| - S0rin (QUIT: Ping timeout: 260 seconds) (~S0rin@user/s0rin) | 00:29 | |
| + S0rin (~S0rin@user/s0rin) | 00:30 | |
| - S0rin (QUIT: Ping timeout: 264 seconds) (~S0rin@user/s0rin) | 00:56 | |
| + S0rin (~S0rin@user/s0rin) | 01:02 | |
| - mtm (QUIT: Ping timeout: 272 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 01:04 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 01:08 | |
| + XYZ (~XYZ@89-24-40-5.nat.epc.tmcz.cz) | 02:04 | |
| - XYZ (QUIT: Ping timeout: 272 seconds) (~XYZ@89-24-40-5.nat.epc.tmcz.cz) | 02:11 | |
| + XYZ (~XYZ@89-24-40-5.nat.epc.tmcz.cz) | 02:16 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 03:09 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 04:00 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 04:07 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 04:11 | |
| - XYZ (QUIT: Ping timeout: 265 seconds) (~XYZ@89-24-40-5.nat.epc.tmcz.cz) | 04:51 | |
| + eery[m] (~eerymatri@2001:470:69fc:105::2:e4bc) | 04:55 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 05:00 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 05:02 | |
| + XYZ (~XYZ@89.24.40.5) | 05:13 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 05:53 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 05:57 | |
| - joeyh (QUIT: Quit: ZNC 1.8.2+deb2+b5 - https://znc.in) (joeyh@2600:3c03::f03c:91ff:fe73:b0d2) | 06:59 | |
| + joeyh (~joeyh@kitenet.net) | 06:59 | |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 07:17 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 07:27 | |
| minute | https://iwp9.org/ | 11:50 |
|---|---|---|
| bluerise | minute: frohe Weihnachten :) | 11:59 |
| + bgs (~bgs@212-85-160-171.dynamic.telemach.net) | 12:09 | |
| josch | Linux reform 6.1.0-reform2-arm64 #1 SMP Debian 6.1.1-1~exp1+reform20221223T050906Z1 (2022-12-23) aarch64 GNU/Linux | 12:21 |
| josch | seems like 6.1.1 does not only work on the imx8mp :) | 12:22 |
| josch | this means i'm going to experiment with h265 decoding on hantro with gstreamer now as well | 12:22 |
| - mtm (QUIT: Ping timeout: 248 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 13:02 | |
| Boostisbetter | bluerise: minute: also from me! | 13:29 |
| + MajorBiscuit (~MajorBisc@31-23-159.netrun.cytanet.com.cy) | 13:50 | |
| minute | apparently the mnt shop is broken. investigating | 14:42 |
| c-keen[m] | <josch> "this means i'm going to experime..." <- That's awesome | 14:47 |
| josch | just getting a green screen only so far -- i'm not an expert XD | 14:49 |
| minute | i personally don't celebrate christmas, but thanks and a merry one to y'all | 14:57 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 15:10 | |
| minute | turns out this wayland branch of xfwm4 is pretty hard to build. | 15:30 |
| - qbit (QUIT: Quit: WeeChat 3.7.1) (~qbit@h.suah.dev) | 16:16 | |
| minute | it's also not very functional | 16:28 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 17:21 | |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 17:32 | |
| bkeys | The trackball in my Reform has become terrible | 17:35 |
| bkeys | It's smooth but 80% of the time it barely moves the cursor | 17:35 |
| sigrid | the sensor got dirty? | 17:36 |
| bkeys | Even if I put a bit of pressure on it, it still does it | 17:36 |
| bkeys | Maybe, it hasn't even been a month and it's not in a very dirty/dusty environment | 17:36 |
| sigrid | i usually blow a bit of strong air at the ball and it works fine afterwards | 17:36 |
| sigrid | a small crumb can make "wonders" if it falls there :D | 17:37 |
| Boostisbetter | bhkeys, sounds like you got dust or lint in there and the sensor is being obstructed. | 17:37 |
| bkeys | Probably, do I need to take the trackball out? | 17:37 |
| Boostisbetter | yes | 17:38 |
| bkeys | Can I do that without taking the entire module out? | 17:38 |
| sigrid | you can try blowing a strong burst of air in it first | 17:38 |
| Boostisbetter | Yes! totally | 17:38 |
| sigrid | the ball automatically goes up a bit when you do that. stuff flies out | 17:38 |
| bkeys | If only I had access to an air compressor or the like | 17:39 |
| sigrid | i use my mouth | 17:39 |
| sigrid | works every time | 17:39 |
| bkeys | Just used my mouth and now it doesn't move the cursor at all | 17:40 |
| sigrid | in one year of owning a reform i took the ball out only once | 17:40 |
| sigrid | haha, oops. | 17:40 |
| Boostisbetter | same year | 17:40 |
| Boostisbetter | I'm going on 2 years now | 17:40 |
| bkeys | Oh wait, it does now? | 17:40 |
| Boostisbetter | still only done it once | 17:40 |
| bkeys | That improved things | 17:40 |
| sigrid | magic | 17:41 |
| bkeys | I will probably still take it out and clean things | 17:41 |
| josch | bkeys: you just need to remove the two round trackball buttons on each side of the trackball to take it out | 17:53 |
| josch | after taking out those two buttons you can loosen the screws and take the plastic frame and the ball out | 17:53 |
| josch | i also just blow into it with my mouth and everything works well again | 17:54 |
| bkeys | Well I blew on it as dry as I could and the ball got coated in a thin layer of saliva | 17:54 |
| bkeys | Whatever is in there is in there pretty good | 17:55 |
| bkeys | josch: Any strategy to pull those keys out without scoring the chassis? | 17:55 |
| josch | bkeys: there are tools for that but it's easy with a thin wire or thread that one puts on the top and bottom of each key to pull them out straight up | 17:57 |
| bkeys | Yeah I used a bent zip tie and they came right off | 17:57 |
| bkeys | Yeah the sensor is obstructed | 17:59 |
| josch | bkeys: when you put it back in, take care of how tight you fasten both screws -- it makes a big difference to how easy the ball moves | 18:03 |
| bkeys | I didn't read that until I put it back in; but it seems to be fine now | 18:06 |
| bkeys | There was a piece of dust right on top of the sensor | 18:08 |
| Boostisbetter | yeah same thing happened to me as well. Moved it away and like butter trackball action ever since. | 18:09 |
| bkeys | The keys came off much easier than I thought; I was thinking it'd be akin to cherry mx keycaps | 18:10 |
| bkeys | I got a new ratcheting screwdriver so I am always looking for a reason to use it | 18:11 |
| bkeys | I like how the culture here is "take it apart, blow on it, etc." which would almost never be the case with commodity hardware | 18:19 |
| bkeys | Taking it apart would be a massive pain in the ass | 18:19 |
| Boostisbetter | yep, the Framework community is very similar as well, but the Reform crowd is just extra awesome, and you have the distinct feeling that it is so by design. | 18:19 |
| bkeys | I wish framework would have coreboot support | 18:20 |
| bkeys | But I also like that the Reform is thick and isn't another thin ultrabook laptop | 18:22 |
| bkeys | It'd be great if we had Vulkan or GL3 on the Reform | 18:26 |
| bkeys | GL3/Vulkan and 8GB of RAM would be great | 18:36 |
| minute | vulkan will be possible with the cm4 and later with rk3588 (that one is gonna take a bit more time though) | 18:45 |
| minute | 8gb ram is also possible with imx8mplus or ls1028a | 18:45 |
| minute | (no vulkan there for now) | 18:45 |
| Boostisbetter | I read that the imx8plus could also do 16gb RAM, is that true? | 18:58 |
| Boostisbetter | I really wonder what the real world differences are going to be between 2 fast cores and 4 more efficient cores. | 18:58 |
| minute | Boostisbetter: where did you read that? | 19:13 |
| Boostisbetter | minute, I'm looking now, but I did see that it can also come 4 x processors instead of 2. You didn't go with that because of the power and thermal constraints? | 19:20 |
| Boostisbetter | I can't seem to find the RAM thing, so I think I just misread that. | 19:21 |
| bkeys | Yeah that is why I am excited about cm4 | 19:21 |
| Boostisbetter | yeah maybe it was the CM4 that can do 16gb. | 19:22 |
| Boostisbetter | although I must say, I was skeptical of 4gb being enough, and the Reform has proven that it is just fine, especially if coupled with a fast NVME drive. | 19:23 |
| minute | Boostisbetter: probably we're talking about two different chips? | 19:24 |
| bkeys | I have been brought to my knees cause of lack of RAM a few times but the pros of the Reform far outweigh it | 19:24 |
| minute | Boostisbetter: the imx8mplus we're using has 4 cores. | 19:24 |
| minute | Boostisbetter: the ls1028a has only two cores (but they are cortex-a72) | 19:24 |
| minute | just published an update article! https://mntre.com/media/reform_md/2022-12-24-december-update.html | 19:25 |
| Boostisbetter | ahh and the ls1028a can have up to 16gb. I think that is what I read. | 19:30 |
| bkeys | minute: What is your estimate of once the crowdsupply is launched for the pocket reform and them starting to ship? | 19:39 |
| minute | Boostisbetter: correct, it could have even more, but not on that little board | 20:05 |
| minute | bkeys: 6 months | 20:05 |
| minute | bkeys: the longest lead time is that of the imx8mplus modules | 20:05 |
| Boostisbetter | minute, when do you think you'll be launching the crowdfunding for the Pocket? | 20:13 |
| minute | Boostisbetter: in jan or feb... probably feb because shooting a little movie will take a bit of time | 20:16 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 20:16 | |
| rafostar[m] | <josch> "just getting a green screen only..." <- Have you looked for any errors/warnings when run with `GST_DEBUG=3`? Also, a pipeline graph (`GST_DEBUG_DUMP_DOT_DIR="/path/to/some/dir" and paused-playing file from it) would be useful to confirm what exactly is being used during playback. Green screen often occurs when GL fails at upload or colorspace conversion. | 21:58 |
| minute | interesting https://github.com/LBCrion/sfwbar | 21:59 |
| josch | rafostar[m]: i first tried backporting the h265 patches to gstreamer 1.20 which is the current version in debian but now i'm just packaging the current git instead | 21:59 |
| minute | josch: huh, i also got a green video yesterday with clapper+a h265 file, but later there was even some green tinted content in there. i was pretty sure it was software decoded though. so probably a colorspace problem like rafostar[m] says | 22:00 |
| minute | so maybe it's a bug in clapper or gtk | 22:00 |
| josch | doing this will be a bit easier once the gstreamer 1.22 release really happened ^^ | 22:00 |
| minute | when we can play h265, we can do something my ps4 can't ;) | 22:01 |
| c-keen[m] | Sfwbar Looks interesting | 22:02 |
| rafostar[m] | josch: For an easy way of testing latest GStreamer with different devices, Flatpak Nightly build is available at Clapper github actions. It has latest GStreamer from git built every 24h. AArch64 version is also available there. | 22:03 |
| minute | chuckled > Copy sfwbar.config from /usr/share/sfwbar/ to ~/.config/sfwbar/ If you prefer to start with something more like tint2 bar, you can copy t2.config into ~/.config/sfwbar/sfwbar.config instead. If you want something like waybar, you can copy wbar.config and if you prefer something from the darker side, w10.config could be for you | 22:03 |
| josch | rafostar[m]: uuuh, nice! thanks for the tip! | 22:04 |
| josch | rafostar[m]: the flatpak nightly works!!! | 22:36 |
| josch | it's definitely not software decoding because cpu usage is at around 13% :) | 22:36 |
| minute | josch: ohh :3 | 22:36 |
| josch | what an amazing xmas present :D | 22:36 |
| josch | there is a very thin green bar at the bottom of the h265 10bit video but it's very tiny and i guess this artifact is another indication that the hardware decoder is really used :) | 22:37 |
| minute | josch: can we bump the kernel to 6.1 already? | 22:37 |
| josch | minute: yes, we can but it will be a bit more work than the hack i have done locally | 22:38 |
| minute | ahh alright then | 22:38 |
| josch | minute: are we in a rush? | 22:38 |
| minute | nope | 22:38 |
| minute | all in due time | 22:38 |
| josch | good, then we will just wait until the kernel gets uploaded to unstable | 22:38 |
| josch | since this works with clapper flatpak, i'm more motivated to get the gstreamer 1.22 packaging done :D | 22:39 |
| minute | josch: just to clarify, is 6.0 sufficient for h265? i mean, i do have > [ 27.459624] hantro-vpu 38310000.video-codec: registered nxp,imx8mq-vpu-g2-dec as /dev/video1 | 22:39 |
| josch | minute: according to phoronix, 6.1 is needed for h265 | 22:39 |
| minute | ah, alright | 22:40 |
| josch | minute: if you are on 6.0, you could try the flatpak nightly and see what happens | 22:41 |
| minute | josch: this needs to install the whole org.gnome.Platform etc, is that correct? | 22:45 |
| josch | yes | 22:46 |
| josch | at least it did for me | 22:46 |
| minute | ok | 22:46 |
| josch | (i never used flatpak before) | 22:46 |
| minute | i did, a longer time ago... to run chromium | 22:46 |
| minute | luckily those times are over :D | 22:47 |
| minute | the install TUI is neat though | 22:47 |
| josch | it certainly looks fancy :) | 22:47 |
| minute | it works! | 22:50 |
| minute | so, 6.0 is enough | 22:50 |
| josch | awesome :D | 22:50 |
| minute | it has a few glitches though! | 22:53 |
| minute | i wonder if these are also there on 6.1 | 22:53 |
| minute | i used this test file: https://dl5.webmfiles.org/video-h265.mkv | 22:54 |
| minute | > Austrian Alps Time-lapse | 22:54 |
| minute | (5 MB) | 22:54 |
| josch | let me try it out.. | 22:55 |
| minute | imx8mplus also has a video _en_coder. but idk if there is any driver effort for it | 22:55 |
| josch | minute: i see some glitches as well | 22:56 |
| minute | ah! so depends on the file maybe | 22:57 |
| josch | i see some on other files as well | 23:00 |
| minute | aha! | 23:00 |
| minute | so there's still some tweaking required | 23:00 |
| josch | indeed -- but at least i can watch h265 at all! :D | 23:01 |
| josch | zero glitches in fast-paced anime intro -- some glitches in slow-pan scenes | 23:01 |
| rafostar[m] | <josch> "it's definitely not software..." <- In the Clapper GUI when you open video track selection popover (next to progress bar) you can see name of the video decoder used during playback to confirm if HW decoder is used. In your case it should be v4l2slh265dec. | 23:02 |
| josch | rafostar[m]: confirmed! that's what it's showing: https://mister-muffin.de/p/0PSs.png | 23:05 |
| rafostar[m] | <josch> "since this works with clapper..." <- The GL context selection was improved/fixed on Clapper git some time ago (but after last stable release), so it is also possible that green screen issue is caused by that and that it might not be present in Clapper git. Maybe also be worth trying to confirm if you already have local gstreamer with h265 build. | 23:11 |
| josch | rafostar[m]: aha, also a good point, thanks! Are you planning to make a release with those improvements some time soon? | 23:13 |
| rafostar[m] | <josch> "rafostar: aha, also a good point..." <- It was too big of a change to be put in next patch version release, and for the next minor version I had planned to rewrite quite a lot of stuff that is causing some other problems, but due to being recently busy with work and live, I did not have time to do so this year. I hope to resume working on it soon through. | 23:18 |
| josch | rafostar[m]: okay, no worries! thank you for all the help here! :D | 23:20 |
| rafostar[m] | josch: Sure, always happy to help 😄 | 23:23 |
| minute | yep, greatly appreciated! | 23:38 |
| rafostar[m] | <minute> "josch: just to clarify, is 6.0..." <- Kernel API is considered stable since 6.0, although hardware support for different media drivers is kernel dependent (i.e. some might see support added in later then 6.0 versions). | 23:54 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
