 240p or 262.5P is a misnomer. 

Reread my post more carefully because until the last sentence you are agreeing with me, you just don't know you are. And 240p is NOT a misnomer, as I already said there is no difference between "standard definition" SNES video and "true low definition" video intended for a fixed 240 line display. If you don't believe 240p displays exist, that's your problem.

[Alcahest]

I think it's your TV's DSP, but probably the console's fault. Have you tried using a NES via composite? I wonder if it has the same problem. It appears the TV isn't locking fast enough on every other line for the first half of the screen.

I don't own a NES. However here is the SFC connected via S-Svideo:
http://videoff7.free.fr/snes_svideo.jpg

The flaw is there too, hardly noticeable (and I had to put tv sharpness to max to emphasize it), but it is there!
Connection via RGB seem to aggraviate it badly because the signal is much cleaner.

looking closer, the problem goes all the way down to the bottom... Does the cable have a resistor and/or cap on the sync line? I guess it could be a RC constant though logically, all the lines would gradually shift, not every other line. Every other line = the TV is doing something with the signal, perhaps "fixing" it.

I have 2 cables.
First one is no-brand cable with CVBS going directly from SFC pin 9 to Scart pin20 (no components).
Second cable is brand new official Nintendo Gamecube RGB cable, it takes also CVBS signal from SFC(pin 9) and passes through a 220 uF/ 6.3v capacitor before reaching scart pin 20.
The problem exhibits with both cables in the exact same way.

Again, the flaw is apparent when the SFC outputs low-res, but hi-res output is OK, even via RGB!

The fact that S-video exhibits the same flaw as RGB rules out a problem of Sync, doesn't it?
(since SVHS doesn't use the CVBS pin or the CSYNC pin of SFC)
Therefore the cable(s) would be OK.

Also since the problem is there (slightly but still) with S-Video and that, correct me if i'm wrong, the SFC itseft is doing the 240p to 480i conversion internally when S-Video is used, shows that it isn't a TV issue either.

The more and more it seems that the Super Famicom itself is responsible for the flaw.. not the cable or the TV.

I'm going nuts over this. If it was really a Super Famicom internal issue, certainly i wouldn't be the only one noticing the problem!? 
See you,

Alcahest

[albino_vulpix]

Something funky is definitely going on inside the SFC. I think a new one is the best bet.

[NFG]

Edit: Also could the AC Adapter i'm using be somehow responsible for this..  if the SFC is a bit underpowered (9v instead of 10), has it any influence on video output?

I've seen low-power and failing power supplies do all kinds of weird things.  My NeoGeo would have a white-fading-to-normal overlay from the top of the screen, and eventually it started to wobble before finally failing.  Back then (early 90s) I paid over $100 CDN to get a new power supply from SNK US.  =(

It MAY be the problem, but I doubt it.  Luckily it doesn't cost much to try a different PSU.  =)

[Alcahest]

Something funky is definitely going on inside the SFC. I think a new one is the best bet.

The problem is that this damn issue is there on 2 different SFCs already
The probability for both of them to be busted in the exact same way must be very slim..

I've seen low-power and failing power supplies do all kinds of weird things.  My NeoGeo would have a white-fading-to-normal overlay from the top of the screen, and eventually it started to wobble before finally failing.  Back then (early 90s) I paid over $100 CDN to get a new power supply from SNK US.  =(
It MAY be the problem, but I doubt it.  Luckily it doesn't cost much to try a different PSU.  =)

Ouch, quite a sum indeed for a PSU!
Actually I have 2 power adapters, 1 unofficial chinese AC/DC adapter specifically made for SFC, just "1000 mA" written on it with a SFC logo, and an old Master System 9V - 1A - 9VA adapter.
They, very much unfortunately, both produce the same results.
However, what exactly is "VA" in 9VA?
The Jpn Super Famicom official specs are 10V - 850mA -19VA..
9VA versus 19VA hmm..
Thanks for your comments guys, much appreciated.
Later,

Alcahest
Edit: btw, would anyone know what this is for (called "TC1" on the board) ? http://videoff7.free.fr/whatsfc.jpg

[kyuusaku]

Again, the flaw is apparent when the SFC outputs low-res, but hi-res output is OK, even via RGB!

That's because it's not confused by the SFC's interlaced video. Are you sure the TV isn't treating it as interlaced video and deinterlacing it? Are you POSITIVE the PS games that work are truly "low resolution" and that the PS2 isn't scaling them to 480i? Do you have any other old consoles to try?

The fact that S-video exhibits the same flaw as RGB rules out a problem of Sync, doesn't it?
(since SVHS doesn't use the CVBS pin or the CSYNC pin of SFC)

Well, it proves that the problem *is* sync since it happens with all video signals, but it's not caused by the cable. I'm afraid it's just your TV, there isn't anything you can easily do about it except maybe complain to Sony. Is there a game mode on the TV?

[Alcahest]

That's because it's not confused by the SFC's interlaced video. Are you sure the TV isn't treating it as interlaced video and deinterlacing it? Are you POSITIVE the PS games that work are truly "low resolution" and that the PS2 isn't scaling them to 480i? Do you have any other old consoles to try?

No I'm not totally certain if my PS2 was outputting 240p or 480i... indeed.. I don't have my PS1 handy but this is definitely worth a re-try.

New test on my USA PS2 SCPH-50001 with FF9 USA (240p output)
http://videoff7.free.fr/ps1_240p_RGB.jpg
It looks as good as it possibly can imho! ^^
Is it known for a fact that PS2 outputs 240p for PS1 games or is it speculation?

Well, it proves that the problem *is* sync since it happens with all video signals, but it's not caused by the cable. I'm afraid it's just your TV, there isn't anything you can easily do about it except maybe complain to Sony. Is there a game mode on the TV?

No there's no game mode and I already de-activated all image enhancement processes.
It still has a year warranty though.
But wait a sec, isn't S-video delivering 480i directly to the TV? In which case the problem shouldn't show up at all if it was only an issue of the tv handling 240p, wrong thinking on my end?
See you,

Alcahest

[albino_vulpix]

But wait a sec, isn't S-video delivering 480i directly to the TV? In which case the problem shouldn't show up at all if it was only an issue of the tv handling 240p, wrong thinking on my end?
See you,

Alcahest

S-video delivers '240p' just as the RGB lines do. And sorry, I missed the "tried 2 consoles" bit.

[Silver]

Silver, what is this half scanline thing? TVs toggle a flag every line, and because there are an odd number of lines, at retrace this will indicate whether the last field was even or odd. TVs are tricked into progressive mode by simply using a consistent number of lines every field (262/263 depending on whether you want to display on odd/even lines).

In standard PAL video, each field has 287.5 scanlines worth of active picture in each field (312.5 scanlines total - active picture+blanking). Since the adjacent fields will get interlaced together when drawn on the screen, we have a total of 574 complete scanlines and two half scanlines in each video "frame"  of active picture.  (This will be 624 complete scanlines and two half scanlines in each video "frame"  of active picture+blanking=625 lines - the full PAL spec).

(In standard NTSC video, each field has 262.5 scanlines of active picture + blanking) The fields are paired off into frames, where the first field of each frame ends with a half scan line and the second field starts with a half scan line.

This is achived by having different vertical sync pulses for each field of video. Field 1 and Field 2 have different vertical syncs (of different timings) which produces an offset - which in turn means the 2nd field is drawn offset from the first by a set amount = interlaced video.

A lot of old systems - such as the SFC - step outside the PAL/NTSC specs to produce a non-interlaced output. They do this by using the same vertical sync for each field. e.g. They just repeat the vertical sync for Field 1 over and over, no Field 2 vertical sync is used.

Thus there is no offset between fields, the odd scanlines remain permanently black, the even scanlines are updated at 50Hz not 25Hz (PAL - 30/60ish for NTSC) and there is no interlacing. This now produces repeated fields of 312 lines, not 312.5 lines. This changes the frame rate in PAL from 50Hz to 50.0801Hz - but this is well within spec/tolerance for CRT Tv's that these systems were designed for.

All this has the advantage of removing interlace flicker (nice stable image) at the expense of perceived vertical resolution (which is halved).

.....

Soooo... Where this all gets really messy is when modern LCD/Plasma flatscreens internally process video signals into something they can display.

I'm afraid I'm probably with kyuusaku - most likely problem lies with how your TV is processing this signal. The only confusing aspect is that you say it works a treat with a ps1 at 240p. (I have read elsewhere that someone hooked up a ps2 to a 'scope and confirmed the above process was used playing a non-interlaced game).

It's possible there are slight differences - maybe the sfc is alternating between 312/313 lines from field to field to keep the framerate locked at 50Hz (in PAL terms) but I doubt it.

I don't *think* your TV is de-interlacing the signal, as you would expect it produce black scanlines for the odd lines field component. Unless, of course, it is deinterlacing it using all the even lines (so you would get a solid pic, as you do) but I still would not expect your area (or I'd expect a much worse offset).

[Alcahest]

Soooo... Where this all gets really messy is when modern LCD/Plasma flatscreens internally process video signals into something they can display.

I'm afraid I'm probably with kyuusaku - most likely problem lies with how your TV is processing this signal. The only confusing aspect is that you say it works a treat with a ps1 at 240p. (I have read elsewhere that someone hooked up a ps2 to a 'scope and confirmed the above process was used playing a non-interlaced game).

Ok with ps2 confirmed as outputting ps1 games in "240p", there is no doubts anymore, the TV is fine.

It's possible there are slight differences - maybe the sfc is alternating between 312/313 lines from field to field to keep the framerate locked at 50Hz (in PAL terms) but I doubt it.

But then everyone on HDTV would have/see the problem with SNES, right?
(well only if people are using RGB as the flaw is very hard to see with SVHS/Composite)

I don't *think* your TV is de-interlacing the signal, as you would expect it produce black scanlines for the odd lines field component. Unless, of course, it is deinterlacing it using all the even lines (so you would get a solid pic, as you do) but I still would not expect your area (or I'd expect a much worse offset).

Yes i think the TV is fine.
Here are the thoughts of Lyris, HDTV expert and owner of the same TV model as me.

Yes, the progressive nature of the TV is showing the problem, but it's not causing it as far as I'm aware. It looks like the scanlines from one field are out of sync/somehow distorted, but the other is fine.
It's quite likely that the analogue timing on these old consoles is a little slack. That wouldn't have mattered on an analogue TV, but since a panel has to digitise the signal first, it will have a lower tolerance for this sort of thing.

That's exactly my thoughts too since Playstation "240p" display is just fine on this TV.
It looks as if the Super Famicom just isn't sending a totally correct signal in the first place.

Therefore I would love to hear feedback regarding you guys' actual experiences with SNES connected to HDTV via RGB.
If it looks 100% fine, totally free of the defect I'm having towards the top of the screen, can you give your system's origin (JPN SFC / USA SNES) & maybe serial # ?

Better yet, If any of you has a SFC on sale which you have confirmed to be sending a totally accurate signal in RGB, I would be willing to buy it from you. (even if badly yellowed or case broken, as long as it hasn't been modded, with a clean board). I can give ebay refs and have paypal too.

I'm taking a bet thinking these 2 early SFCs I own don't produce a perfect signal to start with, but at this point this is the only position that makes sense to me.
Thanks again for these precisions Silver 
See you,

Alcahest

[kyuusaku]

In standard PAL video, each field has 287.5 scanlines worth of active picture in each field (312.5 scanlines total - active picture+blanking). Since the adjacent fields will get interlaced together when drawn on the screen, we have a total of 574 complete scanlines and two half scanlines in each video "frame"  of active picture.  (This will be 624 complete scanlines and two half scanlines in each video "frame"  of active picture+blanking=625 lines - the full PAL spec).

(In standard NTSC video, each field has 262.5 scanlines of active picture + blanking) The fields are paired off into frames, where the first field of each frame ends with a half scan line and the second field starts with a half scan line.

This is achived by having different vertical sync pulses for each field of video. Field 1 and Field 2 have different vertical syncs (of different timings) which produces an offset - which in turn means the 2nd field is drawn offset from the first by a set amount = interlaced video.

Sorry I misunderstood, I know what you're talking about, but what does it have to do with "non-interlaced" SNES video?

But then everyone on HDTV would have/see the problem with SNES, right?
(well only if people are using RGB as the flaw is very hard to see with SVHS/Composite)

It's your specific TV's locking ability/logic that's failing. Every brand/series has different algorithms and different performance.

That's exactly my thoughts too since Playstation "240p" display is just fine on this TV.
It looks as if the Super Famicom just isn't sending a totally correct signal in the first place.

Of course it doesn't exactly follow any standard, nothing does (even broadcast doesn't necessarily), the TV should have the same or even better tolerance than an analog TV though.

I'm taking a bet thinking these 2 early SFCs I own don't produce a perfect signal to start with, but at this point this is the only position that makes sense to me.

That makes no sense to me, I don't think there's anything wrong with the consoles, and it's even less likely that the raster state machine changed between revisions.

[Alcahest]

It's your specific TV's locking ability/logic that's failing. Every brand/series has different algorithms and different performance.

I can understand that, but then that would mean that the "240p" generated by SNES/SFC is different than the "240p" generated by PS1.

Of course it doesn't exactly follow any standard, nothing does (even broadcast doesn't necessarily), the TV should have the same or even better tolerance than an analog TV though.

Then it'a another mystery :/

That makes no sense to me, I don't think there's anything wrong with the consoles, and it's even less likely that the raster state machine changed between revisions.

As unlikely as it is, there's gotta be an explanation for this issue!
As far as I know there are many different PPU revisions too..
My question still stands: does anyone here have a perfect image with US SNES or JP SFC outputting RGB to an HDTV ?
Of course interlaced RGB monitors are exluded from this question, as the flaw is hidden.
Proof that SNES and PS1 output the same "240p" or a different one would be very welcome too.
See you,

Alcahest

[Silver]

In standard PAL video, each field has 287.5 scanlines worth of active picture in each field (312.5 scanlines total - active picture+blanking). Since the adjacent fields will get interlaced together when drawn on the screen, we have a total of 574 complete scanlines and two half scanlines in each video "frame"  of active picture.  (This will be 624 complete scanlines and two half scanlines in each video "frame"  of active picture+blanking=625 lines - the full PAL spec).

(In standard NTSC video, each field has 262.5 scanlines of active picture + blanking) The fields are paired off into frames, where the first field of each frame ends with a half scan line and the second field starts with a half scan line.

This is achived by having different vertical sync pulses for each field of video. Field 1 and Field 2 have different vertical syncs (of different timings) which produces an offset - which in turn means the 2nd field is drawn offset from the first by a set amount = interlaced video.

Sorry I misunderstood, I know what you're talking about, but what does it have to do with "non-interlaced" SNES video?

In theory, not much - except I was considering that the implementation of non-interlaced is going to be 1/2 scanline shorter than "expected". Again it comes back to how the modern to processes the vertical sync pulses that are coming in out of spec....

In reality I suspect they don't spend much time testing their latest HDTV flatscreen on incoming 240 line verical signals...


@Alcahest

To get the bottom of why your ps2 at 240p works and your snes doesn't, you really need to look at the signals. Sadly this probably means knowing someone who has access to 'scope at home or work...

[Alcahest]

@Alcahest
To get the bottom of why your ps2 at 240p works and your snes doesn't, you really need to look at the signals. Sadly this probably means knowing someone who has access to 'scope at home or work...

Arg, indeed I don't know such person 

[RGB32E]

Alcahest,

Build yourself a sync stripping circuit (look up LM1881 or http://www.gamesx.com/misctech/lm1881.htm for starters).  Feed CVBS, LUMA (someone else thought they were tapping chroma in a previous post), or CSYNC as the input to the circuit.  Take the CSYNC output and that will likely solve your problem (i.e. generate a new sync signal).  Scope that!  

[Alcahest]

Alcahest,

Build yourself a sync stripping circuit (look up LM1881 or http://www.gamesx.com/misctech/lm1881.htm for starters).  Feed CVBS, LUMA (someone else thought they were tapping chroma in a previous post), or CSYNC as the input to the circuit.  Take the CSYNC output and that will likely solve your problem (i.e. generate a new sync signal).  Scope that!  

Ohhh great idea indeed!
I need to get my hands on one of these things, it may be my last resort. 
Just wondering if it does indeed generate a new signal or just strips the signal as only operation. Because i already tried feeding CSYNC (sfc pin 3) to scart pin 20 and it didn't help :/
CSYNC is already "pure" isn't it?
See you,

Alcahest

[RGB32E]

Ohhh great idea indeed!
I need to get my hands on one of these things, it may be my last resort. 
Just wondering if it does indeed generate a new signal or just strips the signal as only operation. Because i already tried feeding CSYNC (sfc pin 3) to scart pin 20 and it didn't help :/
CSYNC is already "pure" isn't it?
See you,

Alcahest

No, CSYNC isn't necessarily "pure"  ....  From what I recall when reading through the datasheet for this IC, it generates a new CSYNC/HSYNC and VSYNC from the input signal containing a composite sync.  One example is the Street Fighter: The Movie arcade board.  Even though this is a standard jamma game (with CSYNC), the game would just not display correctly due to a funky CSYNC.  I ran the RGBS from the arcade board through an Extron PA 250, and the problem was solved (strips the sync, even though it I was using RGBS in and out).  Alternatively I could have used a LM1881 circuit, but the PA 250 was more convenient (cables ready to go).  Point being... that even if you have CSYNC it may need some massaging (sync stabilization/stripping)... 

http://www.national.com/mpf/LM/LM1881.html
I've built several of these according to the diagram on this page.  So, they use more components than the minimal ones described by the gamesx link (and the basic operation section).

[Alcahest]

No, CSYNC isn't necessarily "pure"  ....  From what I recall when reading through the datasheet for this IC, it generates a new CSYNC/HSYNC and VSYNC from the input signal containing a composite sync.  One example is the Street Fighter: The Movie arcade board.  Even though this is a standard jamma game (with CSYNC), the game would just not display correctly due to a funky CSYNC.  I ran the RGBS from the arcade board through an Extron PA 250, and the problem was solved (strips the sync, even though it I was using RGBS in and out).  Alternatively I could have used a LM1881 circuit, but the PA 250 was more convenient (cables ready to go).  Point being... that even if you have CSYNC it may need some massaging (sync stabilization/stripping)... 

Great info, giving me hope! 

http://www.national.com/mpf/LM/LM1881.html
I've built several of these according to the diagram on this page.  So, they use more components than the minimal ones described by the gamesx link (and the basic operation section).

pm sent 

[viletim]

Alcahest,

I just had a skim though the thread and I can say the problem is most likely with your Famicom AND your HDTV. The video output from the Famicom doesn't conform to the standard, and your TV's digitiser is poor and unable to handle this. It's really that simple.

A case in point: I have a old TV pattern generator (I think it's a built kit from a magazine or something) which I used to test repaired TVs with. I have connected to what must be hundreds of analog TVs without any problem but when connected to a TV with digital processing (especialy with Plasma/LCD) there's about a 1 in 10 chance of the picture being either distorted or ignored altogether (TV says "No Signal"!!!). I don't use this pattern generator any more.

[Alcahest]

viletim, thanks for your input. Do you think "sync stripping" would help then? there's a chance if it's a problem on the famicom side i guess?

If with sync stripping the signal becomes within specs, my TV would be able to deal with it, like it does for PS1 "240p", I really hope because after all I won't have my TV replaced only for super famicom. this range of tv have "clouding" problems and mine is perfect in that aspect so i won't be tempting the devil too much 

And again, i still wonder, does ANYONE have a perfect image with US SNES or JP SFC outputting RGB to an HDTV/progressive display ?
See you,

Alcahest
(btw viletim, could it be that these strange components on the EURO scart cable "correct" this flawed signal..?)

[viletim]

Alcahest,
The problem is with the timing of the video signal, not the electrical properties (which is what the componets in the various cables are to correct). The timing of the pulses which tell the TV when to start a new line, new field, etc are a bit off. Nintendo (as well as all the others) allowed this because it reduces complexity and they knew they could get away with it - it worked with all TVs of the period.

This kind of thing has happened before. Anybody tried to connect an old pong "TV game" to an auto-tuning TV (any made in the last 25 years)? The tuner won't lock on to the signal half the time. But it works fine on the old clunker with a turret tuner - what it was designed to work with.

Technology marches forward, tries to be backward compattible, and succeeds most of the time... Same thing with PC/Windows software, light bulbes, etc

I don't think the LM1881 sync stripper is going to help much, it just a comparitor which looks at the bottom of the video signal and reproduces that part on the output. The only correction done is to fix an elongated vertical pulse. It couldn't hurt to try, I suppose...