OSSC Pro

[Josh128]

In any case, if theres some technical reason as to why this mode would induce blur on a CRT ( I dont believe it does with MISTer or Retroarch), the tradeoff would be worth it to me.

Doubling the refresh rate induces both motion incoherency (2 strobes/scans per frame) and latency (half a frame at minimum) so it's not without tradeoffs. You can do motion interpolation to mitigate the former (remember the 100Hz CRT TVs) but that bumps the lag even more. That said, 240p 120+Hz output should be doable on the Pro model.

I'd have to see the result of the 2 scans per frame @120Hz in person on a CRT before I would decide that it would need motion interpolation. The old LCDs used to interpolate 24fps / 30 fps video into something that looked like 60fps (or 120 fps), and the interpolation looked like crap to me (soap opera effect), but on many of these sets I never saw an option to turn it off.

That said, interpolating 30fps video, which is already very easily discernible to the human eye, into 60 or 120 fps is a different ballgame than simply doubling the draw of each 60fps frame on a CRT. I have a feeling I would prefer the raw doubling of each drawn frame as is vs. adding some interpolation to it, so the only added lag would be about 8ms. That would be perfectly acceptable to me.

I'd love to see it in person, I have several VGA monitors in my possession. Ive heard the new Retroarch is capable of the exact same thing with certain video cards that can still output analog video, if anyone is familiar with it please point me in the right direction, as it would be a good indicator of what to expect from this OSSC Pro.

This looks absolutely fantastic to me. https://www.youtube.com/watch?v=C_pYmNG89RQ https://www.youtube.com/watch?v=C_pYmNG89RQ https://www.youtube.com/watch?v=EZNmxTGmZBo

[orange808]

In any case, if theres some technical reason as to why this mode would induce blur on a CRT ( I dont believe it does with MISTer or Retroarch), the tradeoff would be worth it to me.

In any case, if theres some technical reason as to why this mode would induce blur on a CRT ( I dont believe it does with MISTer or Retroarch), the tradeoff would be worth it to me.

In any case, if theres some technical reason as to why this mode would induce blur on a CRT ( I dont believe it does with MISTer or Retroarch), the tradeoff would be worth it to me.

[energizerfellow‌]

Selection of the video ADC was a hard choice between ISL51002 and ADV7842 - I hope I made the right call this time. ISL51002 has its own quircks but so far it has operated way better than TVP7002.

If you dont mind me asking, i'm kind of curious to know what the pros/cons were of each ADC chip?

I'm curious on the gritty details of the ISL51002 vs ADV7842 situation as well, especially considering the ISL51002 + ADV7611 + ADV7513 is a 3-chip solution, but the ADV7842 + ADV7511 is a 2-chip. The ADV7842 + ADV7511 also lets you do things like a built-in HDMI receiver with a 10/12-bit color pipeline at 225 MHz, 2D/3D comb filtering, frame TBC, CVBS/YC input support, CVBS monitor output, etc. You can even try the ADV7842 + ADV7511 combo with one of Analog Device's EVAL-ADV7842 boards (for about $220).

FYI that Magewell's analog capture cards, e.g. Pro Capture AIO, use the ADV7842 and Datapath's VisionRGB-E1S/VisionRGB-E2S cards use the older AD9887A. I've heard some rumblings that the Magewell cards may not have 240p support on CVBS/YC, which may be a limitation of the ADV7842 as the ADV7842 apparently forces the composite/S-video inputs into ITU-R BT.656?

Any reason for the PCM1862 over the PCM1863 other than to save a few pennies? Since this device has the kind of performance to be interesting to the video guys, you may want to consider the PCM1863 as VHS Hi-Fi audio has 90+ dB of dynamic range, never mind LaserDisc. Everybody loves a nice low noise floor too.

Wanted to add one thing:

Markus and Matt: please at least make some calculations on the feasibility of a proper case. Personally I REALLY wouldn't mind to swallow a - let's say - $50 price increase over whatever the base price might be, in order to get a proper housing for the thing. Aluminum injection molding tools aren't this expensive.

This. I'd love to see something like a steel case.

[Elrinth]

Not sure if anyone already asked this. but I will ask it.
Will the HDMI be able to pass thru hdmi 2.0B of for example: Xbox One X?
Edit: After having read: "the ADV7513 HDMI transmitter of the OSSC Pro is limited to a 165MHz TMDS clock and officially only supports HDMI 1.4"
By that line I assume it's not going to support passthru for the modern consoles. Meaning, the OSSC Pro can't be the final destination of the PS4Pro/XBoxOneX.
But for all other consoles prior to that, it should be able to be the final destination. Even Composhite and S-Video limited ones such as the PC-FX.

[strayan]

My ancient plasma TV has a “vertical shift” setting which seems to shift lines up or down by (maybe) a single line or so. Hard to describe what it’s doing but whatever it does it eliminates the combing artefacts present when displaying content rendered at 480i but being presented to the TV as a progressive signal when using an UltraHDMI modded N64. I can try capture some footage if it helps.

Does anyone happen to know what the vertical shift setting might be doing and whether the OSSC pro would be able to replicate it?

[kloow]

Wow! This looks like it's going to be an amazing product! I specially like the HDMI input and the adaptive line multiplier. I'll be sure to keep an eye on this!

[H6rdc0re]

Seeing how this product is still in development. Couldn’t there be a switch made to a chip supporting at least a TMDS of 340MHz? This way the OSSC Pro can support 1440p. That would allow for 6x240p, 3x480p and 2x720p. Should be a real game changer. I’m sure most of us are willing to pay the premium.

[Josh128]

In any case, if theres some technical reason as to why this mode would induce blur on a CRT ( I dont believe it does with MISTer or Retroarch), the tradeoff would be worth it to me.

In any case, if theres some technical reason as to why this mode would induce blur on a CRT ( I dont believe it does with MISTer or Retroarch), the tradeoff would be worth it to me.

In any case, if theres some technical reason as to why this mode would induce blur on a CRT ( I dont believe it does with MISTer or Retroarch), the tradeoff would be worth it to me.

Did you mean to reply about this?

[headlesshobbs]

* enables some new output configurations such as 4x240p/2x480p in standard 1080p frame

I did ask about framing 920p in a 1080p window frame awhile ago, so could this be what I reading?

If so, then this is the next step forward beyond collecting bvm's when they're no longer available. Getting those scanlines to display on a pixel perfect scale is vital for this and we can't fully trust a display's own scaler to do the job perfectly. (close though)

[naz]

* enables some new output configurations such as 4x240p/2x480p in standard 1080p frame

I did ask about framing 920p in a 1080p window frame awhile ago, so could this be what I reading?

If so, then this is the next step forward beyond collecting bvm's when they're no longer available. Getting those scanlines to display on a pixel perfect scale is vital for this and we can't fully trust a display's own scaler to do the job perfectly. (close though)

I kinda like upscaling 960 to 1080, scanlines aren't a hard transition from color to black, there is a bit of bleeding in them. Upscaling from 969 softens the edges making the transition closer to what i see on a crt.

[headlesshobbs]

Well some tv's aren't accurate with this as I learned from being an LG owner. Some lines end up thicker/thinner in patterns and getting anything that helps natively will offset the issue.

Doing light scanlines with hybrid has it's uses, but for that real hard look will require giving up going "full screen" if you want to beat the bvm.

[DiegoPonga]

Wonderful news!

There are a couple of things I'd like to ask though:

1) So far, with my OSSC v1.6, whenever I play classic games it is ×4'd, then I have to chose manually 4:3 in my TV menu and pray for it not destroy a 960p signal. If I understood correctly, now we'll have a black background in an 1080p frame, right? So no problems with resolution nor aspect ratio...

2) Since many classic consoles had a different refresh rate than the stantarized 59.94 fps, I was wondering... Is there any possibility of taking advantage of FreeSync/GSync technologies through new HDMI standards?

[orange808]

In any case, if theres some technical reason as to why this mode would induce blur on a CRT ( I dont believe it does with MISTer or Retroarch), the tradeoff would be worth it to me.

In any case, if theres some technical reason as to why this mode would induce blur on a CRT ( I dont believe it does with MISTer or Retroarch), the tradeoff would be worth it to me.

In any case, if theres some technical reason as to why this mode would induce blur on a CRT ( I dont believe it does with MISTer or Retroarch), the tradeoff would be worth it to me.

Did you mean to reply about this?

Nope.

[nmalinoski]

1) So far, with my OSSC v1.6, whenever I play classic games it is ×4'd, then I have to chose manually 4:3 in my TV menu and pray for it not destroy a 960p signal. If I understood correctly, now we'll have a black background in an 1080p frame, right? So no problems with resolution nor aspect ratio...

The way it reads, you'll have the option for the original 960p output, 960p windowboxed to 1080p, or 960p scaled to 1080p; it's not going to force you into windowboxed 1080p.

But that brings an interesting question: will we have aspect ratio controls, or will the OSSC Pro be able to automatically respond to changes in frame PAR?

As far as I can tell, the current OSSC can't tell the difference between a typical 4:3 frame and an anamorphic widescreen frame, so I need to switch between the two on my LCD when switching between full-screen and widescreen-capable games on PS2 or Xbox; but, when I connect my original Xbox to my FV310 CRT (via YPbPr), and then switch between 4:3 and 16:9 modes in the Xbox's main settings, the CRT automatically switches between normal and 16:9 aspect ratio adjustments, which means automatic widescreen detection is possible. I'm not entirely sure how, though; there's got to be some kind of metadata that the Xbox transmits that my CRT is picking up on.

2) Since many classic consoles had a different refresh rate than the stantarized 59.94 fps, I was wondering... Is there any possibility of taking advantage of FreeSync/GSync technologies through new HDMI standards?

If the HDMI TX chip is only 1.4-compatible, I would say that's doubtful. I think we'd need at least an HDMI 2.0 TX chip, and then FreeSync support would sit on top of that. GSync is likely out entirely due to its licensing and part cost.

[Link83]

when I connect my original Xbox to my FV310 CRT (via YPbPr), and then switch between 4:3 and 16:9 modes in the Xbox's main settings, the CRT automatically switches between normal and 16:9 aspect ratio adjustments, which means automatic widescreen detection is possible. I'm not entirely sure how, though; there's got to be some kind of metadata that the Xbox transmits that my CRT is picking up on.

I dont know what your setup is like, but if your using an Xbox SCART cable its likely using the SCART standard switch signal to inform the TV if the image is 4:3 or 16:9:-
https://en.wikipedia.org/wiki/SCART#Bla ... _switching
In other words it changes the voltage present on SCART pin 8 to signal the selected aspect ratio.

Of the three analog only consoles that provide a system menu aspect ratio option (PS2, Xbox, Wii) I think the Xbox might be the only console that uses this pin correctly and actually adjusts the voltage depending on the aspect ratio setting.

[energizerfellow‌]

Seems like it would be possible to build a good open source comb filter as well. Even better, it would share it's frame buffer with the deinterlacing and scaler.

The ADV7842 does an amazing job, but needs RAM directly attached to the ADV7842 for either the built-in 3D comb or frame TBC to function. Not really the architecture of the OSSC and kinda spendy. We're probably better off doing CVBS/YC -> YPbPr in the analog domain, then processing from there. I get why the ISL51002 is being used.

Speaking of the analog domain, composite/S-video to component conversion, and the Koryuu, we arguably need both a IRE 0.0/7.5 and a IRE 100/107.5 switches. There's a number of posts over at the VideoHelp forum talking about all kinds of devices with broken IRE range support, including having IRE 7.5, but rather than keeping white at IRE 100, shift the entire analog range up so white is 107.5 instead.

Pretty sure that's what the second scaling mode is, integer scaled then framed inside a 1080p window.

it's actually the third mode. OBS doesn't restrict its scaling to integer factors. "Free scaling" will essentially be the same as what you get on the Analogue FPGA units or using the Pixellate shader in Retroarch. I'm sure we'll get the option to combine integer on one axis with free scaling on the other.

Not sure if we're talking about the same thing here. Area scaling is basically bilinear scaling without filtering; it's functionally the same as point scaling one factor too high, and then doing a bilinear resample back down to the target, except in one pass.

I would love to see this.

[squarealex]

Amazing News

[fernan1234]

The ADV7842 does an amazing job, but needs RAM directly attached to the ADV7842 for either the built-in 3D comb or frame TBC to function. Not really the architecture of the OSSC and kinda spendy. We're probably better off doing CVBS/YC -> YPbPr in the analog domain, then processing from there. I get why the ISL51002 is being used.

This may be part of why the other chip combo is not being used, though I would bet onboard composite and S-video inputs would be a big selling point. There is no small amount of people getting Framemeisters and Retrotinks not just for the added features of the former or the lower price of the latter, but because they don't use RGB-modded systems, or use systems with no RGB mods available. A more commercially successful OSSC Pro would be good for everyone. Requiring add-on boards for NTSC decoding makes it less accessible and look more like a complicated DIY for the less tech-savvy users.

Speaking of the analog domain, composite/S-video to component conversion, and the Koryuu, we arguably need both a IRE 0.0/7.5 and a IRE 100/107.5 switches. There's a number of posts over at the VideoHelp forum talking about all kinds of devices with broken IRE range support, including having IRE 7.5, but rather than keeping white at IRE 100, shift the entire analog range up so white is 107.5 instead.

This is important indeed. Since all the old systems I use are Japanese analogue systems, I default to NTSC 0 instead of 7.5. Technically should also use D93 instead of D65 for the same reason, but that's another matter altogether.

[orange808]

The ADV7842 does an amazing job

Does it? I didn't know any new comb filters did a decent job. Color me skeptical.

Anyhow, filtering dot crawl has never been effectively performed without a frame buffer. It may introduce a few artifacts, but I'd still like to see a dot crawl filter that shares the deinterlacer frame buffer.

[Josh128]

Did you mean to reply about this?

Nope.

Cool.

[Hoagtech]

I want to get to the bottom of 4K expansion.

You have you cyclone fpga, ram, flash, and clock generator.

Which of these would need to be upgraded to achieve 4K when prices drop down?

Also instead of soldering the hardware to the pcb, couldn’t you install sockets so we could drop in the upgrades instead of soldering them when they become available?

The possibilities of the pro are already amazing but I thought I would ask @marqs

[maxtherabbit]

Nope.

If you were trying to make some kind of point, it made no sense to anyone but you

[Omnigamer]

I want to get to the bottom of 4K expansion.

You have you cyclone fpga, ram, flash, and clock generator.

Which of these would need to be upgraded to achieve 4K when prices drop down?

Also instead of soldering the hardware to the pcb, couldn’t you install sockets so we could drop in the upgrades instead of soldering them when they become available?

The possibilities of the pro are already amazing but I thought I would ask marqs

4K is first and foremost limited by the HDMI transmitter. Regardless of the power of anything else in the system, the transmitter chip is ultimately responsible for the HDMI clock limits the device as a whole can provide.

As for sockets, there's very little to be gained from socketizing some of the components. Especially for high pin-count parts, commercial sockets for many packages simply don't exist. It's also very unlikely that future parts that would add more features or performance would have an identical pinout or electrical characteristics such that they would be immediately compatible with the system.

Basically, this isn't the kind of device where swapping components makes sense.

[DiegoPonga]

The way it reads, you'll have the option for the original 960p output, 960p windowboxed to 1080p, or 960p scaled to 1080p; it's not going to force you into windowboxed 1080p.

Yes, but you can configure it to windowbox that 960p signal into a 1080p one, so this would be GREAT. For two main reasons: no problem with capture cards and no(t many) artifacts in TVs.

But that brings an interesting question: will we have aspect ratio controls, or will the OSSC Pro be able to automatically respond to changes in frame PAR?

As far as I can tell, the current OSSC can't tell the difference between a typical 4:3 frame and an anamorphic widescreen frame, so I need to switch between the two on my LCD when switching between full-screen and widescreen-capable games on PS2 or Xbox; but, when I connect my original Xbox to my FV310 CRT (via YPbPr), and then switch between 4:3 and 16:9 modes in the Xbox's main settings, the CRT automatically switches between normal and 16:9 aspect ratio adjustments, which means automatic widescreen detection is possible. I'm not entirely sure how, though; there's got to be some kind of metadata that the Xbox transmits that my CRT is picking up on.

As far as I can tell, the OSSC cannot know if it's 4:3 or 16:9 or whatever because there is no such thing. The signal does not tell you the size of each pixel, that's up to the screen. However, what you could do with this is setting pixel aspect ratio for each profile. So you load your Xbox profile and it's square pixels. You load the SNES profile and it may be square or that weird rectangular pixel proportions Nintendo put into that console.

If the HDMI TX chip is only 1.4-compatible, I would say that's doubtful. I think we'd need at least an HDMI 2.0 TX chip, and then FreeSync support would sit on top of that. GSync is likely out entirely due to its licensing and part cost.

Huh, I didn't know it was HDMI 1.4. Anway, this would be an interesting addition. Specially for some super competitive retro games, such as arcade fighting games which have a weird framerate. And also, in the short term, many of us are expecting new TVs with low latency and variable refresh rate.

[orange808]

Nope.

If you were trying to make some kind of point, it made no sense to anyone but you

Well, at least you quoted me and I found something I typed in the quote.

[tongshadow]

Any chance of adding a DB15 output for VGA monitors?
Really excited about this project, Micomsoft better hurry up

[marqs]

Selection of the video ADC was a hard choice between ISL51002 and ADV7842 - I hope I made the right call this time. ISL51002 has its own quircks but so far it has operated way better than TVP7002.

If you dont mind me asking, i'm kind of curious to know what the pros/cons were of each ADC chip?

I'm curious on the gritty details of the ISL51002 vs ADV7842 situation as well, especially considering the ISL51002 + ADV7611 + ADV7513 is a 3-chip solution, but the ADV7842 + ADV7511 is a 2-chip. The ADV7842 + ADV7511 also lets you do things like a built-in HDMI receiver with a 10/12-bit color pipeline at 225 MHz, 2D/3D comb filtering, frame TBC, CVBS/YC input support, CVBS monitor output, etc. You can even try the ADV7842 + ADV7511 combo with one of Analog Device's EVAL-ADV7842 boards (for about $220).

FYI that Magewell's analog capture cards, e.g. Pro Capture AIO, use the ADV7842 and Datapath's VisionRGB-E1S/VisionRGB-E2S cards use the older AD9887A. I've heard some rumblings that the Magewell cards may not have 240p support on CVBS/YC, which may be a limitation of the ADV7842 as the ADV7842 apparently forces the composite/S-video inputs into ITU-R BT.656?

Those are exactly among the concerns I have with ADV7842 - in such all-in-one chip, has the design focus really been on the analog frontend? At least I'm not getting that impression from the datasheet as first half of the general description just touts the HDMI receiver functionality. It'd be nice if all the blocks worked perfectly in there but I'm slightly doubtful despite the millions of dollars spent into development of such ASICs. Another reason is that if there's ever going to be a 4K edition, then that 225MHz HDMI receiver functionality (which doesn't even meet 1440p@60Hz requirement) just becomes dead weight as people certainly would want 4K HDMI input support then too. I rather selected ISL51002 which only has advanced analog frontend which is the most important criteria for the chip. ISL51002 can be also easily included in the DIY add-on board for DE10-Nano etc. as it's TQFP instead of BGA.

Any reason for the PCM1862 over the PCM1863 other than to save a few pennies? Since this device has the kind of performance to be interesting to the video guys, you may want to consider the PCM1863 as VHS Hi-Fi audio has 90+ dB of dynamic range, never mind LaserDisc. Everybody loves a nice low noise floor too.

That's actually a good point even though the difference in SNR is not that big. I might make the upgrade to PCM1863 indeed.

Seeing how this product is still in development. Couldn’t there be a switch made to a chip supporting at least a TMDS of 340MHz? This way the OSSC Pro can support 1440p. That would allow for 6x240p, 3x480p and 2x720p. Should be a real game changer. I’m sure most of us are willing to pay the premium.

Can you point me to such chip? The closest I know is SiI9136-3 with 300MHz, but the programming guide requires a NDA so it's immediately a no-go for this project. I'm not even sure if the FPGA would be capable of driving such high rate. The proper way for generating 1440p/4K 60+Hz would be with a FPGA which has transceivers that can directly output HDMI, but then we're talking about very different price levels (not just the FPGA chips, but development tools as well).

If the HDMI TX chip is only 1.4-compatible, I would say that's doubtful. I think we'd need at least an HDMI 2.0 TX chip, and then FreeSync support would sit on top of that. GSync is likely out entirely due to its licensing and part cost.

I think 1.4 are the last generation of dedicated HDMI transmitters as it's more practical to intergrate the high-speed HDMI IP into ASIC/FPGA itself. HDMI VRR requires v2.1 but Freesync has been a while in DisplayPort specifications.

[strayan]

How are Panasonic able to pull off their 1080p to 2160p (Pixel by 4 Pixel mode) integer scale on their TV’s I wonder?

[Josh128]

Nope.

If you were trying to make some kind of point, it made no sense to anyone but you

Well, at least you quoted me and I found something I typed in the quote.

You quoted what I said 3 times in the same post. You didnt type anything, you just quoted me 3 separate times. All I was doing is asking if you intended to do that. If that was not a mistake, what was your point?

[dead_screem]

Is there any chance to have a model that does away with the SHART connector completely and uses RGBS BNC connectors instead? I have a single RGBS BNC shielded coaxial cable for my Mega Drive but don't really want to invest in more for other systems if I would still have to use a BNC->SHART adapter.

Also will it have internal RF shielding at all?