*** tpb has joined #timvideos | 00:00 | |
*** thaytan_ is now known as thaytan | 03:27 | |
*** ChanServ sets mode: +v thaytan | 03:27 | |
*** rohitksingh has joined #timvideos | 03:35 | |
mithro | xfxf: I'm back now - just going through the backlog of things | 04:18 |
---|---|---|
xfxf | ack | 04:28 |
mithro | xfxf: BTW Did I forget to mention that I would be away (and mostly uncontactable) Thursday->Today? | 04:32 |
mithro | cr1901_modern: ping? | 04:32 |
xfxf | mithro: neg, but not a big deal, we've still progressed on stuff | 04:32 |
mithro | xfxf: great | 04:33 |
xfxf | mithro: i assume you'll be busy until this weekend given you'll be off from work for ~2 weeks | 04:33 |
xfxf | oh, there's a public holiday next week too, so i can make it to at least one day | 04:33 |
xfxf | outside of after hours | 04:33 |
mithro | xfxf: yeah | 04:37 |
mithro | xfxf: one thing I need to organise is wireless internet | 04:37 |
mithro | xfxf: the place doesn't have any good options | 04:38 |
mithro | xfxf: any suggestions? | 04:39 |
xfxf | i'd probably check what vodafone coverage is like | 04:39 |
xfxf | if they have 4g, load up a sim with credit | 04:39 |
xfxf | if you get a telstra sim i have a dongle that works on linux | 04:40 |
xfxf | so iptables etc | 04:40 |
xfxf | i can lend a box to use as a router | 04:40 |
xfxf | fark, mobile data isn't cheap | 04:41 |
xfxf | http://www.vodafone.com.au/personal/mobile-broadband/recharge | 04:41 |
tpb | Title: Mobile Broadband Recharge | Vodafone Australia (at www.vodafone.com.au) | 04:41 |
xfxf | $200 for 22gb | 04:41 |
xfxf | assuming not shit coverage | 04:41 |
xfxf | https://www.telstra.com.au/broadband/mobile-broadband/prepaid | 04:43 |
tpb | Title: Telstra Broadband - Pre-Paid Mobile Broadband (at www.telstra.com.au) | 04:43 |
xfxf | $140 for 16gb with telstra | 04:43 |
xfxf | i'd prob go with telstra | 04:43 |
xfxf | i thought they'd be more expensive | 04:43 |
xfxf | i'd probably lock the router down like anything too to ensure people don't accidentally use up data | 04:43 |
mithro | xfxf: Yeah - I was sure it had gotten cheaper then that... | 04:53 |
xfxf | mithro: poked you in #lca-av, need to respond to an email | 05:14 |
cr1901_modern | mithro: pong | 05:18 |
cr1901_modern | HDMI cables are in | 05:18 |
cr1901_modern | However, trying to actually BUILD a SoC isn't going all that swell lol | 05:19 |
cr1901_modern | mithro: "BTW Did I forget to mention that I would be away (and mostly uncontactable) Thursday->Today?" Not that I am aware of :P | 05:43 |
mithro | cr1901_modern: were are we at? | 05:44 |
cr1901_modern | mithro: Got my cables. Trying to get a SoC working | 05:46 |
cr1901_modern | I got the exact cable you pointed me to | 05:47 |
cr1901_modern | mithro: You said shenki had minispartan6 firmware for HDMI2USB? | 06:33 |
shenki | cr1901_modern: yep | 07:07 |
shenki | cr1901_modern: https://github.com/shenki/HDMI2USB-misoc-firmware/commits/minispartan/testing | 07:08 |
tpb | Title: Commits · shenki/HDMI2USB-misoc-firmware · GitHub (at github.com) | 07:08 |
cr1901_modern | Oh, THAT's why I couldn't find it. Had the wrong branch checked out | 07:13 |
*** rohitksingh has quit IRC | 07:34 | |
xfxf | https://usercontent.irccloud-cdn.com/file/Yr0u7dyx/20160118_182153.jpg | 07:35 |
xfxf | https://usercontent.irccloud-cdn.com/file/EyKCDQpG/20160118_182410.jpg | 07:35 |
xfxf | https://usercontent.irccloud-cdn.com/file/uRWnfiwl/20160118_182645.jpg | 07:35 |
xfxf | mithro: ^^ | 07:35 |
xfxf | packaged well | 07:37 |
*** Bertl_oO is now known as Bertl | 07:43 | |
mithro | He had the MiniSoC working | 07:43 |
cr1901_modern | *** No rule to make target 'divsi3.o', needed by 'libcompiler-rt.a'. Stop. | 07:56 |
cr1901_modern | Of course, this is most likely because of a migen version conflict; I seem to be the only person actively targeting the new version | 07:57 |
cr1901_modern | s/migen/misoc | 07:57 |
cr1901_modern | Even worse, "setup.py develop" ensures that my copies of migen/misoc are seen before yours. Guess I have no choice but to create a conda environment | 07:58 |
cr1901_modern | oh wait, nevermind for now. Looks like submodules didn't initialize | 07:59 |
mithro | https://calendar.google.com/calendar/embed?src=mithis.com_dbrivdjifn9p2qsmtmeaaec928%40group.calendar.google.com&ctz=Australia/Sydney | 08:04 |
mithro | xfxf: Can you send a picture of the other side? | 08:04 |
mithro | xfxf: So, I'm currently planning on crashing at your place on Friday night | 08:06 |
*** se6astian|away is now known as se6astian | 08:09 | |
cr1901_modern | shenki: is the 32MHz clock default for you? | 08:11 |
mithro | bblr | 08:11 |
cr1901_modern | Oh yea, Xilinx ISE is having a field day trying to optimize this lol | 08:12 |
cr1901_modern | shenki: Ignore prev q, got my answer | 08:14 |
shenki | cr1901_modern: :) | 08:14 |
cr1901_modern | You appear to use a PLL to halve the freq and then multiply by 5 | 08:16 |
cr1901_modern | Aaaand, done: http://imgur.com/AB6BRkF | 08:20 |
tpb | Title: Imgur: The most awesome images on the Internet (at imgur.com) | 08:20 |
cr1901_modern | I'd love to know why newlines aren't working properly, but I can deal with that later | 08:21 |
_florent_ | cr1901_modern: you need to configure your terminal with: "implicit CR in every LF" | 08:43 |
xfxf | https://usercontent.irccloud-cdn.com/file/R5Sf9gky/irccloudcapture722209480.jpg | 08:58 |
xfxf | mithro: the sheet says 12v2a, I bought a mix of 1a and 2a adapters as understood a minimum of 1a. did the specs change / is this an issue? | 09:00 |
xfxf | mithro: also i've plugged in a board, gone to the URL on the instructions, can't find any instructions for getting firmware onto this. just shows a test pattern with hdmi2usb written on it when plugged in | 09:26 |
xfxf | lsusb shows a device with no information | 09:26 |
xfxf | Bus 001 Device 012: ID 2a19:5442 | 09:26 |
xfxf | mithro: i'm going to wait a hour then just starting figuring it out from what i see up on github. i'm intentionally pausing here as i assume you want to simulate somebody else getting this + having very clear repeatable instructions | 09:59 |
*** Bertl is now known as Bertl_zZ | 10:03 | |
xfxf | so | 11:27 |
xfxf | what should they come up as? | 11:27 |
xfxf | because i get a test screen and lsusb shows nothing | 11:27 |
xfxf | as in, blank | 11:27 |
xfxf | see above | 11:27 |
xfxf | no device name | 11:27 |
mithro | Dmesg? | 11:27 |
xfxf | oh | 11:28 |
xfxf | hang on | 11:28 |
xfxf | maybe you're right | 11:28 |
mithro | They won't have a device name as we arn't in USB.ids | 11:28 |
xfxf | oh dear | 11:28 |
xfxf | i've been waiting the last 2 hours haha | 11:28 |
xfxf | ok | 11:30 |
xfxf | i can see the test pattern in vocto, so good | 11:30 |
xfxf | it turns out i'm an idiot, big surprise | 11:30 |
xfxf | however | 11:30 |
xfxf | should i have a serial port? | 11:30 |
xfxf | ttyACM0? | 11:31 |
xfxf | my script doesn't appear to work | 11:32 |
xfxf | will try minicom | 11:32 |
xfxf | nope - need help | 11:33 |
xfxf | also it sounds like i should probably finish my switcher GUI so you have something for people to use | 11:33 |
xfxf | considering it's set to a test pattern out of the box | 11:33 |
xfxf | segfault in the uvcvideo driver when i unplugged it just then | 11:38 |
xfxf | i plug it back into the laptop where i've been running the test board + have those drivers installed | 11:38 |
xfxf | the device is not detected + dmesg is full of 'device descriptor read error' messages | 11:39 |
mithro | You mean kernel oops? | 11:39 |
xfxf | sorry, yes | 11:39 |
xfxf | assume ttyACM0 is the device i should be talking to, what are the serial options? | 11:40 |
xfxf | and do i need a special driver? | 11:40 |
mithro | Yes, and no driver needed | 11:40 |
mithro | Just connect and hit enter | 11:40 |
xfxf | connect how? | 11:40 |
mithro | I use flterm | 11:41 |
seaLne | mithro: what do you think the chances of https://github.com/timvideos/HDMI2USB-misoc-firmware/issues/132 happening is? considering getting one | 11:41 |
tpb | Title: Support video capture from Canon HV30 cameras · Issue #132 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com) | 11:41 |
xfxf | sure, but the machine where i have all of that stuff built is having those issues described above | 11:41 |
mithro | seaLne: very high | 11:41 |
xfxf | and i dont' want to install all of the firmware build stuff on this pc - it's what i'll be making an image from | 11:42 |
mithro | Minicom should work | 11:42 |
seaLne | mithro: thanks | 11:42 |
mithro | So should pyserial | 11:42 |
xfxf | neither are | 11:43 |
xfxf | if you think they'll work i'll assume there's an issue, will debug | 11:43 |
xfxf | er | 11:43 |
xfxf | ok | 11:43 |
xfxf | now i'm getting those usb errors on both machines | 11:43 |
xfxf | wtf | 11:43 |
xfxf | this is after a hdmi2usb power cycle too | 11:43 |
xfxf | https://www.irccloud.com/pastebin/luPEPylG/ | 11:44 |
tpb | Title: Pastebin | IRCCloud (at www.irccloud.com) | 11:44 |
xfxf | er | 11:44 |
xfxf | i'm going to reboot | 11:44 |
xfxf | wtf | 11:44 |
xfxf | the board worked fine 5 mins ago | 11:44 |
xfxf | oh, neat, faulty new USB cable | 11:46 |
mithro | xfxf: the device will start out at 2a19:5440 and then switch to 2a19:5442 after a second or two | 11:55 |
xfxf | ok | 11:58 |
xfxf | input1 is working | 11:58 |
xfxf | input0 is not | 11:58 |
xfxf | input0: 0x0 | 11:58 |
xfxf | same cable works fine on input1 | 11:58 |
xfxf | and yes, doing video_matrix etc | 11:58 |
mithro | xfxf: what is the computer showing? | 11:59 |
xfxf | windows seems to see the display but doesn't give any resolutions | 12:00 |
mithro | xfxf: It is really hard to debug Windows | 12:02 |
mithro | xfxf: they should have all been tested btw | 12:02 |
xfxf | okay, i can plug it into a linux box, what can i do to get you info? | 12:04 |
mithro | everything from dmesg and xrandr | 12:04 |
mithro | _florent_: https://github.com/Elphel/gtxe2_gpl | 12:06 |
tpb | Title: Elphel/gtxe2_gpl (at github.com) | 12:06 |
xfxf | [857932.546185] EDID block is all zeroes | 12:07 |
xfxf | [857932.546205] i915 0000:00:02.0: HDMI-A-1: EDID block 0 invalid. | 12:07 |
_florent_ | mithro: hi, I saw that, that's cool | 12:08 |
mithro | xfxf: can you compare to another cable, another input and another board | 12:08 |
xfxf | i just reset the board again and now both inputs are working | 12:08 |
mithro | xfxf: can you also provide a picture of the other side of the opsis board? | 12:08 |
xfxf | this is the same cable i've been testing with the past week | 12:08 |
mithro | xfxf: (not the sheet) | 12:09 |
mithro | xfxf: yes, but please try as anyway | 12:09 |
mithro | xfxf: it helps see if it is a marginal signal problem or something else | 12:09 |
mithro | xfxf: the fact that both works after reset seems to indicate a firmware problem | 12:09 |
mithro | xfxf: but its useful to find out more | 12:09 |
xfxf | to be clear, i had reset the board twice already | 12:12 |
xfxf | now it's working fine, so is another board | 12:12 |
mithro | xfxf: yes, but I'm unclear what you have been doing | 12:13 |
xfxf | what are you unclear on? | 12:14 |
mithro | xfxf: What exactly you did and how you got into a state were you needed a reset | 12:14 |
mithro | xfxf: if you can reproduce the problem, please do log an issue | 12:15 |
xfxf | powered board up, send it the same commands my script does via flterm (my script isn't working for some reason, yet to debug, and minicom is unusable with default settings - all text is on the right of the screen) | 12:15 |
xfxf | https://usercontent.irccloud-cdn.com/file/y7T9eG0o/20160118_231007_HDR.jpg | 12:16 |
mithro | xfxf: great | 12:16 |
mithro | xfxf: those QR codes are unique for each board | 12:16 |
mithro | xfxf: so we should create a tracking spreadsheet which stores information about each board | 12:17 |
mithro | xfxf: did the package contain extra QR code labels? | 12:17 |
xfxf | not that i've seen | 12:17 |
mithro | xfxf: hrm... they should have :( | 12:18 |
xfxf | there's a QR code label stuck on the anti-static bag each board comes in | 12:19 |
xfxf | ah, there's an extra one in each box too | 12:19 |
mithro | xfxf: great! We'll put that on the external of the case | 12:20 |
mithro | xfxf: it gives us an easily scanable tracking number | 12:20 |
xfxf | these power adapters i've got are a bit annoying | 12:20 |
xfxf | the cable is like 0.7m long | 12:20 |
mithro | xfxf: pics? | 12:20 |
xfxf | i'd order more but we won't get them in time | 12:20 |
mithro | xfxf: regarding the 1A verse 2A, my calculations say we only need 1A - the Numato guys wanted a bigger margin, so we went with 2A in the docs | 12:24 |
mithro | shenki: need to change the opsis docs to work like the hdmi2usb.tv website and use travis rather then github | 12:30 |
mithro | shenki: job for tomorrow | 12:30 |
shenki | mithro: ok | 12:31 |
xfxf | ok, i'm now having trouble with the encoder | 12:42 |
xfxf | it's at 0fps no matter what i try | 12:43 |
xfxf | encoder: 1280x720 @ 0fps (0Mbps) from input1 (q: 85) | 12:44 |
xfxf | trying to use the v4l device fails | 12:44 |
xfxf | [860330.960925] uvcvideo: Failed to set UVC probe control : -110 (exp. 26). | 12:44 |
xfxf | dmesg is full of those | 12:44 |
xfxf | i've had zero problems with my atlys the past week, why is this so problematic | 12:46 |
xfxf | is this an older firmwre? | 12:46 |
xfxf | unplugged, replugged back in, encoder briefly showed some data then has gone back to 0 agin | 12:47 |
xfxf | sorry for missing letters, keyboard on this laptpo is horible | 12:47 |
xfxf | ERROR: from element /GstPipeline:pipeline0/GstV4l2Src:v4l2src0: Device '/dev/video1' cannot capture at 1280x720 | 12:49 |
xfxf | Additional debug info: | 12:49 |
xfxf | gstv4l2object.c(3235): gst_v4l2_object_set_format_full (): /GstPipeline:pipeline0/GstV4l2Src:v4l2src0: | 12:49 |
xfxf | Call to S_FMT failed for MJPG @ 1280x720: Input/output error | 12:49 |
mithro | xfxf: this is a newer firmware | 12:51 |
xfxf | okay, because i'm currently unable to use this | 12:51 |
xfxf | just trying to get video into vocto as i have been doing the past week using the atlys | 12:52 |
xfxf | going to try a diff board | 12:52 |
mithro | xfxf: I need a lot more information about the above, I need dmesg and full gstreamer output at high debug levels | 12:52 |
mithro | xfxf: is it working with mplayer? | 12:52 |
mithro | xfxf: we have been testing with mplayer and skype | 12:52 |
mithro | xfxf: not gstreamer, so maybe we have introduce some bugs somewhere | 12:54 |
mithro | xfxf: the encoder FPS tells you how many frames the receiver is pulling | 12:55 |
mithro | it will only go up when you have something receiving on the host | 12:55 |
mithro | xfxf: and it sounds like you have the serial port working now? | 12:55 |
xfxf | yes i know that, i'm doing this when i'm trying to use the v4l device | 12:56 |
xfxf | mplayer works | 12:56 |
xfxf | gstreamer doesn't | 12:56 |
xfxf | mplayer as in your view-hdmi2usb.sh script | 12:56 |
xfxf | and no, same serial port problems | 12:56 |
xfxf | my script doesn't work, haven't debugged it yet | 12:56 |
xfxf | minicom looks terrible | 12:56 |
xfxf | i'm just using flterm on my dev box | 12:56 |
mithro | xfxf: "minicom looks terrible"? | 12:57 |
xfxf | oh hmm | 12:57 |
mithro | xfxf: doesn't really help me :P | 12:58 |
xfxf | i run the view-hdmi2usb.sh script | 12:58 |
xfxf | now gstreamer works | 12:58 |
mithro | xfxf: so, can you see if you can get a reliable repro case for things breaking? | 12:59 |
mithro | xfxf: IE - something I can follow and reproduce the same results | 13:00 |
mithro | xfxf: does it look like minicom isn't doing newlines properly? | 13:00 |
mithro | xfxf: if so, you need to change one of the minicom settings - Can't remember which one..... | 13:01 |
mithro | _florent_: we should make the firmware always return '\r\n' instead of just '\n'.... | 13:03 |
mithro | xfxf: Did you see the spreadsheet I linked in the other channel? | 13:03 |
_florent_ | mithro: ok we'll do that | 13:03 |
mithro | _florent_: is it better to just do a search/replace on the firmware - or a hack in the printf code? | 13:04 |
xfxf | mithro: it's more than newlines, all the text is squished into the last character in the terminal | 13:04 |
mithro | xfxf: that doesn't really mean much to me? | 13:04 |
_florent_ | mithro: just saw you mail about the accomodation, cool :) | 13:04 |
mithro | _florent_: Don't know the weather yet, but it is frequently hot | 13:04 |
xfxf | also sure i can try and get a reproducible case but not tonight, it's too late, i need to go to bed | 13:04 |
_florent_ | mithro: I was wondering about renting a car at the airport | 13:04 |
xfxf | also the encoder is only running at 15fps | 13:04 |
xfxf | will need to continue to debug tomorrow sorry | 13:05 |
xfxf | night | 13:05 |
mithro | xfxf: no worries | 13:05 |
_florent_ | so if you need a second car, I can rent one | 13:05 |
mithro | xfxf: can you make sure that you keep track of what you try on which boards | 13:05 |
mithro | xfxf: incase there is some manufacturing difference or fault (there shouldn't be, but it will save time) | 13:05 |
xfxf | mithro: nod, having a faulty usb cable doesn't help, makes me question the quality of all of them i bought | 13:09 |
xfxf | but this is definitely not that | 13:09 |
xfxf | reproduced issue on two boards so assume firmware thing triggered by certain steps | 13:10 |
mithro | xfxf: we should build kits | 13:10 |
xfxf | that said | 13:10 |
xfxf | i really want to capture overnight | 13:10 |
xfxf | i can get it into voctomix doing some things i didn't have to do on the atlys but why is the encoder claiming 15fps and not 30? | 13:11 |
xfxf | i believe it, the video looks like 15fps | 13:11 |
xfxf | i'm doing the exact same things i've been doing on the atlys to get 30fps video | 13:12 |
xfxf | board is set to 1280x720@60hz mode | 13:12 |
xfxf | the 'things i didn't have to do' refers to having to try your mplayer line before attempting gstreamer + using the hdmi2usb console using flterm instead of my script | 13:13 |
xfxf | same hdmi2usb console commands, same vocto scripts | 13:13 |
mithro | xfxf: I don't know why it would be running at 15fps not 30 - but if you can give me test instructions I can look into it | 13:14 |
mithro | xfxf: the cases are in Australia and clearing customs | 13:14 |
mithro | xfxf: they ended up in 3 seperate packages | 13:15 |
mithro | xfxf: I have stand offs heading to you, but they won't arrive till next week apparently | 13:16 |
xfxf | mithro: 1) plug computer/monitor into input0/output0. 2) use flterm to do the 'INIT' and 'CAMERA' lines (but video_mode 9, not 10) in https://github.com/xfxf/video-scripts/blob/master/ryan/hdmi2usb-control.py | 13:16 |
tpb | Title: video-scripts/hdmi2usb-control.py at master · xfxf/video-scripts · GitHub (at github.com) | 13:16 |
mithro | xfxf: the ones I ordered for me were suppose to arrive today but didn't | 13:16 |
xfxf | 3) try to run a variant of (that doesn't require voctomix) https://github.com/xfxf/video-scripts/blob/master/ryan/lca/hdmi2usb-v4l.sh | 13:16 |
tpb | Title: video-scripts/hdmi2usb-v4l.sh at master · xfxf/video-scripts · GitHub (at github.com) | 13:16 |
xfxf | that's cool, i'm not sure how much time i'm going to get to assemble things before fri - in fact i'm pretty sure i won't | 13:17 |
xfxf | i just want to test a pair of the devices now | 13:17 |
xfxf | to ensure no issues | 13:17 |
mithro | xfxf: yeah, I was hoping you'd get them before the weekend :( | 13:17 |
xfxf | i.e. leave two devices with an external sound card capturing while i'm at work | 13:18 |
xfxf | also re the last script, i saw your hdmi2usb detection stuff in your scripts, will use that | 13:18 |
xfxf | just trying to get 30fps video capture now to leave overnight | 13:19 |
xfxf | but | 13:19 |
xfxf | encoder: 1280x720 @ 15fps (131Mbps) from input0 (q: 85) | 13:19 |
mithro | xfxf: we can also setup udev scripts to always make it /dev/hdmi2usb/video and /dev/hdmi2usb/control or something too? | 13:21 |
mithro | xfxf: is that lying? | 13:22 |
xfxf | the video actually looks like it is 15fps | 13:22 |
xfxf | this is why i investigated to see what the encoder said | 13:22 |
xfxf | i'm testing with movies/videos | 13:22 |
xfxf | as it's the best way to dertermine sync issues (even 1 frame out with lipsync is noticable as hell) + ensure it's running at a smooth 30fps | 13:22 |
xfxf | am using the *exact* same gst-launch line that gave me 30fps on the atlys | 13:23 |
mithro | xfxf: it's better to use a test source which has each frame a different color and then looking for the colors on output | 13:23 |
mithro | xfxf: I'll investigate further tomorrow | 13:24 |
xfxf | cool, ta - going to head to bed for real now | 13:25 |
xfxf | may as well let it capture at 15fps overnight | 13:25 |
*** Bertl_zZ is now known as Bertl | 13:42 | |
mithro | https://twitter.com/TimVideosUs/status/689080039237922820 | 13:45 |
*** rohitksingh has joined #timvideos | 13:52 | |
*** se6astian is now known as se6astian|away | 16:19 | |
*** sb0 has joined #timvideos | 16:26 | |
*** se6astian|away is now known as se6astian | 17:14 | |
*** sb0 has quit IRC | 17:18 | |
*** rohitksingh has quit IRC | 17:21 | |
cr1901_modern | So like where is everyone going? | 17:36 |
cr1901_modern | I see a lot of planning/travel arrangements | 17:36 |
seaLne | i think its lca stuff they are talking about | 17:53 |
seaLne | linux.conf.au | 17:53 |
*** CarlFK has quit IRC | 18:21 | |
*** CarlFK has joined #timvideos | 19:27 | |
*** ChanServ sets mode: +v CarlFK | 19:27 | |
*** sb0 has joined #timvideos | 19:41 | |
*** sb0 has quit IRC | 20:09 | |
xfxf | doh, power went out overnight | 21:46 |
xfxf | i thought it was the hdmi2usb but appears the whole house, phew :) | 21:46 |
*** se6astian is now known as se6astian|away | 22:53 | |
*** Bertl is now known as Bertl_zZ | 23:16 | |
mithro | seaLne / cr1901_modern: yes Linux.conf.au | 23:55 |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!