Accurate & High-quality RGB SNES Captures: The Struggle

The place for all discussion on gaming hardware
User avatar
Rulumi
Posts: 114
Joined: Thu Jul 15, 2021 12:40 am

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by Rulumi »

Component is taking the RGB signal and making it YPbPr, so the ghosting may still be there.
User avatar
NewSchoolBoxer
Posts: 369
Joined: Fri Jun 21, 2019 2:53 pm
Location: Atlanta, GA

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by NewSchoolBoxer »

DragonQuarter wrote: I appreciate all butting in in this thread! So I recorded that footage with FFV1 codec with framerate set to 60 in OBS. I'm assuming that issue is due to the SNES's weird 60.09Hz / 60.10Hz framerate? But I'm not sure how to fix that. Appreciate any feedback there, especially since my end goal is capturing SNES longplays, which = long video files.
For what it's worth, recording and streaming Composite and S-Video in OBS works flawlessly with GV-USB2 capture card. I realize you're digging down the RGB rabbit hole like me but I think quality capture of the video formats that most people played in is important too. Also, sorry I was hard on you about the SCART -> Datapath device. Stuff you can't sell 1000x a year of for bulk supply prices tends to be expensive.

NES and SNES have same off spec framerate. That is so close to 60 Hz that I'm not sure any frame skip or jitter is noticeable. Is why video footage of CRT shows scrolling though since we record at 60 Hz. I'd be surprised if OSSC / Framemeister or whatever don't "fix" to 60.00 Hz since HDMI is very specific in what is allowed, being a digital format.
DragonQuarter wrote: I've ended up buying the HD Retrovision Component Cables to see if that would work out better for me. That way, I could connect my OSSC to the Datapath with HDMI to DVI...
HD Retrovision Component Cable has a good reputation and it almost certainly fixes sync jitter from reading over their website. You add some noise from transcoding RGB to YPbPr and low quality RGB will yield low quality YPbPr. I am very curious of this approach though to confirm how much better it looks than S-Video and how little worse it is to RGB.

I don't know HDMI upscaler settings but try to confirm you're outputting HDMI at RGB or 4:4:4 YCbCr. The 4:4:4 means 100% color information (chroma) versus subsampling to save bandwidth. YouTube is typical internet 4:2:0, which will wipe out scanlines without proper integer scaling. May as well start at 5X.
DragonQuarter wrote: Unfortunately, CSync SCART cables are sold out everywhere, so I can't try getting a replacement SCART cable. There is a Sync on Luma in stock at Insurrection Industries, but as I understand it, that wouldn't be the most versatile cable for me, right? Like, I know it'd work with the OSSC, but if I wanted to connect to PVM as well...
PVMs 100% accept every kind of sync you throw at them at both TTL and 75 ohm levels. I think I saw one datasheet of a very obscure and early model that needed csync. Analog RGB capture cards, however, need csync or sync on green and at TTL level. NTSC SNES and I think every console with native csync output TTL. It's higher voltage than "75 ohm" csync that SCART televisions want. I wrote above I needed to test E1S but this website confirms sync requirements. Actually, your low quality RGB captured video would be related to not sending TTL csync to the E1S. SCART cables meant for NTSC usually add 330-470 ish ohm resistor to drop TTL to 75 ohm.

If you're set on transcoding to and capturing in HDMI though, then sync over luma cable that PAL typically uses is perfectly fine and $10 more expensive csync cable will not give better video (unless cable is garbage). Costs $10 more because sync cleaning chips cost $5.
DragonQuarter wrote: - I have no circuitry knowledge, so if I were to open up my SNES Jr., I don't know what I'd need to look for to figure out if the C11 cap is faulty or something
Not to brag but I have circuitry knowledge. Fortunately, it's called C11 because C11 is literally printed on the PCB next to it. I found a pic of it in FBX's post half way down on page 6: viewtopic.php?f=6&t=61072&start=150

Is a surface mount (SMD) capacitor so would be harder to tell if is bad than the capacitors you see raised on metal pins. New capacitor isn't likely to die and I doubt they used the wrong value. But basically, if capacitor has yellow or clear fluid that dried around it (from leaking), has an excessive amount of dried white-silver metal around it (used to solder it in place), is loose or is swollen or otherwise looks "bad", then it probably is. Pro move would be to use a $60-120 ESR meter to check to test but SNES has so few capacitors compared to, say, a PVM, it's fine to proactively replace them all as a preventive measure.

C11 Change Explanation:
1Chips use a capacitor with 47 nF but, I guess to save money, it has a 20% tolerance rating, meaning the actual value is +/- 20% from 47 nF. As the console and the capacitors age, they don't work as well and near end of an electrical component's life, the actual value can go well outside of the tolerance. If it drops too low, you get ghosting. People replace it with a value 10x higher and with 10% tolerance as a safety measure. Better to be too high than too low. I would have suggested not to take the console out of spec and just use a tighter tolerance 47 nF, but I wasn't there.

Edit: I read rest of the C11 thread and the excessively high 470 nF capacitor causes black bars at the top of some popular games like Street Fighter II and Aladdin. Coming from ikari_01 who coded a ROM patch to fix, black bars are due to the game's code switching the brightness value midscreen, which is too soon for the DAC to adjust. I believe increasing the RC time constant by the new C11 is the reason the DAC can't adjust in time. Interestingly, people decided to use a 220 nF capacitor instead to take partially fixed versions of ghosting and introduced black bars. Related fix discussed in thread is changing R3 to 1.74k ohms but probably screws up Composite and S-Video. Ask what C11 they used and if they changed R3.

A near-dead voltage regulator can cause video issues too. Is a < $1 7805 that is beginner soldering task to replace. Could get that swapped out with superior < $1 78S05. I see people also tend to replace the related fuse and capacitor. By extension, a bad power supply feeding the SNES can mess with video too.
User avatar
Rulumi
Posts: 114
Joined: Thu Jul 15, 2021 12:40 am

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by Rulumi »

NewSchoolBoxer wrote:
DragonQuarter wrote: NES and SNES have same off spec framerate. That is so close to 60 Hz that I'm not sure any frame skip or jitter is noticeable. Is why video footage of CRT shows scrolling though since we record at 60 Hz. I'd be surprised if OSSC / Framemeister or whatever don't "fix" to 60.00 Hz since HDMI is very specific in what is allowed, being a digital format.
The OSSC outputs 60.10 or so with the SNES as the signal isn't buffered or anything like that, is up to the device to accept that and the jittery signal.
The Framemeister should be even 60 I think, although it does weird things for not 60 FPS games.
User avatar
DragonQuarter
Posts: 31
Joined: Sun Jul 03, 2016 9:30 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by DragonQuarter »

So I finally got some HD Retrovision component cables and a Component to DSUB adapter attached to a DSUB to DVI-I adapter going into my Datapath Vision E1s, and sadly I'm still getting the ghosting issue. Sounds like I need to hit up RetroFixes to figure out what's going on.

Super Mario World FFV1 clip captured using Vision window, upscaled 4X via OBS: https://1drv.ms/v/s!ApUHztm-6jEsiJEvQeG ... A?e=fmQfZx

Having fiddled with the Datapath Vision E1s and Vision software for hours upon hours, weeks upon weeks, I'm not entirely sure the Datapath is a good solution. The phase settings alone make it an enormous pain and it constantly feels like none of the phase settings are good, and none of the tutorials really help indicate what a "good" phase is. I guess I'll still try OSSC to Datapath using an HDMI to DVI-I dongle to see if I get better mileage, but my patience is kind of wearing thin with this. Direct lossless 240p capture is ideal for me for conserving space, but it's been such an awful user experience that I'm starting to think it's not worth it.
paulb_nl
Posts: 340
Joined: Sat Feb 20, 2016 5:05 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by paulb_nl »

DragonQuarter wrote:So I finally got some HD Retrovision component cables and a Component to DSUB adapter attached to a DSUB to DVI-I adapter going into my Datapath Vision E1s, and sadly I'm still getting the ghosting issue. Sounds like I need to hit up RetroFixes to figure out what's going on.
If you take a screenshot of Street Fighter in game then the shadow on the top lines will indicate if the capacitor has been replaced and with which one.

Judging by the ghosting on your Mario World capture I would guess that it hasn't been replaced or with a low value.
User avatar
DragonQuarter
Posts: 31
Joined: Sun Jul 03, 2016 9:30 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by DragonQuarter »

paulb_nl wrote:
DragonQuarter wrote:So I finally got some HD Retrovision component cables and a Component to DSUB adapter attached to a DSUB to DVI-I adapter going into my Datapath Vision E1s, and sadly I'm still getting the ghosting issue. Sounds like I need to hit up RetroFixes to figure out what's going on.
If you take a screenshot of Street Fighter in game then the shadow on the top lines will indicate if the capacitor has been replaced and with which one.

Judging by the ghosting on your Mario World capture I would guess that it hasn't been replaced or with a low value.
Thanks for this. I've taken a screenshot from Street Fighter II. I noticed the top row of pixels here was pretty flickery, but I take it it's from the phase/noise and my non-luck with not getting a single "good" phase setting.

Image
paulb_nl
Posts: 340
Joined: Sat Feb 20, 2016 5:05 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by paulb_nl »

I can't view your screenshot. Access denied.
User avatar
DragonQuarter
Posts: 31
Joined: Sun Jul 03, 2016 9:30 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by DragonQuarter »

paulb_nl wrote:I can't view your screenshot. Access denied.
Weird, here it is through imgur:

Image
paulb_nl
Posts: 340
Joined: Sat Feb 20, 2016 5:05 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by paulb_nl »

DragonQuarter wrote: Weird, here it is through imgur:
This looks like the C11 capacitor has not been changed. I am assuming this was not captured with an SD2SNES with 1-CHIP transient fix enabled? That could avoid the shadow lines.

Look here to see the result of different value capacitors with Street Fighter courtesy of FirebrandX: viewtopic.php?p=1287152#p1287152

Scroll down a bit for 220nF+.
User avatar
DragonQuarter
Posts: 31
Joined: Sun Jul 03, 2016 9:30 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by DragonQuarter »

While RetroFixes didn't see the ghosting I mentioned in Super Mario World (may have to circle it for them), they did show me a Mega Man X screenshot in the Chill Penguin stage where ghosting would be apparent, and I didn't see that happening with mine.

Note that I am using a SD2SNES on all of this.

I have a bigger issue I'm running into, however, involving the HD Retrovision SNES component cables when used through OSSC into the Datapath Vision E1s.

SNES Jr. (RGB-modded by RetroFixes) CSync SCART (2016 Retro_GamingCables eBay) to OSSC to HDMI to DVI-I Adapter into Datapath E1s
I get some horizontal pixel flickering here and there, which I am thinking is due to the SCART cable itself. It's too bad you can't get CSync SCART cables anywhere right now; all sold out!

SNES Jr. (RGB-modded by RetroFixes) HD Retrovision Component to OSSC to HDMI to HDMI Audio to RCA splitter to DVI-I Adapter into Datapath E1s
I get red lines flickering in and out with this setup, and I'm thinking it may have to do with the HDMI audio splitter.

SNES Jr. (RGB-modded by RetroFixes) HD Retrovision Component to OSSC to HDMI to DVI-I Adapter into Datapath E1s
The picture looks great, but in 2X mode, 4 columns of pixels are getting chopped off by what I presume is the OSSC when using the FirebrandX profile. So instead of 512x448, I'm getting 508x448.

Image

When I adjusted the H. Backporch from FirebrandX's 38 to 36, I got 512 pixels horizontally again, however the first two columns appear to be brighter than the rest of the columns to the right of it. I'm getting a feeling that the OSSC and HD Retrovision Component Cables don't play nicely 100%. Or, my advanced timings are off, but I have nothing to go by other than FirebrandX's 256 profile that is SCART-minded. Google searches have brought up nothing regarding ideal timings for HD Retrovision SNES component cables and OSSC.

Image

Zoomed in, note the first two columns of pixels.

Image
User avatar
kitty666cats
Posts: 1270
Joined: Tue Nov 05, 2019 2:03 am
Location: Massachusetts, USA

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by kitty666cats »

@retrorgb do you think a SCART2DVI would help out with this? I know that the SNES is finicky with its weird vertical frequency and many variants, but aside from that am not super familiar with SNES's internal hardware...

Insurrection Industries unfortunately no longer has any left, but I believe there's still a few stores online who have some(?). The PCB has a THS7374 on it - but I don't believe it is whatsoever for amplification, but rather strictly the LPF. But then again, if you're already having issues even with an OSSC (which obviously has a LPF on 2/3 of the inputs) then maybe a SCART2DVI would not be of any help...
User avatar
DragonQuarter
Posts: 31
Joined: Sun Jul 03, 2016 9:30 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by DragonQuarter »

kitty666cats wrote:@retrorgb do you think a SCART2DVI would help out with this? I know that the SNES is finicky with its weird vertical frequency and many variants, but aside from that am not super familiar with SNES's internal hardware...

Insurrection Industries unfortunately no longer has any left, but I believe there's still a few stores online who have some(?). The PCB has a THS7374 on it - but I don't believe it is whatsoever for amplification, but rather strictly the LPF. But then again, if you're already having issues even with an OSSC (which obviously has a LPF on 2/3 of the inputs) then maybe a SCART2DVI would not be of any help...
I bought one of those "Voultar Double Penetration SD Datapath E1/E1S" which is like it's own version of a SCART2DVI but I've struggled to get any good phase out of it. Part of it is probably due to my old SCART cable, but no websites have CSync SCART cables in stock, so I'm a little SOL.

So far the best picture I've been getting has been through SCART to OSSC to the Datapath, but even then I'm struggling with phase, and the SCART cable is causing artifacts to appear, so if I can get a new SCART cable I think I'd be in a better spot, but I'm still struggling with figuring out phase. RetroRGB's guide mentions to basically check each value until there are no artifacts, but I feel like there's ALWAYS a fuzzy/noisiness particularly with solid color backgrounds in the picture. The R3 Discord mentioned to me that some bit of noise is unavoidable, which makes me question why I'm bothering with all of these dongles if the Analogue Super Nt would provide the best possible picture out of all of these methods. And even when the 240p Test Suite checkerboard pattern looks OK in the Vision window, it doesn't necessarily mean the correct phase has been dialed in.

It's a bit frustrating that phase values change depending on every single piece of equipment you use. It requires so much effort just to try to get the best possible picture and it's really making me feel like getting a clean image is next to impossible with my RGB modded SNES Jr.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by FBX »

DragonQuarter wrote:While RetroFixes didn't see the ghosting I mentioned in Super Mario World (may have to circle it for them), they did show me a Mega Man X screenshot in the Chill Penguin stage where ghosting would be apparent, and I didn't see that happening with mine.
That's likely from the screenshot testing work I was doing over the c11 fix. In that example, I negative-image the colors in a paint program, and the ghosting becomes blatantly obvious (near the health meter). You can make it out without flipping it, but it makes it easier to demonstrate the artifacts when you do.

When I adjusted the H. Backporch from FirebrandX's 38 to 36, I got 512 pixels horizontally again, however the first two columns appear to be brighter than the rest of the columns to the right of it. I'm getting a feeling that the OSSC and HD Retrovision Component Cables don't play nicely 100%. Or, my advanced timings are off, but I have nothing to go by other than FirebrandX's 256 profile that is SCART-minded. Google searches have brought up nothing regarding ideal timings for HD Retrovision SNES component cables and OSSC.
There's no such thing as 'ideal timings' since the only variables that change are ones that have to be set on a case-by-case basis anyway. That said, I get near perfect results from Voultar's RGB bypass mod in the SNES Junior with HD-Retrovision cables:

https://i.imgur.com/0YX65OC.png

Image
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by FBX »

DragonQuarter wrote:
So far the best picture I've been getting has been through SCART to OSSC to the Datapath, but even then I'm struggling with phase, and the SCART cable is causing artifacts to appear, so if I can get a new SCART cable I think I'd be in a better spot, but I'm still struggling with figuring out phase. RetroRGB's guide mentions to basically check each value until there are no artifacts, but I feel like there's ALWAYS a fuzzy/noisiness particularly with solid color backgrounds in the picture. The R3 Discord mentioned to me that some bit of noise is unavoidable, which makes me question why I'm bothering with all of these dongles if the Analogue Super Nt would provide the best possible picture out of all of these methods.
If all you're after is digital SNES screenshots, then yeah, why not use emulation? You could just use Higan and screenshot from that, or get a Super Nt and capture from that. keep in mind the Super Nt uses the same color ramp (after it was reported being off balance in early firmwares) as Higan, so they should look effectively identical.
paulb_nl
Posts: 340
Joined: Sat Feb 20, 2016 5:05 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by paulb_nl »

DragonQuarter wrote:While RetroFixes didn't see the ghosting I mentioned in Super Mario World (may have to circle it for them), they did show me a Mega Man X screenshot in the Chill Penguin stage where ghosting would be apparent, and I didn't see that happening with mine.
It's easiest to see when darkening the gamma. Here are captures of the Striped Sprite test of the 240p test suite with the gamma extremely darkened (Set to 7.99 in Paint.net->Adjustment->Levels):

Original C11:
Image

Replaced C11 with 460nF:
Image
Note that I am using a SD2SNES on all of this.
Is the 1-CHIP transient fix enabled? If it is disabled and C11 has been replaced then you should see similar results as FirebrandX's Street Fighter captures .
SNES Jr. (RGB-modded by RetroFixes) CSync SCART (2016 Retro_GamingCables eBay) to OSSC to HDMI to DVI-I Adapter into Datapath E1s
I get some horizontal pixel flickering here and there, which I am thinking is due to the SCART cable itself. It's too bad you can't get CSync SCART cables anywhere right now; all sold out!
How are you capturing those OSSC captures with OSD? I see some strange compression on the OSD which should be perfect blue & white & yellow if you are capturing 4:4:4 (or 8-8-8 as the Vision software calls it) but there are shadows and the OSD blue color is 239 while it should be 255.
User avatar
DragonQuarter
Posts: 31
Joined: Sun Jul 03, 2016 9:30 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by DragonQuarter »

FBX wrote: That's likely from the screenshot testing work I was doing over the c11 fix. In that example, I negative-image the colors in a paint program, and the ghosting becomes blatantly obvious (near the health meter). You can make it out without flipping it, but it makes it easier to demonstrate the artifacts when you do.
Here's my example of Mega Man X. Appreciate you chiming in btw, thanks for making all those videos and OSSC profiles!

Capture: Mega Man X via SDSNES (1-CHIP transient fix disabled); SNES Jr. to CSync SCART to OSSC HDMI to DVI-I Adapter to Datapath E1s
2X Scale with 256 FirebrandX Profile; OSSC Sampling Phase 247; Cropped to 512x448 in OBS; RGBEasy Default Settings


Image

Negative Adjustment in Photoshop:
Image
FBX wrote: There's no such thing as 'ideal timings' since the only variables that change are ones that have to be set on a case-by-case basis anyway. That said, I get near perfect results from Voultar's RGB bypass mod in the SNES Junior with HD-Retrovision cables:
I'm starting to wonder whether I should've gone with sending my SNES Jr. to Voultar instead of RetroFixes. Are you keeping the same Advanced Timings on the OSSC as your 256 profile? I have to shift either H synclen or H backporch to restore the 4 columns of pixels I'm losing when simply loading your OG SNES 256 profile.

Capture: Mega Man X via SDSNES (1-CHIP transient fix disabled); SNES Jr. to CSync SCART to OSSC HDMI to DVI-I Adapter to Datapath E1s
2X Scale with 256 FirebrandX Profile; OSSC Sampling Phase 247; Cropped to 512x448 in OBS; Using RGBEasy direct capture


Image

Capture: Mega Man X via SDSNES (1-CHIP transient fix disabled); SNES Jr. to HD Retrovision Component to OSSC HDMI to DVI-I Adapter to Datapath E1s
2X Scale with 256 FirebrandX Profile; OSSC Sampling Phase 247; Cropped to 512x448 in OBS; Using RGBEasy direct capture


Notice the 2 columns of pixels missing on either side:

Image

Here's the same settings above except with H backporch dropped from 38 to 36:

Image
FBX wrote: If all you're after is digital SNES screenshots, then yeah, why not use emulation? You could just use Higan and screenshot from that, or get a Super Nt and capture from that. keep in mind the Super Nt uses the same color ramp (after it was reported being off balance in early firmwares) as Higan, so they should look effectively identical.
I'm actually after longplay footage, particularly of RPGs like Breath of Fire I and II, capturing at 240p or 480p then nearest neighbor scaling to 4K or whatever other resolution I want (4K would go to YouTube). Phase makes this a pain as I feel I'd have to adjust for it every single time I boot things up. And there's still noise that you can't avoid. I guess the Super Nt with Zero Delay buffer really is the best way. I was hoping to go with original hardware to make it "authentic", but I guess as someone earlier mentioned, RGB modding and all these other dongles hardly make even analog capture truly "authentic".
paulb_nl wrote: Is the 1-CHIP transient fix enabled? If it is disabled and C11 has been replaced then you should see similar results as FirebrandX's Street Fighter captures .
1-CHIP transient fix has never been enabled, I wasn't even aware of this setting nor do I know what it does. Here's another capture of SFII with transient fix remaining disabled:

Capture: Street Fighter II via SDSNES (1-CHIP transient fix disabled); SNES Jr. to CSync SCART to OSSC HDMI to DVI-I Adapter to Datapath E1s
2X Scale with 256 FirebrandX Profile; OSSC Sampling Phase 247; Cropped to 512x448 in OBS; Using RGBEasy direct capture


Image
paulb_nl wrote: How are you capturing those OSSC captures with OSD? I see some strange compression on the OSD which should be perfect blue & white & yellow if you are capturing 4:4:4 (or 8-8-8 as the Vision software calls it) but there are shadows and the OSD blue color is 239 while it should be 255.
I was using Direct Capture in OBS and cropping to 512x448. With these screenshots I am now using the RGBEasy plug-in with default settings, but still at the 512x448 crop.

Capture: 240p Test Suite via SDSNES (1-CHIP transient fix disabled); SNES Jr. to CSync SCART to OSSC HDMI to DVI-I Adapter to Datapath E1s
2X Scale with 256 FirebrandX Profile; OSSC Sampling Phase 247; Cropped to 512x448 in OBS; Using RGBEasy direct capture


Image
paulb_nl
Posts: 340
Joined: Sat Feb 20, 2016 5:05 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by paulb_nl »

DragonQuarter wrote: 1-CHIP transient fix has never been enabled, I wasn't even aware of this setting nor do I know what it does. Here's another capture of SFII with transient fix remaining disabled:
Short summary: the transient fix tries to override a brightness register write to avoid the shadow lines after replacing C11.

Below is your 240p test suite Striped sprite test with the gamma darkened. It has massive ghosting similar to my SNES capture with stock C11. With that and your Street Fighter capture it is 100% sure that C11 in your SNES hasn't been replaced.

Image
User avatar
DragonQuarter
Posts: 31
Joined: Sun Jul 03, 2016 9:30 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by DragonQuarter »

Thanks, I'll share this with the RetroFixes person.

I noticed that, both with SCART and HD Retrovision Component cables, I'm seeing weird horizontal lines of pixels flicker at the top of the screen. See this video with Breath of Fire as an example. This is definitely not normal...

Video: https://www.dropbox.com/s/hewrbfg55suvh ... e.avi?dl=0

EDIT: I tested this on my 4KTV and I'm NOT getting those pixel artifacts...sounds like a Datapath Issue?
paulb_nl
Posts: 340
Joined: Sat Feb 20, 2016 5:05 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by paulb_nl »

DragonQuarter wrote: I noticed that, both with SCART and HD Retrovision Component cables, I'm seeing weird horizontal lines of pixels flicker at the top of the screen. See this video with Breath of Fire as an example. This is definitely not normal...

EDIT: I tested this on my 4KTV and I'm NOT getting those pixel artifacts...sounds like a Datapath Issue?
Those flickering pixels look like they are coming from your SNES. The flickering pixels in the 512x448 capture are 2x2 meaning they are double size SNES pixels. If it was caused by the Datapath then you would also see 1x1 pixel glitches.

Are you 100% sure you don't see these pixels on your TV? Did you connect the OSSC to your TV? Can you test original carts? Another SNES?

Do you get the same issue when you skip the OSSC and connect the SNES to the Datapath directly?
User avatar
DragonQuarter
Posts: 31
Joined: Sun Jul 03, 2016 9:30 pm

Re: Accurate & High-quality RGB SNES Captures: The Struggle

Post by DragonQuarter »

So in a crazy twist, I determined that the pixel artifacts in this particular "Spyre" dungeon of BoF only occurred on two of my own ROM rips on the SD2SNES. On my 3rd cart rip, this did not occur. This leads me to think it could be a finicky rom rip in conjunction with the SD2SNES. However, this artifacting does NOT occur on the Analogue Super Nt with any of my cart rips. And it doesn't happen on the original carts, either.

In any case, I've sent the SNES Jr. back to RetroFixes so they can look into adding a stronger C11 cap. Fingers crossed!
Post Reply