Thursday, 2017-08-10

*** tpb has joined #timvideos00:00
*** whatnick has joined #timvideos00:25
whatnickhi all00:40
whatnickjust ordered one of the numato dev boards, point me to the toolchain to get stuff onto it00:40
*** whatnick has left #timvideos01:11
*** CarlFK has joined #timvideos01:41
*** ChanServ sets mode: +v CarlFK01:41
*** sb0 has quit IRC03:34
*** shenki has quit IRC03:41
*** shenki has joined #timvideos03:47
*** ChanServ sets mode: +v shenki03:47
*** rohitksingh_work has joined #timvideos03:54
mithropaddatrapper: pong?04:04
paddatrappermithro: Just wondering if you had had a chance to look at the Opsis set_interface issue with the audio firmware?04:05
mithropaddatrapper: Sadly no04:05
mithropaddatrapper: I'm back at work today04:05
mithropaddatrapper: Currently trying to catch up on everything with that...04:06
paddatrappermithro: Ok no worries. I haven't had much chance to do much in the last few days either04:06
mithropaddatrapper: Got a weekly update post?04:10
mithropaddatrapper: When is a good time to have our weekly meeting?04:10
paddatrappermithro: I'll have to write it tomorrow (it's past midnight at the moment), though there isn't much to report.04:14
paddatrapperMeeting wise - I'm not sure, I guess it will have to be something around this time on Friday for me (tomorrow is the conference dinner, so I'll be busy the whole night)04:16
mithropaddatrapper: you've done a bunch of testing on the Atlys / Opsis hardware now, right?04:17
paddatrappermithro: yup and found the audio firmware to work with no problems on the Atlys, while the Opsis is still a work in progress04:18
paddatrapperThe lights example also works correctly on both04:19
mithropaddatrapper: Oh great!04:21
mithropaddatrapper: Can you upload some videos of it working on all the boards you have?04:21
paddatrappermithro: yeah, it was better than I expected.04:21
paddatrapperUpload videos of the lights example, I take it? If so sure04:22
mithropaddatrapper: Basically, it would be good if we could go - "And the results of loading this you should see [this on your board](link to a video of it working)"04:22
paddatrappermithro: yeah. The audio firmware doesn't have much actually happen on the board (aside from the lights indicating start up), so I don't think that will be very useful, but can certainly do the lights example04:24
mithropaddatrapper: Yeah - I was only really thinking the lights example04:24
paddatrappermithro: I'll do it tomorrow when I can liberate tumbleweed Atlys from playing hackers on repeat :)04:25
mithropaddatrapper: Okay :-P04:25
mithropaddatrapper: Some videos of a working HDMI2USB setup and how it is connected up would be kind of nice too04:25
mithropaddatrapper: More videos which show things working are always good...04:26
paddatrappermithro: ok. I can also through in some links to how Debconf uses the opsis boards which may help. Will try get some video before the end of the conference too04:27
mithropaddatrapper: Basically we want things which are useful in two different ways04:27
mithropaddatrapper: Stuff which gets people excited about doing things04:27
mithropaddatrapper: Stuff which shows how things are setup and can be included in tutorials, documentation and similar.04:28
paddatrappermithro: ok cool. I shall keep that in mind04:28
mithrorohitksingh_work: ping?04:29
*** CarlFK has quit IRC04:32
*** sb0 has joined #timvideos05:29
*** CarlFK has joined #timvideos05:32
*** ChanServ sets mode: +v CarlFK05:32
*** whatnick has joined #timvideos05:46
mithroHi whatnick07:26
whatnickhey mithro07:32
whatnickgot my numato coming on 14th07:32
whatnickwas going to get tools07:33
mithrowhatnick: I assume I met you at PyCon AU?07:33
whatnicktisham ...07:33
whatnicki was doing esp32 stuff with joel07:33
whatnickI am thinking of making high sampling energy monitors with NN inference with the FPGA fabric07:34
whatnickwould be good learning on how such asics are built07:34
whatnickNN inference for appliance disaggregation07:35
whatnickAnyway not video related interest but ... seems like a good place to hang out07:40
whatnickthe only video related interest I have is HDCP decryption07:40
*** sb0 has quit IRC08:24
*** whatnick_ has joined #timvideos08:50
mithrotumbleweed: Ping? Wanted to ask about your pull request09:17
mithrohey jea10:54
jeahi mithro11:03
mithrojea: doing some hacking on hdmi2usb?11:12
jeamithro: not at the moment :( looking into work production issues11:13
jeaI will be taking my opsis to work tomorrow though to see if/how we could integrate it with our stuff11:15
*** rohitksingh_work has quit IRC12:42
mithroxfxf: So, can I get a copy of PyCon AU lightning talk?13:12
*** tsglove has quit IRC13:14
Ishan_Bansalmithro : ping13:26
mithroIshan_Bansal: pong13:27
mithroIshan_Bansal: I'll try and get to reviewing your change list tomorrow13:27
*** tsglove has joined #timvideos13:27
Ishan_Bansalmithro, Also mithro I am working on indentation of every module, as I have made separate branches for each module so if I am doing my changes in some module in say "Huffmancore" but those change are not made up in the next branches such as "bytestuffer" so is it a problem while merging the two.13:30
*** rohitksingh has joined #timvideos13:50
*** sb0 has joined #timvideos13:55
*** sb0 has quit IRC14:29
*** sb0 has joined #timvideos14:56
*** CarlFK has quit IRC17:08
*** CarlFK has joined #timvideos17:32
*** ChanServ sets mode: +v CarlFK17:32
*** rohitksingh has quit IRC18:39
*** cyrozap has quit IRC19:02
*** cyrozap has joined #timvideos19:02
*** cyrozap has quit IRC19:05
*** cyrozap has joined #timvideos19:05
*** CarlFK has quit IRC19:33
mithroCarlFK[m]: ping?19:33
CarlFK[m]mithro: pong19:58
mithroSo, PyCon AU just finished up a couple of days ago19:59
mithroMost videos were up within a couple of hours of the presentation \o/19:59
mithroCarlFK[m]: Thanks for your help with the veyepar side of things19:59
mithroCarlFK[m]: Also, the funding for the LiteSDI core finally got approved!20:02
*** felix_ has joined #timvideos20:03
CarlFK[m]Neat.  Um...  What does that mean?20:03
*** CarlFK has joined #timvideos20:11
*** ChanServ sets mode: +v CarlFK20:11
*** CarlFK has quit IRC20:24
mithroCarlFK[m]: We will have gateware code for doing SDI in FPGAs - means we can do our own version of the BlackMagic Decklink cards (fully open source!)20:25
* felix_ will work on the sdi gateware and axi/wishbone adaperts from next month on for half a year20:26
CarlFK[m]hi felix20:29
froztbyte(hello world)20:40
froztbyteso here's a small pandora's box: if I wanted to buy some boards, is that a thing I can do atm?20:40
paddatrapperHey froztbyte20:41
froztbytepaddatrapper: I was thinking about poking you on slack earlier20:42
froztbytepaddatrapper: didn't know if you were student-ing atm20:42
froztbytepaddatrapper: we should chat about venue stuff soon (although I prefer next week, this week is a bit of a mess)20:43
*** CarlFK has joined #timvideos20:43
*** ChanServ sets mode: +v CarlFK20:43
paddatrapperfroztbyte: I'll be studenting (kinda) from next week. Then works for me too, currently I'm busy with Debconf20:43
froztbytethere's a debconf? is this the one that I couldn't really mission to because reasons?20:44
CarlFK[m]froztbyte  - hey!20:44
froztbytehow are you, CarlFK[m]20:45
paddatrapperfroztbyte: yup. In Montreal this year20:45
felix_for the people who are interested in the litesdi project: we have created an irc channel for that: #litesdi currently not much traffic, but i'm planning to mainly use that channel for discussions about the sdi gateware project20:46
CarlFK[m]hanging in - just did PyOhio last week - "the venue staff wants to handle audio, lets let the have it.  it may be a disaster and then for 2018 we will tell them to stuff it."20:46
CarlFK[m]guess what?20:46
froztbytestill hunting equipment?20:46
froztbytefelix_: I'm "interested" but I don't know if I have braincycles to really allocate in the beginning. I'll wallflower and be willing to trial things out or so, if I can get my hands on things.20:47
froztbyte(interested in the sense that I'm totally keen on such a thing existing, but would be useless at general meaningful participation)20:47
* froztbyte --honest20:48
felix_well, you can just join that channel and read the stuff that is discussed there :)20:48
froztbyteyeah, done so20:49
felix_the work on the litesdi gateware will begin next month and i'm currently still more or less busy with my many other projects20:50
CarlFK[m]I need to make the the little blooper real out of the sound check, then the venu guy  touching the mic and then no more sound BUT WE DIDN'T NOTICE because they didn't give me the mixers they promised20:52
mithrofroztbyte: What type of boards?20:56
*** CarlFK has quit IRC20:56
froztbytemithro: something like opsis or such21:04
mithrofroztbyte: What do you want to do?21:04
froztbyteevery time we've been doing the stuff in .za, we've been bringing CarlFK's kit over21:04
mithrofroztbyte: Ahh - can you fill out the Opsis interest spreadsheet I sent out a couple of months ago?21:05
froztbyteI could, where is it?21:05
*** CarlFK has joined #timvideos21:06
*** ChanServ sets mode: +v CarlFK21:06
froztbyteCarlFK[m]: lululul21:06
tpbTitle: Google Groups (at
froztbyteCarlFK[m]: it's _comical_ at this point :|21:06
froztbyteCarlFK[m]: or I guess maybe like taxes; just one of those things you can rely on21:06
froztbytemithro: thanks21:07
froztbytemithro: for allowing me to adequately/truthfully gauge the number I put down here, what's a (very) rough ballpark of what the per-board cost could be?21:07
froztbyte$150? $200? $500?21:07
mithrofroztbyte: Probably around $420 USD21:08
froztbytemithro: is this a list I should be following?21:09
mithrofroztbyte: Yes21:09
froztbytemithro: can I trick you into mentioning things on IRC in general? ;)21:10
mithrofroztbyte: Hrm?21:10
froztbytejust being difficult, don't mind me.21:10
froztbyte(it's way too late and my day has run way too long, brain unclutching)21:10
mithroIRC is probably where most discussion ends up happening21:11
CarlFK[m]froztbyte: my guess is what you are fishing for... gets you an old boot.    mithro is trying to find out about the next production run and hasn't heard back.21:14
froztbyteCarlFK[m]: nah, I've answered the list entry already21:15
froztbyteI figure there's no immediate hope, that's okay21:15
froztbytefigured so at the onset of my original question, even :)21:15
paddatrapperfroztbyte: we _may_ have some to use for pyconza this year, as I plan on bringing one of Debian's back and tumbleweed may be able to organise too21:21
tumbleweedmithro: are you still interested in looking at the atlys?21:23
mithrotumbleweed: In what manner?21:23
mithroI would like paddatrapper to have an Opsis and Atlys if possible21:23
tumbleweedmithro: that the USB streaming is all broken21:24
mithrotumbleweed: Yes, I want to fix that but I'm still waiting a bug report?21:29
tumbleweedmithro: sure21:32
tumbleweedis there a bug about 2 outputs not working?21:33
* tumbleweed doesn't see one21:33
mithrotumbleweed: Dunno, there should be21:34
mithrotumbleweed: Actually, its 2 outputs, 2 inputs and the encoder that things break21:34
mithrotumbleweed: I believe that xfxf was using 1 input, 2 outputs and encoder at PyCon AU? (xfxf was that correct?)21:34
tumbleweed1 input off didn't seem to help21:34
mithrotumbleweed: Yeah - I'd log a bug about that...21:34
CarlFKI'm in a coffee shop at the moment, but I can setup tests at home in 30 in or so21:35
tpbTitle: enabling a second output makes output unstable · Issue #346 · timvideos/HDMI2USB-litex-firmware · GitHub (at
mithrotumbleweed: Need as much information as you can provide21:37
mithrotumbleweed: And some output of the debugging would also be very useful21:38
CarlFK[m]tumbleweed (IRC): I'll need 2 hdmi displays, right?21:38
mithroCarlFK[m]: well, nothing really needs to be attached to the hdmi outputs21:39
CarlFK[m]I was wondering bout that.  good.  they are in boxes.21:39
mithroCarlFK[m]: if you see the underflow number be non-zero then it is a problem21:39
olasdmithro: our electronics people don't do ice based designs, so there's not much interest in the toolchain21:42
mithroolasd: Hrm?21:42
paddatrappermithro: we use/want to use ice for the fgpa?21:43
tumbleweedmithro: and
tpbTitle: USB streaming is corrupted on Atlys · Issue #347 · timvideos/HDMI2USB-litex-firmware · GitHub (at
tumbleweedmithro: I need to know what information you need21:44
mithropaddatrapper: I doubt it will be ever useful for HDMI2USB, but it is useful for MicroPython on FPGAs which is a good gateware drug to getting people able to contribute to HDMI2USB21:44
paddatrappermithro: ah I see21:44
mithrotumbleweed: I need the output of "version", a picture of how things are wired, info on how the device is configured, dmesg output on capture computer, output from what ever software is doing the capture, lsusb output info21:45
CarlFK[m]froztbyte (IRC): and anyone that wants a chuckle...
tpbTitle: Veyepar: Video eyeball produce and review. (at
tumbleweedwe tried various wirings21:46
mithrotumbleweed: You sure you are loading the Atlys hdmi2usb.hex? The hdmi2usb.hex is board specific (so don't load the one from the opsis)21:46
tumbleweedand the setup that we got that from is long gone21:46
tumbleweedmithro: yes I'm sure21:46
tumbleweedalso, the atlys is still online for you to play with21:47
mithrotumbleweed: Okay great21:47
tumbleweedit's been up for about a wek now :P21:47
CarlFK[m]click the bump in the image that takes you to 11:00:00 where the presetner says stuff... then click the last little peeks that take you to 16:00:00 where staf guy touches the mic and ... thats the end of audio :p21:47
mithrotumbleweed: I'm afraid that I also have to do the work I'm actually paid for :-P21:48
tumbleweedmithro: of course21:48
tumbleweedand I have lots of conference orga work to do21:48
mithrotumbleweed: All the above info would be super useful21:48
froztbyteCarlFK[m]: great crowd noise tho21:50
CarlFK[m]froztbyte: left is sound system, right is camera mic.  which is what got used for the talk cuz the other thing is or whatever21:51
CarlFK[m]see ya, bb in 2021:51
*** CarlFK has quit IRC21:51
*** CarlFK has joined #timvideos22:14
*** ChanServ sets mode: +v CarlFK22:14
xfxfmithro: thats correct. i thought it was only the two plenary rooms but at some point MCEC added off stage confidence monitors to the other rooms too and hooked them to output1. so even venue av people have confidence in the opsis's now :P22:14
mithroxfxf: Ha22:15
mithroxfxf: I would really like to find out why Debconf is having issues while you are not22:15
xfxfwierd. no idea - they are def using redmere cables? what length?22:16
mithroxfxf: The problem is underflows - so it has got something to do with the memory bandwidth22:18
paddatrapperxfxf: all the redmere cables we are using are about 1.5m22:18
xfxfmanafacturing differences? i assume i ordered mine earlier than debconf, was all stock made in one go?22:18
* tumbleweed should get some longer ones...22:18
xfxfpaddatrapper: should be fine, we're using 3ft to 12ft ones no problems. i did notice that they get damaged easily tho, i keep having to buy new ones22:19
xfxfand by damaged you don't see physical damage but you start having issues with capture. not underflows tho, just WER errors22:20
mithroWER is very different to underflow22:20
mithroxfxf: I wonder if it's the 50Hz verse 60Hz difference22:20
mithroxfxf: 60Hz does require more bandwidth I think?22:21
xfxfim happy to try some tests tonight, but can't now as on train to work22:21
CarlFK[m]mithro (IRC): is xenial lts your preference ?22:21
mithroxfxf: You mean you can't do FPGA development on the train? What a slacker?22:22
xfxfis debconf in us or eu?22:22
tumbleweedxfxf: neither, ca22:22
xfxfi thought eu was 50hz generally22:22
mithroCarlFK[m]: Latest Ubuntu LTS is probably the best choice at the moment22:22
tumbleweedthe country north of us22:22
xfxfoh right. i think of ca as California. did they secede when i wasn't looking? :P22:23
mithroCanada :-P22:23
tumbleweedso yeah, metric system + 110v @ 60Hz22:24
xfxf1 of 3 is right at least22:24
paddatrapperwell "metric system" I still had to give cable lengths in feat...22:24
*** twoolie has joined #timvideos22:25
xfxfwierd you're having issues tho, things were fairly smooth at pycon au for us. i think we had a few situations where we had to reboot a unit but it was rare22:25
xfxfthat said22:26
xfxfmost of our presenters used macs22:26
xfxfi bet thats not the case there22:26
tumbleweedhaven't had any input issues22:26
mithroAt 16bit YUV -- 720p50 is 737MBit/s while 720p60 is 885Mbit/s22:26
paddatrapperThinkpads were the most common. I think I have seen 1 mac and that was running Debian22:27
tumbleweedand a number of old vga thinkpads22:27
xfxflul vga22:27
*** twoolie has quit IRC22:32
CarlFK[m]did I bring my xatron? vga->hdmi box to LCA?22:50
CarlFK[m]I can't find it.22:50
xfxfnot sure but i don't have it22:57
mithrocr1901: so, SPI xmodem? :-P23:13
cr1901mithro: Debugging it now23:13
cr1901mithro: You have seen my recent commit?23:14
mithrocr1901: Probably not23:14
cr1901mithro: Oh, well current status is I hacked up an xmodem impl that supports all modes, that naturally doesn't work. But I have qemu-litex up and running and I'm testing it against minicom's xmodem impl23:15
cr1901After that I'll do sfl (which should be easier, tbh), confirm it works on qemu-litex. Then we get mimasv2 support "for free"23:16
mithrocr1901: Okay23:16
cr1901And my opsis is running23:16
mithrocr1901: \o/23:17
cr1901mithro: Has anyone besides me played w/ opsis on windows?23:17
cr1901to make things quick: hdmi2usb-mode-switch is not going to be trivial to make  work on Windows.23:18
cr1901So I build my firmware on Windows, and then send it to an ARM SBC which talks to the opsis for now23:19
cr1901mithro: We can talk more about this after the SPI stuff, just wanted to give a heads-up. Monday was the first time I actually tried playing w/ Opsis _on Windows_.23:20
mithrocr1901: So I'm considering meson verse waf for replacing our current mishmash of shell, make and python23:20
mithrocr1901: Hrm? libusb should work fine on Windows....23:20
*** whatnic74 has joined #timvideos23:21
mithrocr1901: One big missing part is a Python only implementation (using libusb) of the fxload tool23:21
cr1901mithro: That's not the problem. The problem is that on Windows, you either can use the libusb or the CDC driver. You can't hot-swap between them23:21
cr1901*nix OSes (including the BSDs) have a mechanism for "hot swapping" between the serial/CDC driver and "direct" access using libusb23:22
cr1901Windows does not have that mechanism, so the solution is to install a special signed driver called libusbK23:22
cr1901mithro: libusbK makes libusb work on Windows, but you lose access to the driver you have previously23:22
*** whatnick_ has quit IRC23:23
cr1901unless you uninstall  libusbK and reinstall the CDC driver23:23
cr1901it sucks23:23
cr1901mithro: I've heard good things about meson.23:23
mithrocr1901: Looking at the docs, there seems a number of ways around this23:24
cr1901mithro: link?23:25
mithroLooking at ...23:27
tpbTitle: libusb-win32 – libusb (at
mithroAnyway, that is not important for now...23:28
cr1901mithro: I agree- just a heads up23:29
cr1901mithro: The way _I_ thought  about doing it was to use the free USB endpoint in the fx2 as a write-only implicit mode switch. But that means modifying the firmware and the numato driver (is their driver even open source?)23:30
mithrocr1901: Definitely an option...23:31
* cr1901 /end minddump23:31
*** whatnic74 has quit IRC23:36
mithrocr1901: The big problem I'm trying to solve at the moment is why the mor1k is locking up when executing our firmware23:37
*** whatnick_ has joined #timvideos23:37
cr1901mithro: Have you tried compiling it to a simulator?23:38
mithroYes, it all runs fine in qemu23:38
cr1901mithro: I mean icarus23:38
mithroI haven't tried the verilator one recently23:38
cr1901Well, icarus is only gonna work if it locks up early, but I've successfully found a lock up w/ lm32 in one of my designs that way23:39
mithrocr1901: If I haven't figured it out by when you finish the xmodem SPI stuff, I'll probably get you to take a look at it23:40
cr1901mithro: Okay23:40
mithroxfxf: What is the current status of your GUI thingy?23:48
xfxfmithro: the tk version works for basic use, started rewriting using toga, haven't finished23:48
xfxfit's pretty high up the TODO list to complete after pycon was done23:49
mithroxfxf: Is your code up anywhere?23:49
xfxfyeah, not sure if it's the latest version tho23:49
xfxflet me check23:49
tpbTitle: hdmi2usbmon/ at master · xfxf/hdmi2usbmon · GitHub (at
xfxfthere's a later version than that, but it's on a device not near me atm23:51
mithroI'll dump that onto the github issue23:51
xfxfiirc that version lets you choose inputs/outputs, and dumps debug output into a text frame, not much more23:51
mithroxfxf: Added that info to the GitHub issue about the GUI23:59

Generated by 2.13.1 by Marius Gedminas - find it at!