| *** tpb has joined #timvideos | 00:00 | |
| *** loadstar81 has quit IRC | 00:10 | |
| *** loadstar81 has joined #timvideos | 00:12 | |
| *** skay has quit IRC | 00:37 | |
| *** skay has joined #timvideos | 00:38 | |
| *** skay has joined #timvideos | 00:38 | |
| *** ChanServ sets mode: +v skay | 00:38 | |
| *** ChanServ sets mode: +o CarlFK | 00:42 | |
| *** ChanServ sets mode: +o shenki | 00:42 | |
| *** ChanServ sets mode: +v ldlework | 00:42 | |
| *** mithro sets mode: -o techman83 | 00:44 | |
| *** mithro sets mode: -v techman83 | 00:45 | |
| *** mithro sets mode: -v thaytan | 00:45 | |
| *** FeltonChris has joined #timvideos | 00:54 | |
| *** ChanServ sets mode: +v FeltonChris | 01:28 | |
| mithro | Hey FeltonChris! | 01:28 |
|---|---|---|
| FeltonChris | mithro: hey | 01:31 |
| mithro | How are things? Did some HDMI2USB work over the weekend | 01:33 |
| mithro | FeltonChris: did your Atlys turn up? | 01:33 |
| FeltonChris | not yet, FedEx indicate it was supposed to arrive on sat. but it did not, check the tracking again today and no updates? | 01:35 |
| CarlFK | the fpga class I told ended with "lets bang out a VGA signal .. pretty sure it was outputs to resistors to a 15pin vga connector | 01:38 |
| mithro | CarlFK: yeah, you can generate a VGA signal just using the pins on a digital chip | 01:39 |
| mithro | Capturing is much harder :) | 01:39 |
| CarlFK | do we have a "send vga" project ? | 01:39 |
| *** puck has quit IRC | 01:43 | |
| *** puck has joined #timvideos | 01:44 | |
| mithro | no | 02:16 |
| mithro | it's call a video card :) | 02:16 |
| mithro | shenki: it turns out that my motherboard has an embedded arm chip running Linux on it :P | 02:44 |
| Joelw | mithro: Did you see the project where someone got Linux running on one of the ARM chips on their hard drive? | 02:44 |
| mithro | Joelw: yeah | 02:44 |
| mithro | Wow, it has 64 megabytes of memory | 02:46 |
| shenki | mithro: the BMC? | 03:36 |
| shenki | all servers have them these days | 03:36 |
| shenki | not always running linux tho | 03:36 |
| shenki | mithro: did you see this? https://plus.google.com/111282580643669107165/posts/Uwh9W3XiZTQ | 03:37 |
| mithro | shenki: nope! | 03:38 |
| *** mparuszewski has joined #timvideos | 03:43 | |
| *** mparuszewski has quit IRC | 03:50 | |
| *** Niharika has joined #timvideos | 04:28 | |
| *** sarwarc has quit IRC | 05:16 | |
| *** slomo has joined #timvideos | 06:23 | |
| *** slomo has joined #timvideos | 06:23 | |
| shenki | mithro: ping | 07:19 |
| mithro | shenki: just going to bed, is it quick? | 07:19 |
| shenki | mithro: it can be. i want to set up a vpn on a mithis server | 07:20 |
| shenki | mithro: is that ok? if so, which one? | 07:20 |
| mithro | for? | 07:20 |
| shenki | connecting to while travelling | 07:20 |
| mithro | Not really sure if any are that suitable | 07:27 |
| mithro | gerty would have been before I turned it off | 07:28 |
| shenki | okay. i'll find somewhere else then | 07:32 |
| *** rohitksingh has joined #timvideos | 07:34 | |
| *** sarwarc has joined #timvideos | 08:07 | |
| *** Niharika has quit IRC | 08:09 | |
| *** Niharika has joined #timvideos | 08:11 | |
| *** Niharika has quit IRC | 08:20 | |
| *** Niharika has joined #timvideos | 08:24 | |
| *** Niharika has left #timvideos | 09:08 | |
| *** Niharika has quit IRC | 09:09 | |
| *** rohitksingh has quit IRC | 09:09 | |
| *** FeltonChris has left #timvideos | 11:18 | |
| *** Niharika has joined #timvideos | 12:41 | |
| canci | I experience an interesting graphical artifact in images I capture with hdmi2usb | 13:42 |
| canci | it is easily reproducible with this pattern http://nowhere.ws/dump/testpattern.png | 13:42 |
| canci | for each color of r,g,b there is a sample in the image consisting of four quadrants. | 13:43 |
| canci | each of those quadrants representing one possible possition of a colored pixel in a 2x2 part of otherwise black pixels | 13:44 |
| canci | the artefact is that for each of the colors, one of the quadrants is distinctly more visible than the other three | 13:44 |
| canci | and which quadrant that is changes when the test pattern is moved around the screen | 13:45 |
| canci | I'll make a sample capture this evening, maybe somebody has an idea what might be the problem | 13:46 |
| *** shenki has quit IRC | 13:52 | |
| *** shenki has joined #timvideos | 13:56 | |
| *** ChanServ sets mode: +v shenki | 13:56 | |
| *** mparuszewski has joined #timvideos | 14:03 | |
| canci | (a result in real world data is visible e.g. here http://nowhere.ws/dump/stream.jpg - notice the strange coloring in the braces of the green text and the weird smushed font on the red text) | 14:06 |
| CarlFK | canci: (forgive my ignorance here, I am not very into the details) isn't this expected given the high jpg compression? | 14:08 |
| canci | CarlFK: I don't think so, as it's 100% | 14:11 |
| canci | even when I compress it to the same level with another encoder the outcome is not the same | 14:11 |
| CarlFK | k - never mind ;) | 14:12 |
| canci | It /might/ be some issue with chroma subsampling, maybe the implementation tries to be more efficient there, skipping some optimizations a full blown encoder does | 14:12 |
| *** mparuszewski has quit IRC | 14:32 | |
| *** sarwarc has quit IRC | 15:02 | |
| *** mparuszewski has joined #timvideos | 15:33 | |
| *** slomo has quit IRC | 17:16 | |
| mithro | canci: there are two modes, mjpeg and raw | 17:36 |
| mithro | canci: I assume you are using mjpeg there | 17:36 |
| mithro | canci: can you take a screenshot and look at what happens if you do jpeg compression of it? | 17:37 |
| *** mparuszewski has quit IRC | 17:38 | |
| mithro | canci: I guess it means you have a working board now? | 17:39 |
| *** Niharika has left #timvideos | 17:49 | |
| *** Niharika has quit IRC | 17:49 | |
| canci | mithro: yea, it works when I flash it with the usb stick | 18:38 |
| canci | mithro: https://github.com/cfra/HDMI2USB/wiki/Digilent-Atlys-Board%3A-Loading-Firmware#usb-flash-drive-method | 18:39 |
| tpb | Title: Digilent Atlys Board: Loading Firmware · cfra/HDMI2USB Wiki · GitHub (at github.com) | 18:39 |
| mithro | Can you try the libFPGALink method? | 18:39 |
| canci | mithro: I will probably try it eventually, but right now my focus is more about trying a few scenarios with hdmi2usb and figuring out whether I want to use gst-switch or use something more simplistic | 18:41 |
| *** slomo has joined #timvideos | 18:41 | |
| *** slomo has joined #timvideos | 18:41 | |
| mithro | The HDMI2USB is the priority project at the moment | 18:42 |
| canci | mithro: btw, I think the "issue" is that the jpeg encoder does 4:2:2 subsampling | 18:42 |
| canci | and does the subsampling by just dropping every second pixel | 18:42 |
| mithro | gst-switch is on the back burner | 18:42 |
| mithro | canci: that is quite probable | 18:43 |
| canci | Unless that has been tried and thrown out before, I'd have a look whether I can figure out how to average the pixels and check whether that improves the captured data | 18:44 |
| CarlFK | canci: I would love to help out with gst-switch, but currently the install is more than I want to figure out. If you will figure out how to install it, I'll try working with it | 18:45 |
| CarlFK | canci: the current install is very travis-ci laden, which may sound like a good thing, and I expect anyone will do what I did and will later agree, or later tell me I am nuts. | 18:46 |
| canci | CarlFK: I (think I) have installed it yesterday following the .travis-setup.sh | 18:46 |
| canci | I'm not exactly sure what the target platform is, I was doing it on debian testing and the tools are too new for the version of gstreamer referenced in the install script | 18:47 |
| canci | (although that fixed quite easily by cherry-picking a later commit that adds support for bison >= 3.0) | 18:48 |
| CarlFK | that sounds about as far as I got. | 18:48 |
| CarlFK | I have to run and get food - I'll be back in an hour or so. | 18:48 |
| canci | in the end I was able to build everything, however I was not really sure how to proceed from there, as in how to add inputs or process output | 18:50 |
| CarlFK | I forget if I was able to build. but I want to hear someone actually got it working | 18:52 |
| CarlFK | I was holding out for someone producing a real talk, but that may never happen if no one is first ;) | 18:52 |
| CarlFK | bbl | 18:52 |
| *** CarlFK has quit IRC | 18:52 | |
| mithro | canci: take a look at the python API | 18:53 |
| mithro | canci: we need to create a good test suite | 18:53 |
| *** mparuszewski has joined #timvideos | 19:46 | |
| *** mpar has joined #timvideos | 20:22 | |
| *** mparuszewski has quit IRC | 20:25 | |
| *** slomo has quit IRC | 20:27 | |
| *** CarlFK has joined #timvideos | 21:42 | |
| *** ChanServ sets mode: +v CarlFK | 21:42 | |
| canci | so, this is the testimage actually used | 22:05 |
| canci | http://nowhere.ws/dump/flicker.png | 22:05 |
| canci | and this is a recording of a screen displaying that image http://nowhere.ws/dump/flicker.mjpeg.xz | 22:05 |
| canci | the screen itself does not show this artefact, only the mjpeg recording | 22:06 |
| canci | other than that I also encountered these interesting effects :) http://nowhere.ws/dump/foo.png http://nowhere.ws/dump/foobar.png | 22:20 |
| *** CarlFK has quit IRC | 23:04 | |
| *** CarlFK has joined #timvideos | 23:19 | |
| *** ChanServ sets mode: +v CarlFK | 23:19 | |
Generated by irclog2html.py 2.12.1 by Marius Gedminas - find it at mg.pov.lt!
