February 18, 2017 at 5:26 PM #11374
Just updated to the new firmware today, and it’s fantastic – on every console I have apart from my NES. I have an AV Famicom (the revision of the original, same design as the NES top loader) that I RGB modded. On the old firmware, it worked flawlessly in x3 mode with a crystal clear image in both 4:3 and 1:1 native 256px modes
On the new firmware though, the top portion of the image is unstable, warped and blurry. I’ve tried every combination of scaling options and every variable on advanced timing, and nothing fixed it. This is only an issue with my NES (my other consoles are all fine, including the SNES).
I’ve uploaded a video here to show what the issue is (ignore the latter half of the video though!) https://youtu.be/3j7DATj0aeo
Any idea of a fix? If not, I’ll have to sadly roll back to the last firmware – which will be a shame.
and here’s an image if you can’t view the video:February 18, 2017 at 5:58 PM #11376
It is normal for the top part to be unstable for the NES/Famicom. It happens on all scalers even the Hidef NES.
Can you make a video of the same part of the game on the previous firmware? It should have a similar effect.February 18, 2017 at 6:11 PM #11377
It’s normal?! That’s bizarre. It’s a completely stable image throughout on my CRT and on the previous firmwares. I’ve just rolled back to 0.75 and taken a video to prove this – make sure you watch it in 1080p60 mode:
So no, it was literally PERFECT on the old firmware. Perfect image, no unstable top portion – just perfect.
I’ve only ever experienced the unstable top part of the image on the 0.76 firmware. Nowhere else. It was also completely stable on my Framemeister (which I have since sold).
In any case, now I’ve rolled back to 0.75 and NES is flawless again. Hm. I’ll wait and see what happens with the next firmware, but the NES is the deciding factor in my case. There’s no point in updating if 1 out of my 8 consoles hooked up to the OSSC is worse off.February 18, 2017 at 6:47 PM #11381
I mean its normal for it to be slightly unstable. On your old firmware video its mostly hidden. On your video on the 0.76 firmware you can see more lines at the top and bottom. So the two videos do not show the same area of the signal.
Either your TV is changing the overscan somehow or the V.active, V.backporch values are different. You can try V.active 224 and check your TV zoom settings on 0.76 firmware.
Are you using linetriple on both firmwares?February 18, 2017 at 6:50 PM #11382
That’s not true. On the original video I was using the EXACT SAME scaling mode, and the EXACT SAME screen portion was visible on both. On 0.76 it’s unstable, on 0.75 it’s not. I could post yet another video proving this using the intro of SMB3, but do I really have to? The values in adv timings are the same too, with no effect.
And yes, line triple on both videos. And no, JUST SCAN/pixel perfect mode on my TV on EVERY input, including the OSSC.
So yeah, either way – it’s fucked on 0.76 and not-fucked on 0.75. that’s it, really.February 18, 2017 at 6:57 PM #11383
Also tested with the 240p suite via my everdrive. The screen ratio is EXACTLY THE SAME on both, and EXACTLY the same on both 0.75 and 0.76, although these photos are from 0.75:
So, whatever you assume is the case with my NES, isn’t the case. It’s perfect on 0.75 and ruined on 0.76. That’s all there is to it.February 18, 2017 at 7:17 PM #11385February 18, 2017 at 7:49 PM #11386
.February 18, 2017 at 9:42 PM #11388
Nevertheless that picture I posted proves that in your two videos the overscan was different. Good luck with your issue.February 19, 2017 at 9:46 AM #11395
Is it really so that the result with Line3x generic modes differs between 0.75 and 0.76? Implementation of optimized modes has indeed changed in 0.76, but there should be no relevant changes in how generic modes are processed. You might try turning “Allow TVP HPLL2x” off in sampling menu, but it probably just makes things worse.
I still have my NESRGB kit on the shelf awaiting to be installed, so I’ll keep an eye if I can notice similar phenomena once I have it running.February 19, 2017 at 3:13 PM #11403
You probably tried this already, but did you play with the analogue sync LPF at all?February 28, 2017 at 8:51 PM #11585
I finally installed my NESRGB kit, and can notice the same wobble on few first scanlines on PC monitor (my TV crops just enough to hide the effect). I didn’t go back to 0.75 yet, but disabling “Allow TVP HPLL2x” clearly helped.April 2, 2017 at 5:02 AM #12148
Just got my OSSC today with firmware 0.76, hooked up my NES (with NESRGB) and got this:
It flickers between straight vertical and that distorted bend.
Disabling “Allow TVP HPLL2x” makes a very slight improvement, if any. Adjusting analog sync LPF makes no noticeable improvement, and just shifts the image horizontally.
EDIT: I downgraded to 0.75 and still have the same problem.April 2, 2017 at 5:46 PM #12161
Don’t know if this means anything – but the exact same thing happens to me when running FCE Ultra GX (NES emulator) on the Wii, via RGB SCART with “original” rendering mode set. The top line is slightly distorted to the left, which is usually hidden when I set my TV to overscanApril 4, 2017 at 7:17 PM #12205
Nes PPU skips a tick every other frame just before first visible scanline is drawn, resulting to non-periodic hsync which is bad reference for any digitizer. An universal fix would be to alter Nes clocking circuitry (e.g. via CPLD/MCU) so that one oscillation cycle on the uneven scanline is gated, thus stabilizing hsync period.
You must be logged in to reply to this topic.