XRGB-3

[Fudoh]

So in summary, these depend on the target display supporting non VESA modes, right?

you mean the right thing, I just wouldn't call it VESA timings. We're talking more or less standard TV timings (50 and 59,94Hz). VESA includes mostly PC timings (60, 75Hz etc) and is missing most of the SD/HD modes (50Hz for SD, 720p HD etc)...

[viletim]

Ultimately, I just use the audio outputs on the Sega CD 2, instead of the 9 pin mini din, so that I can reduce the noise.

This is what I wound up doing.... I found an easy way to put it in the loop with cabling I had lying in the house. I was able to do it very tidy so I am happy now.

Although I did notice that the RGB coming from my megadrive 2 is arranged slightly more to the right than my megadrive 1 (the picture displays a bit more to the right in other words.

Any ideas on what causes this? I can live with it but I am curious if there is a way to correct it or just knowing what causes this.

The difference in horizontal position between the MD1 and 2 is related to the SYNC signal. I own devices that will allow you to change the horizontal shift. One such device is an Extron sync separator that has a horizontal position setting (others are RGB interfaces). It takes composite sync or video, and outputs composite sync. So, that type of adjustment operates on the sync signal. I know others in the past have wanted to build their own horizontal shift/position circuit, but I never found such a circuit...

http://nfggames.com/forum2/index.php?topic=3017.0

I don't remember if it was ever actually tested.

[em0ti0n]

b) XRGB-3 in B1 -> Gefen -> Optoma HD3000 (the Gefen rebuilds the sync completely)

This could be helpful information for Konsolkongen having problems with Optoma & XRGB3

[Konsolkongen]

Sorry but i already have an Extron 190F on the way. Hopefully that will solve my problem!

[Fudoh]

Make sure to report back once you got it How much was it ?

[Konsolkongen]

50eur (shipping included). Will probably be over 100eur if it gets opened in the customs... :/

I should get it early next week.

EDIT: About 80eur if it get caught in customs

[Fudoh]

Some notes on the "bad scaling" problem (which should actually be the same problem as demonstrated with the Phantasy Star screenshots before):

I tried a few cabling combinations today involving a XSelect unit, the XRGB and a bunch of other scalers around. The whole problem is very likely caused by the fact that the XRGB does *NOT* really output 640x480 in B1 mode. While it's a 31,5khz signal with a RGB color signal, it's actually a 720x480 signal. This makes sense as the XRGB accepts a 720x480i signal from a PS2 and does nothing else than linedoubling. The problem now is that there are hardly any TV sets out there which can differentiate between a 640x480 and a 720x480 signal *ON THE VGA PORT*. My Sony TV for example states that the incoming signal is 640 pixels wide. The consequence is a wrong sampling of the incoming signal. This causes differences in the pixel width and blurs some areas of the screen. It can be described as broad columns, some softer and some sharper.

The effect is much more noticeable on 480p signals (720x480 incoming AND output, but the TV downsamples to 640 pixels in width), but you can see it in 240p signals as well. If I use the manual phase/pitch setting of the TV and maximize all values I can nearly get rid of the effect, but I'm short a tick or two. The problem of course is that I loose the proper aspect ratio as the TV still assumes 640x480 signal with square pixels.

I retested the theory on the HDMI ports of the TV. Here the TV can actually tell the difference between a 640x480 and 720x480 signal, since the former is interpreted as a PC Signal and the later one as a DTV signal offering a different set of options (Aspect Ratio Control etc). Once the TV recognizes the 720 pixel width, all the problems are gone. Every pixel perfectly the same width and not a single soft spot anywhere. I even switched between 640 and 720 output to the TV and it proved the theory quite well. The problem is that many devices interpretate a VGA signal as 640x480 without further checking. My TV does and so does the Optoma. If I feed the same signal to the Optoma directly (component instead of transcoding to VGA first), the scaler gets it right at once.

This not only affects the XRGB, but basically all analogue transcoders. If I use another transcoder, e.g. the Burosch Con-1 or connect the XSelect directly to the TV using VGA, I get the same result every time - a 720 pixel wide source gets displayed as a 640 pixel wide. I'll try to check some affordable VGA to HDMI converters to see which are worth using. While the more expensive ones should be able to tell the difference between a 720 and 640 pixel wide input, I suspect the cheaper ones to fail in this matter as well. We'll see.....

PS: The effect of "bad scrolling" which I described in the past, should actually be the same. Like a non-uniformity when paying close attention the backgound. This is quite noticeable in horizontally scrolling 2D games. The non-uniformity is blown away, once the signal gets correctly interpretated as 720x480 instead of 640x480.

[Artemio]

That should be in the wiki as well, can you add it Fudoh?

[Fudoh]

I'll add a shorter version over the weekend.

[Konsolkongen]

Very interesting. When using my VGA to HDMI converter my TV still recognizes the signal as a 640x480 so that pretty much supports your theory. Is this also the case with Dreamcast and VGA like i asked you about in a PM some time ago?

So for the best possible picture from Dreamcast and XRGB-3 on my Optoma i should use RGB-HV to component transcoders instead?

Should i remove the part about "bad scaling" from the Wiki then?

[em0ti0n]

The problem is that many devices interpretate a VGA signal as 640x480 without further checking. My TV does and so does the Optoma. If I feed the same signal to the Optoma directly (component instead of transcoding to VGA first), the scaler gets it right at once.

Interesting that Optoma interpretate a VGA signal incorrectly as 640x480 and not as 720x480. Honestly I never noticed anything weird in the picture when fed the Optoma with a VGA signal (while Optoma says there is an Analog 640x480/60 signal on the input)

The bad intepreration of a VGA signal explains why my TV fails to display correctly a VGA signal from XRGB3 in B1 (640x480 vs 720x480) and has no problem in B0 (1024x768). Actually this was the reason why I purchased the Gefen.

Just checked today (to be sure):

XRGB3 -> Gefen VGA to DVI = Gefen detects correct signal 720x480 on its input

[Fudoh]

Interesting that Optoma interpretate a VGA signal incorrectly as 640x480 and not as 720x480. Honestly I never noticed anything weird in the picture when fed the Optoma with a VGA signal (while Optoma says there is an Analog 640x480/60 signal on the input)

the resampling (scaling) of the intput resolution is very good on the Optoma. This masks most of the effect. The Gefen unit differentiates 640x480 from 720x480. It will show the correct input. This is what I used to test the effect. The output of the Gefen hast to be set to 480p, not VGA though, since otherwise the Gefen will again downsample from 720 pixels on the input to 640 pixels on the output.

Very interesting. When using my VGA to HDMI converter my TV still recognizes the signal as a 640x480 so that pretty much supports your theory.

can you provide me with a link to which VGA/HDMI converter you're using ?

Is this also the case with Dreamcast and VGA like i asked you about in a PM some time ago?

Could be it. I believe though that the DC outputs native VGA (which is 640x480) and not a DTV 480p signal. I'll check it out.

So for the best possible picture from Dreamcast and XRGB-3 on my Optoma i should use RGB-HV to component transcoders instead?

On the Optoma it makes very little difference (see answer above)

Should i remove the part about "bad scaling" from the Wiki then?

not yet

[lordnikon]

Hi all, I had spoken with Fudoh quite a few times via email and am an avid XRGB-3 user (though in my setup right now I am mainly using it for its transcoder functions).

I wasn't sure where to post this link, but since I did quite a bit of testing using the XRGB-3, I figured this would be as good a place as any.

Below is an extensive article of mine covering the problem of display lag, as well as other problems associated with rendering classic games in HD:
http://www.benpekarek.com/video_games_i ... ition.html

I constantly run into people who simply don't understand how lag can affect one's playing experience. Short replies to forums never seem to do the trick, and people who don't know much about the technology have a hard time grasping various technical concepts. So I set out to define these complications in a very straight forward way. For many here, some of the information in the article may seem like beginners information (such as explaining the difference between interlaced and progressive) however, at the very least this can be something to pass off to others who have yet to understand these concepts.

Anyways, thanks for reading and feel free to leave your comments or questons. If you spot any glaring inaccuracies please do let me know.

As a side note, I have been working on a rather indepth extension/offshoot to Fudoh's scanline article. Though this is still a work in progress and will be another month or so before it is completed.

[Fudoh]

nice article. I'll make sure to link to it when I get around updating my page(s).

Can you tell us which game you used for the counter display shown on your pictures ? Would be interesting to learn which games available in 240p, 480i and 480p can be used for such testing. A timer with two decimal digits is needed.

You point out to the Eizo FX2431 in your article. While it's a nice approach indeed, it's not perfectly implemented. The display lags 34ms when fed over VGA, DVI or HDMI, no matter what's the input resolution. The "thru" mode is limited to HDMI 1080p only and the source has to deliver video levels, so it's basically limited to PS3 and 360 gaming. Have you found a review actually measuring the "thru-mode" ? I have found good opinions about it with input lag between 7ms and 1-frame, but no site actually measuring it.

[lordnikon]

nice article. I'll make sure to link to it when I get around updating my page(s).

Thanks, I appreciate the positive feedback.

Can you tell us which game you used for the counter display shown on your pictures ?

Virtual On for both the Dreamcast and the Saturn was used for testing, which is mentioned in the article. It is true that 2 decimals are needed for ultra accurate testing, but the main goal of the counter tests were to show non-believers that lag does indeed exist. Especially for those in the general Dreamcast community that have been preaching with conviction that the Dreamcast VGA output eliminates lag completly. So it was much more about showing proof of the existance of lag and giving people a worthwhile model for testing it, rather than determining exactly how much lag the XRGB-3 performs in B0 scaling mode.

Also, make note of where I discuss the direction the numbers are counting. Virtual On counts down backwards in milliseconds rather than forwards. I could have chosen other games to test, but I guess I was going for uniformity in the time counters across platforms, and wanted to get 2 games that ran at their peak performance in terms of framerate.

Would be interesting to learn which games available in 240p, 480i and 480p can be used for such testing. A timer with two decimal digits is needed.

I didn't include a compatibility list for testing various output modes in milliseconds, because the key genre of key games to use would be of course racing games. Titles such as Virtua Racing Deluxe for the 32X and Ridge Racer 64 would be prime options, but once you get into the Saturn and Playstation, there are a plethora of available racing games with millisecond counters on those platforms(the list would be enormous). At step 2 of the lag test steps, I mention the recommendation of using racing games. Though maybe I should elaborate on this slightly with some of the text I have already written in this posting.

You point out to the Eizo FX2431 in your article. While it's a nice approach indeed, it's not perfectly implemented. The display lags 34ms when fed over VGA, DVI or HDMI, no matter what's the input resolution. The "thru" mode is limited to HDMI 1080p only and the source has to deliver video levels, so it's basically limited to PS3 and 360 gaming. Have you found a review actually measuring the "thru-mode" ? I have found good opinions about it with input lag between 7ms and 1-frame, but no site actually measuring it.

For the FX2431, I state that it "still has quite a bit of lag". Also I was tryng to convey that the Foris is "is frighteningly close to what we should expect from future HD displays". Its not ideal yet and does have problems. The main point is, it is one of the first displays to actually address the problem of processing pixels, not just in regards to technology, but also by publically advertising this point. We want companies to both solve the problems of lag, as well as video signals with pixel based sprite art. The FX2431 is headed in that direction, but of course its not there yet, as it still has quite a few problems. I should tweak that paragraph a bit maybe in the article so this point is clearer.

There is only 1 test that I know of that tests the lag for the Eizo FX2431:
http://www.bit-tech.net/hardware/2009/0 ... r-review/2

The problem is, they use a Dell3007WFP as their control, which is innaccurate. They should be using a zero lag CRT. So your assessement of about 34ms of lag is close to what it would have.

[Fudoh]

tested the DC: it does output real VGA at 640x480 and the soft spots seem to be rendered into the output, so what probably happens is that the graphics chip renders in 720 pixels, but then downscales to 640 pixels before the output. Anyway, it doesn't make a difference if the DC is connected via VGA or HDMI, results are the same.

[Konsolkongen]

can you provide me with a link to which VGA/HDMI converter you're using ?

This cheap-o unit:
http://cgi.ebay.com/PC-HDTV-VGA-Audio-t ... 255827debb

So for the best possible picture from Dreamcast and XRGB-3 on my Optoma i should use RGB-HV to component transcoders instead?

On the Optoma it makes very little difference (see answer above)

tested the DC: it does output real VGA at 640x480 and the soft spots seem to be rendered into the output, so what probably happens is that the graphics chip renders in 720 pixels, but then downscales to 640 pixels before the output. Anyway, it doesn't make a difference if the DC is connected via VGA or HDMI, results are the same.

So no reason to use a VGA to Component transcoder since the DC messes the picture up and outputs in 640x480. And a transcoder for the XRGB-3 is pointless as it will look pretty good on the Optoma as is. Right?

[Fudoh]

I will compare XRGB->Optoma to XRGB->XSelect->Optoma when I get around and report back if I can spot a difference.

[Konsolkongen]

Thanks

Reading lordnikon's page now. Great stuff

[Konsolkongen]

I don't think the "soft spots" in the Dreamcast signal is something made by the Dreamcast itself. They can be moved around using the Phase setting on the Optoma.

Looks like the Extron got through customs just fine. I should receive it tomorrow

[Fudoh]

I don't think the "soft spots" in the Dreamcast signal is something made by the Dreamcast itself

I'll look into that again - need more appropriate games to test this. Nevertheless the output is real 640x480, so transcoding to 480p YUV will likely introduce more problems than solving any. We'll see...

[HDgaming42]

I don't think the "soft spots" in the Dreamcast signal is something made by the Dreamcast itself

I'll look into that again - need more appropriate games to test this. Nevertheless the output is real 640x480, so transcoding to 480p YUV will likely introduce more problems than solving any. We'll see...

I'm not calling into question your methodology, but I'm curious how you determined that the DC outputs 640x480?

[Konsolkongen]

The Extron doesn't work either. Must have tried every dipswitch combination with the XRGB-3's FULLHD setting LOW/HIGH and Vsync ON/OFF. The Extron works fine with Dreamcast and Optoma. I'm pretty sure that something is wrong with either the Optoma or XRGB-3 now

[em0ti0n]

The Extron doesn't work either. Must have tried every dipswitch combination with the XRGB-3's FULLHD setting LOW/HIGH and Vsync ON/OFF. The Extron works fine with Dreamcast and Optoma. I'm pretty sure that something is wrong with either the Optoma or XRGB-3 now

That's bad
After all suggestions what to do/try/check I am starting to suspect your Optoma has a problem with 720x480 VGA signal (even when VGA->HDMI transcoder is used) since your Dreamcast works nice as the console outputs real 640x480.

[Fudoh]

I'm not calling into question your methodology, but I'm curious how you determined that the DC outputs 640x480?

The Gefen VGA to DVI Scaler will show the input resolution. It shows 720x480 when connected to the XRGB-3 in B1 mode and it shows 640x480 when connected to a Dreamcast VGA Box (original Sega). The 50pro will tell the difference as well and you can "ask" your Edge - it should state the correct input format as well.

[SGGG2]

I'm not happy with the picture quality of the Gamecube via component (D1) through the XRGB. It seems a step below the Xbox and PS2. Would the RGB mod make any difference? VGA or SCART?

http://gamesx.com/wiki/doku.php?id=av:gamecube_rgb
http://web.archive.org/web/200212122032 ... /Main.html

[Fudoh]

No, it does not. The GC's component and RGB (VGA) Signal are based on the same source signal. If you run your GC's YUV signal (480p) through the XSelect and output RGBHV to your TV directly, you get the exact same quality as with the cable mod. Scart is out of the question anyway, since I would imagine you play most games in native 480p.

[HDgaming42]

I'm not calling into question your methodology, but I'm curious how you determined that the DC outputs 640x480?

The Gefen VGA to DVI Scaler will show the input resolution. It shows 720x480 when connected to the XRGB-3 in B1 mode and it shows 640x480 when connected to a Dreamcast VGA Box (original Sega). The 50pro will tell the difference as well and you can "ask" your Edge - it should state the correct input format as well.

Thanks Fudoh. Somehow I'd glossed over your enormous post where you explained your process in detail. Thanks for the recap!

[Fudoh]

The past days I tried running the usual candidates through the XSelect before running into the XRGB. Most people in this thread probably know the XSelect from Lawrence's review a few years back. It was released as a D-Terminal switch to work in conjunction with the XRGB-2plus. The funny thing about the XSelect is that it has very good analogue input filtering. Exactly what the XRGB-3 is missing and what's causing all the trouble with the noise in the PS2 signal. The XSelect offers three component inputs (RCA and D-Terminal), a 21-pin RGB socket, a VGA input, a VGA output and a component output. You can run one cable from the XSelects RGBHV output to the VGA input of the XRGB. The input on the XRGB-3 should be set to seperate sync.

With this combination I don't have to use the LPF anymore on a PS2 component signal. The PS2 signal gets filtered and transcoded in the XSelect and is send to the XRGB-3. While the sharpness & clarity is not 100% the same as when directly fed to the XRGB-3, but it's still noticeably sharper than when using the LPF filter on the XRGB-3. 480p looks the same as when directly fed, but without any noise in the signal. The XSelect also works as a Syncstripper and sync seperator which works very nice on my PC Engine. Where I had extreme trouble before, I can now route the PCE's RGB signal through the XSelect. No more need for ultra-high AFC settings and no more sync-instability either.

I will test more systems with this combination sometime in the future....

[HDgaming42]

With this combination I don't have to use the LPF anymore on a PS2 component signal. The PS2 signal gets filtered and transcoded in the XSelect and is send to the XRGB-3. While the sharpness & clarity is not 100% the same as when directly fed to the XRGB-3, but it's still noticeably sharper than when using the LPF filter on the XRGB-3.

Any chance this makes the optoma HD3000 happy when using PS2?