| *** tpb has joined #timvideos | 00:00 | |
| mithro | The HDMI2USB can do 1080p@24fps | 00:01 |
|---|---|---|
| tumbleweed | I don't know if that telecine hack is something other cameras do, if so, one may want the HDMI2USB to be able to reverse it | 00:02 |
| mithro | telecine hack? | 00:03 |
| mithro | tumbleweed: you mean the camera is doing pull-down | 00:03 |
| mithro | ? | 00:03 |
| tumbleweed | yeah, they call it pf24. e.g. https://forums.adobe.com/thread/1170716 | 00:05 |
| tpb | Title: What's the best way to deal with Canon pf24 files? | Adobe Community (at forums.adobe.com) | 00:05 |
| mithro | tumbleweed: yeah - it should be actually pretty trivial to take that input and convert it to real 1080p@24fps output | 00:15 |
| *** CarlFK has joined #timvideos | 00:46 | |
| *** ChanServ sets mode: +v CarlFK | 00:46 | |
| mithro | seaLne: FYI -> https://github.com/mithro/HDMI2USB-TOFE-HDMI4 | 00:58 |
| tpb | Title: mithro/HDMI2USB-TOFE-HDMI4 · GitHub (at github.com) | 00:58 |
| CarlFK | mithro: flterm --port /dev/ttyVIZ0 --speed 115200 | 01:05 |
| CarlFK | Unable to open serial port: No such file or directory | 01:05 |
| CarlFK | sure enough, it doesn't exist | 01:05 |
| mithro | CarlFK: I've explained how /dev/ttyVIZ0 works a bunch of times | 01:05 |
| CarlFK | yeah, this was working a few days ago | 01:06 |
| mithro | CarlFK: check dmesg | 01:07 |
| CarlFK | mithro: er.. hang on. git pull is pulling new stuff, so trying with current files now | 01:08 |
| mithro | CarlFK: this has nothing to do with the stuff in git | 01:08 |
| CarlFK | I was wondering about that | 01:08 |
| mithro | CarlFK: It's provided by https://github.com/mithro/exar-uart-driver | 01:09 |
| tpb | Title: mithro/exar-uart-driver · GitHub (at github.com) | 01:09 |
| mithro | CarlFK: and has very little to do with the HDMI2USB project really | 01:09 |
| CarlFK | mithro: carl@twist:~/src/tv$ apt-cache policy vizzini-dkms | 01:11 |
| CarlFK | Installed: 1.0.0-5~ppa2~ubuntu14.04 | 01:11 |
| CarlFK | whoops. wrong host | 01:12 |
| CarlFK | juser@trist:~/HDMI2USB-firmware-prebuilt/atlys/firmware/unstable$ apt-cache policy vizzini-dkms | 01:13 |
| CarlFK | Installed: 1.0.0-5~ppa2~ubuntu14.04 | 01:13 |
| CarlFK | 500 http://ppa.launchpad.net/timvideos/fpga-support/ubuntu/ trusty/main amd64 Packages | 01:13 |
| tpb | Title: Index of /timvideos/fpga-support/ubuntu (at ppa.launchpad.net) | 01:13 |
| CarlFK | ok, fresh box, it is working | 01:33 |
| tumbleweed | CarlFK: I'd be interested if you have a box where it isn't working | 01:52 |
| CarlFK | tumbleweed: I do. and it was working a week ago | 01:52 |
| CarlFK | i think... | 01:52 |
| CarlFK | it might have a public ipv6 | 01:53 |
| tumbleweed | also, I probably need to get my atlys to work before the debconf video sprint :P | 01:53 |
| CarlFK | give me 20 min to hack up a schedule consumer | 01:53 |
| tumbleweed | but right now is not the time for *that* one | 01:53 |
| CarlFK | na, I got that one down (well, on a trusty box) | 01:53 |
| CarlFK | I hear mithro's openocd patches landed. you don't even want to look at my packaing job, it will make you cry | 01:54 |
| tumbleweed | hehe | 01:54 |
| * tumbleweed needs to go grab something to eat. Back in half an hour or so | 01:55 | |
| CarlFK | k | 01:55 |
| mithro | I have openocd in conda now | 01:55 |
| mithro | If you take a look at the openocd branch, that makes openocd the default programmer | 01:56 |
| tumbleweed | . | 02:45 |
| CarlFK | tumbleweed: I'll be home in 30 min and can get you ssh access to the box that isn't ttyVIZing | 03:00 |
| tumbleweed | that works | 03:01 |
| mithro | CarlFK: have you looked at dmesg yet? | 03:07 |
| *** CarlFK has quit IRC | 03:07 | |
| *** CarlFK has joined #timvideos | 03:25 | |
| *** ChanServ sets mode: +v CarlFK | 03:25 | |
| CarlFK | home.. give me a moment to unpack | 03:27 |
| *** automatical has quit IRC | 03:46 | |
| tumbleweed | getting anywhere? | 04:00 |
| mithro | tumbleweed: Carl seems to have yet to check dmesg or if even the module is loaded / loadable | 04:04 |
| CarlFK | I got ... recruited ;) | 04:06 |
| CarlFK | https://2015.osdc.com.au/schedule/conference.json | 04:06 |
| CarlFK | name": "Opportunities in Openness. Driving positive change in local communities.", | 04:06 |
| CarlFK | room": "", | 04:06 |
| CarlFK | it is happening ... outside! | 04:06 |
| CarlFK | "Thunder Talks" ?! | 04:09 |
| CarlFK | trist booting | 04:22 |
| CarlFK | tumbleweed: what is your ghithub username? | 04:24 |
| tumbleweed | CarlFK: stefanor | 04:25 |
| CarlFK | juser@trist:~$ ssh-import-id gh:stefanor | 04:26 |
| CarlFK | do you know about that? | 04:26 |
| tumbleweed | yep | 04:26 |
| CarlFK | ssh -A -p 2227 [email protected] | 04:27 |
| xfxf | hey CarlFK are you running any >1tb drives in your laptops? | 04:29 |
| xfxf | i can only find two 2gb models that'll fit most laptops; one is a WD green so it's disqualified automatically, the other is a seagate whose reliability i question | 04:30 |
| tumbleweed | CarlFK: seems firewalled | 04:31 |
| tumbleweed | you said something about ipv6? | 04:31 |
| CarlFK | tumbleweed: that was at ps1... | 04:32 |
| CarlFK | tumbleweed: sorry, got the port wrong. try 2722 | 04:32 |
| CarlFK | xfxf: no, just 1T | 04:33 |
| tumbleweed | that's more like it :) | 04:33 |
| CarlFK | xfxf: osdc expected:492 g | 04:33 |
| CarlFK | so 1t should be fine | 04:33 |
| CarlFK | tumbleweed: this is now just forensics | 04:35 |
| CarlFK | if I want it working I will reinstall the OS and start over | 04:36 |
| CarlFK | but I think it is worth figuring out why it is broken | 04:36 |
| xfxf | CarlFK: this isn't for this, it's for future events where i expect i'll be capturing in HD | 04:37 |
| CarlFK | ah | 04:37 |
| xfxf | i have a dedicated encode/storage PC but it's not practical to bring to interstate events i can't drive to | 04:37 |
| tumbleweed | CarlFK: so, you have a vizzini.ko for the previous kernel, but not the current one | 04:37 |
| CarlFK | tumbleweed: what should happen when apt-get upgrade pulls down a new kernel? | 04:38 |
| tumbleweed | yeah, it should | 04:38 |
| tumbleweed | can haz root? | 04:38 |
| tumbleweed | I assume I can sudo but I can't remember your sudo password | 04:39 |
| tumbleweed | virtualbox, 4.3.10, 3.13.0-65-generic, x86_64: installeddkms.conf: Error! No 'BUILT_MODULE_NAME' directive specified for record #0. | 04:40 |
| tumbleweed | guessing that's it | 04:40 |
| tumbleweed | Error! Bad conf file. | 04:40 |
| tumbleweed | can I kill virtualbox-dkms ? | 04:40 |
| CarlFK | sure | 04:40 |
| CarlFK | virtualbox... this sounds familiar... like "let me bla bla up a vm..." | 04:41 |
| tumbleweed | no, it wasn't virtualbox | 04:42 |
| tumbleweed | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=651973 I guess | 04:47 |
| tpb | Title: #651973 - BUILT_MODULE_NAME[#] required if linux-headers-* is installed/updated after the foo-dkms package - Debian Bug report logs (at bugs.debian.org) | 04:47 |
| tumbleweed | mithro: https://github.com/mithro/exar-uart-driver/pull/3 | 05:07 |
| tpb | Title: DKMS fixes by stefanor · Pull Request #3 · mithro/exar-uart-driver · GitHub (at github.com) | 05:07 |
| tumbleweed | CarlFK: I'll upload that to the PPA now | 05:07 |
| CarlFK | oh cool, I didn't realized there was a fix | 05:07 |
| tumbleweed | there wasn't :P | 05:08 |
| CarlFK | lol - yay bugfix! | 05:08 |
| CarlFK | I thought it was because this schedule data is making me rage | 05:09 |
| CarlFK | 2799 if event['conf_key']==21: | 05:09 |
| CarlFK | 2801 event['location'] = 'Derwent 1' | 05:09 |
| tumbleweed | :) | 05:09 |
| CarlFK | I was able to figure that out by looking at how FF renders the html | 05:10 |
| tumbleweed | hah | 05:10 |
| CarlFK | and hope that is how the presenter figures out where to stand and talk | 05:10 |
| tumbleweed | mithro: ta | 05:10 |
| * tumbleweed was amused to discover that dkms won't automatically build against new kernels, by default | 05:11 | |
| CarlFK | cuz there are a handful of items in the json that don't have a room specified, and I don't see any other clue ... so .. | 05:11 |
| xfxf | the optics and lens in those consumer cameras is going to suck for recording non-well lit venes | 05:18 |
| xfxf | there's a reason i splash $2k on my cameras | 05:18 |
| xfxf | and for what they are/do they're cheap | 05:18 |
| xfxf | s/is going/are going/ | 05:18 |
| xfxf | CarlFK: i got a dude fixing the source data now | 05:19 |
| tumbleweed | yeah, you can't beat a large lens | 05:19 |
| xfxf | mostly just filling in wrong/missing rooms | 05:19 |
| xfxf | CarlFK: i paid 2.5k new for one of my dvx200b's back in 2009 when they still cost that much new | 05:20 |
| xfxf | so you're probably going to end up dissapointed looking at new cameras under that | 05:20 |
| xfxf | relative product performance to price doesn't tend to change over time | 05:21 |
| xfxf | relative in terms of tech available at the time | 05:21 |
| CarlFK | I figured it was something like that | 05:21 |
| xfxf | if you want to go cost effective | 05:22 |
| xfxf | i'd be looking at an older HDV camera that can do 720p out | 05:22 |
| xfxf | HDV is 1080i max | 05:22 |
| CarlFK | the $300 one is that | 05:22 |
| xfxf | a lot of people will be getting rid of good cameras with good optics because they want full 1080p + newer formats/codecs | 05:22 |
| xfxf | probably worth looking at that then, as long as it will actually do 720p out via the HDMI | 05:23 |
| CarlFK | the $300 may be good for me to experiment with. | 05:23 |
| xfxf | yeah for $300 sounds like a no brainer | 05:23 |
| xfxf | worst case scenario sell it again | 05:23 |
| xfxf | oh mlt | 05:23 |
| xfxf | this is a useful error | 05:23 |
| xfxf | LOG: MLTDISABLED <b>swfdec</b> : this is an <b>optional</b> module that provides ... dunno ... . To enable it, try installing a package called something like ... dunno .... | 05:24 |
| xfxf | ERROR: One or more required MLT modules could not be enabled | 05:24 |
| CarlFK | LOL | 05:24 |
| CarlFK | I love that error ;) | 05:24 |
| CarlFK | I have stopped building and use the one that comes from the shotcut tarball | 05:24 |
| xfxf | i'm updating your mkmlt.sh script to actually install everything it needs - i'll send PR back so i don't have to do this dance agin | 05:24 |
| tumbleweed | CarlFK: PPA package published, and installed on that box of yours | 05:25 |
| CarlFK | tumbleweed: thanks tons. that thing was bugging me in a I am getting stressed out not understanding what could be going on | 05:26 |
| xfxf | CarlFK: you probably want to update your install documentation, it's quite out of date now | 05:26 |
| CarlFK | xfxf: well.. I kinda want to bust it up into two or 3 separate parts: web server, encoder, and something else I can't remember | 05:27 |
| xfxf | yeah, nod, makes sense | 05:28 |
| xfxf | it's just eating time trying to get an instance working, working through errors and poking through code | 05:28 |
| xfxf | and yeah i know i can bug you but i need to understand the changes, so i'll do that if i get stuck | 05:29 |
| CarlFK | you should poke me when you bump so I can tell you what's going on | 05:29 |
| xfxf | if i get too stuck i will - just again helps me to dig into this to understand i | 05:37 |
| xfxf | rather than asking you the same dumb questions every 2-3 months | 05:37 |
| CarlFK | $800 Sony HVR-HD1000U http://chicago.craigslist.org/wcl/pho/5288456113.html http://www.bhphotovideo.com/c/product/518555-REG/Sony_HVRHD1000U_HVR_HD1000U_Digital_High_Definition.html#Specification | 05:39 |
| tpb | Title: Sony HVR-HD1000U Digital HD HDV Camcorder (slightly used) (at chicago.craigslist.org) | 05:39 |
| CarlFK | hmm, no xlr in | 05:39 |
| tumbleweed | minijack, though | 05:41 |
| xfxf | no xlr in shouldn't matter - hdmi2usb isn't getting the audio from the camera | 05:43 |
| xfxf | we're unfortunately still going to have to drag around mixers, which i was hoping to eradicate | 05:43 |
| tumbleweed | xlr in is usually from a mixer. It means the audio and video are naturally in sync | 05:44 |
| xfxf | fortunately you can get quite light ones these days | 05:44 |
| CarlFK | reviews" it's not great in low light rooms. Camera is great for shooting outside and in great lighting" | 05:44 |
| tumbleweed | lol | 05:44 |
| xfxf | tumbleweed: sure, but we typically usually feed xlr in from a venue mixer to the camera | 05:44 |
| xfxf | and capture audio/video from firewire | 05:44 |
| tumbleweed | yep | 05:44 |
| xfxf | we can't do that with hdmi2usb, it'll need to be injected into a sound card/device | 05:44 |
| xfxf | which changes the setup a little | 05:45 |
| tumbleweed | "can't" is presumably a bit strong there | 05:45 |
| tumbleweed | bandwidth the issue? | 05:45 |
| xfxf | i asked the other day and was told no | 05:45 |
| xfxf | if we can - that would be amazing | 05:45 |
| xfxf | means much simpler cabling/setups/less equipment | 05:45 |
| CarlFK | mithro said "later" | 05:45 |
| xfxf | right | 05:46 |
| xfxf | okay, good | 05:46 |
| CarlFK | xfxf: I am thinking to run xlr into the camera, use the camera's limiter and gain and vu meter, then analog out into a usb audio device | 05:46 |
| xfxf | i'm not aware of cameras that will do 'analog out' | 05:46 |
| CarlFK | or maybe firewire ... which makes me cry | 05:46 |
| xfxf | if you mean the hadphone jck you probably don't want do to that, it'll sound rubbish | 05:46 |
| xfxf | and if you mean the audio/video out jacks they're usually non-existant on HD cameras | 05:47 |
| CarlFK | the dvx100 does it. that one I just posted: Audio L/R Output: RCAx2 | 05:47 |
| xfxf | some of the HDV models might be a diff story tho as they're sort of transitional between SD and HD | 05:47 |
| xfxf | dvx100 isn't a HD camera | 05:47 |
| xfxf | i'm talking HD specifically, sorry | 05:48 |
| xfxf | er, HDMI | 05:48 |
| CarlFK | that one I just posted: Audio L/R Output: RCAx2 | 05:48 |
| xfxf | it's not a standard thing rhough iirc | 05:48 |
| xfxf | that said happy to be corrected | 05:48 |
| xfxf | xlr into camera with 2x RCA out sounds workable | 05:49 |
| xfxf | bbl, this laptop is running out of disk i/o and is lagging hard | 05:49 |
| xfxf | i'll keep typing liek an idiot | 05:49 |
| CarlFK | hmm, looking over the $1200 panasoncs I have been looking at, no audio out | 05:51 |
| xfxf | ya | 05:51 |
| xfxf | you may just need to get a small USB sound card that takes USB in and has a compressor onboard | 05:52 |
| xfxf | there's a bunch of them around | 05:52 |
| mithro | tumbleweed: did you upload to the PPA? | 05:52 |
| xfxf | usually fairly light, too | 05:52 |
| tumbleweed | mithro: yep | 05:52 |
| CarlFK | http://www.bhphotovideo.com/c/product/905777-REG/panasonic_ag_af100apj_ag_af100a_digital_cinema_camera.html#Specification Audio Output Connector: RCA pin jack x2 (Ch1/Ch2, 316 mV, 600 Ω) | 05:53 |
| tpb | Title: Panasonic AG-AF100A Digital Cinema Camcorder AG-AF100APJ B (at www.bhphotovideo.com) | 05:53 |
| CarlFK | I knew there was a reason to spend an extra $500 on that one ;) | 05:53 |
| tumbleweed | don't think there's a huge value in going through a camera just for its compressor and vu meter | 05:54 |
| tumbleweed | but then hdim2usb also isn't really targetted at camera capture | 05:54 |
| tumbleweed | don't forget you need a lens :P | 05:54 |
| CarlFK | chicago.craigslist.org/chc/pho/5258082657.html $1750 | 05:55 |
| CarlFK | hey.. wheres the http?! | 05:55 |
| CarlFK | http://chicago.craigslist.org/chc/pho/5258082657.html | 05:55 |
| tpb | Title: Panasonic AF100 (with 60p UPGRADE) camera + Fujinon lens FULL PACKAGE (at chicago.craigslist.org) | 05:55 |
| tumbleweed | that's a weird looking body | 05:55 |
| CarlFK | right! it looks like a ray gun | 05:56 |
| mithro | tumbleweed: copied to the other distro versions too? | 05:56 |
| tumbleweed | mithro: yep | 05:57 |
| CarlFK | I like that putting audio into the camera is a known working everything so that when things are wonky it helps isolate where the problem is. like being able to plug headphones into the camera has become my normal place to monitor audio | 05:57 |
| xfxf | tumbleweed: if hdmi2usb isn't targeted at camera capture, how are cameras intended to be captured? | 06:16 |
| CarlFK | it is targeted at camera capture | 06:19 |
| tumbleweed | I'd have said it's targetted at laptop capture | 06:20 |
| tumbleweed | and could do camera capture, as many cameras output HDMI | 06:21 |
| tumbleweed | not doing audio is a bit of an issue for camera capture :P | 06:21 |
| tumbleweed | you don't need HDMI-out again, so you do have many more capture options | 06:21 |
| tumbleweed | (even if they're all awful) | 06:22 |
| CarlFK | I was tempted to look into others for Nodeshow, but meh. there isn't a mixer ready | 06:23 |
| seaLne | xfxf: i've used 2x HGST Travelstar 1Tb laptop drives for the last 2 years and they have been fine | 07:23 |
| seaLne | mithro: would it make sense to collect kicad libs/mods that we create in the tofe template repository? | 07:25 |
| seaLne | mithro: your "stick it on github" for the mini-dp was a bit vague :) | 07:25 |
| mithro | seaLne: I mean - just put it up in a temporary github repo or gist or something | 07:28 |
| seaLne | ah | 07:28 |
| seaLne | mithro: https://gist.github.com/sealne/d50c3411aabfb9e28305 | 07:31 |
| tpb | Title: TE-2129320 · GitHub (at gist.github.com) | 07:31 |
| mithro | Guess I should order some of the parts so I can do a paper fit | 07:31 |
| CarlFK | seaLne: I haven't fogotten i owe you some debugging help. but not now. I need to get to bed. | 07:34 |
| seaLne | for a pleasant change its datasheet was pretty understandable | 07:34 |
| mithro | seaLne: BTW how good is your git? | 07:36 |
| seaLne | pretty basic | 07:37 |
| mithro | seaLne: okay, let me construct a special repo for the VGA board | 07:38 |
| mithro | I'll basically merge the TOFE template and Rohit's repo | 07:40 |
| seaLne | i had some problems with the rename script when i had a quick look yesterday btw https://paste.kde.org/p2mmkquh6 | 07:42 |
| tpb | Title: KDE Paste (at paste.kde.org) | 07:42 |
| seaLne | not looked at that properly though | 07:43 |
| xfxf | having problems trying to do PROG=fpgalink make load | 08:02 |
| xfxf | on latest git | 08:02 |
| xfxf | https://www.irccloud.com/pastebin/0RWnY5w4/ | 08:02 |
| tpb | Title: Pastebin | IRCCloud (at www.irccloud.com) | 08:02 |
| xfxf | and i uninstalled the ixo-usb-jtag package | 08:02 |
| xfxf | unsure whether this is how i should still flash the device? | 08:03 |
| seaLne | i think its needed for getting to the jtag stage? | 08:03 |
| xfxf | i'll brb in about 2 hours to try setup again (have a test setup ready for last day of osdc, keen to attempt recording) | 08:03 |
| seaLne | mithro: i'm assuming high speed tofe boards will need to be 4 layer. due to the connections on both sides? | 08:07 |
| mithro | You can't use ixo-usb-jtag with PROG=fpgalink | 08:07 |
| mithro | I'll be around in 30 ish minutes | 08:09 |
| mithro | Bblr | 08:09 |
| mithro | seaLne: if your careful you might be able to get away with 2 layer but it isn't worth the effort, the price difference is just not worth the time. | 08:10 |
| seaLne | mithro: k, thats what i thought. and trying to debug random problems later wouldn't be fun | 08:12 |
| mithro | seaLne: ask Rohit :-) | 08:32 |
| seaLne | heh | 08:34 |
| *** se6astian|away is now known as se6astian | 08:35 | |
| *** rohitksingh has joined #timvideos | 08:43 | |
| mithro | speaking of the devil, Hi rohitksingh | 08:46 |
| *** tvCommitBot has joined #timvideos | 08:46 | |
| tvCommitBot | [HDMI2USB-vmodvga] mithro pushed 2 new commits to master: http://git.io/vWH2q | 08:46 |
| tvCommitBot | HDMI2USB-vmodvga/master 2dfaa90 Tim 'mithro' Ansell: Removing files which shouldn't be committed. | 08:46 |
| tvCommitBot | HDMI2USB-vmodvga/master a9cdafc Tim 'mithro' Ansell: Removing README.md from unused directories. | 08:46 |
| *** tvCommitBot has left #timvideos | 08:46 | |
| rohitksingh | mithro: hi! | 08:47 |
| *** tvCommitBot has joined #timvideos | 08:48 | |
| tvCommitBot | [HDMI2USB-vmodvga] mithro pushed 1 new commit to master: http://git.io/vWH26 | 08:48 |
| tvCommitBot | HDMI2USB-vmodvga/master 472ffa1 Tim 'mithro' Ansell: Rename PCB to board. | 08:48 |
| *** tvCommitBot has left #timvideos | 08:48 | |
| rohitksingh | what am i missing? :) | 08:48 |
| seaLne | rohitksingh: preperation for tofe i think | 08:49 |
| rohitksingh | seaLne: oh! awesome! and hi! | 08:50 |
| seaLne | hiya | 08:51 |
| seaLne | rohitksingh: 07:38:37 @mithro | I'll basically merge the TOFE template and Rohit's repo | 08:51 |
| rohitksingh | good idea! i haven't looked at TOFE template in kicad...lemme check | 08:53 |
| seaLne | https://github.com/timvideos/HDMI2USB-TOFE-kicad-template | 08:54 |
| tpb | Title: timvideos/HDMI2USB-TOFE-kicad-template · GitHub (at github.com) | 08:54 |
| rohitksingh | thanks! | 08:54 |
| mithro | rohitksingh: cleaning up your repository a bit | 08:58 |
| mithro | seaLne / rohitksingh: https://github.com/timvideos/HDMI2USB-TOFE-VGA | 09:01 |
| tpb | Title: timvideos/HDMI2USB-TOFE-VGA · GitHub (at github.com) | 09:01 |
| mithro | seaLne: just need to clean up the board directory and move the things which should be libraries into the ../libraries directory | 09:04 |
| mithro | seaLne: let me rephrase that, you'll just need to up the board directory and move the things which should be libraries into the ../libraries directory | 09:05 |
| seaLne | k | 09:06 |
| seaLne | mithro: thanks | 09:07 |
| *** rohitksingh has quit IRC | 09:25 | |
| *** rohitksingh has joined #timvideos | 09:40 | |
| mithro | seaLne: you should have push access to that repo | 09:42 |
| seaLne | i was assuming i should do things in a fork and file pull requests though as a workflow? | 09:44 |
| seaLne | mithro: how are you supposed to have connectors on the edge and have the keepout area? i'd assume connectors are ok in there? | 09:46 |
| *** rohitksingh has quit IRC | 09:51 | |
| seaLne | that seems to be the way the pci(e) cards i have lieing around do it | 09:52 |
| mithro | seaLne: merging doesn't work so well with schematics / pcb design | 09:53 |
| mithro | seaLne: so found its better to just use a central repository | 09:53 |
| mithro | seaLne: follow what PCI cards do and you should be fine | 09:54 |
| *** rohitksingh has joined #timvideos | 10:05 | |
| seaLne | rohitksingh: did the vga pass through output work ok? | 10:16 |
| rohitksingh | seaLne: yeah, we have tested it...works well! | 10:17 |
| rohitksingh | atleast with the monitors we tested :) | 10:18 |
| seaLne | cool | 10:19 |
| seaLne | previously when i'd looked at similar circuitry it was recomended to use transistors and pull up resistors which takes up lots more room | 10:21 |
| seaLne | i guess the output isn't really essential for this and if it caused problems you'd just disconnect it | 10:21 |
| xfxf | mithro or any others, still around? | 10:25 |
| xfxf | I'll head back to the area with the av gear if somebody can help me debug why hdmi2usb isn't working | 10:26 |
| xfxf | 19:05:20 <@mithro> You can't use ixo-usb-jtag with PROG=fpgalink | 10:26 |
| xfxf | I'm not - uninstalled it | 10:26 |
| mithro | xfxf: okay | 10:30 |
| mithro | xfxf: what is the error? | 10:30 |
| mithro | I'm currently tidying up my place so when my mum arrives tomorrow I can pretend I'm an adult | 10:31 |
| xfxf | pasted above earlier - went out for dinner, just got back | 10:31 |
| xfxf | haha | 10:31 |
| xfxf | should I be using fpgalink or the new method by using your branch? | 10:32 |
| xfxf | the name of which escapes me | 10:32 |
| mithro | xfxf: use the fpgalink one | 10:41 |
| mithro | xfxf: did you power cycle the Atlys board after uninstalling ixo-usb-jtag | 10:41 |
| xfxf | ok, what am i doing wrong? | 10:41 |
| mithro | xfxf: other open is openocd | 10:41 |
| xfxf | not sure, but i'll try aain | 10:42 |
| xfxf | apparently not; it's working not. not sure why i didn't think of that | 10:43 |
| mithro | Please do | 10:43 |
| xfxf | er, it's working now rather, this keyboard sucks | 10:44 |
| xfxf | https://www.irccloud.com/pastebin/dPrzZYhw/ | 10:44 |
| tpb | Title: Pastebin | IRCCloud (at www.irccloud.com) | 10:44 |
| xfxf | load-lm32 same | 10:45 |
| xfxf | https://www.irccloud.com/pastebin/HAuxBSdv/ | 10:46 |
| tpb | Title: Pastebin | IRCCloud (at www.irccloud.com) | 10:46 |
| xfxf | https://www.irccloud.com/pastebin/o1qt2CpF/ | 10:46 |
| tpb | Title: Pastebin | IRCCloud (at www.irccloud.com) | 10:46 |
| xfxf | there is /dev/ttyACM0 though | 10:47 |
| mithro | xfxf: see the issue tumbleweed and CarlFK were discussing earlier | 10:50 |
| mithro | xfxf: update the exart package | 10:50 |
| mithro | xfxf: http://logs.timvideos.us/%23timvideos/%23timvideos.2015-10-28.log.html#t2015-10-28T01:09:12 | 10:52 |
| tpb | Title: Wednesday, 2015-10-28 (at logs.timvideos.us) | 10:52 |
| xfxf | right, ta | 10:53 |
| xfxf | https://www.irccloud.com/pastebin/meYP87xs/ | 11:15 |
| tpb | Title: Pastebin | IRCCloud (at www.irccloud.com) | 11:15 |
| xfxf | yay | 11:15 |
| xfxf | something funky happening though | 11:15 |
| xfxf | i'm having to unplug and plug in devices and reset the output resolution on the camera a few times before hdmi2usb sees it | 11:15 |
| xfxf | assuming something going wierd with negotiation | 11:16 |
| xfxf | woo | 11:19 |
| xfxf | ghetto mixing works | 11:19 |
| xfxf | hmm | 11:20 |
| xfxf | as soon as i enabled the encoder, the little LCD i'm using with output0 starts freaking out | 11:20 |
| xfxf | disable encoder, picture back to normal again | 11:20 |
| xfxf | reproducable | 11:21 |
| _florent_ | xfxf: the atlys does not have enough DDR bandwidth to enable all in/out ports together and the encoder | 11:22 |
| _florent_ | you have to disable at least one input or output port | 11:22 |
| xfxf | this is input0, input1, output0, encoder | 11:22 |
| xfxf | not using output1 | 11:22 |
| xfxf | hmm, is this something we can overcome, or a hardware limitation btw? | 11:22 |
| _florent_ | but is your output1 just unplugged or disabled? (in your pastebin it's enabled) | 11:23 |
| xfxf | disabled | 11:23 |
| xfxf | 'output0 off' | 11:24 |
| _florent_ | ok | 11:24 |
| _florent_ | and if you disconnect one input, same behaviour? | 11:24 |
| xfxf | oh hmm, unable to reproduce now | 11:25 |
| xfxf | perhaps i did have both outputs enabled | 11:25 |
| _florent_ | this is something that is not happening on the Opsis (more DDR bandwidth), but we need to try higher frequencies for the DDR controller to get that working on the atlys. | 11:25 |
| xfxf | right, makes sense | 11:25 |
| xfxf | ta for explanatin | 11:26 |
| xfxf | ion* | 11:26 |
| xfxf | neat, will probably write up a basic pyserial script tonight to automate swapping inputs | 11:26 |
| xfxf | this isn't critical for capture tomorrow, just taking the opportunity while here to test this | 11:26 |
| mithro | xfxf: if you do, link your stuff from https://github.com/timvideos/HDMI2USB-misoc-firmware/issues/41 | 11:30 |
| tpb | Title: Create a HDMI2USB GUI controller · Issue #41 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com) | 11:30 |
| xfxf | will do | 11:31 |
| xfxf | you're handy with gstreamer, know a working pipeline to save the hdmi2usb stream to disk? | 11:31 |
| xfxf | preferably encoding to an intermediate format given raw mjpeg frames will be huge | 11:31 |
| xfxf | figuring it out now, you may know quicker | 11:31 |
| *** rohitksingh has quit IRC | 11:33 | |
| mithro | xfxf: I'd try saving raw mjpeg to disk first and see how much space it is taking up | 11:35 |
| mithro | xfxf: then go to something like webm | 11:35 |
| xfxf | we did the math the other day and it'll be huge, but nod, doing that first | 11:35 |
| mithro | xfxf: #gstreamer is a good place to get help with that too btw | 11:35 |
| *** rohitksingh has joined #timvideos | 11:48 | |
| *** Bertl_zZ is now known as Bertl | 12:05 | |
| mithro | xfxf: can we close https://github.com/timvideos/HDMI2USB-misoc-firmware/issues/114 then? | 12:16 |
| tpb | Title: input0 appears to be broken · Issue #114 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com) | 12:16 |
| *** se6astian is now known as se6astian|away | 12:54 | |
| *** se6astian|away is now known as se6astian | 12:57 | |
| *** rohitksingh has quit IRC | 13:05 | |
| *** rohitksingh has joined #timvideos | 13:21 | |
| *** rohitksingh has quit IRC | 15:06 | |
| CarlFK | wakes up, checks osdc data.. still empty rooms. | 15:16 |
| *** rohitksingh has joined #timvideos | 15:17 | |
| GChriss | xfxf: I'm handy w/ gstreamer pipelines. https://wiki.xiph.org/GST_cookbook is a good place to start | 15:29 |
| tpb | Title: GST cookbook - XiphWiki (at wiki.xiph.org) | 15:29 |
| CarlFK | GChriss: do you have one of the hdmi2usb boards? (Atlys or Opsis) | 15:41 |
| GChriss | No, but usually videotestsrc can be used as a stand-in for testing purposes | 15:47 |
| CarlFK | GChriss: the data coming over usb needs to be.. um.. taken into account? | 15:51 |
| GChriss | the brunt of the work is figuring out what the pipeline you want should look like and stitching it together, editing as needed for CPU stability. assuming the usb data transfer is not CPU-intensive it should be OK | 15:54 |
| CarlFK | GChriss: this pipeline works for reading from hdmi2usb and converting to dv | 15:54 |
| CarlFK | http://paste.ubuntu.com/12990485/ | 15:54 |
| tpb | Title: Ubuntu Pastebin (at paste.ubuntu.com) | 15:54 |
| CarlFK | can you remove the dv and replace with filesink? | 15:55 |
| GChriss | a channel PSA for anyone interested in gstreamer pipelines: it's important to place a queue (with increased buffer size defaults -- see GST cookbook) between every element that does something with the datastreem. without queues elements will block on each other which is akin to running a car engine without oil | 15:58 |
| * GChriss looks at hdmi2usb pipleine | 15:58 | |
| CarlFK | you can chuck the audio - or if you can read in from a usb sound device (or alsa or whatever layer) that would be great | 16:01 |
| GChriss | going for lunch but see if the following does what you're expecting on your system: | 16:30 |
| GChriss | gst-launch-1.0 videotestsrc is-live=true ! video/x-raw,framerate=\(fraction\)30000/1001,width=1280,height=720 ! queue max-size-bytes=100000000 max-size-time=0 ! videorate ! video/x-raw,pixel-aspect-ratio=\(fraction\)32/27 ! queue max-size-bytes=100000000 max-size-time=0 ! videoscale ! video/x-raw,width=720,height=480 ! queue max-size-bytes=100000000 max-size-time=0 ! videoconvert ! queue max-size-bytes=100000000 max-size-time=0 ! avenc_dvvideo ! queu | 16:30 |
| GChriss | I'll need to build dvswitchsink and such | 16:30 |
| CarlFK | I don't think you understand | 16:32 |
| CarlFK | this is to run without dvswitch. without dv anything. so no width=720,height=480 | 16:33 |
| CarlFK | also, your paste got cut off at: "... ! avenc_dvvideo ! queu" | 16:34 |
| *** Bertl is now known as Bertl_oO | 16:58 | |
| GChriss | oops, ok. https://wiki.xiph.org/GST_cookbook#hdmi2usb_capture_.28timvideos.us.29 | 17:16 |
| GChriss | ? | 17:16 |
| tpb | Title: GST cookbook - XiphWiki (at wiki.xiph.org) | 17:16 |
| CarlFK | GChriss: thanks.. checking... | 17:19 |
| *** se6astian is now known as se6astian|away | 17:22 | |
| GChriss | I guess the 'DV NTSC' par is a DV NTSC thing, can add on-screen display if needed | 17:23 |
| GChriss | 30FPS needed or is a reduced framerate ok? | 17:24 |
| GChriss | 32/27* | 17:24 |
| CarlFK | good Q... | 17:25 |
| CarlFK | running on a Intel(R) Core(TM)2 Duo CPU T7500 @ 2.20GHz | 17:28 |
| CarlFK | top shows both cores 50% id | 17:28 |
| CarlFK | so current pipeline is ok | 17:28 |
| GChriss | anther question: do the Opsis and/or hdmi2usb boards drop/add samples as needed to stay synced to real-world time? or all frames are processed/sent through to v4l2? | 17:29 |
| CarlFK | I don't know | 17:29 |
| CarlFK | gst-launch 80% | 17:30 |
| GChriss | gnome-system-monitor is a bit better than top for visualizing what's going on: you're looking for 'smooth/flowing' cpu utilization lines | 17:31 |
| GChriss | as opposed to ones that stick to 100% for more than a second or two, sharp jumps, hills and valleys | 17:31 |
| CarlFK | gnome-system-monitor ... both cpu are around 50% =- at most 5% | 17:33 |
| CarlFK | no spikes | 17:33 |
| GChriss | that's a stable pipeline (but monitor over several hours and also switch to real video input) | 17:33 |
| GChriss | with complex motion | 17:34 |
| GChriss | not really an issue if it's just jpeg encoding but it's possible that video encoders can stall under certain conditions (slow pans seem to be the most problematic) | 17:35 |
| GChriss | on the "drop/add samples" question: if the answer is 'no/not currently' it might be possible to implement a simple rtsp server on the boards; gstreamer can pull in frames using rtspsrc | 17:36 |
| GChriss | it's important for a/v sync | 17:37 |
| CarlFK | rtsp server on the boards - can you bring that up on the list? | 18:05 |
| CarlFK | sounds like you know what you are talking about. I kinda know, but don't really have anything to offer the conversation | 18:06 |
| CarlFK | bb in 10 | 18:08 |
| *** CarlFK has quit IRC | 18:12 | |
| *** CarlFK has joined #timvideos | 18:42 | |
| *** ChanServ sets mode: +v CarlFK | 18:42 | |
| *** rohitksingh has quit IRC | 19:11 | |
| *** CarlFK has quit IRC | 19:30 | |
| *** CarlFK has joined #timvideos | 19:42 | |
| *** ChanServ sets mode: +v CarlFK | 19:42 | |
| xfxf | GChriss: ta, there's some extra magic needing to get it working with v4l2src though | 22:25 |
| xfxf | assuming some yuv stuff based off what v4l-conf is telling me | 22:26 |
| xfxf | ah, neat, got something working | 22:39 |
| GChriss | in theory recording from hdmi2usb can be done w/ embedded hardware: ~$40 Allwinner-based SoC board, jpeg encoding w/ Cedrus-based hw acceleration (working I believe, but no GStreamer integration as of yet), and SATA connection to a HDD | 22:50 |
| CarlFK | GChriss: all this recording stuff is temporary | 22:55 |
| CarlFK | the ultimate plan is to feed a live video mixer like http://dvswitch.alioth.debian.org/wiki/ | 22:56 |
| tpb | Title: DVswitch (at dvswitch.alioth.debian.org) | 22:56 |
| xfxf | GChriss: you still around? | 23:07 |
| xfxf | gst-launch-1.0 v4l2src device=/dev/video1 ! jpegparse ! queue max-size-bytes=100000000 max-size-time=0 ! matroskamux name=mux alsasrc device='hw:1,0' ! audio/x-raw,channels=2,rate=48000 ! audioconvert ! vorbisenc ! queue ! mux. mux. ! queue max-size-bytes=100000000 max-size-time=0 ! filesink location=hdmi2usb_recording_`date +%s`.mkv | 23:07 |
| xfxf | that works... sort of | 23:07 |
| xfxf | as in, i'm getting video, audio captured, but the video plays back super slow, as in, 10fps | 23:07 |
| xfxf | and yes i've intentionally gutted out the jpegenc stuff, i'm just trying to save the raw mjpeg from hdmi2usb for now | 23:08 |
| xfxf | i'll add in encoders etc once that's working | 23:08 |
| xfxf | i couldn't find a working video=x-raw/whatever line, even matching info from v4l-conf, hence jpegparse | 23:09 |
| xfxf | i'm assuming it's probably inferring the framerate or something wrong? | 23:09 |
| xfxf | (or maybe it's something else) | 23:09 |
| xfxf | oh and with a jpegdec after jpegparse, and with your original jpegenc, same issue | 23:10 |
| xfxf | CarlFK: i'm running hdmi2usb through a vga > hdmi scaler, output is 1280x800 (16:10 projectors here), 720p@50 forced out of the scaler. even with rescaling the capture of this talk (small terminal text) looks so much better than dv it's not funny | 23:21 |
| xfxf | as soon as i get a working gstreamer pipeline i'll upload some actual capture | 23:22 |
| CarlFK | xfxf: did you try https://wiki.xiph.org/GST_cookbook#hdmi2usb_capture_.28timvideos.us.29 | 23:30 |
| tpb | Title: GST cookbook - XiphWiki (at wiki.xiph.org) | 23:30 |
| CarlFK | it worked for me, but hdmi2usb static test pattern. so I can't tell if playback is slow | 23:31 |
| xfxf | it's not capturing from hdmi2usb | 23:31 |
| xfxf | note 'videotestsrc' | 23:31 |
| xfxf | you need extra bits in the pipeline to deal with the specific format data coming out of the hdmi2usb | 23:32 |
| xfxf | what i pasted is that modified for hdmi2usb | 23:32 |
| xfxf | which works. just again, video plays back at a suuuper slow speed and out of sync with the audio | 23:32 |
| CarlFK | dammit. I didn't play that close attention | 23:33 |
| xfxf | if somebody sees this - would appreciate help. at an event, everything's plugged into the hdmi2usb, great opportunity to actually test this at an event | 23:37 |
| xfxf | just need a working gstreamer pipeline to actually save the video + mux in alsa audio from an external sound card | 23:37 |
| mithro | CarlFK: that thing in the xiph.org cookbook doesn't use the mjpeg pathway? | 23:58 |
| mithro | xfxf: take the one out of the dvsource-v4l-other | 23:59 |
| CarlFK | mithro: um.. du know. | 23:59 |
| mithro | videotestsrc is-live=true ! video/x-raw <--- | 23:59 |
| CarlFK | it didn't use v4l. I haven't tired anything else. | 23:59 |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!