did you recompile the project after making the change?AndehX wrote:man, I tried that on the line above, and it didnt work, so I figured there was more to itmaxtherabbit wrote: change line 1152 to:
//vsyncActive = GBS::STATUS_SYNC_PROC_VSACT::read();
GBS 8200/8220 CFW Project
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
Re: GBS 8200/8220 CFW Project
Yeah, thats what I mean. I compiled it and it spit out a massive error. Working now though.
Re: GBS 8200/8220 CFW Project
So after testing both the saturn and the ps1 i get an unstable signal, nothing like before but i guess the saturn fix messed up something overall.
https://streamable.com/rspyh
From 1:25 you can see the effect, it's present multiple times during the capture.
https://streamable.com/rspyh
From 1:25 you can see the effect, it's present multiple times during the capture.
Re: GBS 8200/8220 CFW Project
Meh.. That's the overflow protect feature kicking in when it shouldn't.
I really want it to be active, as it helps with more serious sync issues (Mega Drive HSync width dropping, 240p/480i switching, etc).
For some reason it seems to trigger on your regular sync signal.
Is this an early or late PSX? (Early: SCPH-1000 to SCPH-5500, everything else is late)
I really want it to be active, as it helps with more serious sync issues (Mega Drive HSync width dropping, 240p/480i switching, etc).
For some reason it seems to trigger on your regular sync signal.
Is this an early or late PSX? (Early: SCPH-1000 to SCPH-5500, everything else is late)
Re: GBS 8200/8220 CFW Project
It's a late one, the same happens on saturn.
Re: GBS 8200/8220 CFW Project
Can you take a (sharp) picture of this section of your GBS?
Re: GBS 8200/8220 CFW Project
Now i can't but it's a V.4 8200 with black buttons
Re: GBS 8200/8220 CFW Project
Okay, I've got that hardware here.
I can't get it to glitch but I'll disable the feature with the next update.
I can't get it to glitch but I'll disable the feature with the next update.
Re: GBS 8200/8220 CFW Project
Thank you.
Can't you make the feature selectable ?
Can't you make the feature selectable ?
Re: GBS 8200/8220 CFW Project
Nah, there's no point in it randomly working.
I need it to be reliable on all boards, or else it'll just confuse people.
When it works, you only notice it by the sync glitches in some consoles / some situations being less severe.
It mainly affects Mega Drive 256px games and progressive / interlace switches of some consoles.
These situations are handled okay even without the feature, but they look more jarring.
I need it to be reliable on all boards, or else it'll just confuse people.
When it works, you only notice it by the sync glitches in some consoles / some situations being less severe.
It mainly affects Mega Drive 256px games and progressive / interlace switches of some consoles.
These situations are handled okay even without the feature, but they look more jarring.
Re: GBS 8200/8220 CFW Project
The picture can jump more :p
Somehow it also depends on the active video content, with lots of blacks being more jumpy.
None of that disturbs the output stream, so it's just a visual annoyance, and doesn't lead to loss of signal on the receiver.
Somehow it also depends on the active video content, with lots of blacks being more jumpy.
None of that disturbs the output stream, so it's just a visual annoyance, and doesn't lead to loss of signal on the receiver.
Re: GBS 8200/8220 CFW Project
Sounds like a good compromise, can't wait to test the new update.
Re: GBS 8200/8220 CFW Project
Quick question rama, I've been using the 720p preset on my 1440p monitor because 720p scales up perfectly since it's exactly half of 1440p (and none of the other presets seem to work on this monitor, except 1080p)
This monitor won't allow me to adjust the aspect ratio, so I've been trying to scale the image horizontally in GBS control. It works great, although I get slight artifacts now and then, which I'm guessing happens when you scale the image too much.
Any way to fix this? or is this something that is unavoidable when scaling the image to a 4:3 aspect?
This monitor won't allow me to adjust the aspect ratio, so I've been trying to scale the image horizontally in GBS control. It works great, although I get slight artifacts now and then, which I'm guessing happens when you scale the image too much.
Any way to fix this? or is this something that is unavoidable when scaling the image to a 4:3 aspect?
Re: GBS 8200/8220 CFW Project
Scaling involves a few limitations that I can't explain well.
It can definitely be made better in the future though (as I understand the parameters better).
For now, you may get good results with a faster memory clock (as one limitation is the SDRAM streaming performance).
In general though, I don't really like the 720p format.
It doesn't fit well with the source vertical requirements: 3x 240 lines makes for uneven scaling.
1280x1024/x960 work much better. Pitty, if the display doesn't work with those.
The display obviously can deal with 1080p, which is a few more lines..
It can definitely be made better in the future though (as I understand the parameters better).
For now, you may get good results with a faster memory clock (as one limitation is the SDRAM streaming performance).
In general though, I don't really like the 720p format.
It doesn't fit well with the source vertical requirements: 3x 240 lines makes for uneven scaling.
1280x1024/x960 work much better. Pitty, if the display doesn't work with those.
The display obviously can deal with 1080p, which is a few more lines..
Re: GBS 8200/8220 CFW Project
Yeah switching memory clocks does affect the artefacts (some better, some worse) are memory clock settings saved in the custom presets?rama wrote:Scaling involves a few limitations that I can't explain well.
It can definitely be made better in the future though (as I understand the parameters better).
For now, you may get good results with a faster memory clock (as one limitation is the SDRAM streaming performance).
In general though, I don't really like the 720p format.
It doesn't fit well with the source vertical requirements: 3x 240 lines makes for uneven scaling.
1280x1024/x960 work much better. Pitty, if the display doesn't work with those.
The display obviously can deal with 1080p, which is a few more lines..
Curious as to why you don't like 720p though. I thought it would be a decent resolution for scaling, as 240p scales perfectly 3x into 720p...
The problem with 1280x1024/960 or 1080p (on this monitor) is that they're not integers of the 1440p resolution, and as a result, scanlines can look uneven and a little janky, which is why I prefer 720p. I have a real OCD for integer scaling and crisp pixels
Of course this isn't an issue on my 1080p projector, as i'd just use 1080p, but yeah, just thought I'd mention that.
Being able to scale the image to 4:3 on the 16:9 presets without artefacts, would be very handy though, as i'm sure my monitor isn't the only one that doesn't allow you adjust aspect ratio
Re: GBS 8200/8220 CFW Project
Okay, so 720p does look fine on some displays. I'll keep it in mind
Memory clocks should get saved with custom presets.
About the scale to fit aspect ratio thing, I agree that this is useful in some cases.
Horizontal scaling has a few "ideal" steps, where pixels get duplicated evenly and everything in-between should best be avoided.
I'm not sure yet how to implement this nicely.
Iraito:
The update with SP_H_PROTECT disabled is out.
I hope this makes your setup happy, as I can't replicate the problem here.
Must be some kind of hardware difference..
Memory clocks should get saved with custom presets.
About the scale to fit aspect ratio thing, I agree that this is useful in some cases.
Horizontal scaling has a few "ideal" steps, where pixels get duplicated evenly and everything in-between should best be avoided.
I'm not sure yet how to implement this nicely.
Iraito:
The update with SP_H_PROTECT disabled is out.
I hope this makes your setup happy, as I can't replicate the problem here.
Must be some kind of hardware difference..
Re: GBS 8200/8220 CFW Project
Update is out that ignores VSync if CSync is decoding into a known signal.
This should fix the gscart issue.
This should fix the gscart issue.
Re: GBS 8200/8220 CFW Project
720p is useful for me because I setup a recording setup and the box I use only accept common hdmi resolution up to 1080i, and for youtube 60fps is enabled with 720p and above.
https://www.youtube.com/watch?v=TJPgdHwlB1U
there's audio delay that I figured out and is fixed now.
Speaking of that box, it has an HDMI out but for some reason that output is very flacky (loss of picture) only when recording for some reason, in either HDMI or component input (note that the record itself is fine). The loss of picture doesn't happen on more standard sources (tested HDMI PS4 and component PS3). I understand it's because 240p sources not being standard refresh (frametime lock didn't help unfortunately), still it's weird that it only happens while recording...
https://www.youtube.com/watch?v=TJPgdHwlB1U
there's audio delay that I figured out and is fixed now.
Speaking of that box, it has an HDMI out but for some reason that output is very flacky (loss of picture) only when recording for some reason, in either HDMI or component input (note that the record itself is fine). The loss of picture doesn't happen on more standard sources (tested HDMI PS4 and component PS3). I understand it's because 240p sources not being standard refresh (frametime lock didn't help unfortunately), still it's weird that it only happens while recording...
Re: GBS 8200/8220 CFW Project
awesome! Not entirely sure how many people would be affected by this, but it affected me, so I guess it'll affect some others too. Hopefully it's a non-issue now.rama wrote:Update is out that ignores VSync if CSync is decoding into a known signal.
This should fix the gscart issue.
720p definitely looks great on a 1440p monitor. There is a wavy, scrolling interference pattern on it, although i'm not worried about it right now, as it seems specific to the 720p preset, so i'm sure you could address it software when you get some time. Cycling RAM speeds seems to change the direction/speed of the interference pattern, but it's always there.
Also, having the 720p 1080p presets is important for people with modern capture cards that don't accept VGA resolutions like 1280x960 etc...
It's also cool to know that there's potential for horizontal scaling to maybe look as crisp as vertical scaling. If it doesn't happen though, it won't bother me too much, as softer horizontal look kinda resembles a CRT when scanlines are enabled.
This is 720p with scanlines on my 1440p monitor
Re: GBS 8200/8220 CFW Project
Thanks for the clarification, I look forwards to further experimentation! I just updated to the latest version and everything is still working as it was for me, for what it's worth.rama wrote:I'm not sure what each source does, but it manifests in slightly different vertical refresh times when decoded.
I'll look into adding htotal adjustments to aid in dialing in accepted HDMI timings.
When you get a "Fieldrate" log, that's the source, yep.
One quirk I've had is each button in the web interface only working once on my phone before becoming completely unresponsive, even after refreshing the page they wouldn't work until I came back to the web interface some time later. At first I thought this was another IE issue so tried an Android handset that I could install Chrome and Firefox on and that showed the same issue. I think it's due to the use of GET requests which the mobile browsers are caching aggressively (I don't have this issue on my desktop, for what it's worth). Changing to POST would seem more correct for requests that change a setting, however my quick and hacky fix was the old-fashioned appending of a dummy variable that changes on each request in webui.html:
Code: Select all
// ... snip ...
var queuedText = '';
function loadDoc(link) {
var xhttp = new XMLHttpRequest();
xhttp.open('GET', 'sc?' + link + '&nocache=' + new Date().getTime(), true);
xhttp.send();
}
function loadUser(link) {
var xhttp = new XMLHttpRequest();
xhttp.open('GET', 'uc?' + link + '&nocache=' + new Date().getTime(), true);
xhttp.send();
// ... snip ...
Re: GBS 8200/8220 CFW Project
I still get the same issue, this started only after the SS fix.
Re: GBS 8200/8220 CFW Project
Iraito:
Please connect your pin 20 sync signal to the green RCA again, making sure it's not connected to the other input.
Easiest to do this would be if you had a yellow plug standard composite cable.
Either the PSX or Saturn are fine, as long as they normally show the sync problem when on the RGBS input.
If you now get a black and white picture, please proceed to test for sync instability.
It should be stable.
benryves:
That's odd. Maybe something else is going on there, as I can't reproduce this here.
But sure, I can add that hack.
I liked get requests better as they produce far less traffic than posts.
Please connect your pin 20 sync signal to the green RCA again, making sure it's not connected to the other input.
Easiest to do this would be if you had a yellow plug standard composite cable.
Either the PSX or Saturn are fine, as long as they normally show the sync problem when on the RGBS input.
If you now get a black and white picture, please proceed to test for sync instability.
It should be stable.
benryves:
That's odd. Maybe something else is going on there, as I can't reproduce this here.
But sure, I can add that hack.
I liked get requests better as they produce far less traffic than posts.
Re: GBS 8200/8220 CFW Project
The composite is stable.
Re: GBS 8200/8220 CFW Project
Okay, then the RGBS path should work as well.
Are there any logs when it becomes unstable?
Are there any logs when it becomes unstable?
Re: GBS 8200/8220 CFW Project
As a bit of clarification what's going on:
The YUV input on every GBS is fixed to standard properly.
Each signal is 75 Ohm terminated and the Y channel is branched off into the SOG1 input on the scaler chip.
The RGBS path varies across different GBS models.
Usually, the S signal is 1000 Ohm or 500 Ohm terminated (hence the need to install the additional 100 Ohm to ground, to bring the total to about 75 Ohm).
Also, the S signal sometimes branches to the Myson 230 controller via a 1000 Ohm resistor, and sometimes via direct connection.
If it's a direct connection, the signal is badly affected.
V4 models should have the resistor at R58 (iirc) so they should be fine.
Lastly, on each GBS, the S signal then goes to the SOG0 input on the scaler chip.
Due to all the uncertainty on what the hardware is, the "S" input is usually not as reliable as the RCA green input.
The YUV input on every GBS is fixed to standard properly.
Each signal is 75 Ohm terminated and the Y channel is branched off into the SOG1 input on the scaler chip.
The RGBS path varies across different GBS models.
Usually, the S signal is 1000 Ohm or 500 Ohm terminated (hence the need to install the additional 100 Ohm to ground, to bring the total to about 75 Ohm).
Also, the S signal sometimes branches to the Myson 230 controller via a 1000 Ohm resistor, and sometimes via direct connection.
If it's a direct connection, the signal is badly affected.
V4 models should have the resistor at R58 (iirc) so they should be fine.
Lastly, on each GBS, the S signal then goes to the SOG0 input on the scaler chip.
Due to all the uncertainty on what the hardware is, the "S" input is usually not as reliable as the RCA green input.
Re: GBS 8200/8220 CFW Project
Is there a solution to this issue\quirk ?
Re: GBS 8200/8220 CFW Project
You can avoid all hassle by building a sync stripper.
The sync stripper takes your CVBS sync signal, strips all the video and outputs an amplified CSync signal that's easy for the GBS to understand.
But it should work with a direct connection as well.
You should get some logs on the web ui. The minimum I output on bad sync is a single dot ('.').
The sync stripper takes your CVBS sync signal, strips all the video and outputs an amplified CSync signal that's easy for the GBS to understand.
But it should work with a direct connection as well.
You should get some logs on the web ui. The minimum I output on bad sync is a single dot ('.').
Re: GBS 8200/8220 CFW Project
But i'm using a sync stripper