OSSC (DIY video digitizer & scandoubler)

[marqs]

Just looked through the new entries of the compatibility list (https://docs.google.com/spreadsheets/d/ ... edit#gid=0).

Actually, what is it with the jitter? I've already seen these artifacts in one of the Chrono Cross videos posted earlier and I've never experienced anything like that with any of the video capture devices I've used.
Is being prone to jittering inherent to the OSSC's hardware design or is this something that can be fixed for good (respectively for anything but rare exceptions)?

If you mean the horizontal jitter in the Chrono Cross video in the middle of the screen, that is a result of a design flaw in the video digitizer IC which appears in a certain sync mode. Fortunately that mode is quite rare - other sources that seem to employ it are Wii/Jaguar in 288p mode, and a few Taito arcade board games. So far the only way to get jitter-free picture from those sources is to connect them to digital sync input (RGBHV/RGBS of VGA connector) which has slightly different processing chain.

If you're talking about jitter in more general fashion, it's a sum of several factors. The stability of source hsync signal and quality/filtering of sink PLL which locks to it have a big impact, especially with systems of which dot clock rate does not match sampling rate (read: pretty much all pre-DC consoles). You can easily reproduce this by any digital desktop monitor which has VGA input: just change the 'clock' setting to an unoptimal value and you'll see jitter in some column(s). When sampling rate cannot be set freely, LPF is a good way to mitigate jitter as it slows down video transition times and reduces overshoots.

4) Remote
I am using a Harmony universal remote. Is there a commercial device that has the same remote so I can use it to easily parameter in the Harmony software?
If not I can still use the IR learning mode.

The bundled remote is pre-programmed to Yamaha DVD player codes. You could either program Harmony to use those codes or alternatively, select another code set from Harmony and then program OSSC to acceot those codes.

[BuckoA51]

Latest test firmware seems to have almost completely eliminated the jitter, apart from the unusual sync mode bug (I'll try the sources connected via VGA soon). It has introduced some new bugs but I know Markus is working on them.

[NormalFish]

Marqs you're using csync to test the PS2?

[Fudoh]

PS2 doesn't output csync without a mod. I used a cable which is wired for luma as sync.

[BuckoA51]

I tried Csync from PS2 via Extron Interface and it worked great.

[NormalFish]

PS2 doesn't output csync without a mod. I used a cable which is wired for luma as sync.

Which is why I asked. Unless I'm misreading (entirely possible given I woke up a little more than an hour ago), he tested the PS2 with a csync cable. But as long as you tested with luma it's a nonissue.

[Fudoh]

I tried Csync from PS2 via Extron Interface and it worked great.

doesn't the PS2 just offer internal H/V sync ? How's the sync internally combined ?

[marqs]

I tried Csync from PS2 via Extron Interface and it worked great.

doesn't the PS2 just offer internal H/V sync ? How's the sync internally combined ?

At the the model I have offers internal csync too. In that case though, I believe Bucko is using Extron to generate csync.

[NormalFish]

I tried Csync from PS2 via Extron Interface and it worked great.

doesn't the PS2 just offer internal H/V sync ? How's the sync internally combined ?

At the the model I have offers internal csync too. In that case though, I believe Bucko is using Extron to generate csync.

You lost me. to the best of my knowledge no PS2 outputs Csync natively. You have a modded console?

[bobrocks95]

You lost me. to the best of my knowledge no PS2 outputs Csync natively. You have a modded console?

That's why marqs said internal csync. So csync is available on some chip somewhere to be pulled and wired up to the output port if you like.

[werk91]

RetroGamingCables sells Csync PS2 Scart, I guess it has a chip inside that combines the H and V?

[Fudoh]

no, it uses a stripper IC inside the cable and most likely luma as a base.

[ZellSF]

Uh, what's this talk about PS2 H/V sync? isn't it SoG?

[Guspaz]

H and V sync are both available internally, but AFAIK are not exposed.

[bobrocks95]

I was going to point out that the PS2 Linux kit came with a VGA cable, but after looking it up it was apparently SoG too, which is really weird and I wonder what monitors even supported SoG on their VGA inputs (just Sony's probably).

[Guspaz]

Probably many? My Dell monitor (U2711) lists the following sync support:

separate horizontal and vertical,
3.3V CMOS or 5V TTL level, positive or negative sync.
SOG (Sync on green)

Then again, my monitor also supports component video, and sync-on-luma is the same thing as sync-on-green electrically.

A bigger problem is probably that it's very hard to find VGA at all, at least in high-end monitors. I don't think Dell put VGA on any of their UltraSharp monitors past ~2012, but it looks like the 2012 models also supported SoG, and they didn't have component.

[BazookaBen]

I was going to point out that the PS2 Linux kit came with a VGA cable, but after looking it up it was apparently SoG too, which is really weird and I wonder what monitors even supported SoG on their VGA inputs (just Sony's probably).

Actually tons of monitors do (or should I say did).

A database was compiled a while back: http://ps2linux.no-ip.info/playstation2 ... m/sog.html

I actually have three non-sony CRT's that support SoG: Samsung, LaCie, and NEC

[Edward_Tz]

Is there a FAQ set up for the OSSC or a wiki?

For example if csync is needed? 240p to 480i transition times? Does the sound drop up with 240p to 480i transitions? Can it be set up for JP-21 cables?
I'm sure this has been covered but the thread is getting a bit long and it's tough to see what is the final product and what was said in the development.

[BuckoA51]

Wiki is here http://junkerhq.net/xrgb/index.php/OSSC

[Edward_Tz]

Thanks BuckoA51

Wasn't coming up when I tried googleing. I'll keep an eye on it while it's updated along with this thread.

[lev11]

If sampling phase is the feature to tweak for different sources is there opportunity to fit a good old manual knob to turn? Just so no remote is needed.

[jamesfmackenzie]

Will the OSSC support adjustable horizontal sampling rate?

I often use Atari ST medium resolution which renders 640 columns. Most scalers don't sample horizontally at a high enough rate and I get column drop.

Thanks!

Jimmy

[jamesfmackenzie]

Will the OSSC support adjustable horizontal sampling rate?

I often use Atari ST medium resolution which renders 640 columns. Most scalers don't sample horizontally at a high enough rate and I get column drop.

Thanks!

Jimmy

For an example of what I mean, see the following:
http://www.jamesfmackenzie.com/img/post ... amdisk.png

The capture is missing columns due to horizontal undersampling

[marqs]

If sampling phase is the feature to tweak for different sources is there opportunity to fit a good old manual knob to turn? Just so no remote is needed.

Phase tweaking is generally relevant only with newer consoles and PC sources. The setting has 32 discrete steps (N * 11.25 degrees), so an analog knob would be a bit overkill.

Will the OSSC support adjustable horizontal sampling rate?

I often use Atari ST medium resolution which renders 640 columns. Most scalers don't sample horizontally at a high enough rate and I get column drop.

Yes, although I haven't yet exposed the setting on the menu (besides 2 pre-defined modes for 480p). For best possible quality, you'll need to find out dots/line rate of ST output and then use it as horizontal sampling rate. I've dug out this number for some more common consoles already - probably should put that info on a dedicated wikipage...

[BazookaBen]

If anybody has used this on a PC CRT, I'd love to hear about your experience.

Like with 480p line-doubled to 960p (if possible). I find that's the only drawback to playing on my PC CRT is the scanlines on 480p material. I'd rather play at 960 to eliminate the scanlines.

[NJRoadfan]

For an example of what I mean, see the following:
http://www.jamesfmackenzie.com/img/post ... amdisk.png

The capture is missing columns due to horizontal undersampling

Is this output from a GBS-8220? Its not exactly friendly with 640x200 modes on old computers. A real test of horizontal sampling would be the Amiga's rarely used 1280x200 modes.

[Raen]

Hi all,

Not sure if this is the most appropriate topic to post my problem, but I'll give it a shot and hopefully I can get some help from you guys.

I recently acquired a Sony Trinitron Multiscan CPD-G500 21'' CRT Monitor, which has VGA + RGBS BNC inputs. I use it with an old desktop PC and with my Dreamcast + Toro VGA box combo. I eventually bought a SCART to BNC cable from retrogamingcables.co.uk to avoid having to swap VGA cables every time I wanted to switch between PC and DC, leaving the VGA input for the PC and the RGBS BNC inputs for the Dreamcast (not a big deal, since the Toro VGA Box can also output 480p through SCART).

Everything was going great, until I tried to play a game that apparently is not VGA compatible, and cannot be trickied into playing in the 31 khz VGA mode. That, allied to the fact that the monitor doesn't support 15,6 kHz RGB mode (says "out of scan range"), poses my problem.


I searched a little bit and the solution is to get a scandoubler (aka line doubler, upscan converter, etc) to convert the 15,6 kHz signal into a 31 kHz one that the monitor supports, and by doing so, even the games that can't be trickied into VGA mode should be playable in this monitor in RGB mode, which was initially my plan.

Initial search attempts led me to a scandoubler called CV3D Scandoubler, but that is for the Amiga and nothing that works with the Dreamcast came up. Then I saw the OSSC review on Fudoh's website and thought that I'd found the solution, until I saw that it has only digital outputs, not analog ones.

Someone said to me that I could buy a cheap GBS-8220 CGA to VGA board, that it would do the trick. I just wanted to hear a second opinion, since these boards don't seem to be of the highest quality and don't have SCART IN and SCART OUT (maybe a mod lies ahead, just like fagin's?).
Is this GBS-8220 CGA to VGA board really the best option for the job (scan doubling 15 khz to 31 khz), or is there a better alternative?

[donluca]

I believe this was already asked at some point but I can't seem to find it:

Is it possible for the OSSC to generate a 1080p signal with a centered line quadrupled 240p signal (so 960p) and double thickness scanlines?

This is actually what I use on my iMac for MAME, a quadrupled resolution with 2px wide scanlines (50% transparency) and it looks great.

[jamesfmackenzie]

If sampling phase is the feature to tweak for different sources is there opportunity to fit a good old manual knob to turn? Just so no remote is needed.

Phase tweaking is generally relevant only with newer consoles and PC sources. The setting has 32 discrete steps (N * 11.25 degrees), so an analog knob would be a bit overkill.

Will the OSSC support adjustable horizontal sampling rate?

I often use Atari ST medium resolution which renders 640 columns. Most scalers don't sample horizontally at a high enough rate and I get column drop.

Yes, although I haven't yet exposed the setting on the menu (besides 2 pre-defined modes for 480p). For best possible quality, you'll need to find out dots/line rate of ST output and then use it as horizontal sampling rate. I've dug out this number for some more common consoles already - probably should put that info on a dedicated wikipage...

For the ST, the official ST dots/line are:

ST Low = 320
ST Medium = 640

However this doesn't include the overscan areas.

From this source:
http://aldabase.com/atari-st-fullscreen-demos-history/

It looks like the actual numbers are:

ST Low = 416 dots
ST Medium = 800 dots

However, from my own analysis with my Micomsoft capture card, I'm seeing exactly 25% of the line reserved for overscan, which yields:

ST Low = 427 dots
ST Medium = 853 dots

Hope this helps! I'm not sure how to make it determine the sampling rate more accurately.

[jamesfmackenzie]

For an example of what I mean, see the following:
http://www.jamesfmackenzie.com/img/post ... amdisk.png

The capture is missing columns due to horizontal undersampling

Is this output from a GBS-8220? Its not exactly friendly with 640x200 modes on old computers. A real test of horizontal sampling would be the Amiga's rarely used 1280x200 modes.

No, it's actually from a Micomsoft sc-512n1-l/dvi.

Despite offering custom horizontal sampling configuration options, the actual capture is always locked to 720 pixels. Once you count the overscan areas, that's not enough to capture ST Medium @ 640 pixels wide - it's always undersampled.

Amiga 1280x200 would be even worse