XRGB-mini Framemeister

[TheShadowRunner]

Therefore, as a starting point, I suggest setting it as follows: Output_Color: AUTO, Output_Range: Full

I'm not sure what you want to achieve with this.
Output_Color: AUTO will select RGB 99% of the time unless the Mini falls back in failsafe mode (YCbCr).
Output_Range: Full, only applies to RGB (at least it's supposed to).

FBX: while sharing your perfect geometry makes sense (since we all have 1080p displays), I think colors / brightness should be left to default in the profiles you share and for users to set themselves.

Because there are so many variables that affect the picture, backlight power-save, picture tone (warm, neutral, etc), the panel themselves, etc.. that the only person who'd get your intended picture would be a person using the exact same TV model and TV settings as you (and even same room lighting if ambient light sensing is in use as well!).

Brakels: I believe "LOAD" will just (re)load whatever profile was last used, ie current profile while "LOAD_SELECT" just lets you select which one you want to use.

[Brakels]

Brakels: I believe "LOAD" will just (re)load whatever profile was last used, ie current profile while "LOAD_SELECT" just lets you select which one you want to use.

That was not my experience. LOAD presented me with a choice of files to load from, and I definitely was able to load first the SNES 4x and then the SNES 5x profiles with it.

[TheShadowRunner]

Ok, after testing, the difference is simple:
LOAD_SELECT lets you chose to which input you want to apply the profile to:
either V/S (composite or svid + sound) / D (component) / RGB / HDMI
LOAD will apply the selected profile to all inputs (LOAD_ALL OK/Cancel).

[RGB32E]

FBX: while sharing your perfect geometry makes sense (since we all have 1080p displays), I think colors / brightness should be left to default in the profiles you share and for users to set themselves.

Because there are so many variables that affect the picture, backlight power-save, picture tone (warm, neutral, etc), the panel themselves, etc.. that the only person who'd get your intended picture would be a person using the exact same TV model and TV settings as you (and even same room lighting if ambient light sensing is in use as well!).

I think you hit the nail on the head there! Aside from the previously known Zoom 90 setting to get square pixels from most 240p sources, most settings will vary by console/revision and cabling. The most useful thing about the profiles is being able to backup and restore settings based off of your own particular equipment chain. I'd find it much more useful to include HDMI screen caps so that the intended look of a particular set of settings can be conveyed independent of a particular equipment chain and improve the S/N ratio of posts!

[Bobster]

I signed up just to thank you, FBX, for all the effort you've put into these profiles.

I was in the middle of playing Star Ocean: Till the End of Time when 2.0 came out. The difference between my settings before, and your settings was staggering. I'm glad there are people out there who know more about this than I do .

Anyways, I'm surprised you're getting so much flak about attempting to set up the colors correctly. It makes sense to me. If you match the color settings between direct feed and the feed from the XRGB Mini, the only variable at that point is the color settings of the TV itself. I'm no pro so having someone else do that calibrating for me sounds excellent. At that point, it's up to each person to figure it out from there.

In your notes in the profile package you do recommend the LOAD_SELECT option. Just straight LOAD worked better for me. It actually keeps all your settings like Zoom. Not sure if that's the intention, but maybe the notes need to be updated?

But yeah, thanks again for your hard work/obsession.

[Smashbro29]

FBX: while sharing your perfect geometry makes sense (since we all have 1080p displays), I think colors / brightness should be left to default in the profiles you share and for users to set themselves.

Because there are so many variables that affect the picture, backlight power-save, picture tone (warm, neutral, etc), the panel themselves, etc.. that the only person who'd get your intended picture would be a person using the exact same TV model and TV settings as you (and even same room lighting if ambient light sensing is in use as well!).

I think you hit the nail on the head there! Aside from the previously known Zoom 90 setting to get square pixels from most 240p sources, most settings will vary by console/revision and cabling. The most useful thing about the profiles is being able to backup and restore settings based off of your own particular equipment chain. I'd find it much more useful to include HDMI screen caps so that the intended look of a particular set of settings can be conveyed independent of a particular equipment chain and improve the S/N ratio of posts!

So more like the ones I shared the other day of direct feed less pictures of TVs so we can improve the base signal and allow people to tweak for their own TVs/set ups?

[12345]

Therefore, as a starting point, I suggest setting it as follows: Output_Color: AUTO, Output_Range: Full

I'm not sure what you want to achieve with this.
Output_Color: AUTO will select RGB 99% of the time unless the Mini falls back in failsafe mode (YCbCr).
Output_Range: Full, only applies to RGB (at least it's supposed to).

No, it doesn't on D-terminal. AUTO is always detected as YCBCR on this input from the Framemeister, while on RGB it's claiming to output RGB. If those names are really true to the formats being processed to the TV, I don't know, the fact is though that the YCBCR setting is more accurate to the source, as previously described. And again, the separate Output_Range setting (full/limited) is broken as of know.

Because there are so many variables that affect the picture, backlight power-save, picture tone (warm, neutral, etc), the panel themselves, etc.. that the only person who'd get your intended picture would be a person using the exact same TV model and TV settings as you (and even same room lighting if ambient light sensing is in use as well!).

No, this is why people should calibrate their TVs to have a matching reference. If everybody adjusts their picture by their subjective, personal feelings we will never achieve a uniform setting. There are certain standards to be met, and with the limted settings we have on the Framemeister, it doesn't even have to be the most expensive equipment you have to use to accomplish this.

[Fudoh]

It makes sense to me. If you match the color settings between direct feed and the feed from the XRGB Mini, the only variable at that point is the color settings of the TV itself.

it makes sense if your TV is properly calibrated. If it isn't like 99% of all sets, you just mimic the wrong color settings you're seeing with the Mini. After all a video processor is about making corrections beyond what your TV is capable of.

[12345]

It makes sense to me. If you match the color settings between direct feed and the feed from the XRGB Mini, the only variable at that point is the color settings of the TV itself.

it makes sense if your TV is properly calibrated. If it isn't like 99% of all sets, you just mimic the wrong color settings you're seeing with the Mini. After all a video processor is about making corrections beyond what your TV is capable of.

In addition to that, I doubt every TV handles the A/D conversion equally. I'm almost 100% sure that direct feed component will look slightly different on every TV, no matter if it has been calibrated or not.

[CkRtech]

I don't really know that these are multi-user viable options for profiles due to the variables involved. The user's display is basically what ultimately defines the color. Zooming and scaling would seem to be much more of a universal setting to tweak that can be shared. Colors/brightness/gamma & A/D level add many cogs to the machine. If you take those settings for the FM + the user's display & its settings + possible cable and console variants, that is a lot of variables that will never be equal across all users - mostly due to the display panel variable.

The point of my new calibration settings is that what comes out of the Framemeister is now the same as what goes into it (as far as the official Nintendo component cables are concerned) in terms of brightness, saturation, etc. Yes, everyone's TV is different. My calibrations don't worry about that. Users can make adjustments on the TV set to their liking the same way they would if the Gamecube had been hooked directly to it. This was a needed job as people have often mentioned the default Framemeister settings are way too saturated for GC component input.

Sorry I didn't follow up on this. It sounds like it is quite relevant to the discussion in the thread at the moment.

Perhaps component (D-terminal) input could possibly use some "ballpark balancing" due to the general saturation reports for component input on the FM (I don't use that input, myself). However, I still believe that color settings shouldn't really be tweaked in these profiles due to the display variables. Also note that eyeballing the Gamecube output via component on a TV's native component vs the xrgb's HDMI output is most likely apples to oranges, in my opinion.

There are still the variables of how the current display handles component input as well as what the TV's display settings are for color temp/brightness/contrast on that component input vs the same settings on the HDMI input the xrgb is plugged into. Heck - even a TV that has matching settings with numbers and terms matching (i.e. "Vivid, Brightness = 50, Contrast = 80, etc") between two different types of inputs (Component vs HDMI) might be completely different despite having the same unit of measurement between the two inputs. Is 50 contrast on Component the same as 50 contrast on HDMI? Maybe...but who knows?

That aside, the save option is a full snapshot, right? (I haven't used it yet) So if you tweak the settings or not, those settings are still saved to the profile. Either way - everyone will probably have to tweak the colors and re-save them on their xrgb regardless of if they are default or preferred.

[TheShadowRunner]

No, this is why people should calibrate their TVs to have a matching reference. If everybody adjusts their picture by their subjective, personal feelings we will never achieve a uniform setting. There are certain standards to be met, and with the limted settings we have on the Framemeister, it doesn't even have to be the most expensive equipment you have to use to accomplish this.

... you think all the dudes currently using FBX's profiles calibrated their sets to have some "matching reference"?
I must have skipped that part of the conversation..

[12345]

No, this is why people should calibrate their TVs to have a matching reference. If everybody adjusts their picture by their subjective, personal feelings we will never achieve a uniform setting. There are certain standards to be met, and with the limted settings we have on the Framemeister, it doesn't even have to be the most expensive equipment you have to use to accomplish this.

... you think all the dudes currently using FBX's profiles calibrated their sets to have some "matching reference"?
I must have skipped that part of the conversation..

In a forum like this? -yes
Since we are still having the discussion and it seems many people here don't get the point? -no

[Fudoh]

In a forum like this? -yes

I'm all about proper calibration when it comes to movie playback, but for videogames I don't see it. For example I really enjoy the live color function on the Sony sets.

[Oliversum]

So I got me a Super Famicom with 1CHIP-03 today for some bucks.
Before this I used a PAL SNES, modded with 50/60Hz (switch) and a modded original Gamecube RGB cable.
As I know I need new cables and such for the SFC, which should I buy? I heard the 1CHIP-03 is lacking CSYNC?
Is retrogamingcables.co.uk a good place to buy these?

[FBX]

If I could explain what seems to be some confusion about how I calibrated the colors on the Gamecube:

When BOTH the inputs on the TV are set to the EXACT same color values, it removes the TV from the equation.

Think about it: Whether the TV is set to 25% saturation or 75% saturation, as long as both inputs are the SAME setting, it becomes possible to match the Framemeister to the direct feed image. Imagine if you will two glasses of water. One glass is 3/4 full and the other is 1/4 full. You pour a teaspoon of dye in each. Obviously each glass with be a different shade of the dye due to difference in dilution. Now imagine both glasses being 3/4 full. You pour in that teaspoon of dye, and now both glasses look exactly the same shade. That's the same idea with both inputs on the TV being set to the same levels. It makes it to where the only difference is between the Framemeister and the Gamecube, thus making it possible to calibrate the Framemeister to closely match what the Gamecube outputs directly.

I hope I'm making sense here because it's perfectly logical when you really think about. Color accuracy can be universal, and it effectively makes any variance the user wants something they can change on the TV itself.

[12345]

In a forum like this? -yes

I'm all about proper calibration when it comes to movie playback, but for videogames I don't see it. For example I really enjoy the live color function on the Sony sets.

Colors, apart from clipping, are probably the least important part and I can understand if somebody prefers a more vivid look, for whatever reason. It's not even necessary to talk about colors if people don't even want to set brightness or contrast right, not speaking of greyscale.
In the end the resulting image is not only a matter of the individual TV everybody has. If your source (Framemeister) is set incorrectly, you won't get back any detail you've lost before, no matter how you much you tweak your TV-internal settings.

@FBX
IF everybody had the same TV, then I'd say your theory is correct. But as I said before, we can't be sure if every set handles the A/D conversion from a direct component-in the same way. How can we be sure that a TV doesn't make a similar weird conversion the Framemeister does with the D-terminal?
So as long as you are comparing analog (component to TV) vs digital (Framemeister to TV) your comparison is lacking accuracy. The only way I see this working would be through a WII2HDMI dongle instead, for example, making the conversion directly at the beginning of your signal chain. At least we are getting comparable results then...

[FBX]

@FBX
IF everybody had the same TV, then I'd say your theory is correct. But as I said before, we can't be sure if every set handles the A/D conversion from a direct component-in the same way. How can we be sure that a TV doesn't make a similar weird conversion the Framemeister does with the D-terminal?
So as long as you are comparing analog (component to TV) vs digital (Framemeister to TV) your comparison is lacking accuracy. The only way I see this working would be through a WII2HDMI dongle instead, for example, making the conversion directly at the beginning of your signal chain. At least we are getting comparable results then...

Ah I see what you're getting at now. That my Sony display might not be 'perfectly' handling A/D from the direct feed. I can agree to that being a logical variable.

Does the WII2HDMI dongle convert the component signal into HDMI and, if so, how do we know that's properly calibrated?

[12345]

@FBX
IF everybody had the same TV, then I'd say your theory is correct. But as I said before, we can't be sure if every set handles the A/D conversion from a direct component-in the same way. How can we be sure that a TV doesn't make a similar weird conversion the Framemeister does with the D-terminal?
So as long as you are comparing analog (component to TV) vs digital (Framemeister to TV) your comparison is lacking accuracy. The only way I see this working would be through a WII2HDMI dongle instead, for example, making the conversion directly at the beginning of your signal chain. At least we are getting comparable results then...

Ah I see what you're getting at now. That my Sony display might not be 'perfectly' handling A/D from the direct feed. I can agree to that being a logical variable.

Does the WII2HDMI dongle convert the component signal into HDMI and, if so, how do we know that's properly calibrated?

As far as I've heard it's not properly calibrated, but that's not the point. If we are feeding both, the TV and the Framemeister, with the same source (the digital signal that comes e.g. from the WII2DMI) we can expect that our results will be at least comparable, even if they are not equal. Then, theoretically speaking, you could compare the different outputs from your TV and try to match them.

[FBX]

PS2 RGB 480i, processed via Framemeister:

http://i.imgur.com/pRNSjkI.jpg

PS2 RGB 480p, processed via Framemeister:

http://i.imgur.com/yl4CyRe.jpg

PS2 RGB 480p, directly fed to my monitor (don't mind the weird aspect ratio):

http://i.imgur.com/ZP7BSN8.jpg

It's strange that I would get the polar-opposite result on my display with the false contours still showing up on direct-feed. Is somehow my Sony using the same or similar upscaling technology as the Framemeister? Further still, the Framemeister seems to handle Gamecube 480p without such issues. I just took a new pic of the Wind Waker heart container, only this time I left the aspect ratio the same as "SMART-2X". Here it is next to the original source graphics:

Spoiler

Other than analogue color bleed, there's no contours that shouldn't be there. So either the Framemeister has a specific issue with the PS2 component signal, or there's more going on than we fully understand. I'm going to contact Ste and see if he can do some scope tests of the PS2 component signal to find out what's actually coming out of that thing. Might help resolve the issue anyway.

[12345]

Finally got the calibration done. I can do a small write-up in a couple of hours if anybody is interested.

[FBX]

Awesome! Would love to check it out!

On the big PS2 480p debate, I heard back from Ste and he is in agreement with me thus far about it being from the PS2, and I'm thinking the reason why the blurring shows up or not is based on the sensitivity of the equipment used to the upscale the image. This also explains why Gamecube 480p appears to be free from such artifacts. Here's an excerpt from his email explaining what is going on:

"First off, I agree with your hunch that this is a problem inherent with the PS2 and not introduced by the Framemeister or other equipment. Based on the photos and your research thus far, it seems like there is bandwidth issue with the video DACs (or amplifiers) in the PS2, similar to what happens on the SNES DACs (older revisions vs 1CHIP revisions). A 480p video signal is twice as fast as a 480i signal. Because it has twice as many lines to draw, it has to draw them twice as fast.

So let's consider the 480i SSX image. First, let's get the vertical blurring out of the way. That's simply a common deinterlacing artifact. I'm sure if you put it on a different deinterlacing mode (I forgot which one does strict line-doubling), that artifact will go away but then anything that is moving in the image will look terrible.

So now look at the horizontal edges of the logos. From an electrical signal perspective, an edge like that is where the voltage changes quickly (from left to right). For kicks, let's assume a transition like that takes 100ns (nanoseconds) which the PS2 is obviously capable of doing. Switch to 480p mode, and now the PS2 has to make that same transition in half the time (50ns). Most likely, since the blurring is happening, the PS2 video circuitry can't hit a transition that fast."

I've wrote him back supplying him with links to the other pics posted about how direct-feed isn't showing those artifacts on some displays. My own personal hypothesis is what I explained above about different equipment 'picking up on it'. But we'll see if he has an opinion similar to mine on it being an equipment sensitivity issue.

[Thomago]

NGC Component 480i:

http://i.imgur.com/Li3enTL.jpg

NGC Component 480p:

http://i.imgur.com/44rcU8l.jpg

Same 480p "horizontal contouring" here; it's just not as obvious as the Gamecube's Component output is fairly soft.

[FBX]

Same 480p "horizontal contouring" here; it's just not as obvious as the Gamecube's Component output is fairly soft.

I discussed this before, but the Gamecube 480i output is EXTREMELY soft, while the 480p output is actually showing the correct detail level (as in my heart containers example). The problem here is the extra sharpness will be improperly scaled by the Framemeister if you use anything other than "SMART_2X" or a custom zoom setting of ZOOM_SIZE: 100 & ZOOM_OVERSCAN: 89. Any other screen size/mode will introduce visible scaling artifacts (unless you use scaler blurring like H_SCALER: 11 and V_SCALER: 5)
So to recap if you want to see the pixels in perfect clarity:

H_SCALER: 4
V_SCALER: 7
ZOOM: ON
ZOOM_SIZE: 100
ZOOM_OVERSCAN: 89
ZOOM_WIDTH: 28 (this is optional if you want uncorrected AR)

Make sure your pics are using the above settings when doing comparison shots, but as I said, GC 480i cannot be compared to because the picture is too soft to see the source detail that we need.

In the PS2 480p case, the artifacts are present even in "SMART_2X" mode. That's what I'm concerned about.

So the next step suggested by Ste is I'm going to hook up a line doubler to the PS2 tomorrow and run it in 480i mode. Then from there, I'll feed the upconverted 480p signal to the Framemeister. If the horizontal blurring shows up, we'll know it's the Framemeister for certain.

Edit: For the record, these issues really don't matter in the long run in the sense that I use scaler blurring in my 480 profiles to antialias anyway, but I just like knowing for certain what's going on there. In all honesty, even if we could have digital-perfect pixels, the scaler blurring would still look better for polygonal games, much as I hate to say it since I love razor-sharp graphics.

[Thomago]

NGC Component 480i; SMART_2X scaling mode:

http://i.imgur.com/mfFRF9V.jpg

NGC Component 480p; SMART_2X scaling mode:

http://i.imgur.com/toB7DAr.jpg

Nothing's changed; despite using SMART_2X mode, the artifacts are still there. I really don't know what you are talking about.

Btw if you feed the Framemeister with a 480p HDMI picture (or any other more-than-SDTV resolution), you will see the same artifacts:

PS3 HDMI 1080p direct:

http://i.imgur.com/aVWqO0S.jpg

PS3 HDMI 1080p via Framemeister:

http://i.imgur.com/4XG7U6H.jpg

Kind of illustrates that this problem isn't caused by the A/D-conversion and therefore - logically - not limited to any particular analog source.

[Fudoh]

Btw if you feed the Framemeister with a 480p HDMI picture (or any other more-than-SDTV resolution), you will see the same artifacts:

hardly. I would say this falls into the H_SCALER tolerance. Some systems just require different settings. BTW: don't you think that you're using terrible V_SCALER settings for your 480i/p sources ? I mean these zick-zack patterns just hurt:



The contouring bug isn't limited to 480p by the way. It can happen on some 240p sources as well, for example on a few N64 revisions, see here (and needless to mention that these contours DON'T show on a RGB CRT).

So the next step suggested by Ste is I'm going to hook up a line doubler to the PS2 tomorrow and run it in 480i mode. Then from there, I'll feed the upconverted 480p signal to the Framemeister. If the horizontal blurring shows up, we'll know it's the Framemeister for certain.

you need a very good linedoubler for that test. Most linedoublers won't leave you with anything better than what your Cube what output (in terms of horizontal sharpness).

[Thomago]

hardly. I would say this falls into the H_SCALER tolerance

Now that you say it... I never tried to adjust H_SCALER for HDMI sources. I don't use the HDMI inputs anyway

BTW: don't you think that you're using terrible V_SCALER settings for your 480i/p sources ? I mean these zick-zack patterns just hurt:

Hehe, indeed. I usually use 960p DVI output for NTSC sources, but as the artifacts weren't as pronounced there, I temporarily switched to 1080p without adjusting the scaling settings

[Thomago]

hardly. I would say this falls into the H_SCALER tolerance. Some systems just require different settings.

Just tested... nope. There's no H_SCALER setting that allows pixel-perfect horizontal sharpness with 1080p HDMI sources.

[Fudoh]

here's no H_SCALER setting that allows pixel-perfect horizontal sharpness with 1080p HDMI sources.

still probably more of a chroma down/up-sampling problem and not the same as what we're seeing on 480p component signals.

[Thomago]

still probably more of a chroma down/up-sampling problem and not the same as what we're seeing on 480p component signals.

A chroma sampling problem... wouldn't the softness be confined to screen content in colour then? Cause it isn't.

[Oliversum]

Sorry to ask again, but maybe my posting on the last site was overlooked.

So I got me a Super Famicom with 1CHIP-03 today for some bucks.
Before this I used a PAL SNES, modded with 50/60Hz (switch) and a modded original Gamecube RGB cable.
As I know I need new cables and such for the SFC, which should I buy? I heard the 1CHIP-03 is lacking CSYNC?
Is retrogamingcables.co.uk a good place to buy these?