May 29, 2019 at 9:26 AM #26387
Is there a possibility of providing a compatibility mode for those of us stuck on fw 0.76?
“For some reason those reported displays seem to accept 628-line signal if sync/enable signal is formatted somewhat incorrectly (as can be done with the posted settings on 0.76). I’d need check resulting timings and then perhaps provide a compatiblity option to enable such deranged sync on later fw.”
The problem is detailed in this thread:June 2, 2019 at 12:03 PM #26461
It’s on TODO list, maybe I have some time to check after getting v0.83 released. Unfortunately I don’t have my old VT50 anymore which could be used to test this, but maybe I can first make a variant out of the display test firmware which would have the necessary hack enabled.June 28, 2019 at 12:32 AM #26808
Tonight I checked which kind of output v0.76 produces with the mentioned settings. There are a couple things off with resulting sync, but most prominent one is that data enable signal ends up overlapping vsync which cannot occur in later firmwares. I made a temporary change enabling this on the latest fw and uploaded a test image here. Please try that PSX 288p in line2x mode with V.active set to 300 as mentioned in the linked thread. You may then also need to adjust V.backporch to make the signal recognizable by the TV.June 28, 2019 at 2:02 PM #26820
Thanks for trying to resolve this. I’ve just tried the test firmware with my GT50 Panasonic. Sadly it didn’t resolve the issues 🙁
I tested with a PAL PS1, a PAL PS2 and a UK 60hz modded Neo-Geo AES. Unfortunately I cannot get a picture when modifying the settings to those which are known to work with fw0.76
Upon returning to fw 0.76-aud these systems started displaying ok again on this tv.
Just to clarify – the PS needs v.active of at least 300. The 60hz Neo geo aes requires a v.active of 256
Is there anything else which you think could be different about post-0.76 that could be stopping pal working on these types of tv?
Thanks again for your time and efforts!June 28, 2019 at 11:46 PM #26824
Ok, maybe it’s be better to debug this the other way. I made a couple additional changes on the firmware (uploaded to the same file) so that it should now behave just like v0.76 when you use PSX 288p and set V.backporch to 16 and V.active to 300.June 29, 2019 at 8:44 AM #26830
That seems to have fixed it! Nice one 🙂
Ps1 and ps2 start to get a picture once v.active is 297 and higher. The neogeo aes also now produces a picture too when v.active is 253 or higher
Thanks so much for looking into thisJune 29, 2019 at 10:43 AM #26831
That’s not a final fix yet but a starting point. The next step is to normalize the signal step-by-step to get an idea of minimum number of hacks needed for a compatibility mode option. I just corrected hsync-vsync alignment and re-uploaded to the same file – can you try if it still works with PSX?
Ps1 and ps2 start to get a picture once v.active is 297 and higher.
Was that with default V.backporch (19) or with adjusted one (16)?June 29, 2019 at 11:20 AM #26833
This was with default v.backporch of 19. All that needed to change was increasing v.active.
I will test the new upload nowJune 29, 2019 at 11:28 AM #26834
Ok, that new upload breaks compatibility. Returning to the previous test firmware and I get a picture again.June 29, 2019 at 3:32 PM #26836
I see. I rolled back that change and made another one. If that restored compatibility, could you check what’s the smallest accepted V.active with V.backporch 16 and 10?June 29, 2019 at 4:33 PM #26837
Ok, this version breaks compatibility.June 29, 2019 at 6:15 PM #26839
Another version uploaded for test.June 29, 2019 at 6:36 PM #26840
This new version fixes compatibility, and I can lower v.backporch from default 19 right to 1 and I still get a picture.
And… I can reduce v.active down from the 300 I normally use to get a ps1 picture right down to 260 and lower and I still get a picture!
The same for neogeo 60hz – I can reduce v.active from 256 back down to 240 and I get a pictureJune 30, 2019 at 9:18 AM #26845
Finally in the right direction, then. It’s good to hear the default sampling settings are now accepted with the current method which consists of 2 separate hacks (DE cut on last active line and vsync offsetting). I still wonder if the former alone is sufficient so I uploaded one more test version for you to try. After that it’s pretty clear what you have to do with the Panasonics.June 30, 2019 at 10:03 AM #26846
That new version works great too, on both ps1 and Neo-geo and even with v.active turned back down to default, and with all ranges for v.backporch.
Can I just check – what is the actual default value for v.active on 240p and also for 288p?
- You must be logged in to reply to this topic.