Cloning the GameCube component cable

Portables, case replacements, mods etc, all in here!
User avatar
Unseen
Posts: 190
Joined: Fri Jul 04, 2014 11:52 am

Re: Cloning the GameCube component cable

Post by Unseen » Thu Jul 16, 2015 5:55 pm

andre104623 wrote:I wanted to ask as well when you first came out with the GCvideo-DVI code did you have a other PCB you were running the code on besides the pluto like you Video-lite board.
Yes, the code was initially developed on a Digilent Atlys. I haven't hooked it up for quite some time though because re-wiring the cube from one board to the other is quite a hassle.
I was looking at the datasheet for the 200A and the pinouts are quite different then the 50A and was the reason I didn't order it.
The datasheet (DS529) seems to claim on page 70 that the pinout of the 50A and 200A are the same except for a few differential signals. The only thing that uses differential signalling on the Shuriken Video board is the HDMI output, but that uses pins that are identical, so my current assumption is that the 200A should work on that board.
Asking for support by PM is anti-social. Ask in an open forum instead, so other people can benefit from the answers!
happy_bunny
Posts: 106
Joined: Mon Mar 09, 2015 10:57 pm

Re: Cloning the GameCube component cable

Post by happy_bunny » Thu Jul 16, 2015 6:40 pm

The input pin config would need changing for the shuriken video board though. By the way i really like the way you design your vhdl code its really moduler its great.
meneerbeer
Posts: 212
Joined: Wed Sep 03, 2014 9:13 am

Re: Cloning the GameCube component cable

Post by meneerbeer » Sat Jul 18, 2015 2:16 pm

I am trying out the p2x code. I get a red screen. Most likely something is wired up incorrectly, because I have the idea to yuv to rgb is always active?

Currently looking what is going wrong, but perhaps you have an idea, Unseen?

Edit: never mind it is working now. The inversion of the TMDS was probably incorrect. I already tried changing the code to set it correctly to try and fix the issue. I guess I flashed an old file to the SPI.

Gotta go somewhere now, can't wait to play around with this. The OSD is really cool. Thanks Unseen! :D
Attachments
20150718_155008.jpg
(1.29 MiB) Not downloaded yet
User avatar
Unseen
Posts: 190
Joined: Fri Jul 04, 2014 11:52 am

Re: Cloning the GameCube component cable

Post by Unseen » Sat Jul 18, 2015 2:41 pm

meneerbeer wrote:I am trying out the p2x code. I get a red screen. Most likely something is wired up incorrectly, because I have the idea to yuv to rgb is always active?
Yes, since the p2x outputs a DVI signal it has to convert the output to RGB.
Currently looking what is going wrong, but perhaps you have an idea, Unseen?
My first guess would be something wrong on the input side because as you mentioned the colors do look more like YUV problems than RGB problems. Since there is a stable picture, at least Clock, VData 4 and VData 5 are connected correctly (HSync/VSync, muxed within the pixel data). CSel might be involved in the problem, but it would be connected to a wrong signal and not just open or shorted to ground/supply. Another possibility could be a mis-wiring of the highest VData lines (6/7) so that it is seen as constant/mostly-constant 1 during the picture area, although it can't be miswired to the adjacent 3.3V pin because then you wouldn't get a picture at all (all-zero Y component is used to mark that the cube is transmitting Sync+Flags instead of picture).

If you want to check that it's not a problem of the board, you could try to flash KNJNs Pong demo from the StartupKit zip, subfolder "FPGA project - HDMI". It doesn't matter if the board is connected to the cube or not to run that demo, choose the method that is more convenient for you.

(which reminds me that this might be a good test for andre104623's board too...)
Asking for support by PM is anti-social. Ask in an open forum instead, so other people can benefit from the answers!
andre104623
Posts: 694
Joined: Wed May 07, 2014 2:24 pm

Re: Cloning the GameCube component cable

Post by andre104623 » Sat Jul 18, 2015 8:42 pm

Unseen wrote:
meneerbeer wrote:I am trying out the p2x code. I get a red screen. Most likely something is wired up incorrectly, because I have the idea to yuv to rgb is always active?
Yes, since the p2x outputs a DVI signal it has to convert the output to RGB.
Currently looking what is going wrong, but perhaps you have an idea, Unseen?
My first guess would be something wrong on the input side because as you mentioned the colors do look more like YUV problems than RGB problems. Since there is a stable picture, at least Clock, VData 4 and VData 5 are connected correctly (HSync/VSync, muxed within the pixel data). CSel might be involved in the problem, but it would be connected to a wrong signal and not just open or shorted to ground/supply. Another possibility could be a mis-wiring of the highest VData lines (6/7) so that it is seen as constant/mostly-constant 1 during the picture area, although it can't be miswired to the adjacent 3.3V pin because then you wouldn't get a picture at all (all-zero Y component is used to mark that the cube is transmitting Sync+Flags instead of picture).

If you want to check that it's not a problem of the board, you could try to flash KNJNs Pong demo from the StartupKit zip, subfolder "FPGA project - HDMI". It doesn't matter if the board is connected to the cube or not to run that demo, choose the method that is more convenient for you.

(which reminds me that this might be a good test for andre104623's board too...)
I will give it a try when I get around to it.
meneerbeer
Posts: 212
Joined: Wed Sep 03, 2014 9:13 am

Re: Cloning the GameCube component cable

Post by meneerbeer » Sun Jul 19, 2015 2:38 pm

Unseen wrote:
meneerbeer wrote:I am trying out the p2x code. I get a red screen. Most likely something is wired up incorrectly, because I have the idea to yuv to rgb is always active?
Yes, since the p2x outputs a DVI signal it has to convert the output to RGB.
Currently looking what is going wrong, but perhaps you have an idea, Unseen?
My first guess would be something wrong on the input side because as you mentioned the colors do look more like YUV problems than RGB problems. Since there is a stable picture, at least Clock, VData 4 and VData 5 are connected correctly (HSync/VSync, muxed within the pixel data). CSel might be involved in the problem, but it would be connected to a wrong signal and not just open or shorted to ground/supply. Another possibility could be a mis-wiring of the highest VData lines (6/7) so that it is seen as constant/mostly-constant 1 during the picture area, although it can't be miswired to the adjacent 3.3V pin because then you wouldn't get a picture at all (all-zero Y component is used to mark that the cube is transmitting Sync+Flags instead of picture).

If you want to check that it's not a problem of the board, you could try to flash KNJNs Pong demo from the StartupKit zip, subfolder "FPGA project - HDMI". It doesn't matter if the board is connected to the cube or not to run that demo, choose the method that is more convenient for you.

(which reminds me that this might be a good test for andre104623's board too...)
Ouch, quite a long post, when it was just me flashing the wrong file.. Sorry about that.

I have made a custom PCB, which is why I had to modify the code. I uploaded an album with some images here. I still need to get screws with ring nuts to secure it better. Right now I just took two screws from the DVD drive and stuck them through to at least keep it a bit in place. Also, I kind of screwed up the opening for the HDMI connector..

Everything seems to work fine. Some games look really good (Smash Bros and Mario Kart for instance). Other games do not really look that great. In Wind Waker I notice a pattern with stripes. My guess is that this is normal (dithering?). With analog cables I suppose you do not really notice this. It looks kind of bad though. I added a zoomed in image of Wind Waker to the album, so you can see what I mean.

My GC is PAL and not modded, so I can only try interlaced video modes. Interlaced video over HDMI works great for my TV luckily. I am eager to try out true 480p and 576p. If you go as far as doing this mod, it is a bit stupid to only use interlaced video.

I tried out the line doubler. It seems to give a "jumpy" image. As if every other frame the image jumps down by one line. It definitely happens at the top and bottom line, but I have the idea that the whole image is jumping up and down. I think this can be fixed, but then we first need to be certain about what causes it.

Scanlines are working great for me for linedoubled 576i. For linedoubled 480i they are barely noticeable. They only seem to be noticeable when there is movement on the screen. I think this also has to do with the above issue with the jumpy image.

I do not really care too much about the linedoubling issues, since I do not really plan on using it.

Oh, one other thing. It seems you do not set the active video you send over HDMI as 720x480/720x576. Perhaps some TVs get confused by this because they expect a line to have 720 active pixels (that will give black bars). Although it is definitely not the problem that is causing the p2x to not work with so many TVs.

Other than that, it is really nice. I would have never put this much effort into it. I probably would have only decoded the video signal to HDMI and been too lazy to add all these features.
novenary
Posts: 1756
Joined: Mon Dec 30, 2013 7:50 am

Re: Cloning the GameCube component cable

Post by novenary » Sun Jul 19, 2015 3:48 pm

If you have a spare atmega8 (or a raspberry pi :P) you can make a xeno pretty easily to try out Swiss. Otherwise I'd recommend that you make yourself an IPL replacement, it's not hard and it's much nicer than the xeno.
User avatar
Unseen
Posts: 190
Joined: Fri Jul 04, 2014 11:52 am

Re: Cloning the GameCube component cable

Post by Unseen » Sun Jul 19, 2015 6:13 pm

meneerbeer wrote:In Wind Waker I notice a pattern with stripes. My guess is that this is normal (dithering?).
Interesting. I just ran around the starting island in the PAL version a bit and I didn't notice any artifacts like that, but I'm not sure if I had that particular shade of brown on my screen while doing that.
I tried out the line doubler. It seems to give a "jumpy" image. As if every other frame the image jumps down by one line. It definitely happens at the top and bottom line, but I have the idea that the whole image is jumping up and down. I think this can be fixed, but then we first need to be certain about what causes it.
That's normal - it is only a simple line doubler which is not really appropiate for interlaced images and thus creates a slightly "jumpy" image. Proper deinterlacing would require an external RAM chip to store a complete field, but the deinterlacing algorithms in modern TVs are likely better than anything I can easily implement on my own.
Scanlines are working great for me for linedoubled 576i. For linedoubled 480i they are barely noticeable. They only seem to be noticeable when there is movement on the screen. I think this also has to do with the above issue with the jumpy image.
If you want noticable scanlines with interlaced modes, try to disable the "alternating scanlines" option. It's on by default because the resulting effect is closer to the behaviour of a real CRT, where you usually don't notice scanlines that much with interlaced content.
Oh, one other thing. It seems you do not set the active video you send over HDMI as 720x480/720x576.
That's slightly more complicated than it may appear. The video signal that is sent for 15kHz modes has an active area of 1440 pixels instead of 720. Some monitors/scalers/etc show the actual horizontal resolution, some claim to see 720 pixels. The size of the active area would be exactly the same with a "proper" HDMI signal, the only difference is that HDMI would signal in an info frame that each pixel is sent twice. Pixel doubling is required in this case because HDMI and DVI have a minimum pixel clock of 25 MHz, but non-doubled 480i/576i needs just 13.5MHz.

For games/apps running in 480p/576p you should see a "proper" 720x480/720x576 active area.
Other than that, it is really nice. I would have never put this much effort into it. I probably would have only decoded the video signal to HDMI and been too lazy to add all these features.
Linedoubling went in because some of my HDMI sinks don't like 720x240p signals and the interpolation on those that do tends to look quite ugly with pixel art. Optional scanlines were a no-brainer after that and the OSD went in because the jumper configuration (in the initial release) for those two features was hard to explain in the documentation.
Streetwalker wrote:Otherwise I'd recommend that you make yourself an IPL replacement, it's not hard and it's much nicer than the xeno.
Got a link for that? The only one I know about is the old tmbinc release, but IIRC that didn't come with any Cube-side software.
Asking for support by PM is anti-social. Ask in an open forum instead, so other people can benefit from the answers!
bobrocks95
Posts: 161
Joined: Fri Jul 26, 2013 11:19 pm

Re: Cloning the GameCube component cable

Post by bobrocks95 » Sun Jul 19, 2015 6:45 pm

The vertical lines in Wind Waker are there with the official component cable, and through a Wii with component cables. I think Skyward Sword has the same problem, though I haven't played more than an hour or so of it.

I noticed it when I played on a CRT and I noticed it when I played on a 480p plasma. My best guess is that it's some weird engine bug. I mostly only saw it on very solid colors, which Wind Waker happens to have a lot of.
meneerbeer
Posts: 212
Joined: Wed Sep 03, 2014 9:13 am

Re: Cloning the GameCube component cable

Post by meneerbeer » Sun Jul 19, 2015 7:48 pm

Streetwalker wrote:If you have a spare atmega8 (or a raspberry pi :P) you can make a xeno pretty easily to try out Swiss.
I have both. ;) I suppose I still need to burn a CD with Swiss on it? I am kind of new to GC modding. I do have a modded GC lying here, but it is a white model. I guess I could swap drives, but then I still need to get CDs. I'd rather not go the CD way.
Streetwalker wrote: Otherwise I'd recommend that you make yourself an IPL replacement, it's not hard and it's much nicer than the xeno.
How does that work, is it expensive? If I can somehow boot Swiss it should be a piece of cake, I still have a SD gecko that I used with my Wii back in the day.
Unseen wrote:Interesting. I just ran around the starting island in the PAL version a bit and I didn't notice any artifacts like that, but I'm not sure if I had that particular shade of brown on my screen while doing that.
When I google for "Wind Waker dithering" I find similar complaints. I guess it is normal then. I notice it in Mario Sunshine too. The intro movie looks like crap too. Of course I am still only using interlaced input video, so with progressive it might get a bit better again.
Unseen wrote:That's normal - it is only a simple line doubler which is not really appropiate for interlaced images and thus creates a slightly "jumpy" image. Proper deinterlacing would require an external RAM chip to store a complete field, but the deinterlacing algorithms in modern TVs are likely better than anything I can easily implement on my own.
I would need to look better into this. I think it is possible to at least get rid of the jumpy image. I suppose both the even and odd field have an equal amount of lines, so I think for one of the fields you need to delay by one line. I think that means you need an extra buffer to hold a line, so it may not fit anymore.
Unseen wrote:If you want noticable scanlines with interlaced modes, try to disable the "alternating scanlines" option. It's on by default because the resulting effect is closer to the behaviour of a real CRT, where you usually don't notice scanlines that much with interlaced content.
That explains the problem. The effect with alternating scanlines was quite interesting though. But I think it would also give me a headache at some point. ;)
Unseen wrote:That's slightly more complicated than it may appear. The video signal that is sent for 15kHz modes has an active area of 1440 pixels instead of 720. Some monitors/scalers/etc show the actual horizontal resolution, some claim to see 720 pixels. The size of the active area would be exactly the same with a "proper" HDMI signal, the only difference is that HDMI would signal in an info frame that each pixel is sent twice. Pixel doubling is required in this case because HDMI and DVI have a minimum pixel clock of 25 MHz, but non-doubled 480i/576i needs just 13.5MHz.
I can put my TV to 4:3. I have a setting on my TV that I can make the resulting bars grey. I was expecting to still see at least some black bars around the picture, because the games do not use the full 720 pixels width. There were no black bars, so my guess is that you send a different horizontal active video area than 720 pixels (I believe the GC sends a flag that holds whether there is active video data). It is not really an issue, but maybe there are TVs that expect 720 active pixels per line.
novenary
Posts: 1756
Joined: Mon Dec 30, 2013 7:50 am

Re: Cloning the GameCube component cable

Post by novenary » Sun Jul 19, 2015 7:53 pm

meneerbeer wrote:
Streetwalker wrote:If you have a spare atmega8 (or a raspberry pi :P) you can make a xeno pretty easily to try out Swiss.
I have both. ;) I suppose I still need to burn a CD with Swiss on it? I am kind of new to GC modding. I do have a modded GC lying here, but it is a white model. I guess I could swap drives, but then I still need to get CDs. I'd rather not go the CD way.
Streetwalker wrote: Otherwise I'd recommend that you make yourself an IPL replacement, it's not hard and it's much nicer than the xeno.
How does that work, is it expensive? If I can somehow boot Swiss it should be a piece of cake, I still have a SD gecko that I used with my Wii back in the day.
For the xeno, yes you need to burn swiss to a dvd. A cd won't work in the GC. The ipl replacement route is a little more complex but on the hardware side it's really dumb : you can do it with a cpld and some flash. Drop me a pm and I'll send you some details.
User avatar
Unseen
Posts: 190
Joined: Fri Jul 04, 2014 11:52 am

Re: Cloning the GameCube component cable

Post by Unseen » Sun Jul 19, 2015 8:47 pm

meneerbeer wrote:I would need to look better into this. I think it is possible to at least get rid of the jumpy image. I suppose both the even and odd field have an equal amount of lines, so I think for one of the fields you need to delay by one line. I think that means you need an extra buffer to hold a line, so it may not fit anymore.
One line is easy - the current linedoubler already introduces a one-line delay anyway. The problem starts when you have sent that line to the display, you need to send another one. A proper deinterlacer would use the corresponding line from the previous field (for higher resolution) or an interpolation from the previous and next lines of the current field (when there was motion in this location and using the previous field would introduce "jaggies"). Since my linedoubler doesn't have the data from the previous field, it just repeats the previous line, optionally dimming it when (alternating) scanlines are turned on. Since that repeated line isn't in the correct Y-position, the image looks like it's jumping up and down by one line all the time. The other extreme would be to show a black line instead (just set scanlines to maximum) which leaves the job of combining both fields to a complete image to your eyes and brain, but that produces a flickering image.
I was expecting to still see at least some black bars around the picture, because the games do not use the full 720 pixels width. There were no black bars, so my guess is that you send a different horizontal active video area than 720 pixels (I believe the GC sends a flag that holds whether there is active video data). It is not really an issue, but maybe there are TVs that expect 720 active pixels per line.
Indeed, I saw that on your screenshot. Maybe your TV detects that the image is surrounded by a constant area of pure black and extends its grey bars over that too? There is indeed a module in the video pipeline that increases the active video area from whatever the cube outputs (shown in the popup whenever it changes) to 720xWhatever. Since you seem to know your way around VHDL, you could either bypass it in toplevel_p2xh.vhd (replace the Inst_Reblanking block with "video_444_rb <= video_444;") or change the color of the active area extension in Blanking_Regenerator_Fixed.vhd (search for "GC is blanked" and change the PixelY assignment from "00" to "e0") to experiment with it.

Some TVs indeed don't like non-standard resolutions that much, for example I have one that will always zoom such an image to the full screen area and ignore the aspect ratio button on the remote.
Asking for support by PM is anti-social. Ask in an open forum instead, so other people can benefit from the answers!
Voltz
Posts: 28
Joined: Thu Apr 16, 2015 8:13 am

Re: Cloning the GameCube component cable

Post by Voltz » Sun Jul 19, 2015 10:07 pm

Sorry if this is a dumb question, however, what type of code is used in the GC Video Lite.
tueidj
Posts: 564
Joined: Fri May 03, 2013 6:57 am

Re: Cloning the GameCube component cable

Post by tueidj » Sun Jul 19, 2015 10:32 pm

The vertical stripes are from the GC's horizontal scaler, they appear on a lot of games.

For the jumpy image problem with the linedoubling I believe what meneerbeer is suggesting is to add one extra blank line to the top of the bottom field image to properly align the original scanlines with their intended display position, rather than implement some sort of deinterlacing algorithm.
bobrocks95
Posts: 161
Joined: Fri Jul 26, 2013 11:19 pm

Re: Cloning the GameCube component cable

Post by bobrocks95 » Sun Jul 19, 2015 10:43 pm

tueidj wrote:The vertical stripes are from the GC's horizontal scaler, they appear on a lot of games.
Interesting, I guess the Wii uses the same scaler? Is there any more info I can find out there on this? Is it used to upscale a lower res back buffer to the standard output resolution or for some other reason?
tueidj
Posts: 564
Joined: Fri May 03, 2013 6:57 am

Re: Cloning the GameCube component cable

Post by tueidj » Sun Jul 19, 2015 11:29 pm

bobrocks95 wrote:Is there any more info I can find out there on this? Is it used to upscale a lower res back buffer to the standard output resolution or for some other reason?
Basically yes, the unscaled YUV data is resampled before being fed to the D/A converter.
http://hitmen.c02.at/files/yagcd/yagcd/ ... tml#sec5.3
HSR controls the stepping size, where each "step" is one tick of the 13.5MHz clock used to drive the output signal. The FCT registers may hold the weights used for the resampling (they are NOT used for anti-aliasing), in any case the values used are hardcoded in the SDK and not configurable by games.

A similar artifact in gamecube/wii games is visible dithering when the internal framebuffer has the alpha buffer enabled or anti-aliasing is active, due to reduced color depth (6 bits per channel when alpha is used, 5-6-5 when SSAA is used). This is normally very visible in Super Mario Sunshine due to lots of flat colored surfaces being used.
meneerbeer
Posts: 212
Joined: Wed Sep 03, 2014 9:13 am

Re: Cloning the GameCube component cable

Post by meneerbeer » Mon Jul 20, 2015 7:08 am

tueidj wrote: For the jumpy image problem with the linedoubling I believe what meneerbeer is suggesting is to add one extra blank line to the top of the bottom field image to properly align the original scanlines with their intended display position, rather than implement some sort of deinterlacing algorithm.
Not sure if you mean the same. I know with the current mode the fields are displayed separately, rather than combined. Obviously it is impossible to combine both fields with the current hardware + it is rather useless if you can force progressive video for most(/all?) games.

What I am seeing is that both fields do not start at the same y-value after line doubling. When I look at the top/bottom of the image I can see the image moving up and down. My idea is that one of the fields should be moved down by one line (and still doing the linedoubling). I think that will stop the image from jumping up and down. To do that I think you need an extra line buffer.
tueidj
Posts: 564
Joined: Fri May 03, 2013 6:57 am

Re: Cloning the GameCube component cable

Post by tueidj » Mon Jul 20, 2015 7:58 am

It's the same thing, add an extra blank line to the top of the frame produced from the bottom field (the one that contains only the bottom lines from each top/bottom pair). Otherwise the picture will be shown one line higher than originally intended, causing the jumping effect.
User avatar
Unseen
Posts: 190
Joined: Fri Jul 04, 2014 11:52 am

Re: Cloning the GameCube component cable

Post by Unseen » Mon Jul 20, 2015 8:44 pm

tueidj wrote:For the jumpy image problem with the linedoubling I believe what meneerbeer is suggesting is to add one extra blank line to the top of the bottom field image to properly align the original scanlines with their intended display position, rather than implement some sort of deinterlacing algorithm.
I'm reasonably sure that the alignment is already correct. My method for testing this uses a short video capture using a PEXHDCAP card, with a 480i source signal, linedoubling enabled and maximum-strength alternating scanlines. The scanlines black out every second line in the image and two consecutive fields of the captured video can be combined in Photoshop or similar image editing programs with a "Lighten only" blending option. A test capture from Mario Kart Double Dash is available on imgur and the attachment is a 400% enlarged crop from that.
mkdd-crop.png
(3.25 KiB) Not downloaded yet
Of course I still may have gotten it wrong, but I think the flickering cannot be avoided without proper deinterlacing.
Last edited by Unseen on Mon Jul 20, 2015 8:48 pm, edited 1 time in total.
Asking for support by PM is anti-social. Ask in an open forum instead, so other people can benefit from the answers!
andre104623
Posts: 694
Joined: Wed May 07, 2014 2:24 pm

Re: Cloning the GameCube component cable

Post by andre104623 » Mon Jul 20, 2015 8:46 pm

meneerbeer wrote:
Unseen wrote:
meneerbeer wrote:I am trying out the p2x code. I get a red screen. Most likely something is wired up incorrectly, because I have the idea to yuv to rgb is always active?
Yes, since the p2x outputs a DVI signal it has to convert the output to RGB.
Currently looking what is going wrong, but perhaps you have an idea, Unseen?
My first guess would be something wrong on the input side because as you mentioned the colors do look more like YUV problems than RGB problems. Since there is a stable picture, at least Clock, VData 4 and VData 5 are connected correctly (HSync/VSync, muxed within the pixel data). CSel might be involved in the problem, but it would be connected to a wrong signal and not just open or shorted to ground/supply. Another possibility could be a mis-wiring of the highest VData lines (6/7) so that it is seen as constant/mostly-constant 1 during the picture area, although it can't be miswired to the adjacent 3.3V pin because then you wouldn't get a picture at all (all-zero Y component is used to mark that the cube is transmitting Sync+Flags instead of picture).

If you want to check that it's not a problem of the board, you could try to flash KNJNs Pong demo from the StartupKit zip, subfolder "FPGA project - HDMI". It doesn't matter if the board is connected to the cube or not to run that demo, choose the method that is more convenient for you.

(which reminds me that this might be a good test for andre104623's board too...)
Ouch, quite a long post, when it was just me flashing the wrong file.. Sorry about that.

I have made a custom PCB, which is why I had to modify the code. I uploaded an album with some images here. I still need to get screws with ring nuts to secure it better. Right now I just took two screws from the DVD drive and stuck them through to at least keep it a bit in place. Also, I kind of screwed up the opening for the HDMI connector..

Everything seems to work fine. Some games look really good (Smash Bros and Mario Kart for instance). Other games do not really look that great. In Wind Waker I notice a pattern with stripes. My guess is that this is normal (dithering?). With analog cables I suppose you do not really notice this. It looks kind of bad though. I added a zoomed in image of Wind Waker to the album, so you can see what I mean.

My GC is PAL and not modded, so I can only try interlaced video modes. Interlaced video over HDMI works great for my TV luckily. I am eager to try out true 480p and 576p. If you go as far as doing this mod, it is a bit stupid to only use interlaced video.

I tried out the line doubler. It seems to give a "jumpy" image. As if every other frame the image jumps down by one line. It definitely happens at the top and bottom line, but I have the idea that the whole image is jumping up and down. I think this can be fixed, but then we first need to be certain about what causes it.

Scanlines are working great for me for linedoubled 576i. For linedoubled 480i they are barely noticeable. They only seem to be noticeable when there is movement on the screen. I think this also has to do with the above issue with the jumpy image.

I do not really care too much about the linedoubling issues, since I do not really plan on using it.

Oh, one other thing. It seems you do not set the active video you send over HDMI as 720x480/720x576. Perhaps some TVs get confused by this because they expect a line to have 720 active pixels (that will give black bars). Although it is definitely not the problem that is causing the p2x to not work with so many TVs.

Other than that, it is really nice. I would have never put this much effort into it. I probably would have only decoded the video signal to HDMI and been too lazy to add all these features.
Very cool little board. Very good job thats my next step to make a PCB and see how it goes. But I'm going to start simple does anyone know a good PCB maker for flexible PCB?
Yohanov
Posts: 18
Joined: Sat Jun 20, 2015 10:42 pm

Re: Cloning the GameCube component cable

Post by Yohanov » Tue Jul 21, 2015 12:28 am

Hi guys, do you know if this cloned component board is moddable to vga like the old component/d-terminal official cables could be modified into VGA ? I'd rather play in 480p through a CRT with vga than a HDTV with components.
User avatar
megalomaniac
Posts: 2480
Joined: Sun Aug 21, 2011 5:33 am
Location: Drunk in Texas
Contact:

Re: Cloning the GameCube component cable

Post by megalomaniac » Tue Jul 21, 2015 1:33 am

Yohanov wrote:Hi guys, do you know if this cloned component board is moddable to vga like the old component/d-terminal official cables could be modified into VGA ? I'd rather play in 480p through a CRT with vga than a HDTV with components.
have you read anything in this thread?
emu_kidid wrote: beer is like WD40 for megalomaniac's brain, gets the gears moving
>>> BadAssConsoles.com <<<

Image Image Image
Yohanov
Posts: 18
Joined: Sat Jun 20, 2015 10:42 pm

Re: Cloning the GameCube component cable

Post by Yohanov » Tue Jul 21, 2015 1:50 am

I read a lot of it but I guess I didn't understand some of it as i don't have this kind of technical knowledge hence my beginner question.

The vga mod i'm refering to was never talked about on this thread, only Hdmi or dvi to vga adapter stuff.

Since the original component/d-terminal cable were moddable to vga ,with a direct mod on the cable without adapter, I don't think it's a stupid question to ask, at least for a novice, if you can do the same with a cloned board/cable or if you are forced to use a converter of some kind, but if the answer is so obvious for you could you please ELI5 me the answer ?

Not like when you didn't answer my PM with the same question and now bash me here : (


I'm talking about this kind of modded cable :
http://www.goldenshop.com.hk/AI-trad/gc/vgacable.htm

http://www.ebay.com/itm/Nintendo-Gamecu ... 7675.l2557
User avatar
megalomaniac
Posts: 2480
Joined: Sun Aug 21, 2011 5:33 am
Location: Drunk in Texas
Contact:

Re: Cloning the GameCube component cable

Post by megalomaniac » Tue Jul 21, 2015 3:58 am

are you sure you sent me a PM?
if you did you must have deleted it before i had a chance to see it....
i also didnt get an email notification...
but that doesnt matter...

here is a quote from the very first post:
Unseen wrote:The board can output analog RGB (both CSync and HV-Sync available) or Component video.
here is another from page 3
Unseen wrote:
meneerbeer wrote:If you use the RGB output from the board and the game is 480p, can you then display it on a computer monitor with VGA?
Yes

RGBHV = VGA compatible
emu_kidid wrote: beer is like WD40 for megalomaniac's brain, gets the gears moving
>>> BadAssConsoles.com <<<

Image Image Image
meneerbeer
Posts: 212
Joined: Wed Sep 03, 2014 9:13 am

Re: Cloning the GameCube component cable

Post by meneerbeer » Tue Jul 21, 2015 6:40 pm

Unseen wrote:
tueidj wrote:For the jumpy image problem with the linedoubling I believe what meneerbeer is suggesting is to add one extra blank line to the top of the bottom field image to properly align the original scanlines with their intended display position, rather than implement some sort of deinterlacing algorithm.
I'm reasonably sure that the alignment is already correct. My method for testing this uses a short video capture using a PEXHDCAP card, with a 480i source signal, linedoubling enabled and maximum-strength alternating scanlines. The scanlines black out every second line in the image and two consecutive fields of the captured video can be combined in Photoshop or similar image editing programs with a "Lighten only" blending option. A test capture from Mario Kart Double Dash is available on imgur and the attachment is a 400% enlarged crop from that.
mkdd-crop.png
Of course I still may have gotten it wrong, but I think the flickering cannot be avoided without proper deinterlacing.
I think you are right. Even if you place the second field one line higher, you would still notice the copyright logo changing every frame, as the data is different.
Post Reply