Video probs w/ SNES RGB (Vertical Bands, etc)

The place for all discussion on gaming hardware
User avatar
James-F
Posts: 87
Joined: Fri Mar 23, 2018 11:01 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by James-F »

The 1.2k/750 (or R3) will definitely affect the levels of Composite and S-Video as it is before the S-RGB encoder,
the question is how much and whether Comp and S-Video are out of spec before modification (stock).

Lets leave the question "why I would want to use S-Video or Composite ever again" aside for a moment.

EDIT:
I did some measurement in comparison to my trusty old SFC 2chip, and indeed the SFC Jr. has slightly hotter Composite and S-Video, so adding 1.2k will bring them to spec levels too.

EDIT2:
1.2k with THS7316 overdid it on my SFC Jr. and it is dimmer than my 2chip SFC.
I think I will have to use higher value than 1.2k for slightly brighter image.

EDIT3:
I completely removed the 1.2k resistors and now my 1chip with THS7316 matches with my 2Chip SFC in terms of brightness on the RGB Lines,
but the Comp and S-Video are a touch brighter than on the 2Chip.
Bob from RetroRGB reported that some people did not need the 1.2k at all on their SFC Jr. with THS7314/6 which is my case too.
Last edited by James-F on Sat Mar 31, 2018 4:48 am, edited 4 times in total.
User avatar
James-F
Posts: 87
Joined: Fri Mar 23, 2018 11:01 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by James-F »

Just wanted to share and say a big thank you (including Ultron from AssemblerGames Forum).
Edit: This is with 1.2k still installed.
Spoiler
Image
Image
Image
Image
Image
Last edited by James-F on Sun Apr 01, 2018 3:07 pm, edited 1 time in total.
User avatar
James-F
Posts: 87
Joined: Fri Mar 23, 2018 11:01 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by James-F »

Regarding C11;
I have installed a 470nF value and it fixed the ghosting artifacts, but as we know Capcom games exhibit in the top 1-3 lines some darkening and uneven fade-in/out effect.
I guess it is a compromise beween the two, so I'm interested what value people use and whether Capcom games can be fixed.
rama wrote: Regarding C11: What would the effect of a higher / lower ESR capacitor (of the same size) be?
Anyone has an answer to this?
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by FBX »

James-F wrote:...and whether Capcom games can be fixed.
I asked Voultar this and he says it cannot be fixed. It's a glitch of the 1CHIP line that the cap merely intensifies.
User avatar
James-F
Posts: 87
Joined: Fri Mar 23, 2018 11:01 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by James-F »

Thanks.
It is easy to ignore the single dark line and uneven fade effect in some Capcom games in favor of a completely ghosting free image, so the 470nF stays in.
I can always pull the 2chip if I'm really bothered by the 1chip quirks,,, and vice versa of course. :D
naz
Posts: 71
Joined: Tue May 02, 2017 5:26 pm

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by naz »

Hi, just wanted to ask if those top lines with issues on the capcom games are visible if you use the OSSC on 5x.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by FBX »

naz wrote:Hi, just wanted to ask if those top lines with issues on the capcom games are visible if you use the OSSC on 5x.
Likely yes, because those Capcom games start partway down the screen. Line5X in 1080p mode would only cut the top 4 and bottom 4 pixels, which games like SFII don't even use.
User avatar
Syntax
Posts: 1774
Joined: Wed Aug 09, 2017 12:10 am
Location: Australia

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by Syntax »

FBX wrote:
James-F wrote:...and whether Capcom games can be fixed.
I asked Voultar this and he says it cannot be fixed. It's a glitch of the 1CHIP line that the cap merely intensifies.

If you can find a pin that goes high or low when that Capcom feature kicks in then you could make a 2 way fix.
User avatar
ikari_01
Posts: 27
Joined: Mon Jan 30, 2012 9:07 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by ikari_01 »

Here's a little POC IPS patch for Super Street Fighter II - The New Challengers (U) [CRC32: f16d5ce9] (headerless of course):
https://sd2snes.de/files/misc/1chip-dac ... %20fix.ips

The "glitch" occurs because the game changes the brightness DAC value from 0 (to be precise: forced blanking + brightness 0) to the desired brightness mid-screen. The brightness DAC doesn't have enough time to adapt.

The "fix" is to patch it so when it turns on forced blanking during VBlank it doesn't reset the brightness to 0 but retains whatever brightness was set previously by game logic and just sets the forced blanking bit in addition. This way the brightness DAC doesn't have to adapt to any brightness change at the time the screen is turned back on during active display. The forced blanking bit doesn't affect the brightness DAC but just makes the PPU output black pixels.

This works fine for me during the stage select and fights but there might be other occurrences to cover. It's just a POC ;) Would be nice if anybody else could test it. I took a little bit of free ROM space to store the extra code so it might interfere with other patches.

Old:

Code: Select all

C001A9  LDA #$80
C001AB  STA $2100
New:

Code: Select all

C001A9  JSR $C119
C001AC  NOP
C001AD  NOP

...

C0C119  LDA #$80
C0C11B  ORA $C3  <-- variable for brightness in WRAM
C0C11D  STA $2100
C0C120  RTS
User avatar
James-F
Posts: 87
Joined: Fri Mar 23, 2018 11:01 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by James-F »

It works!

Will it be possible to make a small exe tool that patches the roms directly without having to apply IPS patches?
Something similar to this N64 Antialiasing fix (by saturnu):
https://assemblergames.com/threads/is-i ... ost-879966
It searches for the right memory place and patches the rom directly.

If not, Ikari can you please share how to find the proper addresses and save the changed rom?
I use Lunar IPS to patch but how do I change to original rom?

There are enough games plagued by this to bother one man with requests, that's why I want to do it myself.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by FBX »

Interesting approach of patching the code of the game itself!

I would love to see this done for all the SF and other Capcom titles affected, and a download of the patches posted here!
tzakiel
Posts: 165
Joined: Tue Jan 14, 2014 8:03 pm

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by tzakiel »

How does one replace the c11 cap? I’m new to soldering. Any videos or guides for this type of cap? is it any harder then replacing typical caps with legs thru the holes?
User avatar
Kez
Posts: 818
Joined: Thu Jul 20, 2017 7:09 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by Kez »

It's a little bit fiddly but not particularly difficult.

Removing them is easy - heat it from the side so that the iron is touching the solder at both ends, then just slide it off.

What I normally do after that is clean both pads, then put a bit of solder on one pad, position the cap with tweezers and reflow the solder to tack it into place. Then I add solder to the other side to secure it, then properly solder the first side.

There's some information here:

https://etim.net.au/av-driver/pcebars/

and a video with some SMD soldering here:

https://www.youtube.com/watch?v=C-aOu9IO1FU
User avatar
James-F
Posts: 87
Joined: Fri Mar 23, 2018 11:01 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by James-F »

Been playing with "SNES Professional ASM Development Kit" software which is a very handy tool for this kind of job.

After some fiddling a simply search for "A9808D0021" hex value in "Code" and change the default 80 value.
As a result I found that there are 16 (0-F) values (4 LSB) from 80 to 8F that change the behavior of the fade effect, while the 4 MSB of 80,90,..,F0 are just repetition of the same 16 values.

Please try these IPS for Aladdin (U) [!], and find the most appropriate value.
http://www.mediafire.com/file/ua581d5m5 ... %20IPS.zip

I found that 84 value gave the most accurate fade effect with only the top line slightly darkened.
This is with 470nF Cap.
What's more, this value fits most games that I've tried. :D

Not all games I tried have "filling gibberish" like in SSF2-NC rom, that's why I decided to stay with a simple value change without using the "Jump" command ikari used.

EDIT:
I can see a bright line at the top when Pausing in Yoshi's Island on a 1Chip with 470nF.
So it is not just with Capcom games.

EDIT2:
Stock SFC (fat) 1Chip-02 shows the same darkening in the first few pixels on the top line in all Capcom games.
I can see the uneven fading effect but it is much quicker than with 470nF.

Ghosting is present but very slightly, almost not noticeable with the OSSC, much better than SFC Jr. when stock.
My SFC Jr. had much more ghosting than my SFC 1Chip-02 with stock 47nF cap.
Spoiler
Image
samson7point1
Posts: 58
Joined: Sun Jun 26, 2016 6:44 pm

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by samson7point1 »

I'm just sharing this in case it ends up being helpful to someone. When I installed the RGB bypass in my 01-CHIP-03, I swapped this capacitor out as a matter of course and without first checking to see if I even had the ghosting problem. That was probably stupid, but I *thought* it would be safe to get the circuit into a known good state.

In my case it actually degraded the picture noticeably. Subjectively the colors and brightness all seemed a little more subdued, but the thing that really brought to my attention was the effect it had on my Super Gameboy. After the SGB splash screen loaded, the top row of pixels flickered between light and dark persistently. When I put the original capacitor back in everything went back to being beautiful. I was unable to detect even a hint of the ghosting issue with the OSSC, so it turns out it was not even really necessary.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by FBX »

samson7point1 wrote:I'm just sharing this in case it ends up being helpful to someone. When I installed the RGB bypass in my 01-CHIP-03, I swapped this capacitor out as a matter of course and without first checking to see if I even had the ghosting problem. That was probably stupid, but I *thought* it would be safe to get the circuit into a known good state.

In my case it actually degraded the picture noticeably. Subjectively the colors and brightness all seemed a little more subdued, but the thing that really brought to my attention was the effect it had on my Super Gameboy. After the SGB splash screen loaded, the top row of pixels flickered between light and dark persistently. When I put the original capacitor back in everything went back to being beautiful. I was unable to detect even a hint of the ghosting issue with the OSSC, so it turns out it was not even really necessary.
Easiest place to check for ghosting is the snow stage in Mega Man X. Do a lossless screengrab and look closely to the right of the health meter. If you see faint zebra stripes in the background, you've got ghosting.
User avatar
ikari_01
Posts: 27
Joined: Mon Jan 30, 2012 9:07 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by ikari_01 »

Works fine on my 1chip at least but I'll mention that the sync/shadow patch forces brightness 0 instead of what the game intended to write, so the shadow in A.S.P. will be a bit darker than expected. ;)
fandangos
Posts: 144
Joined: Tue Sep 18, 2012 3:48 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by fandangos »

I'm building my own cables for my SNES 1-Chip and I would like to some clarification on this:

Retrogaming cables sell a C-Sync attenuated:
"Sync output available as CSYNC which is correctly attenuation which uses a 470Ω resistor on the console end of the cable"

And I remember that some cables use 330 ohm resistors with C-sync.

So which one is correct? 330 or 470?
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by FBX »

fandangos wrote:
And I remember that some cables use 330 ohm resistors with C-sync.

So which one is correct? 330 or 470?
For attenuating true TTL csync, I believe 470 is considered the typical rating to use.

However, apparently the NTSC SNES needs only 330 from what I've read.
rama
Posts: 1373
Joined: Wed Mar 08, 2017 3:15 pm

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by rama »

Some SNES consoles output stronger CSYNC than others, so there is a range for the series resistor.
470 Ohm is a good value that should always work though.
fandangos
Posts: 144
Joined: Tue Sep 18, 2012 3:48 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by fandangos »

Thanks to both of you.

I'll work with 470 on my cables.
tzakiel
Posts: 165
Joined: Tue Jan 14, 2014 8:03 pm

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by tzakiel »

Huh. I got a cap for this from console5, specifically marked for my snes jr for this purpose and my picture was all screwed after installing. I then tried a fresh new cap and same result. Then I put the original in and it’s fine again (but of course ghosting not fixed).

Disappointing outcome... I wonder what happened.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by FBX »

tzakiel wrote:Huh. I got a cap for this from console5, specifically marked for my snes jr for this purpose and my picture was all screwed after installing. I then tried a fresh new cap and same result. Then I put the original in and it’s fine again (but of course ghosting not fixed).

Disappointing outcome... I wonder what happened.
Did you put the cap in the wrong location? If you confirmed C11 as being correct location, the next thing you need to do is measure the caps with a digital multimeter. console5 might have sent you the wrong caps by mistake. Reason I say this is because this fix does not garble the graphics. I've done it myself with a 470nF on the Junior at C11 and it works.
tzakiel
Posts: 165
Joined: Tue Jan 14, 2014 8:03 pm

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by tzakiel »

Yeah I’m sure I screwed up somehow, it’s just odd that the 2 new caps had the same issue but the old one worked right away. I’ll check with a multimeter
User avatar
James-F
Posts: 87
Joined: Fri Mar 23, 2018 11:01 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by James-F »

After several months with this mod I decided to go back to stock because there are too many AAA that suffer from the C11 mod.
Starfox, Yoshis Island, Kirby Super Star, Street Fighter, and more.
Basically, the mod does work and fixes the ghosting but also introduces much slower transition time to the analog part of the PPU that responsible for fade-in/out effect, thus resulting in an uneven fade effect or part of the screen will be displayed wrong.
So it is a compromise, completely ghosting free image with wrong transition effects, or correct transition effects with some ghosting in bright scenes.

The SD2SNES now has switchable fix for this but some games run the fade-in/out code in HDMA so the SD2SNES can't affect it and the game will be displayed wrongly with the C11 mod, and obviously real cartridge games transitional effects will always be wrongly displayed (the ones that use this C11 related part of the PPU).
rama
Posts: 1373
Joined: Wed Mar 08, 2017 3:15 pm

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by rama »

I think I remember testing 100nF and this being fine for transitions in affected games.
Maybe 150nF is also fine?
This may be a good compromise.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by FBX »

rama wrote:I think I remember testing 100nF and this being fine for transitions in affected games.
Maybe 150nF is also fine?
This may be a good compromise.
The level of glitch-versus-ghosting will be effected not matter what rating you use. Some time back when this was a hot topic, I posted several pics of what happens when you use lower rated caps. the ghosting becomes detectable until you clear that 400nF range. Conversely, the transition glitch will still be slightly visible even with a stock cap.

But really this is all moot if you get a Super Nt, and especially once Analogue releases the RGB cable adapter for it.
User avatar
James-F
Posts: 87
Joined: Fri Mar 23, 2018 11:01 am

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by James-F »

The SuperNT is an emulator, a good one that is, but certainly not perfect.
https://github.com/SmokeMonsterPacks/Su ... eak/issues

This is still a very much relevant topic for those who use the real hardware 1chip, and the side effects of this ghosting fix are very much real and visible.
I think the 47nF value chosen by Nintendo is not trivial and they were aware of the slower response of the PPU with higher values.
We should no forget that the Super Nintendo was designed to be played on a CRT and nothing else.

I can live with barely visible ghosting on a flat bright colored backgrounds (when playing through the OSSC) but not with wrong transitional effects, personal preference of course.
On a Sony Trinitron 29" with Scart-RGB I can't see ghosting with stock 47nF capacitor, I think mainly because of scanlines and the less sharp nature of CRT overall.
Last edited by James-F on Fri Aug 03, 2018 2:58 pm, edited 1 time in total.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: Video probs w/ SNES RGB (Vertical Bands, etc)

Post by FBX »

James-F wrote:The SuperNT is an emulator, a good one that is, but certainly not perfect.
https://github.com/SmokeMonsterPacks/Su ... eak/issues
That's for the jailbreak to add in ROM loading from SD cards. As a machine to load real carts, it works exactly and perfectly like the original SNES. Do some Googling on it and you'll see it is WELL regarded as the BEST non-original SNES ever made.
Post Reply