norly | he said the G word | 00:34 |
---|---|---|
- xenial-user (QUIT: Remote host closed the connection) (~puppy@user/crabster) | 01:27 | |
- BluRaf (QUIT: Ping timeout: 264 seconds) (bluraf@junkcc.net) | 04:11 | |
+ BluRaf (bluraf@junkcc.net) | 04:14 | |
Claude | South G is awake ! | 05:27 |
- bfarmilo (QUIT: Ping timeout: 258 seconds) (~bfarmilo@203.190.214.11) | 07:10 | |
+ bfarmilo (~bfarmilo@203.190.214.11) | 07:11 | |
- bfarmilo (QUIT: Ping timeout: 265 seconds) (~bfarmilo@203.190.214.11) | 07:20 | |
+ bfarmilo (~bfarmilo@203.190.214.11) | 07:21 | |
- bfarmilo (QUIT: Ping timeout: 265 seconds) (~bfarmilo@203.190.214.11) | 07:32 | |
+ bfarmilo (~bfarmilo@203.190.214.11) | 07:32 | |
mntmn | berlin too ;) | 09:01 |
Chain|Q | berlin is awake, actual footage: https://tenor.com/view/morning-good-good-morning-sunshine-wake-up-gif-13979363 | 09:36 |
- xet7 (QUIT: Quit: Leaving) (~xet7@85-156-127-25.elisa-laajakaista.fi) | 12:41 | |
mntmn | ok v interesting, camlink 4k cannot capture from zz9000! | 12:45 |
mntmn | invalid pixel format / invalid resolution index this that | 12:45 |
_Bnu | Doesn't surprise me. | 12:49 |
_Bnu | Some modes are barely 59Hz and some are above 60, haha. | 12:50 |
_Bnu | I tried adjusting them to get closer to 59.94 and a maximum of 60.10, but then the modes didn't work on my monitor... | 12:51 |
mntmn | i demand that this works!!!!!!!! | 12:51 |
mntmn | lol | 12:51 |
mntmn | i wonder if it's really the freq or the changed blanking | 12:51 |
mntmn | where is my bootdisk | 12:51 |
_Bnu | The vblank stuff didin't change the refresh rate at all as far as I could tell. | 12:51 |
mntmn | yeah, i know, i mean the vblank stuff changed the vblank stuff :D | 12:52 |
_Bnu | It seems entirely dependent on the total number of pixels, the resolution and the frequency set by the clock wizard. | 12:52 |
mntmn | ok | 12:52 |
mntmn | ok so the boot disk is in the amiga actually | 12:52 |
_Bnu | Because me changing the MHz for the modes was what broke it for everyone, haha. | 12:52 |
_Bnu | If you hook up the VisionRGB instead, you can actually check the effective refresh rate of the signal... though it might be hard to get it working on any recent Linux kernel. | 12:52 |
_Bnu | So if you have like a tiny Windows computer with a single PCIe slot, you can use it as a resolution tester... | 12:52 |
mntmn | i only have an LX2160A, i wonder if there are aarch64 drivers for the VisionRGB? ;) | 12:53 |
_Bnu | Nah. And the latest Linux driver is like for kernel 5.2 or something... | 12:54 |
_Bnu | https://www.datapath.co.uk/datapath-current-downloads/vision-capture-card-downloads/vision-drivers/vision-drivers-1/623-linux-visioninstall-v7-21-0 Oh, no. 5.4.0. | 12:55 |
_Bnu | But yeah, the drivers constantly break because of V4L2. | 12:55 |
mntmn | ok | 12:55 |
_Bnu | It really is a nightmare. :D | 12:55 |
mntmn | lol guvcview says 24015/406 fps | 12:55 |
mntmn | interesting, that's > 59.15024630541872 | 12:56 |
mntmn | ohhh | 12:56 |
mntmn | there were a few visible frames when rebooting | 12:56 |
mntmn | how strange | 12:56 |
mntmn | for 1280x720 it shows 6037/100 fps (so 60.37Hz), and it also cannot grabbb the image | 12:59 |
mntmn | it's copy protectino | 12:59 |
_Bnu | Yeah, many capture ICs are like. | 13:01 |
_Bnu | "Oh shit, it's not 59.94 or 60. Do not want!!!" | 13:01 |
Jope | my hdmi switch is also like that :-D | 13:02 |
Jope | zz9000 just doesn't go through it to any display | 13:02 |
mntmn | Jope: also not on 1.7? | 13:03 |
Jope | now that you mention it, I'm no longer sure | 13:03 |
Jope | I'll give it a go in case something changed | 13:03 |
Jope | I should be quite up to date with fw/sw | 13:03 |
mntmn | haven't pressed "generate bitstream" in a while! | 13:23 |
_Bnu | General bipstream reporting in... | 13:24 |
+ xet7 (~xet7@85-156-127-25.elisa-laajakaista.fi) | 13:25 | |
mntmn | ZZ9000OS.bif | 13:27 |
mntmn | biff | 13:28 |
- xet7 (QUIT: Remote host closed the connection) (~xet7@85-156-127-25.elisa-laajakaista.fi) | 13:31 | |
+ xet7 (~xet7@85-156-127-25.elisa-laajakaista.fi) | 13:32 | |
Jope | le bit strem | 13:42 |
Claude | mntmn: many consumer hdmi devices just refuse any mode not in the CTA standard , and even if the mode is 101% CTA without proper matching Avi infoframes set they just go "meh! nope" . Samsung is a prime example of stupidity on this | 13:43 |
mntmn | i see | 13:43 |
_Bnu | Just like strawberries. | 14:10 |
mntmn | highly annoying that i can't capture _any_ of the modes | 14:12 |
- Fate (QUIT: Remote host closed the connection) (~fate@bnc.nordkommission.de) | 14:16 | |
+ Fate (~fate@bnc.nordkommission.de) | 14:20 | |
Claude | then a wild guess : sync polarity wrong or faulty Avi infoframes (when in hdmi mode) | 14:37 |
Claude | the "source product descriptor" | 14:39 |
mntmn | hmhmmhm | 14:52 |
mntmn | aha, i can capture through my cheap chinese hdmi->usb box now | 15:12 |
mntmn | (shows up as "MiraBox") | 15:14 |
mntmn | ok camlink is a bit happier now but still doesn't want to do the job | 15:15 |
mntmn | 50000/1001 fps! | 15:15 |
mntmn | (720x576( | 15:15 |
mntmn | oh there is an image but it is glitching around | 15:15 |
_Bnu | 49.95 ('A`) | 15:17 |
_Bnu | That'd result in some pretty annoying stutters even if it worked, haha. But you probably don't have much of a choice but to go for an even 50 for it to work. | 15:18 |
_Bnu | Unfortunately I don't know what capture card would work well other than the CamLink, not on Linux at least. | 15:19 |
mntmn | interesting because if it's so desynced i can see the blanking area etc sometimes | 15:20 |
mntmn | well, the mirabox gives a stable image but it has bad compression | 15:20 |
mntmn | the camlink has a crisp image | 15:20 |
mntmn | hmm maybe our lines are 1 pixel too long | 15:22 |
mntmn | in terms of blanking | 15:22 |
_Bnu | If it's that simple, that'd be good, haha. | 15:22 |
mntmn | i probably should stick ILA in there and let it count everything | 15:23 |
mntmn | or wait, it's probably much easier to simulate this | 15:23 |
_Bnu | Stimulated FPGA. | 15:25 |
mntmn | lol Resolution > 798x600 | 15:27 |
mntmn | that's two pixels missing | 15:28 |
mntmn | (after i changed something) | 15:28 |
_Bnu | The visible image resolution was always correct for me when checking with the Datapath, at least. | 15:30 |
mntmn | yeah, it was ok | 15:30 |
mntmn | thanks 4 confirming | 15:30 |
_Bnu | Not sure about the horizontal blanking width, but it's probably okay... | 15:31 |
mntmn | hmm! | 15:32 |
_Bnu | Or possibly... | 15:33 |
mntmn | how are the h/v max values to be interpreted... as a count (exclusive) or as a position (inclusive) | 15:33 |
mntmn | like, if we have vmax 628, should we output 628 or 629 lines total | 15:34 |
_Bnu | I think any odd number is probably wrong... | 15:35 |
mntmn | htotal/vtotal are supposed to be counters. so, you are right, they should be even | 15:36 |
Jope | https://c65gs.blogspot.com/2020/10/pulling-out-big-guns-for-improving.html hope this suggestion is not offensice | 15:45 |
_Bnu | I'm not sure what the suggestion is. | 15:48 |
_Bnu | Buy some gigantic and expensive HDMI testing device and spend 700 hours trying to deal with the Xilinx Clock Wizard to fit HDMI specs? :D | 15:48 |
_Bnu | For the RTG modes, I'd be all for having either 60 or 59.94 FPS, but it'd be nice to still have the 49.9189899897/59.98 Hz optionally available for the scandoubler. | 15:48 |
Jope | _Bnu, ah well it is a tl;dr post but the mega65 folks say in there that they're happy to help other retro projects with hdmi compliance | 15:49 |
Jope | so no, no need to buy an expensive tester, they did it and are not hogging | 15:49 |
_Bnu | HDMI compliance is very much a can of worms, though. We need to match one of two refresh rates exactly for any VGA/HDMI resolution, otherwise there will be tearing. | 15:50 |
Jope | yes :-/ | 15:50 |
_Bnu | And from what I could tell, it's simply IMPOSSIBLE to hit these refresh rates with some resolutions. | 15:51 |
Jope | what an annoying world really :-D | 15:51 |
_Bnu | There's just no combination of multipliers and divisors that will result in 59.94 or 60.00. | 15:51 |
Jope | it could be anything but they choose to be picky | 15:51 |
Jope | I mean what the monitors and tvs accept | 15:51 |
mntmn | > 10 year old pre-loved protocol analyser | 15:52 |
_Bnu | It's a somewhat conscious decision on the manufacturer's part. | 15:52 |
Jope | sure | 15:52 |
_Bnu | My ASUS Monitor syncs fine to almost everything ever. And everyone else could do that as well, they just don't want to. | 15:52 |
_Bnu | It has adaptive sync on the VGA port. | 15:52 |
Jope | asus here too | 15:52 |
Jope | surprisingly lenient towards all kinds of things | 15:52 |
Jope | then try samsung .. not so much | 15:53 |
_Bnu | Ban Samsung. | 15:53 |
Jope | they're on my botcott list already :-D | 15:53 |
Jope | but I'm allowing samsung components in devices, else I'd have to move into a hole in the ground | 15:53 |
mntmn | Jope: crazy article, thanks | 15:54 |
_Bnu | I allow them the SSDs in my computer, that's about it. | 15:54 |
mntmn | woah, fixed it | 16:04 |
mntmn | ok, so the article gave a very good hint | 16:04 |
mntmn | hsync+vsync leading edge should coincide | 16:04 |
_Bnu | What about the sleeping edge... | 16:05 |
mntmn | the seeping edge | 16:05 |
_Bnu | Does this mean that I can now adjust the timing for the modes to get closer to 60Hz again... | 16:05 |
mntmn | camlink or guvcview does not like resolution switching though | 16:06 |
mntmn | but interesting, now i have 50.2 hz | 16:06 |
mntmn | grab http://dump.mntmn.com/screenshot-2021-06-22-16-08-52.png | 16:09 |
Jope | ooh, excellent | 16:09 |
mntmn | (in this build there is one line wraparound though because i commented a fix for that out) | 16:10 |
mntmn | while true; do mpv /dev/video1; done | 16:19 |
mntmn | this kind of survives resolution changes ;) | 16:19 |
mntmn | http://dump.mntmn.com/screenshot-2021-06-22-16-22-09.png | 16:22 |
mntmn | http://dump.mntmn.com/screenshot-2021-06-22-16-22-15.png | 16:22 |
_Bnu | Oh. | 16:25 |
_Bnu | Are you sure you're not using the old timings for 1920x1080? | 16:25 |
_Bnu | Because they were wrong and set for 50Hz instead of 60. | 16:25 |
_Bnu | Oh, I can't read. | 16:26 |
_Bnu | In general, neither V4L2 or DirectShow can actually deal with resolution changes, they always need a reinit or for the capture device to scale it to a set output resolution. | 16:27 |
mntmn | stream! https://www.twitch.tv/mntmn666 | 16:27 |
_Bnu | It's too big... | 16:28 |
mntmn | _Bnu: what do you mean! | 16:29 |
_Bnu | My monitor is only 640x290... | 16:29 |
mntmn | lol | 16:29 |
_Bnu | Explodebench! | 16:30 |
_Bnu | Can you do a 200% playthrough of Morton's Revenge... | 16:32 |
mntmn | i have not heard of this game! | 16:33 |
mntmn | ok lunch break... | 16:37 |
Claude | mntmn: ShK happens to bought recently a agilent hdmi analyzer :) I can ping him , he offered to lend it to me but I guess he's fine also lending it to you | 16:41 |
Claude | does the full conformance test | 16:41 |
mntmn | cool!! | 16:43 |
_Bnu | But does it even do the small performance test... | 17:15 |
_Bnu | mntmn needs some PiSCSI to load icons faster. | 17:22 |
mntmn | no it's because this is 68030 | 17:23 |
mntmn | no accelerator | 17:23 |
mntmn | oh red screen :33 | 17:23 |
_Bnu | Small red screen or big red screen... | 17:23 |
mntmn | smol blown up big | 17:23 |
_Bnu | Kickstart red screen. ('A`) | 17:23 |
_Bnu | Sofware fairule... | 17:25 |
mntmn | guru!!! | 17:25 |
_Bnu | DF0 exploded... | 17:26 |
_Bnu | Is this my CDTV... | 17:26 |
mntmn | hehe there are no floppy drives in this a3000 | 17:30 |
mntmn | i should put in an 68060 | 17:30 |
_Bnu | Go all out with a 68070! | 17:33 |
mntmn | hehe | 17:34 |
mntmn | hmm i forgot the a3000 jumper settings for warpengine | 17:34 |
mntmn | http://amiga.resource.cx/manual/WarpEngine_A3000_addendum.pdf | 17:35 |
_Bnu | External internal clock service pack. | 17:35 |
Jope | futuristic amiganess | 17:36 |
mntmn | ok we're back | 17:53 |
mntmn | Jope: you had a picky monitor, right? | 17:54 |
Claude | mntmn: oh fixed it ? 'grats :) | 18:08 |
Claude | nice wb backdrop | 18:09 |
mntmn | thanks yeah! | 18:11 |
mntmn | Claude: i think the main problem was misaligned hsync+vsync edges | 18:11 |
mntmn | but also i ~believe~ the number of total columns and rows was off, but i think it doesn't matter much | 18:11 |
mntmn | (changes the frequency slightly though) | 18:11 |
mntmn | Jope: ping | 18:16 |
mntmn | ok now i need to simplify this algoriddim | 18:17 |
Jope | pong | 18:23 |
Jope | what up | 18:23 |
Jope | I have a picky hdmi switch :-) | 18:23 |
Jope | and IIRC also my tv didn't like the fhd output from zz9000 | 18:24 |
Jope | http://jope.fi/zzfhdsonytv.jpg | 18:25 |
+ isoriano_ (~isoriano@5.147.188.96) | 18:31 | |
isoriano_ | Good evening | 18:31 |
mntmn | Jope: ok, wanna try experimental FW? | 18:41 |
Jope | sure | 18:42 |
mntmn | Jope: cool, here you go http://dump.mntmn.com/BOOT.bin-z3-1.9a-20210622a | 18:42 |
mntmn | Jope: my question would be if your switch or tv like this more | 18:43 |
mntmn | who was the person here with a picky monitor... | 18:43 |
Jope | hmm, it doesn't seem to like this fw | 18:44 |
Jope | black screen via hdmi switch | 18:45 |
Jope | let's see whether it works straight | 18:45 |
Jope | hm, it didn't like it at all | 18:46 |
Jope | sorry, didn't work with my board | 18:48 |
Jope | reverted back to the old boot.bin and that still runs | 18:48 |
mntmn | ok so it doesn't work at all, even with your regular monitor, yes? | 18:59 |
mntmn | Jope: i'm very stupid sorry | 19:00 |
mntmn | Jope: i didn't actually export the working one | 19:00 |
mntmn | my god i should just retire | 19:00 |
mntmn | Jope: if you have the patience, try http://dump.mntmn.com/BOOT.bin-z3-1.9a-20210622b | 19:04 |
Jope | I do :-D son't worry | 19:04 |
mntmn | this one works with both my capture devices | 19:05 |
mntmn | interestingly, the MiraBox can handle the mode switches with no problem | 19:05 |
Jope | let's see what the tv likes | 19:06 |
Jope | it does! | 19:06 |
Jope | testing all modes now | 19:07 |
Jope | all rtg modes show in p96mode | 19:08 |
Jope | ntsc/pal captured modes also work | 19:09 |
mntmn | oh, that's good news | 19:10 |
Jope | hmmmm pal scrolling jerks with this firmware | 19:10 |
mntmn | yeah | 19:10 |
Jope | ok, was to be expected then | 19:10 |
mntmn | this is possible | 19:10 |
mntmn | i can try bnu's modeline again | 19:10 |
Jope | ok | 19:11 |
mntmn | Jope: you have 50hz mode enabled yes? | 19:18 |
mntmn | Jope: and this is what is jerky? | 19:18 |
mntmn | (for pal) | 19:18 |
Jope | yeah I was playing pinball fantasies | 19:18 |
Jope | in pal | 19:18 |
mntmn | and with fw1.8 this was better? | 19:23 |
Jope | yep. not sure actually which fw I had before | 19:25 |
Jope | but 1.8 or some intermediate from bnu | 19:26 |
mntmn | ah :3 | 19:27 |
mntmn | special sauce | 19:27 |
mntmn | oh i have strange mouse pointer jumps | 19:28 |
mntmn | ah but amiga is in an unhappy state | 19:29 |
mntmn | low 5V i guess | 19:29 |
mntmn | i need a more powerful picopsu, this one is getting very hot | 19:30 |
Jope | I'm not totally convinced about the picopsu to be honest | 19:31 |
Jope | in a big box amiga | 19:31 |
Jope | now a modern sfx psu on the other hand.. | 19:31 |
mntmn | hmm | 19:33 |
mntmn | i have a 120W picopsu in there | 19:33 |
mntmn | i mean without the warpengine it was fine | 19:33 |
mntmn | but with it, i'm at 4.6V again | 19:33 |
mntmn | i do have an SFX but naja lets try that one again... | 19:34 |
Jope | 120W should be quite a bit but maybe the WE adds quite a bit of draw on +5due to old simms and such | 19:40 |
mntmn | ah, that one is gutted | 19:40 |
mntmn | the question is how much of that 120W is available on the 5V | 19:41 |
mntmn | i suspect not so much | 19:41 |
Jope | max load 6A on the 5V line | 19:45 |
mntmn | ugh yeah so that's not enough | 19:46 |
isoriano_ | Wow .. new fw for the ZZ .. any improvement to the alpha i was running with Linux? | 19:48 |
mntmn | hmm the main issue is in the crimping of the a3000 power connector plugging into the mainboard | 19:54 |
mntmn | drops 200mV there somewhere | 19:55 |
mntmn | ok with a 5v bypass cable soldered to a capacitor on the motherboard i get now 4.8V using picopsu | 20:23 |
- S0rin (QUIT: Ping timeout: 258 seconds) (~S0rin@user/s0rin) | 21:11 | |
+ S0rin (~S0rin@user/s0rin) | 21:12 | |
mntmn | still stable so far. | 21:44 |
mntmn | testing impact of secondary palette on fpga timing | 22:22 |
- xet7 (QUIT: Quit: Leaving) (~xet7@85-156-127-25.elisa-laajakaista.fi) | 22:30 | |
+ xet7 (~xet7@85-156-127-25.elisa-laajakaista.fi) | 22:32 | |
mntmn | _Bnu: i pushed 2 possibly interesting-for-you things! | 23:06 |
mntmn | _Bnu: prep for secondary split-screen palette https://source.mnt.re/amiga/zz9000-firmware/-/commit/4d054eedb9afe8e755a881bb8e756fb8a20d305c | 23:06 |
mntmn | _Bnu: sync tweaks fixing camlink and possibly other devices/tvs https://source.mnt.re/amiga/zz9000-firmware/-/commit/fc689094bdf455a43ff7e471a7057b7a2e899fe6 | 23:07 |
- xet7 (QUIT: Remote host closed the connection) (~xet7@85-156-127-25.elisa-laajakaista.fi) | 23:19 | |
+ xet7 (~xet7@85-156-127-25.elisa-laajakaista.fi) | 23:23 | |
_Bnu | Sneaky snakes! | 23:34 |
- isoriano_ (QUIT: Ping timeout: 265 seconds) (~isoriano@5.147.188.96) | 23:36 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!