Monday, 2017-01-09

*** tpb has joined #timvideos00:00
jeamithro: I tried to merge the edid improvements, but it didn't keep shenki's commit in the history (it does have it as a parent though). I'm not sure of a way to get his commit in the history without him commiting it in that location himself00:57
jeahttps://github.com/joeladdison/HDMI2USB-litex-firmware/commit/685690907d08691a9616f16a8179772346236a4d00:57
tpbTitle: Add monitor range descriptor to EDID · joeladdison/[email protected] · GitHub (at github.com)00:57
*** CarlFK has quit IRC01:07
mithrojea: I can show you how in a minute01:10
mithrojea: While I finish doing other stuff, do you want to load the edid fixes onto an opsis and confirm you see them on a computer connected to it?01:11
jeamithro: ok, I'll give that a go01:11
*** puck`` has quit IRC01:19
*** ivodd_ has quit IRC01:19
*** shenki has quit IRC01:19
*** shenki has joined #timvideos01:20
*** ChanServ sets mode: +v shenki01:20
*** puck`` has joined #timvideos01:20
*** ivodd has joined #timvideos01:20
xfxfmithro: can I get your input on this pls?  https://github.com/xfxf/lca2017-av/issues/602:06
tpbTitle: Opsis + recording laptop monitoring · Issue #6 · xfxf/lca2017-av · GitHub (at github.com)02:06
*** NP95 has joined #timvideos02:50
NP95Hi02:50
*** Sewar has quit IRC02:51
NP95Who here is looking into the USB sound02:51
*** Sewar has joined #timvideos02:53
*** tims-extra-lapto has quit IRC03:58
mithro_florent_: https://github.com/mithro/HDMI2USB-litex-firmware/tree/nextgen05:11
tpbTitle: GitHub - mithro/HDMI2USB-litex-firmware: Opsis SoC based on LiteX (at github.com)05:11
*** CarlFK has joined #timvideos05:28
*** ChanServ sets mode: +v CarlFK05:28
*** videoteam-r3grab has joined #timvideos05:34
videoteam-r3grabCarlFK, Make sure the following are installed - xchat-gnome, bash-completion, vim-gnome, mplayer, vlc, ffmpeg/avconv, fxload, tmux, git05:37
mithroxfxf: Do you have one of these -> https://www.noelleeming.co.nz/shop/computers-tablets/accessories/cables-chargers-and-adapters/moshi-99mo084202-usb-c-to-hdmi-adpator-silver/prod147865.html06:02
tpbTitle: Moshi USB-C to HDMI Adpator Silver - Noel Leeming (at www.noelleeming.co.nz)06:02
*** CarlFK has quit IRC06:05
*** videoteam-r3grab has quit IRC06:23
*** videoteam-r3grab has joined #timvideos06:23
mithropaddatrapper: ping?06:29
mithrojea: https://github.com/timvideos/HDMI2USB-misoc-firmware/issues/21106:37
tpbTitle: Capture is over bright / over saturated · Issue #211 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)06:37
mithrojea: and the pull request it links to https://github.com/timvideos/HDMI2USB-misoc-firmware/pull/20506:38
tpbTitle: Adjusts gamma value provided in in EDID by xfxf · Pull Request #205 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)06:38
jeacurrent firmware is using that gamma value, and comparing it to previous it is certainly much better. on my mac, the new value (0x78) is still brighter than the Generic RGP profile, but it is probably fine06:44
xfxfi think the value should be lower then 0x7806:45
xfxfit's better than what it was but still looks too bright to me06:45
xfxfyou can even see it in the PR above06:45
xfxf0x78 is what I think I nicked off a few other monitors EDID's06:45
mithroxfxf: Where did you disappear too?06:47
xfxfupstairs06:47
paddatrappermithro: pong06:48
mithropaddatrapper: Where did you get up to?06:48
paddatrappermithro: Pretty much nowhere unfortunately... Realised my knowledge of USB/FX2 is nowhere near up to scratch, so busy trying to improve it at the moment06:49
mithropaddatrapper: The first thing I would look into is just generating a valid USB descriptor in the right format06:50
paddatrappermithro: Yeah, that's what I'm trying06:57
mithropaddatrapper: Do you have any USB Audio devices?06:59
paddatrappermithro: Yup, Scarlett Focusrite07:00
mithroStart decoding that? :-P07:00
tumbleweedo/ paddatrapper07:00
mithroI look at the Logitech webcams for the USB stuff07:00
mithroActually, they probably do USB audio too...07:00
xfxfthey do07:01
paddatrappertumbleweed: \o07:01
paddatrappermithro: Good idea. I shall give it a bash07:01
paddatrapperLogitech webcams should also be a nice one to try. May actually have one lying around somewhere...07:02
mithropaddatrapper: Yeah07:24
mithroshenki: https://keithp.com/blogs/embedded-arm-libc/07:28
tpbTitle: embedded-arm-libc (at keithp.com)07:28
*** CarlFK has joined #timvideos07:32
*** ChanServ sets mode: +v CarlFK07:32
mithrojea: So your working on the brightness fix?08:14
jeamithro: yes. just about to submit a pull request for both edid things08:15
mithrojea: The next thing after that is to look into advertising both 720p60 and 720p50 resolutions as supported08:15
jeamithro: ok. i presume that is edid as well? and if so, would you like one pull request with all of this in it, or separate ones?08:15
mithrojea: With the issue with the brightness we were having last time, I think that should be a seperate commit08:16
mithrojea: Oh, I was going to show you how to merge Joel's repo08:16
jeayes you were08:16
mithrojea: The basic steps are08:17
mithro(a) Checkout Joel's repo08:17
mithro(b) Move everything around in Joel's repo to be where you want them to end up (IE move edid.c to firmware/edid.c, edid-decode to third_party/edid-decode, etc)08:18
mithro(c) Commit that and push to your fork08:18
mithro(d) Do a "git remote add edid-hacks git+ssh://github.com/jae/edid-decode.git" and then a "git fetch edid-hacks"08:18
mithro(e) Then just do a "git merge edid-hacks/master --allow-unrelated-histories"08:19
mithro(f) Fix the conflicts, probably by doing a "git checkout edid-hacks/master <filename>"08:19
mithrojea: Does that make sense?08:20
jeamithro: yes, that makes sense. I'll give it a go08:20
mithroIt would be good to have a "test" which lets you check the edid.c is generating valid EDID data as edid-decode thinks...08:21
mithrojea: But I'm not sure what the best way to do that is08:22
mithro_florent_: Any chance reproducing the encoder issue?08:23
jeamithro: yes, that would be a good idea. I'll see if there is a way to set one up08:23
shenkimithro: interesting post. i hope he upstreams it08:52
mithroshenki: jea is currently working on the edid changes you did at PyCon AU08:53
shenkijea: nice!08:53
jeashenki: your changes are all working nicely! needed a slight fix, but was easy to work out from your tests. thanks for that :)08:55
mithropaddatrapper: Do you want to reply to the mailing list with what you are working and details about things like the fx2lib08:56
paddatrappermithro: Yup, busy doing it09:06
shenkijea: good to hear!09:46
_florent_mithro: the encoder issue should be resolved... doing a P&R now...10:17
mithro_florent_: Awesome10:17
mithro_florent_: Did you end up making a target which is just the encoder without the HDMI stuff?10:17
_florent_mithro: yes kind of10:18
_florent_mithro: it still have a video output10:18
mithro_florent_: Just the single one?10:18
_florent_mithro: just because removing it was confusing ISE timing constraints on usb10:18
_florent_mithro: don't know why10:18
_florent_mithro: but there is no reason we could not be able to get a target with the encoder10:19
_florent_mithro: but do we really need it?10:19
mithro_florent_: It depends on if it changes the compile speed - I can imagine if someone is working on the encoder by itself it could be useful10:20
_florent_mithro: yes, I'll see it I can have that easily10:21
*** NP95 has quit IRC10:27
_florent_mithro: can you describe what you want for the fifo under/over flow10:32
_florent_?10:32
_florent_mithro: do you have others things you wants me to do.10:33
_florent_?10:33
mithro_florent_: Just a CSR which we can poll for to see if an underflow on the TX fifos or an overflow in the RX fifos occurred10:33
_florent_mithro: not sure this is as simple since fifo are using flow control, so could not overflow/underflow10:34
_florent_mithro: the overflow/underflow is defined around the fifo according to the usecase10:34
_florent_mithro: I'll see if I can add things10:34
mithro_florent_: For the HDMI TX, we basically want to know if the output wanted pixel data and it wasn't available10:35
_florent_mithro: ok10:35
mithro_florent_: And on the HDMI input, if it wanted to "store" a pixel, but there wasn't any space, so had to drop it10:36
_florent_mithro: where is the pixel clock counter code from pixel cr1901_modern?10:36
_florent_mithro: I think the overflow is already reported to software on HDMI input10:36
mithro_florent_: Yeah, I think you are right - but I have no idea how to prove that?10:37
mithro_florent_: The pull request which added the frequency counter was at https://github.com/timvideos/HDMI2USB-misoc-firmware/pull/28310:38
tpbTitle: Frequency Counter for Debugging by cr1901 · Pull Request #283 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)10:38
mithro_florent_: gateware/freq_count.py in the old misoc firmware10:38
_florent_mithro: ok thanks10:39
mithro_florent_: It's based off Hamster's design at http://hamsterworks.co.nz/mediawiki/index.php/High_Speed_Frequency_Counter10:39
tpbTitle: High Speed Frequency Counter - Hamsterworks Wiki! (at hamsterworks.co.nz)10:39
mithro"So here is a design that will count over 500 MHz on a fast grade of Spartan 6 FPGA."10:40
mithroIt might be a little overkill, but it 1080p60 is 150MHz pixel clock10:40
mithro_florent_: I haven't started on the I2C stuff at all10:47
mithro_florent_: We want to be able to read the contents out of the I2C eeprom on the Opsis board so we can use the correct MAC address10:47
mithro_florent_: But I think it might make more sense for you to work on seeing if you can figure out what is going on when all inputs+outputs are enable at the same time10:48
mithro_florent_: Another "nice to have" would be to calculate a CRC32 of each of the incoming pixel data in video frames on the HDMI input10:49
_florent_mithro: as we already discussed, I don't think the CRC32 feature will have lots of value10:50
_florent_mithro: we already know there is a problem, it will not allow us to know more10:50
mithro_florent_: Why? I want to be able to detect repeated frames which are of the same contents - we need that for pulldown detection on the HV30 deinterlacing10:51
_florent_mithro: ah ok I thought it was for debug purpose10:52
mithro_florent_: Well, it could also be used for debug purposes, send a known frame and check the crc value is correct10:53
_florent_mithro: also did you reintroduce the spi flash on the opsis?10:53
mithro(known frame from a computer)10:53
mithro_florent_: Hrm, I think I did10:53
mithro_florent_: Yeah it looks like I did -> https://github.com/mithro/HDMI2USB-litex-firmware/blob/nextgen/targets/opsis/base.py#L25210:54
tpbTitle: HDMI2USB-litex-firmware/base.py at nextgen · mithro/HDMI2USB-litex-firmware · GitHub (at github.com)10:54
mithroI haven't actually tested it at all though10:54
mithro_florent_: See https://en.wikipedia.org/wiki/Telecine#Frame_rate_differences for info on the different types of pull-down which could be used - the CRC is to detect the repeated fields uses in such a pull down pattern10:55
tpbTitle: Telecine - Wikipedia (at en.wikipedia.org)10:55
mithro_florent_: Doing deinterlacing for telecine signals with pulldown is *much* simpler than normal deinterlacing too, you basically just do line from field A, line from field B after detecting the pattern10:57
mithro_florent_: We were thinking of storing config info in SPI flash at some point too10:58
jeaone of these would be nice https://www.cnet.com/products/dell-ultrasharp-up3218k/preview/11:36
tpbTitle: Dell Ultrasharp UP3218K could be the world's first 8K monitor, and it has a price to match - CNET (at www.cnet.com)11:36
*** NP95 has joined #timvideos11:55
*** videoteam-r3grab has quit IRC12:23
*** videoteam-r3grab has joined #timvideos12:23
*** videoteam-r3grab has quit IRC12:26
*** videoteam-r3grab has joined #timvideos12:27
*** CarlFK has quit IRC12:28
*** CarlFK has joined #timvideos13:09
*** ChanServ sets mode: +v CarlFK13:09
*** NP95 has quit IRC16:16
*** NP95 has joined #timvideos18:13
*** videoteam-r3grab has quit IRC18:25
*** sb0 has joined #timvideos19:08
mithromyself and tumbleweed are up now21:15
tumbleweedand decrying the lack of coffee21:17
cr1901_modernWhat's coffee?21:17
tumbleweedCarlFK: http://www.dx.com/s/hdmi+extender21:54
tpbTitle: hdmi extender - Free Shipping - DX (at www.dx.com)21:54
mithro_florent_: I remember seeing that the ISE tools can use an existing design to speed up routing a new design?23:19
mithro_florent_: Have you tried using that at all?23:19

Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!