Wednesday, 2016-07-06

*** tpb has joined #timvideos00:00
nuecesthere is something more that I could test here?00:00
mithroHey nueces00:03
nueceshi mithro00:03
mithroI'm currently on my phone, so can't really help debug at the moment00:03
mithroWill be around in an hour00:04
nuecesno problem, I could setup the environment other day00:04
mithronueces: were you following the improved instructions?00:05
nuecesI'm going to try to use a different camera00:05
nuecesmithro, both , the new and the old :)00:06
nuecesswitch the mode and later the order to load the gateware and firmware from the previous00:06
mithroBtw, do you have redmere cables? The fact that you are getting a resolution in the prompt is a good sign that things are working00:07
nuecesmithro, I don't have redmere cables here00:08
mithroCamera's are a lot more fussy / random then computers00:08
nuecesbut the other day I use the debug and see the WRE was 000:08
mithroYeah, zero WER is a really good sign00:09
mithroI'm hoping that after we finish the liteXXX upgrade stuff, we can work on improving the HDMI input debugging situation00:10
mithrocr1901_modern is already working on one part of that00:10
mithro(Detecting and determining the pixel clock frequency.)00:11
*** nueces has quit IRC00:14
mithrocr1901_modern: did I ever link you to my i2c modules I was working on?00:14
mithroAt some point I want to get back to that00:15
mithroEventually I hope to replace the current EDID implementation with that code00:16
mithroAs it has things like test benches00:17
mithroAnd it should be easier to both harden, optimise and have a more flexible boundary between the gateware and the soft code00:19
mithroWe really need to come up with a good way of figuring out the FPGA resource usage of a block so we can better understand design trade offs00:21
cr1901_modernmithro: No you didn't. Perhaps that be part of the trouble with the current code?00:23
cr1901_modernmithro: Forgive me that I'm hesitating/freaking out b/c NVIDIA driver is garbage when I attach my board00:25
mithroWell my phone is already going flat :-(00:27
*** nueces has joined #timvideos00:33
*** G33KatWork has quit IRC00:45
*** G33KatWork has joined #timvideos00:54
tpbTitle: Comparing fallen:shift_register_complex...mithro:shift_register_complex · fallen/i2cslave · GitHub (at
cr1901_modernWhy does I2C in and of itself require a BIOS?01:04
cr1901_modernmithro: ^^01:04
mithrocr1901_modern: what do you mean?01:05
cr1901_modernmithro: wait nevermind... ignore my question01:08
CarlFKmithro: did that code I wrote to do automated tests ever get checked into a repo?01:08
mithroCarlFK: you put it up on a gist01:09
CarlFKcan we put it in a repo so it can be found01:09
tpbTitle: · GitHub (at
mithroFound it via searching in the logs01:10
CarlFKgreat.. except sometimes you sleep ;)01:10
CarlFKmore interesting question: why does my lcd sometimes say '720p' and sometimes '1280x720 60Hz' ?01:12
mithroCarlFK: I can't think of a good repo to put it in01:12
mithroCarlFK: NFI?01:12
cr1901_modernmithro: You've answered this before, but you designed the Opsis hardware, correct? Does that including doing the unenviable PCB layout?01:16
mithrocr1901_modern: No, the Numato guys did the PCB layout under my direction01:17
cr1901_modernThat's cool... was just wondering. I wonder how much it costs* to get a custom board designed...01:20
cr1901_modern*More money than I have01:20
mithrocr1901_modern: They didn't charge anything, they make the money from selling the board01:26
CarlFKmithro: in modeswitch.. what installs unbind-helper so that this finds it:01:48
CarlFK hdmi2usb-mode-switch --load-fx2-firmware hdmi2usb.hex01:48
CarlFKWARNING:root:unbind-helper not found, will have to run as root!01:48
mithromake unbind-helper01:49
CarlFK sudo make unbind-helper01:49
CarlFKmake[1]: Entering directory '/home/juser/HDMI2USB-mode-switch'01:49
CarlFKmake[1]: Leaving directory '/home/juser/HDMI2USB-mode-switch'01:49
CarlFKdid that, still get that error01:49
mithroCarlFK: When did you last update?01:50
CarlFKgit pull01:50
CarlFKAlready up-to-date.01:50
mithroCarlFK: I pushed some changes earlier in the week which fix your breaking it01:50
CarlFKI am guessing I need to "sudo make install something"01:51
CarlFKbut I have to run to meet a friend01:51
mithroCarlFK: have you set your path so it sees the bin directory?01:51
CarlFKcan you add to the README how you think the install should go, I can test on a fresh os when I get back01:52
mithroCarlFK: "make setup", everything works01:52
*** panther has joined #timvideos01:53
*** panther is now known as sab_12301:53
sab_123mithro, what you said yesterday was true. It will interfere .01:54
mithroHas anyone seen this? ?02:00
tpbTitle: B101 HDMI to CSI-2 Bridge (15 p.) (at
*** MaZderMind has quit IRC03:03
*** MaZderMind has joined #timvideos03:04
*** puck has quit IRC03:27
*** shenki has quit IRC03:27
*** shenki has joined #timvideos03:27
*** ChanServ sets mode: +v shenki03:27
*** nueces has quit IRC03:29
*** puck has joined #timvideos03:29
*** sab_123 has quit IRC03:48
CarlFKmithro: make setup, still broken03:59
CarlFKWARNING:root:unbind-helper not found, will have to run as root!04:00
mithroCarlFK: is the bin directory in your path?04:00
CarlFKmithro: echo $PATH04:01
mithroCarlFK: so, you need to put the conda stuff in your path04:01
CarlFKmithro: ok.. next problem ..  AssertionError: /usr/local/lib/python3.5/dist-packages/hdmi2usb.modeswitch-0.0.0+93.gd0eb917-py3.5.egg/hdmi2usb/firmware/hdmi2usb.hex04:07
tpbTitle: Ubuntu Pastebin (at
CarlFKmithro: well. this fixes it..  $PWD/hdmi2usb.hex04:09
mithroCarlFK: I'm surprised that fixes it04:18
CarlFKmithro: I have a crappy vga to hdmi thing that my lcd says '1280x720 60Hz'  and I think it is out of phase a bit (left bit of the desktop is showing up on the right side of the monitor ...04:32
CarlFK HDMI2USB>dvisampler1: ph:  -5   -5   -6 // charsync:111 [1 1 1] // WER:  0   0   0 // chansync:1 // res:36x704:32
CarlFKanything to be learned from garbage input?04:33
mithroCarlFK: what is the source?04:33
tpbTitle: LKV351 VGA / Component Video + 3.5mm Audio to HDMI Scaler (at
mithroThe mangle res value is probably because it is output HDMI signal, while we currently only support DVI signal04:38
mithrothe HDMI data is sent during the vsync/hsync blanking periods04:40
mithroCarlFK: You can find way too much information about how it at
tpbTitle: Google Docs - create and edit documents online, for free. (at
CarlFKmithro: meh.. I'm just looking for stuff to help debug05:01
mithroCarlFK: getting better debugging info is blocked behind finishing _florent_'s merge stuff05:02
*** Bertl_zZ is now known as Bertl05:41
*** ssk1328 has joined #timvideos05:49
*** TeaTime has joined #timvideos06:16
TeaTimeYo anyone online got a question to ask about hdmi2usb.06:17
*** TeaTime is now known as Guest9752706:17
Guest97527Yo anyone online got a question to ask about hdmi2usb.06:17
Guest97527I have a question to ask rather*06:17
*** aom_ has joined #timvideos06:43
mithroGuest97527: ask and hopefully someone will get back to you06:44
Guest97527Well i guess the gist of it is I am in need of a displayport 1.4 frame grabber with a usb 3.0 output, now obviously the opius doesnt support this, and the only currently available alternatives are in the form of commercial pci-e capture cards running upwards of 2 grand a pop06:47
Guest97527Is it possible for me to design my own board and adapt the hdmi2usb firmware to it.06:47
mithroGuest97527: It wouldn't be a small project06:49
mithroGuest97527: And would require a lot of new development06:49
Guest97527I thought as much bugger.06:49
mithroWe are in the process of doing that - but it will be quite a while before we can do something like that06:50
mithroGuest97527: why do you need USB 3.0 output?06:50
mithroGuest97527: I assume you want a very high resolution capture?06:51
Guest97527usb 3.0 has the bandwidth required to output uncompressed 4kp3006:51
Guest97527Around 6-7Gbps if i recall06:51
mithroGuest97527: your computer is unlikely to be able to do anything useful with that data06:51
mithroBut the apertus team is interested in those types of resolutions06:52
Guest97527Ill be pushing it into a dual e5-5645 setup (so 24 cores, x58 platform xeons) and encoding the feed using x.26406:52
mithroGuest97527: you are much better going with PCI-Express rather than USB3.0 - you won't have such high overheads06:53
Guest97527My other idea was to just grab a board with a dp 1.4 input and 4 hdmi outputs and split the stream into 4 1080p60 outputs to be captured individually, which is a little simpler (can use already made products) but a messy solution06:53
Guest97527to be reassembled on the otherside ofc.06:54
mithroGuest97527: You could probably do something like that with the Opsis and a custom expansion board06:54
mithroI've already got a plan for an expansion board which adds more HDMI outputs06:55
mithroGuest97527: but we really need more developers and funds06:55
Guest97527Yeah thats always the issue, well anyway thanks for the help mate06:55
*** Guest97527 has quit IRC06:56
*** nueces has joined #timvideos07:24
*** CarlFK has quit IRC08:36
*** nueces_ has joined #timvideos09:15
*** nueces has quit IRC09:16
*** nueces__ has joined #timvideos09:40
*** nueces__ is now known as nueces09:40
*** nueces_ has quit IRC09:40
*** panther has joined #timvideos10:31
*** panther is now known as sab_12310:31
*** aom_ has quit IRC10:37
*** panther_ has joined #timvideos10:42
*** ssk1328 has quit IRC10:44
*** sab_123 has quit IRC11:06
*** panther_ is now known as sab_12311:06
*** aps_ is now known as aps11:36
*** aom has joined #timvideos12:36
*** CarlFK has joined #timvideos13:38
*** ChanServ sets mode: +v CarlFK13:38
*** aom has quit IRC14:25
*** aom has joined #timvideos14:41
*** Bertl is now known as Bertl_oO15:49
*** aom has quit IRC15:53
*** sab_123 has quit IRC17:17
*** CarlFK has quit IRC21:35
*** CarlFK has joined #timvideos21:46
*** ChanServ sets mode: +v CarlFK21:46

Generated by 2.13.1 by Marius Gedminas - find it at!