OSSC (DIY video digitizer & scandoubler)
-
FinalBaton
- Posts: 4461
- Joined: Sun Mar 08, 2015 10:38 pm
- Location: Québec City
-
FinalBaton
- Posts: 4461
- Joined: Sun Mar 08, 2015 10:38 pm
- Location: Québec City
Re: DIY video digitizer & scandoubler
Is the latency still around 2 scanlines? I remember reading that in the few first pages, but at the time the project was scaling to 480p.
Now that it's doing 720p instead, as the latency increased?
Now that it's doing 720p instead, as the latency increased?
-FM Synth & Black Metal-
Re: DIY video digitizer & scandoubler
2 scanlines is the absolute maximum when interlaced content is processed and when there's need to shift a field by one line in respect to the other field. In normal linedouble, and also with linetriple (240p->720p), latency is less than one scanline. Minimal lag is a top priority in this projectFinalBaton wrote:Is the latency still around 2 scanlines? I remember reading that in the few first pages, but at the time the project was scaling to 480p.
Now that it's doing 720p instead, as the latency increased?
Re: DIY video digitizer & scandoubler
Awesome to hear. Do you have a rough launch date and/or price as of yet? Also is there a wait list I need to get on? Thanks!
-
FinalBaton
- Posts: 4461
- Joined: Sun Mar 08, 2015 10:38 pm
- Location: Québec City
Re: DIY video digitizer & scandoubler
Wow! this is an incredibly interesting projectmarqs wrote:2 scanlines is the absolute maximum when interlaced content is processed and when there's need to shift a field by one line in respect to the other field. In normal linedouble, and also with linetriple (240p->720p), latency is less than one scanline. Minimal lag is a top priority in this projectFinalBaton wrote:Is the latency still around 2 scanlines? I remember reading that in the few first pages, but at the time the project was scaling to 480p.
Now that it's doing 720p instead, as the latency increased?
Good 720p upscaling with no lag? sign me up!
-FM Synth & Black Metal-
Re: DIY video digitizer & scandoubler
There's some discussion on the material costs on the previous pages. I checked PCB assembly price estimates from a few companies, and they are somewhat reasonable if you make at least 50 or 100 boards. I'd estimate that a complete board in a plexiglass case could be initially sold at around $150. I'll put up a waiting list and more details soon to get some idea how many people would be interested to order one.geauxwave wrote:Awesome to hear. Do you have a rough launch date and/or price as of yet? Also is there a wait list I need to get on? Thanks!
I've recently made some tests with different sync types to make the system compatible with as many systems/cables as possible. Below is a short summary of my thoughts and technical details (sorry if there's too much of it) of the possible sync formats at each supported connector. Please comment if you have any ideas or think there's something incorrect.
- 3xRCA connector: Component (YPbPr) or RGB Sync-on-Green (RGsB)
The easiest input to work with. A nominal signal has 0.3Vpp level sync with 0.7Vpp luma or green signal on top of it when connected to a 75ohm load.
VGA connector: RGBHV, RGBS or RGB Sync-on-Green (RGsB)
A bit harder input to work with as there are several possible formats. RGBHV should be fairly standard (5V Hsync+Vsync, high impedance load), but some sources may output Csync+Vsync instead (e.g. Sync strike). That can cause problems in RGBHV mode, and RGBS mode should be used instead. A source outputting Csync only won't cause any issues as it does not get even detected in RGBHV mode. RGsB works as described earlier, so it's not a problem either.
SCART connector: Sync on composite video/luma, Csync or RGB Sync-on-Green (RGsB)
This is the most annoying case, because very different signals may be output to a single pin (excluding RGsB), and the system should be able to accept all of them. Sources outputting sync on composite video/luma are fairly standard (0.3Vpp level sync on 75ohm load) and safe to connect to an analog sync input of TVP7002. However, many people (including myself) prefer to use pure composite sync to get cleanest possible sync and avoid potential noise issues of composite which can be obvious especially with lower quality cables. The problem is that csync outputs of consoles are typically meant for TTL level sync inputs of monitors, and are not designed to be directly connected to 75ohm load found at the SCART inputs of TVs, Framemeister etc. Theoretically, a 5V csync output can damage the sink (TVP7002 datasheet e.g. specifies 2.3V as the absolute maximum), but in practice the effective voltage usually drops to a safe range as the driving strength of the Csync output of consoles is limited. This also causes unnecessary load to the video IC in the consoles. A "correct" way would be to use csync with a series resistor inside cable or console, but many cables and mod guides leave that out. To protect the system, I planned on adding a sync separator chip which can handle up to 5V before TVP7002 on the next PCB revision.
Re: DIY video digitizer & scandoubler
I've got a few questions.
- Will you send out a couple units to beta testers before you send out boards to regular users? I'd buy one instantly for $150, but it would be nice to have someone like Fudoh verify that it syncs correctly with his consoles and displays.
- Will it be possible to update the firmware through the JTAG interface?
- Could you add a passthrough for SCART audio? Just a simple RCA or 3.5mm output next to the HDMI/DVI plug.
- Will you send out a couple units to beta testers before you send out boards to regular users? I'd buy one instantly for $150, but it would be nice to have someone like Fudoh verify that it syncs correctly with his consoles and displays.
- Will it be possible to update the firmware through the JTAG interface?
- Could you add a passthrough for SCART audio? Just a simple RCA or 3.5mm output next to the HDMI/DVI plug.
Re: DIY video digitizer & scandoubler
And someone who's got bitchy pcb's like the Seibu stuff, Irem, Xexex, MK, etc.blizzz wrote:- Will you send out a couple units to beta testers before you send out boards to regular users? I'd buy one instantly for $150, but it would be nice to have someone like Fudoh verify that it syncs correctly with his consoles and displays.
Strikers1945guy wrote:"Do we....eat chicken balls?!"
-
bobrocks95
- Posts: 3472
- Joined: Mon Apr 30, 2012 2:27 am
- Location: Kentucky
Re: DIY video digitizer & scandoubler
On the RCA front, would it be simple enough to use a SCART to 4 RCA adapter (RGBs essentially, but as RCA plugs) then an RCA splitter to make SoG? Would the same concerns with voltages from csync cables still apply or does this alleviate the problem?
I'm all for RCA since the adapters are fairly cheap and it offers YPbPr as another input option. Depending on price, however, I would assume most also want a SCART connector? Ignoring the problems with implementation, what would including both connectors mean from a price/manufacturing standpoint?
I'm all for RCA since the adapters are fairly cheap and it offers YPbPr as another input option. Depending on price, however, I would assume most also want a SCART connector? Ignoring the problems with implementation, what would including both connectors mean from a price/manufacturing standpoint?
PS1 Disc-Based Game ID BIOS patch for MemCard Pro and SD2PSX automatic VMC switching.
Re: DIY video digitizer & scandoubler
Yes, the current board has already been in the hands of 2 testers so that I've received some feedback, and one unit is going to Fudoh quite soon.blizzz wrote:- Will you send out a couple units to beta testers before you send out boards to regular users? I'd buy one instantly for $150, but it would be nice to have someone like Fudoh verify that it syncs correctly with his consoles and displays.
There's both microSD and JTAG interfaces which can be used for updating FW.blizzz wrote:- Will it be possible to update the firmware through the JTAG interface?
Yeah, I've thought about adding that next to the DVI connector in the release version.blizzz wrote:- Could you add a passthrough for SCART audio? Just a simple RCA or 3.5mm output next to the HDMI/DVI plug.
Re: DIY video digitizer & scandoubler
SoG is not an issue as its sync level is only 0.3Vpp. The problem is with SCART pin 20, which should be able to cope with csync (from 0.3Vpp to 5Vpp) and composite video, so an adapter unfortunately does not help there.bobrocks95 wrote:On the RCA front, would it be simple enough to use a SCART to 4 RCA adapter (RGBs essentially, but as RCA plugs) then an RCA splitter to make SoG? Would the same concerns with voltages from csync cables still apply or does this alleviate the problem?
The connectors are fairly cheap and both fit well to the current PCB, so there's no reason not to include both.bobrocks95 wrote:I'm all for RCA since the adapters are fairly cheap and it offers YPbPr as another input option. Depending on price, however, I would assume most also want a SCART connector? Ignoring the problems with implementation, what would including both connectors mean from a price/manufacturing standpoint?
-
bobrocks95
- Posts: 3472
- Joined: Mon Apr 30, 2012 2:27 am
- Location: Kentucky
Re: DIY video digitizer & scandoubler
I replied long after I read the post, so I missed that all 3 are still planned to be on there, and you were just posting technical musings on the three. My bad.
PS1 Disc-Based Game ID BIOS patch for MemCard Pro and SD2PSX automatic VMC switching.
Re: DIY video digitizer & scandoubler
I know that RetroRGB and myself would be interested in reselling these and we could help fund the first batch if it keeps costs down. I'm also interested in being a beta tester, obviously
OSSC Forums - http://www.videogameperfection.com/forums
Please check the Wiki before posting about Morph, OSSC, XRGB Mini or XRGB3 - http://junkerhq.net/xrgb/index.php/Main_Page
Please check the Wiki before posting about Morph, OSSC, XRGB Mini or XRGB3 - http://junkerhq.net/xrgb/index.php/Main_Page
Re: DIY video digitizer & scandoubler
Which input is preferred for Jamma PCBs, which afaik often have 5vpp CSYNC(?). Scart?
Would adding a series resistor onboard, acting as a voltage divider with the internal resistance of the IC work? Maybe enable through a dip switch or something.
Would adding a series resistor onboard, acting as a voltage divider with the internal resistance of the IC work? Maybe enable through a dip switch or something.
Re: DIY video digitizer & scandoubler
VGA input (hsync pin) would be best for 5Vpp CSYNC, because VGA sync pins go directly to a digital buffer without loading the source. A selectable series resistor on the SCART sync pin could help in those rare cases where the effective voltage would otherwise get over the limit where sync separators such as LM1881 work reliably (~2Vpp).Buffi wrote:Which input is preferred for Jamma PCBs, which afaik often have 5vpp CSYNC(?). Scart?
Would adding a series resistor onboard, acting as a voltage divider with the internal resistance of the IC work? Maybe enable through a dip switch or something.
Re: DIY video digitizer & scandoubler
This may be a dumb question, but why would you need to separate the sync with a LM1881 if using SCART input? Can't you just feed the CSYNC (possibly with voltage shifted down to a lower value) directly to SOGIN_1, which I think is what your schematic is doing. Is there a benefit in splitting up the CSYNC?
edit: Ah, just saw your post on Fri Aug 07. Makes sense.
If the main issue there is the voltage level, might a buffer/line driver like 74LVC1G17 operating at 1.9vcc be sufficient? That should map high voltages at about 1v-5v to the vcc level of 1.9vcc which the tvp7002 chip should be able to handle. It wont support 0.3vpp sync though.
edit: Ah, just saw your post on Fri Aug 07. Makes sense.
If the main issue there is the voltage level, might a buffer/line driver like 74LVC1G17 operating at 1.9vcc be sufficient? That should map high voltages at about 1v-5v to the vcc level of 1.9vcc which the tvp7002 chip should be able to handle. It wont support 0.3vpp sync though.
Re: DIY video digitizer & scandoubler
That's the problem. A board should cope with composite video with 0.3Vpp sync (and possibly including subcarrier, Macrovision etc.) if it has a SCART connector, so a digital buffer would not make it very compatible.Buffi wrote:If the main issue there is the voltage level, might a buffer/line driver like 74LVC1G17 operating at 1.9vcc be sufficient? That should map high voltages at about 1v-5v to the vcc level of 1.9vcc which the tvp7002 chip should be able to handle. It wont support 0.3vpp sync though.
Speaking of sync splitters, I've been wondering why exactly adding them at the end of cables help in removing crosshatch/interference patterns many people report. Is it perhaps because the cables including one don't have a 75ohm load resistor in there, in which case less current flows on the composite wire? Or does the composite signal cause noise actually inside the digitizer chip found in displays etc. when no splitter is used?
Re: DIY video digitizer & scandoubler
marqs,
Directly connecting TTL sync to the composite video input is a bad practice and doesn't need to be supported or encouraged. A resistor needs to be inserted in series with the TTL signal at the consoles end of the cable, not the SCART end, otherwise buzz in the audio is pretty much guaranteed (from ground difference + capacitive coupling). If you're worried about damage to the sync separator the usual protection method is a low value series resistor and a voltage clamp after it (zener, TVS, etc).
Directly connecting TTL sync to the composite video input is a bad practice and doesn't need to be supported or encouraged. A resistor needs to be inserted in series with the TTL signal at the consoles end of the cable, not the SCART end, otherwise buzz in the audio is pretty much guaranteed (from ground difference + capacitive coupling). If you're worried about damage to the sync separator the usual protection method is a low value series resistor and a voltage clamp after it (zener, TVS, etc).
If you have a bad ground connection then terminating the video input results in a bit of chroma getting into all the signals via the ground resistance. If you don't terminate, the composite video signal is larger and couples better through the cable capacitance (also, you get reflections and noise from not terminating). Not a good solution either way.marqs wrote: Speaking of sync splitters, I've been wondering why exactly adding them at the end of cables help in removing crosshatch/interference patterns many people report. Is it perhaps because the cables including one don't have a 75ohm load resistor in there, in which case less current flows on the composite wire? Or does the composite signal cause noise actually inside the digitizer chip found in displays etc. when no splitter is used?
Re: DIY video digitizer & scandoubler
Yes, a resistor+diode clamp in parallel to the termination resistor would protect the chip in all but extreme cases (e.g. 5V supply accidentally connected to composite wire).viletim wrote:Directly connecting TTL sync to the composite video input is a bad practice and doesn't need to be supported or encouraged. A resistor needs to be inserted in series with the TTL signal at the consoles end of the cable, not the SCART end, otherwise buzz in the audio is pretty much guaranteed (from ground difference + capacitive coupling). If you're worried about damage to the sync separator the usual protection method is a low value series resistor and a voltage clamp after it (zener, TVS, etc).
That's why I don't see how adding a sync splitter (especially at the display end of the cable) helps. I haven't personally had any such problems (and I usually use csync anyway), but I've heard those kind of cables have solved the issue for many people.viletim wrote: If you have a bad ground connection then terminating the video input results in a bit of chroma getting into all the signals via the ground resistance. If you don't terminate, the composite video signal is larger and couples better through the cable capacitance (also, you get reflections and noise from not terminating). Not a good solution either way.
Re: DIY video digitizer & scandoubler
Out of curiosity, have you experimented anything with the internal 6.5MHz clock reference, instead of using an external oscillator?
For low-res stuff like this, I would guess that it would be possible to just use the internal reference, and save some board space + components.
For low-res stuff like this, I would guess that it would be possible to just use the internal reference, and save some board space + components.
Re: DIY video digitizer & scandoubler
I tried it when I was still using the FPGA development board, and I don't remember any major issues with it. However, the datasheet recommends using an external oscillator so that's why I ended up adding one onboard. It can be also used as a fixed clock for some logic on the FPGA (CPU etc.), so it's not necessarily wasted spaceBuffi wrote:Out of curiosity, have you experimented anything with the internal 6.5MHz clock reference, instead of using an external oscillator?
For low-res stuff like this, I would guess that it would be possible to just use the internal reference, and save some board space + components.
Re: DIY video digitizer & scandoubler
Internal RC oscillators are not as accurate and have a greater drift than external dedicated ones. Stick with an external onemarqs wrote:I tried it when I was still using the FPGA development board, and I don't remember any major issues with it. However, the datasheet recommends using an external oscillator so that's why I ended up adding one onboard. It can be also used as a fixed clock for some logic on the FPGA (CPU etc.), so it's not necessarily wasted spaceBuffi wrote:Out of curiosity, have you experimented anything with the internal 6.5MHz clock reference, instead of using an external oscillator?
For low-res stuff like this, I would guess that it would be possible to just use the internal reference, and save some board space + components.
http://world-of-arcades.net
The future of ST-V rests upon our work and your work
The future of ST-V rests upon our work and your work
Re: DIY video digitizer & scandoubler
marqs: If you need help beta-testing this very interesting device, just let me know.. I'm more than happy to help here in .fi Recently I've been playing quite a lot with properly upscaling oldschool RGB signals using a lot of different scalers/processors.
Re: DIY video digitizer & scandoubler
PM sent.Nrg wrote:marqs: If you need help beta-testing this very interesting device, just let me know.. I'm more than happy to help here in .fi Recently I've been playing quite a lot with properly upscaling oldschool RGB signals using a lot of different scalers/processors.
Based on the feedback from current testers, I've made a few changes on the PCB design which add the following features:
- -2 physical buttons, so basic functions like input selection can be quickly accessed without remote control.
-Component and SCART video signals will be routed via a video buffer which has a configurable low-pass filter (9MHz, 16MHz, 35MHz, bypass). That should help with noisy sources.
-Sync-on-Green support for SCART input. Useful for PS2 users.
Re: DIY video digitizer & scandoubler
Unless the PCB design is finalized already, could you add some nice solder points for these buttons so we can add extension wires if we want to add your device inside a supergun to get these buttons where they're needed (not on the inside of a box, lol)marqs wrote:
- -2 physical buttons, so basic functions like input selection can be quickly accessed without remote control.
| My games - http://www.emphatic.se | (Click) I have YEN stickers for sale
RegalSin wrote:Street Fighters. We need to aviod them when we activate time accellerator.
-
bobrocks95
- Posts: 3472
- Joined: Mon Apr 30, 2012 2:27 am
- Location: Kentucky
Re: DIY video digitizer & scandoubler
*Ahem* any word on distribution partnerships to keep costs down?marqs wrote:Based on the feedback from current testers, I've made a few changes on the PCB design which add the following features:
I'm sending the updated PCB for manufacturing next week. If everything will work as intended, then it's very straigtforward to finally turn that board into public release version .
- -2 physical buttons, so basic functions like input selection can be quickly accessed without remote control.
-Component and SCART video signals will be routed via a video buffer which has a configurable low-pass filter (9MHz, 16MHz, 35MHz, bypass). That should help with noisy sources.
-Sync-on-Green support for SCART input. Useful for PS2 users.
PS1 Disc-Based Game ID BIOS patch for MemCard Pro and SD2PSX automatic VMC switching.
Re: DIY video digitizer & scandoubler
marqs: I live here in Finland and I'm also interested testing the scaler you're developing. I'm retro hobbyist and I collect 8 & 16 bit consoles and computers. Please, send me PM if you need any help with the testing.
Re: DIY video digitizer & scandoubler
Possibly. However, the buttons are on PCB edge and they are through-hole -type, so it's easy to solder extension wires to their legs even without dedicated solder points.emphatic wrote:Unless the PCB design is finalized already, could you add some nice solder points for these buttons so we can add extension wires if we want to add your device inside a supergun to get these buttons where they're needed (not on the inside of a box, lol)
There have been a few proposals. The first batch (~50 boards) is still likely to go without any major partnerships in order to test the waters without putting possible partners at risk. If the release is successful, then distribution partnerships are definitely on agenda.bobrocks95 wrote:*Ahem* any word on distribution partnerships to keep costs down?
Re: DIY video digitizer & scandoubler
black frame insertion could be an option ?
I had good results with groovymame and 120hz displays.
I had good results with groovymame and 120hz displays.
Re: DIY video digitizer & scandoubler
That would require buffering at least half a frame, and there's not enough memory for that. Hopefully in a few years there's low-persistence OLED TVs on the market which would have no motion blur. Ideally they could be driven at 60Hz which would be perfect for retro gamers, but I'm afraid manufactures will limit their internal refresh rates to 120Hz or higher to avoid flicker.akumajo wrote:black frame insertion could be an option ?
I had good results with groovymame and 120hz displays.