17:58:44 #startmeeting 17:58:44 Meeting started Tue Nov 10 17:58:44 2020 UTC. The chair is pollo. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:58:44 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:58:49 #topic Roll Call 17:58:52 o/ 17:59:00 please say hello if you are here for the meeting! 17:59:13 #link agenda: http://deb.li/oNCD 18:00:05 is it just the two of us? 18:00:42 let's wait a little 18:00:44 but that would be sad 18:01:15 and/or fast 18:01:45 im reading 18:02:54 #topic Miniconfs - Infra State 18:03:06 tumbleweed: can you summarise the changes to the OBS VM? 18:03:09 not sure I followed 18:03:26 so, from my understanding 18:03:44 highvoltage bumped the size (can't recall whether that is above the planned size or not) 18:04:20 it was still struggling, so he pre-rendered a bunch of scenes, which helped 18:04:32 I bumped it to one larger than planned, but I also converted some scenes to pre-rendered ones that we would do live otherwise 18:04:42 so back to the original size *might* be fine now 18:04:50 those pre-renderings do contain some artifacts (I think) 18:05:20 #action highvoltage to investigate the OBS VM sizing 18:05:24 yeah due to some earlier experiments they rendered as flv, I'll re-render those as mp4 then they should be better 18:05:39 (colour is off too and I think that's also an flv fault) 18:05:53 flv is just a container. But I presume it implies a preset in your editor 18:06:12 yeah 18:06:14 I'd suggest using the same h.264 settings (or roughly similar) as we use for our recordings 18:06:51 I don't there is much to add to this topic 18:06:54 the rest is more setup 18:07:02 highvoltage: I assume it's the main area that's pre-rendered and the bar at the bottom is still composited on it? 18:07:06 tumbleweed: it already looks a lot better after that vbr change (can't remember what the newer technique is called off-hand right now) 18:07:29 I switched it to crf, which is constant-quality at an unlimited bitrate 18:07:51 tumbleweed: yeah, for the most part. the "you're watching" scenes is compositing a little but it seems fine on the cpu in that case 18:08:05 ah yes, crf. yeah that seemed to have helped 18:08:43 #topic Miniconfs - Setup 18:08:55 I ansibled all the VMs last week 18:09:04 we since fixed a buch of issues 18:09:10 #link setup status: https://pad.online.debconf.org/p/mcdo2_setup_status 18:09:12 I did one of my two actions: better VU meters in vogol 18:09:27 #info the VU meter in vogol has been improved 18:09:31 still want to improve etherpad auth (which means switching to a new auth plugin, now that somebody has forked mine and taken over maintainance) 18:09:42 the nodejs world is quite forky 18:09:56 lol, they forked your stuff since dc20 18:10:23 I think all's left is grafana? 18:10:43 and OBS, but we've talked about this already 18:11:12 grafana is up 18:11:20 but will probably see some more tweaking 18:11:27 but the dashboards aren't configured yet 18:11:27 I'm focussed on content related stuff because that's the most critical 18:11:34 pollo: they are 18:11:42 but I'll get OBS essentials in ansible if no one beats me to it 18:12:18 tumbleweed: I don,t see them then 18:12:32 OBS is going to be a bit of a snowflake for these minidebconfs. ideally I'd like to rip out the current scene sequencer completely, hopefull we can use something completely different / NIH'y in the future 18:13:07 (or perhaps get upstream Advanced Scene Switcher to add some more features, either way not going to happen this month) 18:13:16 ah, I do see the dashboard, they're just hidden 18:13:27 anyway, I think we're in a good spot 18:13:41 pollo: I think you need to configure our own home screen 18:13:48 *your* 18:13:54 #action pollo to configure the default home screen 18:14:06 do we know if SReview is up? 18:14:11 I'm following 18:14:13 I'd assume not 18:14:39 we can try to copy & tweak dc20 config 18:15:01 I think wouter was waiting on the schedule 18:15:06 highvoltage: how final is the one published? 18:15:37 I know one talk is pending confirmation, and I know one was going to be moved (might be already, checking...) 18:15:54 that sounds final enough to not wait 18:15:57 +1 18:16:13 tumbleweed: want to chat with wouter and see if we can get Sreview up before the weekend? 18:16:14 yeah it's pretty much there 18:17:04 I see the jitsi links are there too, so thanks 18:17:28 pollo: sure 18:17:47 #action tumbleweed to follow with wouter and try to get Sreview working before the weekend 18:17:54 I think that's all for setup? 18:18:18 I'm going ahead and creating all the html templates for the jitsi rooms too. jitsi from obs really works just fine so I'll keep that as a back-up in case jibri misbehaves at any point. 18:18:32 sreview also needs title screen SVGs 18:18:36 I think highvoltage is on that? 18:18:44 yep 18:19:08 #action highvoltage to finish the sreview title cards 18:19:16 even to start. 18:19:24 we wanted to get a video through the entire pipeline 18:19:35 yeah, I'd like us to do a dry-run on this stack 18:19:50 we can try something before the training session, friday? 18:20:00 yep 18:20:10 if sreview isn't ready yet, we'll still have a recording that we can put through it later 18:20:31 can someone send me the link to the DC20 SVG? just so that I know what the variables in there should be called and so on 18:20:31 I'm free the entire day, any preference? 18:20:39 if not we can do 18:00 UTC 18:20:44 highvoltage: it's in the dc20-online repo 18:21:19 https://salsa.debian.org/debconf-team/public/data/dc20-online/-/tree/master/artwork/video 18:21:32 found it, thanks 18:21:37 pollo: yeah 16:00 onwards works for me 18:22:05 #agreed Stack Test Run on Friday Nov 13th @ 18:00 UTC 18:22:20 #topic Miniconfs - Training 18:22:26 I'm brewing beer on sunday 18:22:44 but I can schedule something on sat 18:23:47 do we have volunteers yet? do we need to do outreach? 18:24:10 if we agree on training this weekend, I can take care of that 18:24:24 I could do a session on sat and ask paddatrapper if he can do one on sun 18:24:26 yeah we need to do ourreach, I'm planning to at least reach out to the .br and malayalam organisers to send their video volunteers for that 18:24:45 I'm happy to help out this weekend 18:24:52 but again, can only do 16:00 onwards 18:24:53 from br is me and @lenharo 18:24:58 #agreed we'll have training session this weekend 18:25:13 later than 16:00 will probably be better for brazilians anyway 18:25:27 on weekends it's ok after 13h UTC 18:25:31 #action pollo to setup the wiki pages for training and check with paddatrapper and tumbleweed for training sessions 18:25:51 phls: we're on a similar timezone, so it should be ok :) 18:25:57 but after 16h UTC is better indeed 18:26:04 ah, yes :-) 18:26:18 anything else on training? 18:26:24 yes 18:26:44 is this training for operate the room? or something deeper? 18:26:54 just operations 18:26:58 ok 18:26:59 the deeper stuff is all already done 18:27:12 (mostly) :) 18:27:58 #topic Miniconfs - MDCO Brazil 18:28:14 I don't know what would be useful there? but happy to walk you through the stack if it helps? 18:28:18 do you want to chat now or should we setup a meeting later? 18:28:24 phls: ^^ 18:28:41 well, I have a general question 18:29:11 MDCO Brazil will be 12h to 23h UTC. How deep we (br-team) will need to setup things? 18:29:28 or there will be always from you with us? 18:29:40 you need to give us a schedule, preferably in the "pentabarf" format 18:29:51 pollo: they have a wafer instance, so that's easy 18:29:55 ah great 18:30:18 phls: have you seen the current loop? we'll have to set up at least a basic scene collection if you want a loopy for the weekend 18:30:21 phls: we'll try to have someone from the core videoteam at all times 18:30:35 but atm we can't guarantee anything 18:30:39 so basically we need a logo for the bottom (I didn't see one on the website so I could make something up too) 18:30:51 highvoltage: which reminds me. Do we want a non-OBS static loop for mdco2? (or static image) 18:30:55 also bgloop 18:30:57 and if you have any specific music that you want to play or want to find then that's also a todo 18:31:06 (and again, these would apply to mdco.br too) 18:31:16 it was useful for dc20 18:31:22 I think we should, as a backup 18:31:27 tumbleweed: I was thinking that we'd have a standby loop like the backup sponsors list we had during dc20? 18:31:29 ah, we will need to add 2 logos from sponsors on this loopy 18:31:38 highvoltage: ack. Can we action you to create those? 18:31:43 yeah 18:31:59 #action highvoltage to generate statics loops 18:32:03 maybe a bgloop won't be necessary 18:32:08 we only have recordings and jitsi 18:32:09 so side-by 18:32:11 what's the bgloop? 18:32:13 -side won't happen 18:32:24 bgloop is shown behind the voctomix layouts that don't fill the screen 18:32:33 phls: if you have specific questions, I think writing on the mailing list would be simpler? 18:32:45 https://debconf-video-team.pages.debian.net/docs/loops.html#safe-zones-for-the-background-loop 18:32:45 pollo, yes, its ok to me 18:33:16 but, with this simplified voctomix setup, I guess we don't need a bgloop 18:33:27 phls: can you (or someone from mdcobr) put together a slide with the sponsors that we can use? 18:33:47 phls: and also a logo that fits in the bar? I guess the debian brazil logo with minidebconf online or similar next to it would work 18:34:39 highvoltage, ok, I will ask details to you latter 18:34:44 great 18:34:54 thanks 18:34:57 #action phls (or someone from mcdobr) to provide a slide with sponsor logos + a general logo for the conf 18:35:00 anything else? 18:35:05 no 18:35:17 #topic Video archive documentation 18:35:24 I haven,t touched that 18:35:42 #topic Any other business 18:36:02 yeah 18:36:25 tumbleweed: I'm not sure how video.js is usually embedded on the site 18:36:39 could it be added to the home page? 18:36:45 highvoltage: yeah... it's a little messy 18:36:47 it totally could be 18:37:03 let's start by looking at what terceiro has done in the wafer-debconf branch you're using 18:37:07 it may be there already 18:37:21 ok 18:38:06 we have received two pre-recorded talks so far 18:38:20 thursday is the deadline so we should have quite a few by the weekend for testing 18:38:37 #info thursday is the deadlin for pre-recordings 18:38:48 I'm uploading them all the obs1 machine too as a backup and they should rsync really quick from there to their final destination 18:39:00 I assume they're being reviewed for quality? 18:39:00 highvoltage: we're not likely to use them from training though 18:39:09 of course 18:40:17 highvoltage: you want to copy them to voctomix1:/home/videoteam/recorded-talks/ 18:40:45 thanks 18:41:14 nothing else from me 18:41:25 #topic Next Meeting 18:41:43 Tue Nov 17th @ 18:00 UTC ? 18:41:47 +1 18:42:13 we'll have had a dry run + training, so it'll be a good moment to assess what remains to be done 18:42:41 #agreed Next Meeting: 17 Nov @ 18:00 UTC 18:42:44 #endmeeting