| *** tpb has joined #timvideos | 00:00 | |
| G33KatWork | which lm32 toolchain do you use? | 00:32 |
|---|---|---|
| mithro | G33KatWork: The one in upstream gcc/binutils | 00:33 |
| G33KatWork | in default ubuntu repos? | 00:36 |
| G33KatWork | i can't find it in 14.04 | 00:36 |
| G33KatWork | that's why i'm asking | 00:36 |
| mithro | G33KatWork: You are not using the bootstrap scripts? | 00:36 |
| G33KatWork | i do | 00:36 |
| mithro | G33KatWork: They get you a precompiled lm32 gcc/binutils using conda | 00:38 |
| G33KatWork | yeah, i just found the code in the scripts | 00:38 |
| G33KatWork | i'll dig around | 00:38 |
| G33KatWork | thx | 00:38 |
| mithro | G33KatWork: The version of gcc/binutils in ubuntu does support lm32 but they don't compile with that on | 00:38 |
| mithro | G33KatWork: build scripts for the conda packages are at https://github.com/timvideos/conda-hdmi2usb-packages | 00:39 |
| tpb | Title: timvideos/conda-hdmi2usb-packages · GitHub (at github.com) | 00:39 |
| mithro | G33KatWork: I've been suggesting that misoc have a set of conda packages which include lm32, or1k, etc | 00:42 |
| mithro | G33KatWork: please do feel free to ask questions! | 00:43 |
| G33KatWork | get-env-root.sh was executed | 00:44 |
| G33KatWork | get-env.sh not | 00:44 |
| mithro | G33KatWork: ahh! | 00:44 |
| G33KatWork | but i didn't get an error message | 00:44 |
| G33KatWork | most likely it's because of my weird setup | 00:44 |
| G33KatWork | i like to use docker containers for bigger projects | 00:44 |
| G33KatWork | so that i don't pollute my main system | 00:45 |
| G33KatWork | and i'm using arch which isn't always an advantage when you need older versions of packages | 00:45 |
| Bertl | why the hell docker? :) | 00:45 |
| G33KatWork | so i basically created a massive 23GB container with ise *cough* | 00:45 |
| G33KatWork | Bertl: why not? :D | 00:46 |
| Bertl | I can understand LXC, but what advantage does docker give you? | 00:46 |
| mithro | Bertl: the push/pull of images is pretty nice | 00:46 |
| G33KatWork | never used lxc | 00:46 |
| Bertl | most likely docker is using that anyway | 00:46 |
| G33KatWork | cgroups, yes | 00:46 |
| Bertl | I know docker became very popular, never figured why | 00:47 |
| Bertl | mithro: maybe the push/pull is a reason, yes | 00:48 |
| G33KatWork | i just jumped on the hypetrain and i like it | 00:48 |
| mithro | Bertl: it was really weird why docker became very popular quickly | 00:48 |
| Bertl | G33KatWork: fair enough :) | 00:48 |
| G33KatWork | weird... i execute the bootstrap script again and this time it installs conda etc. | 00:50 |
| G33KatWork | let's just assume PEBCAK for now | 00:50 |
| mithro | G33KatWork: I started work on having a docker container which just runs bootstrap.sh | 00:55 |
| mithro | G33KatWork: to test it works on a clean system | 00:55 |
| G33KatWork | it's synthesizing things :> | 00:56 |
| mithro | bblr - lunch! | 00:58 |
| G33KatWork | i *so* need to play more with migen and misoc | 00:58 |
| Bertl | bon apettite! | 00:58 |
| mithro | Bertl: I couldn't figure out why you would say "bon apettite" to G33KatWork's comment then realised that you where saying it to my lunch comment :P | 01:35 |
| mithro | Can everyone reweet https://twitter.com/mithro/status/654466236575711233 ? | 01:36 |
| tpb | Title: Tim mithro Ansell on Twitter: "4k video with FOSS DisplayPort core + the @numatolab Opsis! https://t.co/c1elrCf0RW @crowd_supply @field_hamster http://t.c (at twitter.com) | 01:36 |
| xfxf | mithro: neat! | 01:37 |
| xfxf | how do we capture 4k video tho with usb2? | 01:37 |
| mithro | xfxf: we don't really | 01:37 |
| xfxf | so vhat's ze intended purpose of 4k support? | 01:37 |
| mithro | xfxf: But we might be able to put both the HDMI inputs side-by-side on the 4k output | 01:38 |
| mithro | xfxf: we can downscale the 4k input | 01:38 |
| Bertl | xfxf: for now, it might soon be possible with the AXIOM Beta :) | 01:45 |
| Bertl | (4k video input to the opsis that is) | 01:48 |
| Bertl | mithro: yeah, sorry, was a race condition | 01:49 |
| mithro | Bertl: nah, it was entertaining :) | 01:49 |
| Bertl | I usually drop the nick: prefix when immediately answering to somebody (to reduce cluttering), but under certain conditions that gives suboptimal results :) | 01:56 |
| G33KatWork | hm, is it normal, that the router routes for an hour now? | 02:12 |
| G33KatWork | i guess my fpga projects are all way smaller than HDMI2USB | 02:13 |
| mithro | G33KatWork: it should take about 15-30 minutes | 02:22 |
| mithro | G33KatWork: because ISE is crap, occasionally it gets confused and gets stuck | 02:27 |
| Bertl | maybe limit process time? | 02:28 |
| *** travis-ci has joined #timvideos | 02:46 | |
| travis-ci | [mithro/HDMI2USB-misoc-firmware/opsis-fixes#21] (ae94db2): The build has errored. (https://travis-ci.org/mithro/HDMI2USB-misoc-firmware/builds/85446068) | 02:46 |
| *** travis-ci has left #timvideos | 02:46 | |
| *** CarlFK has joined #timvideos | 02:47 | |
| *** ChanServ sets mode: +v CarlFK | 02:47 | |
| *** shenki_ is now known as shenki | 03:14 | |
| mithro | https://www.reddit.com/r/opensource/comments/3ot9cf/4k_displayport_using_foss_fpga_core_on_foss/ | 03:31 |
| Bertl | hmm, goes to nowhere | 03:40 |
| Bertl | ah, no, goes to somewhere with no content :) | 03:40 |
| mithro | Bertl: people haven't commented yet | 03:41 |
| Bertl | how fast is the memory on the opsis? i.e. what sustained read/write throughput can be achieved? | 03:44 |
| mithro | Bertl: _florent_ would know were we are at with that at the moment | 03:45 |
| Bertl | _florent_ is european time zone, yes? | 03:53 |
| mithro | Bertl: yeah | 04:35 |
| *** Bertl is now known as Bertl_zZ | 05:18 | |
| * seaLne catches up on what he's missed | 06:05 | |
| mithro | hey seaLne! Welcome back | 06:17 |
| *** se6astian|away is now known as se6astian | 07:07 | |
| *** Bertl_zZ is now known as Bertl | 07:38 | |
| *** rohitksingh has joined #timvideos | 07:56 | |
| seaLne | mithro: hi, running get-env.sh in your HDMI2USB-misoc-firmware fork currently so i can build with your opsis-fixes | 08:20 |
| seaLne | mithro is opsis-fixes building for you? https://paste.kde.org/pyohnols1 master in your fork doesn't complain | 08:40 |
| tpb | Title: KDE Paste (at paste.kde.org) | 08:40 |
| mithro | seaLne: guess I should have tested after making that change for _Florent_ | 08:41 |
| seaLne | although is that a complaint from upstream misoc? | 08:42 |
| mithro | seaLne: nope, it was just me being an idiot | 08:48 |
| mithro | seaLne: do a pull | 08:48 |
| mithro | Bertl: you were talking about using mini-DisplayPort for cabling from the camera to the Opsis? | 08:49 |
| Bertl | yep, looks like the best option so far | 08:50 |
| Bertl | was considering usb 3 for a brief moment, but the number of connections is too huge | 08:50 |
| mithro | Bertl: what mini-DP connector would you use? | 09:12 |
| mithro | Bertl: I'm pondering a TOFE-to-TOFE brdige | 09:13 |
| mithro | s/brdige/bridge/ | 09:13 |
| Bertl | that is a really good question, I was looking for some last time | 09:13 |
| Bertl | the cheapest available (not directly from china) is the farnell multicomp receptacle | 09:14 |
| Bertl | http://www.farnell.com/datasheets/1660657.pdf | 09:14 |
| Bertl | it looks reasonable and should not require plated slots | 09:14 |
| Bertl | i.e. they can be round vias :) | 09:15 |
| Bertl | directly from china there are more options, also designs which are lower or even at board level | 09:17 |
| Bertl | but please let me know if you find something nice there :) | 09:17 |
| mithro | Bertl: got a footprint for that? | 09:20 |
| Bertl | not yet | 09:20 |
| mithro | Bertl: you generally work in Eagle it seems? | 09:20 |
| Bertl | yes, it is an old habit ... and you know, old habits die hard :) | 09:21 |
| Bertl | migration to kicad is on our todo list though | 09:22 |
| mithro | Bertl: Do you generally do the footprints yourself, or you managed to get a minion to do that? :P | 09:22 |
| Bertl | I tried, but till now, all minions ended up making more work than it would have been in the first place ... well what minions usually do :) | 09:23 |
| Bertl | fortunately since farnell has bought cadsoft, they are creating footprints and component libraries on a daily basis | 09:24 |
| seaLne | mithro: is the tofe kicad stuff available yet? | 09:25 |
| mithro | seaLne: Yes! | 09:25 |
| mithro | seaLne: I'm about to try and eat my own dogfood and create the HDMI4 board | 09:26 |
| mithro | seaLne: realized I need to add a "rename" script, so writing that now | 09:26 |
| seaLne | cool, where are they? | 09:27 |
| mithro | seaLne: https://github.com/timvideos/HDMI2USB-TOFE-kicad-template | 09:28 |
| tpb | Title: timvideos/HDMI2USB-TOFE-kicad-template · GitHub (at github.com) | 09:28 |
| seaLne | ty | 09:29 |
| _florent_ | hi | 09:32 |
| _florent_ | Bertl: for the memory throughput on the opsis, we have probably something like 5Gbps | 09:34 |
| _florent_ | 16bits*50MHz*8(QuarterRate + DDR)=6.4Gbps | 09:34 |
| _florent_ | so close to 5Gbps with 80% (estimated) efficiency | 09:35 |
| Bertl | that is read or write or read and write together? | 09:39 |
| mithro | seaLne: just pushed the rename script | 09:39 |
| mithro | _florent_: Is the quarter rate the highest we can go? Or do you think there are more tricks you can play with the SERDES to get higher? | 09:43 |
| Bertl | the main question for us is if we can buffer frames at 4k raw there | 09:44 |
| seaLne | mithro: i think it would be useful to have a half width version as that should fit within 100x100mm after which pcbs tend to get way more expensive to make from your generic chinese houses | 09:44 |
| *** travis-ci has joined #timvideos | 09:45 | |
| travis-ci | [mithro/HDMI2USB-misoc-firmware/master#23] (bac3cf5): The build passed. (https://travis-ci.org/mithro/HDMI2USB-misoc-firmware/builds/85487799) | 09:45 |
| *** travis-ci has left #timvideos | 09:45 | |
| mithro | Bertl: 4k at 30fps seems to be ~5Gbits | 09:46 |
| mithro | Bertl: which means you could write to it or read from it - but not do both at the same time it seems? | 09:46 |
| Bertl | yes, it should be about 4.5Gbit/s actually | 09:47 |
| Bertl | but of course, it would require read _and_ write | 09:47 |
| mithro | seaLne: by half width, do you mean half height, or half height + half width? | 09:47 |
| Bertl | but I kind of already expected that the memory would be too slow for this ... | 09:48 |
| seaLne | half width + full height would still give you a decent area to work with | 09:48 |
| mithro | Bertl: yeah - it's why on a V2 we need both the higher speed DDR the Artix-7 can do, plus the dual MCBs | 09:49 |
| mithro | seaLne: looking at it now | 09:49 |
| seaLne | ~$50 v ~$135 for 10x 4 layer boards from elecrow | 09:54 |
| Bertl | what is the area of the boards? | 09:55 |
| Bertl | (bounding box) | 09:55 |
| seaLne | Bertl: https://github.com/timvideos/HDMI2USB-TOFE-kicad-template/tree/master/img | 09:55 |
| tpb | Title: HDMI2USB-TOFE-kicad-template/img at master · timvideos/HDMI2USB-TOFE-kicad-template · GitHub (at github.com) | 09:55 |
| seaLne | 106x167 | 09:55 |
| Bertl | is it intentional that the measurements are unreadable? :) | 09:56 |
| seaLne | the view of the raw image is ok https://raw.githubusercontent.com/timvideos/HDMI2USB-TOFE-kicad-template/master/img/tofe-full-height.png | 09:56 |
| seaLne | well semi ok :) | 09:57 |
| Bertl | with two eyes closed, I see :) | 09:57 |
| Bertl | and we have more than 106 in height | 09:57 |
| Bertl | the fingers are outside | 09:57 |
| seaLne | ah, yeah | 09:58 |
| Bertl | so roughly 115 | 09:58 |
| seaLne | ok well something between half and full height then | 09:58 |
| mithro | seaLne: Just added schematic components for 1x and 4x boards | 09:58 |
| Bertl | so rougly 100 USD per board on OSHpark | 09:59 |
| Bertl | yes, there should definitely be a smaller template as well | 10:00 |
| mithro | Fair enough | 10:04 |
| seaLne | mithro: with opsis-fixes built and your openocd what do i do now? | 10:05 |
| mithro | seaLne: PROG=openocd make load-gateware | 10:05 |
| mithro | Bertl / seaLne: The template actually only marks the "largest" area you can use - so you can make it as small as you need | 10:06 |
| mithro | But yeah, having some smaller suggested versions is a good idea | 10:06 |
| Bertl | no problem with holes and such stuff? | 10:07 |
| mithro | Depends on how securely mounted you want your board to be | 10:08 |
| mithro | If ya happy to have it supported by the PCB fingers, you can do pretty much anything :P | 10:08 |
| Bertl | IMHO half height, min width (i.e. flush with the end of the connector) would be nice | 10:08 |
| Bertl | but yes, of course, fingers should suffice :) | 10:09 |
| mithro | For the mini-DisplayPort adapter, I suggest we have the board be just big enough so removing it from the socket is easy enough :) | 10:10 |
| Bertl | yes, that was my idea | 10:10 |
| Bertl | my scratchpad says 63x24mm including fingers | 10:12 |
| seaLne | mithro: any ideas? https://paste.kde.org/p8lmbighg which openocd confirms that it should be the one in /opt | 10:12 |
| tpb | Title: KDE Paste (at paste.kde.org) | 10:12 |
| Bertl | mithro: which means 8 USD for one board at OSHpark :) | 10:13 |
| _florent_ | Berl: numbers were for read and write together (thus the 80% efficiency) | 10:13 |
| Bertl | okay, thanks, case settled | 10:13 |
| _florent_ | mithro: we can probably improve memory bandwidth to get something near 8Gbps | 10:13 |
| mithro | seaLne: that is quite weird | 10:14 |
| _florent_ | to do read/write with 4K, you can probably use YCbCr 4:2:2 if that's acceptable | 10:15 |
| mithro | seaLne: what does "git status; git describe" show? | 10:15 |
| mithro | _florent_: Cheat and do 4:2:0 :P | 10:15 |
| seaLne | mithro: https://paste.kde.org/pqw24veqy | 10:16 |
| tpb | Title: KDE Paste (at paste.kde.org) | 10:16 |
| mithro | seaLne: hrm - I'm v0.0.0-414-ged4fea9 | 10:18 |
| seaLne | arrggghh i ended up building in master that time | 10:18 |
| mithro | seaLne: You don't appear to be on my opsis-fixes branch? | 10:18 |
| seaLne | i'll try make load-gateware again in about an hour | 10:20 |
| * seaLne leaves it to build | 10:21 | |
| *** travis-ci has joined #timvideos | 10:45 | |
| travis-ci | [mithro/HDMI2USB-misoc-firmware/opsis-fixes#22] (ed4fea9): The build has errored. (https://travis-ci.org/mithro/HDMI2USB-misoc-firmware/builds/85487793) | 10:45 |
| *** travis-ci has left #timvideos | 10:45 | |
| Bertl | mithro: how about a trade, populated 1xHDMI and 3xPMOD Beta plugin modules against 3x miniDP Opsis - Beta link? | 10:59 |
| Bertl | (no cables, just boards) | 10:59 |
| mithro | Bertl: I already have the PMOD Beta plugins on order from OHS park :) | 11:03 |
| Bertl | ah, damn! :) | 11:03 |
| Bertl | and if I ask nicely (pretty pretty please?) would you consider doing the Beta side of the 3x miniDP link together with the Opsis one? :) | 11:07 |
| Bertl | (so that I can order both via OSHpark from the gerbers) | 11:07 |
| mithro | Bertl: I'm happy to do the schematic, but I'm not great at the PCB side | 11:08 |
| mithro | Bertl: I generally get Numato to do that part :) | 11:09 |
| Bertl | hmm, I see ... | 11:09 |
| mithro | Bertl: they would probably to the PCB side of this if we got them a schematic and board edge+layout | 11:09 |
| mithro | s/to/do/ | 11:09 |
| mithro | Bertl: help them get more Opsis boards sold, good for them :-) | 11:10 |
| mithro | Bertl: and I think this would work as an Opsis <-> Opsis link | 11:10 |
| Bertl | well, we have 12LVDS channels which go nicely over 3 DP, leaves at least 6 I/Os (AUX) for medium speed | 11:11 |
| Bertl | you need a lot more if you want to bridge TOFE, no? | 11:11 |
| Bertl | or are your plans to go with 3x DP as well? | 11:12 |
| mithro | Bertl: I was thinking 4x mDP | 11:13 |
| Bertl | okay, so that would of course work fine for us as well | 11:14 |
| Bertl | when do you plan to have such a board ready (i.e. so that it can be ordered)? | 11:15 |
| Bertl | (unpopulated PCB I mean) | 11:15 |
| Bertl | although I wouldn't mind ordering a populated either | 11:16 |
| mithro | Bertl: https://github.com/timvideos/HDMI2USB-numato-opsis-docs/blob/gh-pages/pdfs/NumatoOpsis-schematic-v2-prototype-13-gb1c4312.pdf | 11:25 |
| tpb | Title: HDMI2USB-numato-opsis-docs/NumatoOpsis-schematic-v2-prototype-13-gb1c4312.pdf at gh-pages · timvideos/HDMI2USB-numato-opsis-docs · GitHub (at github.com) | 11:25 |
| mithro | Bertl: the "long pole" would be getting a footprint done and confirmed for that part | 11:28 |
| mithro | I'm going to do some non-work like recreation before going to bed :) | 11:29 |
| mithro | Bertl: actually, why mDP rather than normal size DP? | 11:29 |
| Bertl | because normal size DP is huge | 11:30 |
| Bertl | at least it won't fit on the Beta plugin module | 11:30 |
| mithro | Bertl: Guess so, maybe I don't really understand how small your boards are :) | 11:30 |
| Bertl | well, do you have a MicroZed nearby? | 11:31 |
| mithro | Bertl: nope! Don't own one | 11:32 |
| Bertl | parallella? | 11:33 |
| mithro | Nope, Zybo | 11:36 |
| Bertl | way too large :) | 11:37 |
| Bertl | but you probably got a credit card, no? | 11:37 |
| Bertl | so the small side of the credit card is slightly smaller than the dual slot plugin module is wide | 11:37 |
| Bertl | i.e. two PCIE-36 connectors almost side by side | 11:38 |
| seaLne | mithro: so make load-gateware seems to work fine but i'm not sure load-fx2 is working should i still have nothing in dmesg after the USB-JTAG-IF? | 11:53 |
| mithro | seaLne: you'll need to use HDMI2USB-mode-switch for the moment | 11:54 |
| mithro | seaLne: https://github.com/mithro/HDMI2USB-find-board | 11:54 |
| tpb | Title: mithro/HDMI2USB-find-board · GitHub (at github.com) | 11:54 |
| mithro | seaLne: you'll need to switch to serial mode, configure the firmware and then load the fx2 firmware | 11:55 |
| seaLne | mithro: what is required to configure the firmware, not sure what you are meaning by that step and is loading supposed to be like python hdmi2usb-mode-switch.py --load-fx2-firmware ../HDMI2USB-misoc-firmware-mithro/firmware/fx2/hdmi2usb.hex ? | 12:05 |
| mithro | seaLne: ./opsis-mode-switch --mode=serial; make connect-lm32; play with settings; disconnect; make load-fx2 | 12:06 |
| seaLne | mithro: where is opsis-mode-switch? | 12:08 |
| mithro | seaLne: it's just a symlink to hdmi2usb-mode-switch.py in the ./bin/ directory | 12:08 |
| seaLne | ah ok | 12:08 |
| mithro | still need to fix it so that the serial port works during jtag and hdmi2usb mode | 12:12 |
| mithro | the mode-switch stuff should probably download the .bit firmware too | 12:13 |
| seaLne | changed to ACM1 in Makefile.lm32 due to faffing. but i don't seem to get beyond [FLTERM] Starting... | 12:14 |
| mithro | did you hit enter? | 12:15 |
| seaLne | that would be it | 12:15 |
| * seaLne feels silly failing miserably to disconnect from flterm | 12:21 | |
| seaLne | C-c seemed a bit harsh | 12:23 |
| seaLne | mithro: so i have uvcvideo: Found UVC 1.00 device HDMI2USB (1d50:60b7) in dmesg but no /dev/videoX should that be appearing after the load-fx2? | 12:31 |
| seaLne | ooh got one now | 12:47 |
| *** _florent_ has quit IRC | 13:03 | |
| *** _florent_ has joined #timvideos | 13:03 | |
| *** travis-ci has joined #timvideos | 13:11 | |
| travis-ci | [mithro/HDMI2USB-misoc-firmware/opsis-fixes#22] (ed4fea9): The build has errored. (https://travis-ci.org/mithro/HDMI2USB-misoc-firmware/builds/85487793) | 13:11 |
| *** travis-ci has left #timvideos | 13:11 | |
| seaLne | so if i set encoder on and video_matrix connect pattern output0 i get pretty bars on my monitor but vlc won't play /dev/video0 and i can't seem to reconnect via serial | 13:17 |
| CarlFK | seaLne: video_matrix connect pattern encoder | 13:17 |
| seaLne | also if i plug in the only laptop i have that has hdmi (dell xps m1530) i get dvisampler1: IDELAY busy timeout constantly on the serial | 13:18 |
| seaLne | CarlFK: ah so its a seperate sink, makes sense i guess. thanks | 13:18 |
| seaLne | vlc still isn't seeing anything :( | 13:23 |
| CarlFK | seaLne: did you do fxload... ? | 13:29 |
| seaLne | make fx2-load ? | 13:30 |
| CarlFK | um.. maybe. In the last few days i stopped using make to do everything | 13:32 |
| seaLne | that just calls firmware/fx2/download.sh firmware/fx2/hdmi2usb.hex | 13:33 |
| seaLne | which calls fxload | 13:34 |
| CarlFK | heh.. this won't work for you unless you get the patched fxload.. I have it in my ppa | 13:37 |
| CarlFK | fxload -B libusb -D vid=0x16c0,pid=0x06ad -t fx2lp -I hdmi2usb.hex | 13:37 |
| CarlFK | unpatched need a dev path that you figure out by looking at lsusb | 13:38 |
| seaLne | yeah | 13:38 |
| CarlFK | lsusb | grep "16c0:06ad" | 13:38 |
| CarlFK | fxload -D /dev/bus/usb/002/008 -t fx2lp -I hdmi2usb.hex | 13:38 |
| seaLne | CarlFK: how do you get out of flterm? i'm wondering if C-c is causing problems | 13:41 |
| CarlFK | I use ^c, but yeah, something isn't good | 13:41 |
| seaLne | i should be able to connec by serial after the fxload shouldn't i? which suggests there is something wrong | 13:42 |
| CarlFK | I re-run it, get something like "no serial device" - try a few more times, try random things, something makes it work | 13:42 |
| seaLne | i get [FLTERM] Starting... but no amount of pressing return gets me in | 13:43 |
| CarlFK | hmm, I think I have seen that too | 13:43 |
| seaLne | i got in fine before fxload | 13:44 |
| mithro | seaLne: ignore CarlFK he has an Atlys not an Opsis | 13:44 |
| seaLne | and i still have pattern on output0 | 13:44 |
| mithro | seaLne: you have to use the modeswitch to get back into serial mode | 13:44 |
| mithro | seaLne: as I mentioned using the serial port and webcam at the same time doesn't work on the Opsis | 13:45 |
| mithro | seaLne: target for this weekend is to get that working again | 13:45 |
| seaLne | i tried changing mode but before but i got https://paste.kde.org/prfsecjag | 13:45 |
| tpb | Title: KDE Paste (at paste.kde.org) | 13:45 |
| mithro | seaLne: hrm, I don't think I've actually tested modeswitch with an Opsis and webcam mode | 13:46 |
| seaLne | modeswitch doesn't seem to have a video mode option? | 13:47 |
| mithro | Got distracted by the 4k DisplayPort stuff last weekend | 13:49 |
| seaLne | heh | 13:49 |
| mithro | seaLne: actually come to think about it with the current fx2 firmware it won't be able to tell your board is an Opsis once the webcam is loaded | 13:51 |
| seaLne | dmesg was saying uvcvideo: Found UVC 1.00 device HDMI2USB (1d50:60b7) | 13:51 |
| mithro | Yeah | 13:51 |
| mithro | seaLne: did you turn on the encoder when you were in serial mode? | 13:52 |
| seaLne | yes | 13:52 |
| seaLne | encoder on; video_matrix connect pattern encoder | 13:53 |
| mithro | Okay, and no output on the video? | 13:53 |
| seaLne | nope | 13:53 |
| seaLne | well vlc doesn't show anything | 13:54 |
| mithro | Be back tomorrow | 13:54 |
| seaLne | np | 13:55 |
| *** se6astian is now known as se6astian|away | 15:00 | |
| G33KatWork | hm, i'm routing the third time now and it always takes hours and ends with missed timing constraints and unrouted signals | 15:51 |
| _florent_ | G33KatWork: you are routing for atlys or opsis? | 15:59 |
| G33KatWork | opsis | 15:59 |
| G33KatWork | i have access to an atlys as well, but i figured i'd try the opsis first | 16:00 |
| _florent_ | it's the JPEG encoder that is causing this, if you want you can try a design without it by replacing "default_subtarget = HDMI2USBSoC" with "default_subtarget = VideomixerSoC" | 16:03 |
| _florent_ | on my computer it does not take more that 15 min to build the design (both HDMI2USBSoC and VideomixerSoC) | 16:04 |
| _florent_ | but it seems some versions of ISE are not able to build it... | 16:04 |
| G33KatWork | weird | 16:04 |
| G33KatWork | i'm using 14.7 here | 16:04 |
| _florent_ | we'll have to do some optimizations | 16:04 |
| _florent_ | here I have 14.6 | 16:05 |
| G33KatWork | anyway, that's not that much of a dealbreaker for me as long as i got something to synthesize correctly. i just need something to get familiar with migen | 16:05 |
| G33KatWork | the readme said 14.7+, so i used that: https://github.com/timvideos/HDMI2USB-misoc-firmware/tree/master/scripts | 16:06 |
| tpb | Title: HDMI2USB-misoc-firmware/scripts at master · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com) | 16:06 |
| *** rohitksingh has quit IRC | 16:06 | |
| _florent_ | yes if you want to try something, just use the VideomixerSoC, you won't have issues | 16:08 |
| *** rohitksingh has joined #timvideos | 16:21 | |
| *** se6astian|away is now known as se6astian | 16:29 | |
| CarlFK | _florent_ https://youtu.be/tP6aLJ8_cKY?t=23m52s | 17:20 |
| CarlFK | that dotted line is coming from: | 17:20 |
| CarlFK | video_matrix connect input0 output0 | 17:21 |
| CarlFK | nothing was connected to input0. which I hear doesn't work anyway. | 17:21 |
| CarlFK | but.. I have seen that line before and wonder if it isn't a side effect of a bigger problem | 17:22 |
| _florent_ | the dotted lines comes from a DDR issue I think | 17:22 |
| CarlFK | DDR issue... that doesn't sound like a desirable feature ;) | 17:23 |
| CarlFK | hmm, I guess I should have done output0 off | 17:23 |
| _florent_ | not a desirable issue indeed! | 17:25 |
| _florent_ | so this talk was recorded using hdmi2usb? | 17:25 |
| CarlFK | yep | 17:32 |
| CarlFK | that vid has 4 different presenters, each with their own random laptop | 17:32 |
| CarlFK | _florent_: tumbleweed and I did PyConZA a few weeks ago. 20 talks and 2 sets of 5 talk lighting talks, for a total of 30 random laptops | 17:50 |
| CarlFK | there were 1-3 problems, but for sure one of them didn't work with the hdmi plugged directly into the projector | 17:52 |
| CarlFK | the others we didn't have enough time to properly debug | 17:53 |
| CarlFK | one of them was my laptop that I used for driving the hdmi2usb - hp with full sized display port. | 17:54 |
| CarlFK | tumbleweed brought a dp - hdmi dongle, first time I had ever used it. didn't work then, but works fine now. (thank you tumbleweed for giving me that) | 17:55 |
| _florent_ | OK thanks for the feedback! | 18:02 |
| CarlFK | there is still Jay's chromebook that has had problems in the past - I needed to do video_mode 9 to reset something. maybe newer builds have fixed it, need to get with him to veirfy | 18:14 |
| *** Bertl is now known as Bertl_zZ | 18:27 | |
| *** rohitksingh has quit IRC | 19:03 | |
| *** se6astian is now known as se6astian|away | 20:29 | |
| xfxf | talking about the above, _florent_ did you need any more info from us to debug the input0 not working issue? | 21:20 |
| xfxf | i haven't ever tried driving an output from the micro hdmi, haven't needed to, but can certainly that if you think it'll make a difference | 21:21 |
| xfxf | i'm keen to start recording using two sources + switching using the HDMI2USB console | 21:21 |
| _florent_ | last time I tested, input0 was working | 21:22 |
| _florent_ | I'm going to check my jumper configuration with what is in the README | 21:23 |
| xfxf | rad, that'd be great | 21:23 |
| xfxf | also, i'm looking at what 'video_matrix help' says in terms of jumupers | 21:24 |
| _florent_ | ok | 21:24 |
| xfxf | but i did try alternative jumper configurations by alternating what it said, still didn't work | 21:24 |
| _florent_ | ok JP2 | 21:25 |
| _florent_ | are you jumper horizontal or vertical? | 21:26 |
| _florent_ | your | 21:26 |
| xfxf | i don't have the board near me now apologies | 21:26 |
| _florent_ | jumper(s)... | 21:26 |
| _florent_ | ah | 21:26 |
| xfxf | if input0 works for you could you take a pic of yours? | 21:26 |
| xfxf | i can match the same jumper config on it | 21:26 |
| xfxf | actually i'm not going to have the board near me now for a few days | 21:27 |
| xfxf | CarlFK: can you help _florent_ with the above? | 21:27 |
| CarlFK | sure - and yes, a picture of what works wold be great | 21:28 |
| _florent_ | http://www.enjoy-digital.fr/open_hardware/hdmi2usb/atlys_input0_config.jpg | 21:33 |
| CarlFK | well, that's different. /me moves jumpers around | 21:41 |
| CarlFK | no help. power cycled, reloaded, status: input0: 0x0 | 21:43 |
| CarlFK | input0: 1280x720 | 21:45 |
| CarlFK | woot | 21:45 |
| xfxf | rad, thanks _florent_ | 21:55 |
| xfxf | CarlFK: what's wrong with 0x0 video? it compresses well? :P | 21:56 |
| CarlFK | lol | 21:56 |
| CarlFK | any clue where I should put that jpg? | 22:02 |
| CarlFK | https://github.com/timvideos/HDMI2USB/wiki/Cabling good enough | 22:11 |
| tpb | <http://ln-s.net/-DPL> (at github.com) | 22:11 |
| *** CarlFK has quit IRC | 23:06 | |
| *** scrollback has quit IRC | 23:36 | |
| *** G33KatWork has quit IRC | 23:57 | |
| *** derpeter has quit IRC | 23:57 | |
| *** thaytan has quit IRC | 23:58 | |
| *** miselin has quit IRC | 23:58 | |
| *** ysionneau has quit IRC | 23:58 | |
| *** [d__d] has quit IRC | 23:58 | |
| *** _florent_ has quit IRC | 23:58 | |
| *** mithro has quit IRC | 23:58 | |
| *** wanig_ has quit IRC | 23:58 | |
| *** tumbleweed has quit IRC | 23:58 | |
| *** techman83 has quit IRC | 23:58 | |
| *** xfxf has quit IRC | 23:58 | |
| *** Sewar has quit IRC | 23:58 | |
| *** aps has quit IRC | 23:58 | |
| *** hyades_zombie has quit IRC | 23:58 | |
| *** se6astian|away has quit IRC | 23:58 | |
| *** micolous has quit IRC | 23:58 | |
| *** shenki has quit IRC | 23:58 | |
| *** Niharika has quit IRC | 23:58 | |
| *** CARAM__ has quit IRC | 23:58 | |
| *** cyrozap has quit IRC | 23:58 | |
| *** hozer has quit IRC | 23:58 | |
| *** Bertl_zZ has quit IRC | 23:58 | |
| *** automatical has quit IRC | 23:58 | |
| *** xfxf1 has quit IRC | 23:58 | |
| *** seaLne has quit IRC | 23:58 | |
| *** kaalia has quit IRC | 23:58 | |
| *** Kripton_ has quit IRC | 23:58 | |
| *** Joelw has quit IRC | 23:58 | |
| *** springermac has quit IRC | 23:58 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!
