shmups.system11.org

Yes, the same filter can be optionally used in 480p. The Game Boy Player Start-up Disc is an infamous example of it.

So it is only applied on certain games? I did not test a lot of games, but in all the ones I checked both 480i and 480p, 480p invariably looked less sharp. I don't really notice traces of a 480i blurring filter, except perhaps somewhat on text fonts.

You're likely looking at the analog bandwidth limit. The horizontal resolution is effectively halved.

I find the Wii seems to look sharper in 480p than 480i on CRTs. 480i Pikmin for example looks like quite blurry even on a very low use Trinitron.

I'd also be very interested to see which games use a softening filter in 480p.

My comparisons were on 900- and 1000-line monitors, so perhaps those are just exceptionally good at handling 480i relative to consumer CRTs. I always thought that 480p was perhaps not as well calibrated on them, but it being due to a softening filter on 480p maybe would make more sense.

Sure - they do look sharper on my BVM but I find that the 480i flicker is amplified to the point where I find it a little distracting.

Just found an interesting read regarding the de-flicker/blur-filter: https://ppltoast.wordpress.com/2018/05/ ... py-filter/
It's mainly about Dolphin, but helps shed som light on the topic.

"Swiss, when set to force 480p on games, also disables the deflicker filter"
"FIX94’s Nintendont on Wii also disables the filter when forcing video modes"

So it seems that forcing the sharpest possible picture out of Gamecube games is possible when running on either a Gamecube or Wii using Swiss and Nintendont respectively. However, does anyone know to what extent Wii games running in 480p are affected by this "blur"-filter?

I’m trying to reconcile the above statement (that 4:2:2 is a strict hardware limitation) with the GBI high-fidelity variant’s ability to output 4:4:4, as stated on the GBI specs in the gc-forever thread:



I wanted to see this for myself so I inspected the “color bleed” test pattern in three scenarios (all 240p mode on PAL console with RGB scart cable into OSSC):

1. GameCube 240p test suite
2. GBA 160p test suite running in GBI speed-running version
3. GBA 160p test suite running in GBI high-fidelity version

Using the white (pure luma) pixels to ensure I had optimal sampling for all three, I visually confirmed that #1 and #2 showed the expected 4:2:2 chroma subsampling artifacts, while #3 appeared artifact free, no chroma blurring whatsoever, a convincing 4:4:4 signal.

Is this 4:4:4 possible only because it’s 240p? Or only because the game boy player hardware’s involved? Or both? Or some other unrelated trick/hack?

Anyone know how the ZisWorks monitors handle 480p? They support integer scaling. Would that be an option?

No, the GBI claim is just a creative reinterpretation of the horizontal resolution. On the hardware side, the Cube output is still 4:2:2, so there is only a single color sent for every two pixels. But if you ensure that each logical pixel is physically sent twice (with the correct alignment), you can also interpret the video signal as having half the physical horizontal resolution, but full color information (4:4:4) for each logical pixel.

How the video encoder at the output implements 4:2:2 to 4:4:4 interpolation (if needed) can make things a bit more complicated again, but there are some indications that the Nintendo encoders use chroma hold (same chroma for both pixels in a 4:2:2 group) instead of the linear interpolation that is described in the Gamecube patents.

This is interesting info. GC games certainly look clean and sharp through RGB running on Nintendont forcing PAL60. I haven't really looked carefully at Wii games recently, but I'd imagine they probably would have the same filter since the GC and Wii were so similar. But that means a deflicker could also be implemented through Wii game loaders like USBLoader GX or Cfg USB Loader, which can also force video modes. Maybe they already do?

Fascinating, thanks for the explanation. So in other words, the pixel clock is doubled and duplicate values are used for every pair of horizontal pixels, such that the lossy encoding to 4:2:2 only ends up discarding the duplicated information.

I suppose the reason this trick could never work for 480p rendered content is that the console can’t double the higher (27MHz?) pixel clock for 480p? i.e. homebrew software alone could never achieve 4:4:4 480p from a GameCube, it’d need to be overclocked?

The pixel clock isn't doubled, or it wouldn't work with the internal video encoder.

It could be achieved through a video adapter based on the Hi Speed Port (81 MB/s). Patching games for it would be feasible.

Intriguing. This immediately leads me to wonder what other features such an adaptor could have. Compared to the market for GCvideo products, I’d guess far fewer people care or even know about the 4:2:2 limitation, such that 4:4:4 capability by itself likely wouldn’t be enough motivation for development? Who knows...

Adaptive Sync would be a possibility. However, it must be stressed that it couldn't work in any capacity without software support.

Doesn't the Wii apply a blur filter to 480p? Much like the blur for N64 on 240p.

I'll explain again.

Early video encoders in the Wii are analog bandwidth limited to SDTV. This effectively halves the horizontal resolution for EDTV.

Ironically, WiiDual can only be installed on revisions with the later video encoder.

Really noticed this the other night with my day 1 system.
The screen even flickers horizontal bars like the hdd I have installed is interfering.

And the vertical bars when going thru menus...

So that's probably why people often felt that later Wiis had better video output than early ones, because they were probably only comparing 480p through component. 480i/240p should look the same across all models, I assume. Or are there differences also on SDTV resolutions between early and late models?

Are official Wii component cables worth the premium over third party ones?

You might be able to find a quality third-party cable, but stay far away from that Tomee stuff; it's hot garbage. I don't think I've seen an instance where a first-party cable wasn't worth the premium, unless you're going SCART, in which case you can get way-better-quality cables from sources like Retro-Access and Retro Gaming Cables.

Has anyone ever done a direct comparison video?

Apologies if I'm missing something - I've looked back through the posts and can't quite work out if we're at the point yet where it's worth installing WiiDual. Have the Chroma and "whites" issues been solved yet? I'm a little cautious of getting my US & UK Wiis modded just in case there's still problems to be overcome that might eventually require a new mod revision.

I'll be using component on an NEC XM29 & want to be sure the image is significantly better than stock - has anyone tried this using component on a reasonably good CRT & is it a big improvement over the stock output?

Lastly, is this the best thread to stay current? I see no one's been posting here for a while..

Discussion continued in the main GCVideo thread.

Thanks very much :)

I know where the black crush on the analog output comes from and the chroma shift for certain horizontal image issues is also fixed in my development tree.


I can't give any guarantees, but at least for the digital output part it is extremely likely that all issues can be fixed by flashing a newer version of the firmware. I'm hesitant to give the same guarantees for the analog output because I did not design the boards and I (still...) don't have one hooked up to one of my systems to actually try it.

Thanks for that, very much appreciated. As I'm going down the analogue route I'll hold off for now & keep an eye on the other thread.

Looking at having the WiiDual installed just for the convenience of hdmi output. However, I was wondering if there is a noticeable difference in video quality over the component cables.

Personally I felt that there was a noticeable difference in the HDMI output over the component. Component seemed softer to me.

In what kind of monitor did you make this comparison?