DVDO EDGE - How it performs

[Konsolkongen]

Is the pattern you created a 1 pixel pattern (1 pixel colums in color, 1 in black and so on...) ?

Yes.

I assume Saturn -> Gefen -> Edge with the Gefen set the 720x480p output looks fine ??

Yes I believe so. I didn't save the drawing in Dezaemon but I can recreate it if necessary. The Gefen's colorspace is also detected as RGB.

[Fudoh]

There's one thing you could try for me (doesn't have to be with this pattern though).

If you use the XRGB-2plus and adjust the settings so it's seen as 640x480 by the Edge (instead of 720x480), does the problem persist ?

And also try to exchange the Green connector with on the other ones to see if it's input or color-related.

[fagin]

If you use the XRGB-2plus and adjust the settings so it's seen as 640x480 by the Edge (instead of 720x480), does the problem persist ?

I have done that for my scenario and it makes no difference to the "bleed" on RED and BLUE.

[Fudoh]

Thanks! I've posted this on the AVS board and I will email Simplay Labs next week about it. Maybe we're lucky and I can get ahold of one of the ABT engineers.

[Konsolkongen]

Cool I can try swapping the cables next time I have the time.

[Konsolkongen]

When replacing green and blue color, the blue (what EDGE sees as green) is still messed up. So this is a cable issue, right?

So this is the Dezaemon 2 sprite editor screen. Notice the two areas marked 1 and 2. Both show the same image, unscaled:


This is area 1 up close:


Area 2:

Notice how this looks much better. Not 100% perfect but you can't see it in this picture.

Here is area 1 again, but this time I swapped the green and blue cables:

Green is now blue and vice versa. The EDGE now sees the blue as the green color but it's still blurry.

Sorry I took these with scanlines on

EDIT: Here is the one with swapped colors and no scanlines:
http://i40.photobucket.com/albums/e236/ ... 1304804941

[Fudoh]

When replacing green and blue color, the blue (what EDGE sees as green) is still messed up. So this is a cable issue, right?

it means that it's not an internal processing problem, but a problem on the A/D conversion or signal recognition.

You're sure that the Edge's output is in a 1:1 pixel alignment with your display, yes ? If I saw something like this (soft spots on the left, sharp on the right) on a display I would suspect that the alignment isn't perfect - - - but then again why should it only be affecting certain colors. Very weird after all...

[Konsolkongen]

You're sure that the Edge's output is in a 1:1 pixel alignment with your display, yes ?

Absolutely. The TV is set to no overscan and the checkerboard, vertical and horizontal lines test patterns on the EDGE looks perfect.

[BuckoA51]

Did you definitely try this with the Gefen too?

[Konsolkongen]

Like what exactly? Saturn with Deza 2 > XRGB-3 > Gefen@480p > EDGE?

I can try but I'm not sure I get the point. This sounds like a VGA problem not HDMI. Unless of course this problem is on all 480p sources?

[Konsolkongen]

It's still a little blurry when the Gefen is set to 480p but that's either a Gefen scaling problem, or the ringing added by the EDGE. Anyway it looks much better.

Gefen at 480p:
http://i40.photobucket.com/albums/e236/ ... 1304881171

Gefen at 720p:
http://i40.photobucket.com/albums/e236/ ... 1304881172

[Konsolkongen]

Sorry I'm spamming but I think this is relevant. Just updated the XRGB's firmware and when I was in B0 mode I decided to check how the EDGE saw the XRGB's 640x480 resolution.

Here is the info screen for B0 mode, 640x480, V-sync off:


B0 mode, 640x480, V-sync on:


Both are pretty wrong, right? When set to 1024x768 or higher the EDGE seems to get it right.

[BuckoA51]

Like what exactly? Saturn with Deza 2 > XRGB-3 > Gefen@480p > EDGE?

I actually meant just Saturn with Deza 2 > XRGB-3 > Gefen@480p > TV

[Konsolkongen]

That would probably look very similar to the above pictures. Perhaps with a bit less ringing

[BuckoA51]

So the colour bleed is definitely unique to the Edge, is what I'm getting at, it couldn't just be something else in your chain you never noticed before?

[Konsolkongen]

It's definitely not there when doing XRGB-3 > Gefen@1080p > TV. If the TV upscales from 480p or not doesn't matter.

[BuckoA51]

I'll see if I can find that game and replicate the results when I get a moment (stupidly busy IRL with ebook writing right now)

[Konsolkongen]

Got the BNC to RCA plugs today and it makes no difference. We can rule out the cable

[fagin]

Got the BNC to RCA plugs today and it makes no difference. We can rule out the cable

I could of told you that the other week..... mine does it and there is nothing wrong with my cable.

[Konsolkongen]

Hehe that's okay they were dirt cheap on eBay

[BuckoA51]

Unlocked 1080p output does output at broadcast standard specs which is 59.94Hz EXACTLY. Videogames do unfortunately output at roughly 60Hz only, ranging from 59.1Hz (MVS) over 59.94 (PS2), 60Hz (SNES) up to 60.1Hz.

If you lock the output the input refresh is taken over to the output "locking" both the input and the output to each other. This way no framerate conversion is done and you don't lose any frames (nor are any doubled).

Best is to have it locked, but then you can run into problems with digital displays. Some TVs only handle 59.94Hz, others accept AND show a wider range and others again convert everything to 59.94 (e.g. 60Hz to 59.94Hz). Now depending on your TV and your source the TV might show tearing when it transforms the framerate. If you unlock the Edge you should never get tearing, but you might get a short stutter here and there.

Actually similar to the XRGB-3's VSync setting.....

Just got an e-mail from a reader to my site swears the opposite, that tearing only occurs when he unlocks the output frame rate. Could be that he didn't realise you MUST power cycle the edge on the remote when you change this setting else it does not take effect though.

[Fudoh]

Could be that he didn't realise you MUST power cycle the edge on the remote when you change this setting else it does not take effect though.

I didn't know that either.

Just got an e-mail from a reader to my site swears the opposite, that tearing only occurs when he unlocks the output frame rate.

Both can be right. On my text (you quoted) I was refering to the possibility of tearing created through the framerate conversion. It depends where the tearing is produced. After all tearing is an effect usually produced in the display. Unlocking the output gives exact broadcast specs, so no display should PRODUCE tearing for any reason. Locking the output forwards the input's refresh rate which forbids any framerate conversion (which might as well produce tearing).

[BuckoA51]

Ah ok thanks for clearing that up. Yes I was going around in circles until I realised about the power cycling thing.

[Konsolkongen]

Heh I had this mixed up as well. Thought the unlocked framerate was the one that could cause tearing Still I have only used mine with the auto locked frame rate and I haven't experienced any tearing regardless of source.

[Konsolkongen]

Tried the same Deza 2 experiment, but this time with the XRGB2+. The funny thing about the 2+ is that the EDGE sometimes sees this correctly as VGA.

Here is the first shot where the EDGE recognizes the 2+ as YPrPb (same as the XRGB-3):


And here is VGA:


I wouldn't call that much of an improvement... Actually I don't see any at all. The blue and red colors are still messed up like always but this time the picture is squeezed from 720x480 to 640x480 and it's extremely noticeable.

It looks like it wont solve this problem just by making the EDGE detect the correct colorspace. Actually I'd rather have it detect it as YPbPr, like it does now, and get an image with correct pixel width.

Perhaps it could work somewhat correctly if we could just manually change the input colorspace to YPbPr 444? But I don't know shit about colorspaces and whatnot, I'm just guessing

The only positive thing about the correct VGA detection was that I could actually move the picture horizontally a bit further using the XRGB2+, before the colors got messed up... That's also a very strange EDGE problem :/

[Fudoh]

I'm in contact with one the DVDO/ABT engineers and he's looking into this. Unfortunately it can't be prioritized as much as we want, but it might get fixed after all...

[Konsolkongen]

I'm in contact with one the DVDO/ABT engineers and he's looking into this. Unfortunately it can't be prioritized as much as we want, but it might get fixed after all...

Kewl thanks Fudoh

[BuckoA51]

Been doing some interesting experiments on my DVDO Edge with 240/288p material. I just got a CD32 with a SX32 expansion module which unusually outputs both RGBHV 15khz and RGBC 15khz at the same time from different connectors. This allowed for some interesting experiments....

480i Material
Looks great through RGBHV, picture is slightly blurred through RGBC
240/288p Material
RGBHV looks sharper, but has some unstable elements around word/sprite edges in places
RGBC has no unstable elements, but picture is more blurry.

On the XRGB3 on the other hand, I cannot see any difference between RGBC (Game in) and RGBHV (D in) from the CD32 even with my nose pressed against the screen.

Tried with and without my Extron in the chain, made no difference.

Weird huh? Videoprocessors sure are quirky... However from these experiments I can say for sure that for those of you using the Edge for 240p material it would certainly be worth trying a sync splitter/cleaner to take the RGBC into RGBHV, like Keene's Syncblaster cable or the Sync Strike. I certainly preferred the RGBHV picture to the RGBC, even with the slight wobble on some elements (which isn't forced to be there with another cable).

If anyone wants pictures of this experiment let me know.

Edit - Should add in case anyone was wondering, the Amiga outputs pure sync not composite video for sync.

[Fudoh]

I'm surprised that the Edge takes 15khz RGBHV after all. That's sheer luck, since 15khz should use another sync input than a 31khz+ signal. On the other DVDO units this doesn't work (tried it).

Have you tested this with any other system ? Or in other words, are you sure that it's due to the Edge's processing and not due to the CD32 offering some weird sync re-combination on it's own ?

[BuckoA51]

I've not tried it with any other systems no, I sent my Syncblaster cable back else I would test it, I would love to do more experiments along these lines. I wasn't aware that the sync inputs were different for 31khz signals though (by this do you mean they connect to a different input on the Edge?).