| Wednesday, December 14th, 2011, 00:03 UTC | ||
| [00:03:43] | zombor (zombor!~zombor_@kohana/developer/zombor) has joined #mythtv | |
| [00:08:29] | jarle (jarle!~jarle@70.84-234-133.customer.lyse.net) has quit (Ping timeout: 252 seconds) | |
| [00:10:39] | andreax (andreax!~andreaz@p5089F01F.dip.t-dialin.net) has quit (Read error: Connection reset by peer) | |
| [00:11:41] | zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Remote host closed the connection) | |
| [00:21:13] | jarle (jarle!~jarle@70.84-234-133.customer.lyse.net) has joined #mythtv | |
| [00:36:28] | stf (stf!~stf@dslb-178-009-240-254.pools.arcor-ip.net) has quit (Quit: Verlassend) | |
| [01:08:56] | Jordack (Jordack!~jordack@h69-131-44-221.plmomi.dedicated.static.tds.net) has quit () | |
| [01:09:07] | Mousey (Mousey!~wtfisme@ross154.net) has quit (Ping timeout: 248 seconds) | |
| [01:09:19] | saintdev (saintdev!~saint@unaffiliated/saintdev) has quit (Remote host closed the connection) | |
| [01:11:40] | pheld (pheld!~heldal@cl-5.osl-01.no.sixxs.net) has quit (Quit: Leaving.) | |
| [01:48:02] | natanojl (natanojl!~jonatan@c83-252-237-63.bredband.comhem.se) has quit (Ping timeout: 252 seconds) | |
| [01:50:14] | tstaerk (tstaerk!~quassel@p54A37D92.dip.t-dialin.net) has joined #mythtv | |
| [01:50:27] | tstaerk_ (tstaerk_!~quassel@p54A37CB4.dip.t-dialin.net) has quit (Ping timeout: 240 seconds) | |
| [02:05:53] | davide (davide!~david@host70.16.intrusion.com) has quit (Remote host closed the connection) | |
| [02:06:17] | davide (davide!~david@host70.16.intrusion.com) has joined #mythtv | |
| [02:08:21] | jpabq (jpabq!~jpabq@mythtv/developer/jpabq) has quit (Quit: jpabq) | |
| [02:13:18] | stichnot (stichnot!~chatzilla@192.55.54.40) has joined #mythtv | |
| [02:45:43] | stichnot (stichnot!~chatzilla@192.55.54.40) has quit (Remote host closed the connection) | |
| [02:49:29] | stichnot (stichnot!~chatzilla@192.55.54.42) has joined #mythtv | |
| [03:08:22] | stichnot (stichnot!~chatzilla@192.55.54.42) has quit (Ping timeout: 268 seconds) | |
| [04:01:56] | MythBuild: | build #374 of master-vista-mingw-32bit is complete: Exception [exception setproperty] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . t/builds/374 blamelist: Michael T. Dean <mdean@mythtv.org > |
| [04:14:20] | sphery: | MythBuild: force build master-vista-mingw-32bit now |
| [04:14:21] | MythBuild: | build forced [ETA 1h36m32s] |
| [04:14:21] | MythBuild: | I'll give a shout when the build finishes |
| [04:14:28] | MythBuild: | build #375 of master-vista-mingw-32bit is complete: Exception [exception setproperty] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . t/builds/375 |
| [04:23:36] | LTHorn (LTHorn!~Android@19.sub-174-253-19.myvzw.com) has quit (Ping timeout: 240 seconds) | |
| [04:24:22] | Mkaysi (Mkaysi!Mkaysi@TouchLay/Member/Mkaysi) has quit (Excess Flood) | |
| [04:25:29] | Mkaysi (Mkaysi!Mkaysi@TouchLay/Member/Mkaysi) has joined #mythtv | |
| [04:25:30] | Mkaysi (Mkaysi!Mkaysi@TouchLay/Member/Mkaysi) has quit (Excess Flood) | |
| [04:34:29] | Mkaysi (Mkaysi!Mkaysi@TouchLay/Member/Mkaysi) has joined #mythtv | |
| [04:40:03] | Anduin_ (Anduin_!~awithers@c-71-204-162-27.hsd1.ca.comcast.net) has quit (Remote host closed the connection) | |
| [04:42:20] | sphery: | MythBuild: force build master-vista-mingw-32bit now |
| [04:42:21] | MythBuild: | build forced [ETA 1h36m32s] |
| [04:42:21] | MythBuild: | I'll give a shout when the build finishes |
| [04:44:19] | sphery: | seems much happier this time--after I talked to MythBuild in a PM and had it reboot the machine :) |
| [04:55:09] | wagnerrp: | would that be MythBeirdo? |
| [04:57:07] | sphery: | hehe, yeah |
| [04:58:15] | Beirdo: | hehe. |
| [05:11:49] | danielk22 (danielk22!~danielk@96.57.9.142) has quit (Ping timeout: 248 seconds) | |
| [05:26:15] | [intra]lanman ([intra]lanman!~lanman@freeswitch/developer/intralanman) has left #mythtv ("screw you guys, i'm going home") | |
| [05:38:11] | JoeyJoeJo (JoeyJoeJo!~brian@pool-173-71-223-79.clppva.fios.verizon.net) has joined #mythtv | |
| [05:38:35] | JoeyJoeJo: | Are there any capture cards that have a built-in h.264 encoder? |
| [05:39:21] | wagnerrp: | cards, no... capture devices, yes |
| [05:39:25] | wagnerrp: | Hauppauge HDPVR |
| [05:39:40] | JoeyJoeJo: | I'll look that up, thanks! |
| [05:39:47] | wagnerrp: | note, this is the development channel |
| [05:39:54] | wagnerrp: | you want #mythtv-users |
| [05:40:24] | JoeyJoeJo: | Crap, I always forget about that |
| [05:40:38] | wagnerrp: | please continue there because i do not believe you actually want an HDPVR |
| [06:01:44] | jstenback (jstenback!~jstenback@2620:101:8003:200:224:e8ff:fe39:34c2) has quit (Ping timeout: 252 seconds) | |
| [06:06:54] | jstenback (jstenback!~jstenback@2620:101:8003:200:224:e8ff:fe39:34c2) has joined #mythtv | |
| [06:09:12] | MythBuild: | Hey! build master-vista-mingw-32bit #376 is complete: Success [build successful] |
| [06:09:12] | MythBuild: | Build details are at http://code.mythtv.org/buildbot/builders/mast . . . t/builds/376 |
| [06:24:21] | jmartens (jmartens!~jmartens@s5597ca60.adsl.wanadoo.nl) has joined #mythtv | |
| [06:35:09] | kenni: | Isn't the post-commit hook setup for the mythweb branch? It seems like it didn't catch https://github.com/MythTV/mythweb/commit/4811657f8 |
| [06:35:55] | wagnerrp: | no, trac only supports a single repository |
| [06:36:16] | kenni: | ok, I'll just close it manually then :) |
| [06:36:44] | wagnerrp: | anything in the main repository should work |
| [06:36:48] | wagnerrp: | anything outside, not so much |
| [06:40:48] | jmartens (jmartens!~jmartens@s5597ca60.adsl.wanadoo.nl) has quit (Quit: Leaving.) | |
| [06:42:49] | Goga777 (Goga777!~Goga777@128.71.232.177) has joined #mythtv | |
| [06:56:02] | jmartens (jmartens!~jmartens@109.232.42.33) has joined #mythtv | |
| [07:25:58] | kormoc (kormoc!~kormoc@mythtv/developer/kormoc) has quit (Quit: kormoc) | |
| [07:52:40] | willcooke (willcooke!~will@canonical/willcooke) has joined #mythtv | |
| [08:02:25] | Goga777 (Goga777!~Goga777@128.71.232.177) has quit (Remote host closed the connection) | |
| [08:18:06] | Beirdo: | stuartm: in case you were wondering, I've gotten the BBC weather stuff half-rewritten |
| [08:19:02] | Beirdo: | the search is working again, now I have to get the actual data pull working. They completely reworked the ids. The RSS feeds are only there for the old IDs it seems, and the mapping... is buried in the HTML |
| [08:19:15] | Beirdo: | or I could scrape it from the HTML easily enough |
| [08:21:56] | Beirdo: | Users of existing features like our embedded modules and RSS feeds from the previous site will still work and we will be reinstating the ability to pick them up from forecast pages very soon. |
| [08:22:31] | Beirdo: | hmm, well, I guess I'll hold off for "very soon" if we want to keep using the RSS |
| [08:47:29] | saintdev (saintdev!~saint@c-24-9-201-236.hsd1.co.comcast.net) has joined #mythtv | |
| [08:47:30] | saintdev (saintdev!~saint@unaffiliated/saintdev) has joined #mythtv | |
| [08:47:30] | saintdev (saintdev!~saint@c-24-9-201-236.hsd1.co.comcast.net) has quit (Changing host) | |
| [08:50:12] | pheld (pheld!~heldal@cl-5.osl-01.no.sixxs.net) has joined #mythtv | |
| [08:55:39] | stuarta: | i hate the new bbc weather site |
| [08:57:45] | Beirdo: | yeah, it's sucking |
| [09:04:44] | stuarta: | clearly designed by the teletubby department |
| [09:09:08] | Beirdo: | OK, almost done doing the temporary fix |
| [09:09:31] | markk: | clearly designed by the 'oh my god, must justify job and reinvent webiste before I get downsized' department |
| [09:10:11] | stuarta: | utterly uttery horrible |
| [09:18:01] | Beirdo: | there we go |
| [09:18:38] | Beirdo: | I think the yrnoxml one is potentially hosed too, but I'll look at that another day |
| [09:21:44] | stuarta: | these things look interesting -> http://www.theregister.co.uk/2011/11/28/raspberry_pi/ |
| [09:22:07] | stuarta: | 20 quid and it'll do 1080p they reckon |
| [09:25:34] | stuarta: | plus nokia are making Qt work on it |
| [09:27:43] | Beirdo: | might be a fun toy if nothing else |
| [09:28:11] | stuarta: | when a co-worker mentioned it i had hopes for finally finding a cheap frontend |
| [09:28:27] | stuarta: | suspect it'll be a few years before that's a reality |
| [09:30:28] | Beirdo: | I'm looking forward to scoring a Tegra3 + GPU development board from nvidia |
| [09:30:39] | Beirdo: | they should be available in June or so |
| [09:30:55] | Beirdo: | as in a CUDA-compatible external GPU :) |
| [09:31:11] | stuarta: | shiny |
| [09:31:46] | markk: | stuarta: 256Mb ram is the biggest problem. and I bet the video decode is unsupported. |
| [09:31:48] | Beirdo: | the product manager for the CUDA group was estimating a price of $350US or so, but I bet it will be a bit more |
| [09:32:13] | Beirdo: | but, it's 6 months out regardless, and it would be a fun toy if nothing else |
| [09:33:38] | stuarta: | markk: from what i'm reading the video driver is pure binary, "The GPU is accessed exclusively through an OpenGL ES driver, which talks to the GPU via a sockets-like interface in the kernel. There's not much chance of directly poking registers in the hardware, I'm afraid." |
| [09:34:35] | markk: | but that's graphics (i.e. opengl) – not video decoding. |
| [09:34:51] | markk: | and without some accelerated decoding it's next to useless. |
| [09:34:51] | stuarta: | who knows what they will make available |
| [09:35:10] | stuarta: | http://www.raspberrypi.org/ <- that's the main site |
| [09:36:34] | Beirdo: | only one way to know for sure, but it does sound half useless. :) |
| [09:37:26] | Beirdo: | now, if they put mini-pcie on there, maybe you could put on a CrystalHD |
| [09:39:15] | Goga777 (Goga777!~Goga777@128.71.232.177) has joined #mythtv | |
| [09:39:53] | Beirdo: | If youâre a Qt developer who wants to get their hands on a free board (Iâm looking at you, team Tampere), visit Nokiaâs Qt blog to find out how to apply. |
| [09:40:02] | Beirdo: | bloody UTF-8 paste |
| [09:46:33] | Beirdo: | well, I think I should go to bed. Night all. |
| [09:46:42] | stuarta: | nn |
| [09:50:25] | rsiebert_ (rsiebert_!~quassel@g226061155.adsl.alicedsl.de) has joined #mythtv | |
| [09:50:47] | rsiebert (rsiebert!~quassel@g226063001.adsl.alicedsl.de) has quit (Ping timeout: 240 seconds) | |
| [09:52:29] | Goga777 (Goga777!~Goga777@128.71.232.177) has quit (Remote host closed the connection) | |
| [10:21:42] | GreyFoxx (GreyFoxx!~greg@2607:ae00:fff9::1) has quit (Ping timeout: 252 seconds) | |
| [10:21:49] | GreyFoxx (GreyFoxx!~greg@2607:ae00:fff9::1) has joined #mythtv | |
| [10:54:58] | stuartm: | Beirdo: ok, thanks for that, it's a real shame they've made it harder to use |
| [10:57:18] | stuartm: | Beirdo, stuarta: I've seen no details at all on how it's actually achieved, but it's apparently capable of full H.264 1080p decode – I assume that means hardware accel |
| [10:58:31] | stuartm: | oh right, that's already been said – I should stop reading conversations in reverse |
| [11:01:56] | stuartm: | markk: the article does say that video was demoed and since the whole point of it is to encourage kids to program etc there must be some video decode API accessible |
| [11:02:35] | stuarta: | the GPU is fully capable of OpenMAX and OpenVG and OpenGL-ES |
| [11:02:55] | stuarta: | how much of that is exposed via an API is yet to be determined |
| [11:03:08] | stuarta: | but they have demo'd it doing 1080p playback |
| [11:04:16] | stuarta: | and the XBMC guys are working on getting it working |
| [11:05:02] | mike|2 (mike|2!~mike@c-76-115-119-121.hsd1.or.comcast.net) has quit (Remote host closed the connection) | |
| [11:05:50] | mike (mike!~mike@c-76-115-119-121.hsd1.or.comcast.net) has joined #mythtv | |
| [11:06:16] | mike is now known as Guest70048 | |
| [11:07:50] | stuartm: | right, IMHO they wouldn't tease with the video demo if that wasn't something people could actually use – not when it's being produced purely as a platform to encourage hobbyist programmers |
| [11:19:01] | stuarta: | for now i'm filing it under "wait and see" |
| [11:35:43] | natanojl (natanojl!~jonatan@c83-252-237-63.bredband.comhem.se) has joined #mythtv | |
| [11:54:29] | markk: | gpu video decode is OpenMax based – i.e. not a trivial addition |
| [11:56:38] | stuarta: | does anything else currently use OpenMax, this is the first i've heard of it |
| [11:59:19] | stuartm: | I assume that someone will be releasing an implementation we can work from |
| [11:59:57] | stuartm: | maybe that's one assumption too many |
| [12:00:04] | ** stuarta chuckles ** | |
| [12:00:30] | stuarta: | maybe that's what the xbmc guys are good at |
| [12:05:56] | stuartm: | the Arm OpenMAX DL libs provide implementations of common video decoding steps, so what comes next is someone writing a H.264 decoder on top of that – which obviously has been done already, hence the demo but whether they'll GPL that ... |
| [12:10:01] | stuarta: | guess that's what they mean by the openmax api is a pita to work with |
| [12:11:03] | stuartm: | yeah, it's very low-level, it's not like vdpau where you feed in bytes and get a video frame back |
| [12:11:46] | stuartm: | well, in vdpau you don't even have to deal with the frame, it's a closed pipeline |
| [12:12:52] | stuartm: | even the DL libs are a layer above the OpenMAX api |
| [12:16:46] | stuarta: | three different levels of OpenMax according to wikipedia |
| [12:17:30] | stuarta: | DL, IL and AL |
| [12:17:48] | stuarta: | Development, Integration & Application Layer |
| [12:18:36] | stuarta: | the standards are at www.khronos.org for those who are interested |
| [12:18:45] | stuartm: | hmm, now I know what I'll be reading tonight |
| [12:19:13] | stuartm: | if I can stay awake ... |
| [12:19:28] | stuarta: | found quick reference cards for at least the AL aip |
| [12:19:29] | stuarta: | api |
| [12:24:15] | markk: | iirc you need IL |
| [12:24:30] | stuarta: | yeah, i think so too. AL looks far to high level |
| [12:26:36] | Anssi: | dunno if you are talking about something else, but xbmc has openmax support since May 2010 AFAICS (I don't know anything about openmax either) |
| [12:28:01] | stuartm: | it might be an interesting challenge to get the frontend memory consumption down enough to run it with just 256MB of memory, for a start it would involve a specially crafted theme and less memory caching of theme images |
| [12:28:38] | stuartm: | and you have to share that 256MB with the gpu |
| [12:30:22] | stuarta: | time to get lean and mean |
| [12:31:55] | markk: | stuartm: 256 is probably doable. I started a theme that just used rounded rects (which are accelerated for OpenGL 2.0/ES) – the main memory load then comes from text (for which we could reduce the cache size) |
| [12:34:48] | stuartm: | there are probably other tweaks, stuff that I'm just not able to think of right now – I'd kill for a decent memory profiler |
| [12:36:00] | stuarta: | spose we can get some of that from valgrind |
| [12:36:54] | stuartm: | probably, I've never used it in that way before so I've no idea what it's capable of |
| [12:37:50] | Peitolm: | is it really that memory intensive? |
| [12:44:28] | stuartm: | a fullscreen 1920x1080 image and text heavy UI will chew through a lot of memory, it's not necessary to cache those things but it would be much, much slower if you didn't |
| [12:47:21] | stuartm: | one day perhaps I'll write a guide on optimising themes to minimise memory usage and speed up loading, there are simple steps you can take to make a small difference |
| [12:48:15] | stuartm: | text uses space because it's faster if we render it just once and cache the resulting image |
| [12:49:40] | stuartm: | some very clever stuff could be written to combine cached images into one where that would make sense, that could have a significant impact but it's a lot of work |
| [12:51:11] | stuartm: | or maybe it's not ... |
| [12:51:24] | ** stuartm scribbles some ideas down ** | |
| [12:51:36] | ** stuarta hands over an empty fag packet ** | |
| [12:52:02] | stuarta: | not that i smoke mind |
| [12:53:51] | stuarta: | i've skimmed the openmax il api and died of boredom |
| [12:54:01] | stuartm: | heh |
| [12:54:06] | stuarta: | must be why graphics is not my thing |
| [12:54:08] | ** Peitolm ponders a pure text UI ** | |
| [12:54:29] | ** stuarta suggests that might be pointless for a video player ** | |
| [12:54:59] | Peitolm: | not really, take out all fancy images, render simple text. |
| [12:55:12] | stuarta: | mythcentre |
| [12:55:26] | stuarta: | that was pretty light |
| [12:55:50] | stuartm: | iirc graphite came close |
| [12:56:58] | Peitolm: | if i could work out how to do graphics playback from the console, (in fact, didn't mplayer have a framebuffer output mode), you could do an MythTV API + Curses frontend |
| [12:57:44] | stuarta: | there's that frontend api interface which i have no idea how to use |
| [12:57:49] | stuarta: | s/api/telnet |
| [12:59:33] | stuartm: | Peitolm: sure, there's even a curses video renderer, but who would really want to turn their brand new 1080p TV into the emulation of a 1970s terminal? |
| [13:00:16] | stuartm: | lots of things are possible, but only a few of them are really desirable to the majority |
| [13:00:19] | ** Peitolm prevents himself putting his hand up ** | |
| [13:01:04] | Peitolm: | just for giggles i'd be tempted to do it, but I agree that most people want pretty |
| [13:01:20] | ** stuarta just wants it to work :) ** | |
| [13:02:01] | stuartm: | people like pretty UIs, it might not make a whole lot of practical sense but it's what sells |
| [13:06:03] | davide (davide!~david@host70.16.intrusion.com) has quit (Remote host closed the connection) | |
| [13:06:24] | davide (davide!~david@host70.16.intrusion.com) has joined #mythtv | |
| [13:21:33] | danielk22 (danielk22!~danielk@96.57.9.142) has joined #mythtv | |
| [14:05:47] | simonckenyon (simonckenyon!~simoncken@195.7.61.12) has joined #mythtv | |
| [14:12:46] | Goga777 (Goga777!~Goga777@128.71.232.177) has joined #mythtv | |
| [14:17:04] | Jordack (Jordack!~jordack@h69-131-44-221.plmomi.dedicated.static.tds.net) has joined #mythtv | |
| [14:19:02] | markk: | he shoots, he scores, he goes to make another cup of coffee. VDPAU accelerated transcode done. |
| [14:20:51] | Captain_Murdoch: | nice. |
| [14:20:56] | Captain_Murdoch: | very nice |
| [14:22:04] | Goga777 (Goga777!~Goga777@128.71.232.177) has quit (Read error: Operation timed out) | |
| [14:22:19] | jeff999 (jeff999!~jeff@203-217-51-4.dyn.iinet.net.au) has joined #mythtv | |
| [14:38:23] | j-rod|afk is now known as j-rod | |
| [14:39:22] | zombor (zombor!~zombor_@kohana/developer/zombor) has joined #mythtv | |
| [14:43:53] | anykey_ (anykey_!~guedel@46-126-245-147.dynamic.hispeed.ch) has quit (Ping timeout: 252 seconds) | |
| [14:44:40] | anykey_ (anykey_!~guedel@46-126-245-147.dynamic.hispeed.ch) has joined #mythtv | |
| [14:51:09] | markk: | live streaming test, 1280x720p -> 480x273, Ion frontend: plain ffmpeg 19fps, cpu 150% |
| [14:51:24] | markk: | using vdpau: 47fps, cpu 260% |
| [14:51:54] | markk: | Captain_Murdoch: is mythtranscode 'niced' in some way? |
| [15:08:59] | saintdev (saintdev!~saint@unaffiliated/saintdev) has joined #mythtv | |
| [15:11:50] | Captain_Murdoch: | Only when run from inside the JobQueue. we don't nice() inside mythtranscode |
| [15:12:35] | iamlindoro: | markk: Damn you, now people can go on pretending an ION is a valid backend choice ;) nice work |
| [15:15:00] | Captain_Murdoch: | the scaling currently occurs in the main thread, so you could be hitting that limit. using the HLS transcoding with libx264, we do use multiple threads to encode. currently there's 1 thread for decoding with a 5-frame deep FIFO. main thread grabs a frame off that and scales and then passes to encoder. libx264 encoding is threaded since it uses slices, but all other encoding is single threaded. encoding blocks the main th |
| [15:15:00] | Captain_Murdoch: | read. |
| [15:17:04] | Anduin (Anduin!~awithers@pdpc/supporter/professional/anduin) has joined #mythtv | |
| [15:17:20] | Captain_Murdoch: | I've debated creating a mythtranscodeplayer that would handle the cutlist skipping, scaling, audio reblocking, etc.. there's also no reason that scaling couldn't be done in the current decode thread or another thread. I've debated that as well but wanted to get the other HLS issues worked out first before concentrating on performance too much more. |
| [15:25:27] | jwhite (jwhite!~jwhite@75-146-153-89-minnesota.hfc.comcastbusiness.net) has joined #mythtv | |
| [15:26:20] | Captain_Murdoch: | markk, to test w/ threaded h264 encoding, try "mythtranscode --video -i ${INFILE} --nodblog --hls --width 480 --bitrate 800 --audiobitrate 96" that will spit out HLS files into ~/.mythtv/tmp/hls/ it doesn't update the DB, so you can delete the files when done. FPS rate should get printed every 20 seconds or so by mythtranscode |
| [15:36:51] | markk: | Captain_Murdoch: so presumably threaded encoding is the default? |
| [15:37:44] | Captain_Murdoch: | for libx264 yes |
| [15:38:57] | markk: | I can't quite see why ffmpeg utilisation is lower (or low) for the ffmpeg vs vdpau comparison above. |
| [15:39:07] | Captain_Murdoch: | 2 threads by default, can be overridden for now with "-O HTTPLiveStreamThreads=4" |
| [15:40:51] | Captain_Murdoch: | ffmpeg is single threaded so you're probably spending a core to decode and half a core to encode |
| [15:41:01] | Captain_Murdoch: | ffmpeg decode that is... |
| [15:41:09] | Captain_Murdoch: | unless it's h264 content w/ slices |
| [15:41:18] | Captain_Murdoch: | so you're encoder is starved. |
| [15:43:35] | Captain_Murdoch: | vdpau decode lets you keep the encoder happy, but since you're encoding to mpeg4, you're single threaded there. that's why I suggested testing with the HLS code since it uses libx264 encoding which will use multiple threads and might get you over the 260% if the encoder is the bottleneck in your current test. |
| [15:45:29] | ** iamlindoro assumed since markk said "live streaming test", he had meant HLS ** | |
| [15:46:19] | markk: | Captain_Murdoch: the comparison was with hls – and ffmpeg was using 2 cores for decode. I suspect the bottleneck for vdpau is actually vdpau (i.e. it still operates at a finite rate, even if there is lower cpu impact) |
| [15:47:15] | Captain_Murdoch: | ah, I thought it was a different test like that video decode test in mythavtest |
| [15:47:30] | Captain_Murdoch: | sorry 'bout that. :) I misread. |
| [15:48:54] | Captain_Murdoch: | yeah, makes sense about vdpau. what's the fastest we can play recordings using VDPAU decode? we should be able to get somewhere near that decode speed not counting the penalty for transferring back to main memory. |
| [15:49:04] | Captain_Murdoch: | fastest FPS... |
| [15:51:20] | Captain_Murdoch: | are you sure it was ffmpeg using 2 cores or was one for ringbuffer, one for decode, one for encode/main thread, etc.? we don't spin up the normal decode threads for mythtranscode. I added a simple 5-frame FIFO decode thread in mythtranscode recently to speedup HLS. |
| [15:54:25] | markk: | Captain_Murdoch: it's 'normal' AvFormatDecoder to try and use multiple-threads if enabled in the display profile – which is what the logs confirmed. The fact that video display profiles are being used to determine transcode behaviour is another topic entirely:) |
| [15:55:12] | Captain_Murdoch: | in my HLS tests, I see 4 threads consuming most of the CPU. 2 should be encode, 1 should be decode thread, and haven't looked into what the 4th is. I'm seeing around 90+% on thread and 10–15% on the other three threads encoding 720p down to 480x272 |
| [15:56:25] | Captain_Murdoch: | yes, we configure AVFormatDecoder to support multiple threads, but I believe the only codec that it can use multiple threads to decode currently is h264 w/ slices. I thought that was the big reason for ffmpeg-mt, to enable threaded decoding of other codecs |
| [15:56:54] | Captain_Murdoch: | I'm talking about video decode, not any other threads it may use such as file reading. |
| [15:57:19] | Captain_Murdoch: | can libav* decode mpeg2 using multiple threads? |
| [15:58:21] | iamlindoro: | I believe versions more recent than our own can |
| [15:58:23] | iamlindoro: | in both forks |
| [15:58:35] | Captain_Murdoch: | because of merging in -mt? |
| [15:58:39] | iamlindoro: | yeah |
| [15:59:19] | Captain_Murdoch: | that's what I thought, that ours was still single threaded. I couldn't recall if one of the forks was merging -mt in yet or not. |
| [15:59:57] | iamlindoro: | They both did semi-shortly after the split |
| [16:00:15] | Captain_Murdoch: | and the race is on... |
| [16:00:26] | iamlindoro: | I can't speak to the efficacy of each implementation, though-- at that point they were pretty much identical, but they've diverged a fair amount since then |
| [16:00:47] | iamlindoro: | And the various codecs implementing mt happened over time-- so there are very likely some differences |
| [16:05:07] | sphery (sphery!~mdean@mythtv/developer/sphery) has quit (Quit: leaving) | |
| [16:05:27] | sphery (sphery!~mdean@mythtv/developer/sphery) has joined #mythtv | |
| [16:09:16] | stuartm: | so we'll soon have to stop sitting on the fence ... |
| [16:14:31] | markk: | Captain_Murdoch: does transcode check the size of each frame as it arrives to determine the scaling required? or does it check once up front? |
| [16:16:34] | Captain_Murdoch: | it gets the size from the decoder at startup |
| [16:17:34] | Captain_Murdoch: | calls MythPlayer::GetVideoBufferSize() and gets the width and height from that. |
| [16:17:58] | Captain_Murdoch: | then calculates the new width/height if necessary |
| [16:18:19] | Captain_Murdoch: | neither mythranscode or mythcommflag handle physical frame size changes. |
| [16:25:05] | markk: | Captain_Murdoch: as an aside, it looks like the handling of 1088/1080 needs some work. |
| [16:25:46] | Captain_Murdoch: | yes, that was noticed in one of iamlindoro's demo videos. I'm not on master so I don't know if that's a recent regression or how long it's been there. |
| [16:27:42] | jmartens (jmartens!~jmartens@109.232.42.33) has quit (Quit: Leaving.) | |
| [16:29:45] | Captain_Murdoch: | made sure there's a note on my TODO to look into it. I looked at it once before (probably late at night or early early morning) but didn't see anything wrong. |
| [16:46:42] | jmartens (jmartens!~jmartens@s5597ca60.adsl.wanadoo.nl) has joined #mythtv | |
| [17:10:52] | markk: | Captain_Murdoch: have you any idea what might cause the very blocky transcoded output? only really noticeable with hd material (source) |
| [17:13:28] | stuartm: | btw, I'm not sure whether it's been fixed yet but following the last of the QT3 stuff being replaced in mythtranscode the transition around cuts in SD mpeg2 was blocky |
| [17:17:49] | kth (kth!~kth@unaffiliated/kth) has joined #mythtv | |
| [17:22:37] | kth (kth!~kth@unaffiliated/kth) has quit (Client Quit) | |
| [17:29:25] | Goga777 (Goga777!~Goga777@128.71.232.177) has joined #mythtv | |
| [17:36:41] | dekarl_zZz is now known as dekarl | |
| [17:37:24] | dekarl: | danielk22, shall I write up a patch? I see more instances of PESPacket* instead of PSIPTable* |
| [17:50:37] | stuarta: | i would if i were you |
| [17:50:47] | stuarta: | tho it is up to danielk22 |
| [17:53:55] | dekarl: | just trying to avoid duplicate work should he be about to commit something similar ;) |
| [18:05:46] | davide (davide!~david@host70.16.intrusion.com) has quit (Remote host closed the connection) | |
| [18:06:10] | davide (davide!~david@host70.16.intrusion.com) has joined #mythtv | |
| [18:09:17] | kth (kth!~kth@unaffiliated/kth) has joined #mythtv | |
| [18:14:26] | Mousey (Mousey!~wtfisme@ross154.net) has joined #mythtv | |
| [18:14:38] | gregL (gregL!~greg@cpe-74-76-125-87.nycap.res.rr.com) has quit (Read error: Connection reset by peer) | |
| [18:18:59] | Goga777 (Goga777!~Goga777@128.71.232.177) has quit (Read error: Connection reset by peer) | |
| [18:26:53] | kj2128 (kj2128!~quassel@c-24-128-83-54.hsd1.ma.comcast.net) has joined #mythtv | |
| [18:27:32] | kj2128: | I'm getting constant harddisk access from the mythtv backend. Not sure why. Any ideas? |
| [18:30:22] | gregL (gregL!~greg@cpe-74-76-125-87.nycap.res.rr.com) has joined #mythtv | |
| [18:33:28] | kevin2128 (kevin2128!~ubuntu@c-24-128-83-54.hsd1.ma.comcast.net) has joined #mythtv | |
| [18:34:00] | kevin2128 (kevin2128!~ubuntu@c-24-128-83-54.hsd1.ma.comcast.net) has quit (Client Quit) | |
| [18:37:58] | dekarl: | kj2128: you likely want #mythtv-users |
| [18:38:32] | kj2128: | dekarl: thanks. I'm over there now |
| [18:38:41] | kormoc (kormoc!~kormoc@mythtv/developer/kormoc) has joined #mythtv | |
| [18:40:49] | Captain_Murdoch: | markk, we're still tuning the libx264 options we're setting via libavcodec. iamlindoro was going to play around with them some I think. I haven't looked at tuning them much, I was going for functionality and bugfixing first. one of those probably affects the blockiness at startup as well. |
| [18:47:31] | Beirdo: | stuartm: I'd need examples, logs, etc to dig into the mpeg2 cut stuff much. I haven't seen it misbehaving in my tests, but things act differently for different source materials sometimes |
| [18:55:58] | willcooke (willcooke!~will@canonical/willcooke) has quit (Quit: We will learn more of his wisdom later) | |
| [19:09:49] | danielk22: | dekarl: I started working on it, but I probably won't have time to finish up until next week. If you want to take a stab at before then go ahead.. I didn't get much further than a search and replace. |
| [19:10:59] | danielk22: | Basically, we should be constructing a PSIPTable not a PESPacket.. |
| [19:11:33] | dekarl: | danielk22: I'm at it atm, looking how far the s/PESPacket/PSIPTable/ has to go |
| [19:11:58] | iamlindoro: | Captain_Murdoch: markk: I have some tweaks to the encoding options which to me look a bit better-- but I haven't tested what if any impact they have on encode speed |
| [19:12:09] | stuartm: | Beirdo: yeah, I keep meaning to produce logs/examples but time isn't on my side atm, I mentioned it in case markk/CapM spotted anything obvious while looking to fix the 1080i/p issues and not to nag you :) |
| [19:13:34] | Beirdo: | no prob :) |
| [19:14:01] | Beirdo: | Once we have something to dig into, we really do wanna get that solid if we can. |
| [19:14:22] | Beirdo: | The more eyes on it, the better, too. |
| [19:21:48] | markk: | Captain_Murdoch, iamlindoro: changing the max_b_frames seems to make a huge difference here (from 0 to 2) – but frame ordering seems to go wrong. |
| [19:25:29] | iamlindoro: | markk: Captain_Murdoch: https://github.com/carsonmcdonald/HTTP-Live-V . . . 5b460fbce45a is a pretty good basis for changing our options-- requires minor interpretation of how variables map to command line options, but this is what I used |
| [19:26:22] | iamlindoro: | (Added the 4MV flag, adjusted H.264 Level from 13 to 30, change quality settings, etc) |
| [19:26:35] | iamlindoro: | I can't find my patch, I seem to have accidentally lost it in a git stash disaster |
| [19:33:00] | kj2128 (kj2128!~quassel@c-24-128-83-54.hsd1.ma.comcast.net) has quit (Remote host closed the connection) | |
| [19:33:10] | stichnot (stichnot!~chatzilla@nat/intel/x-zdtdjdamdlsvsobk) has joined #mythtv | |
| [19:38:04] | iamlindoro: | markk: Apple apparently recommends not using B frames for HTTP Live Streaming |
| [19:38:14] | iamlindoro: | At least when targeting IOS |
| [19:41:12] | iamlindoro: | Ref: http://www.google.com/url?sa=t&rct=j& . . . mlcMOoFhU7MA Page 20 |
| [19:41:26] | iamlindoro: | specifically addresses the out of order nature |
| [19:42:15] | iamlindoro: | http://pastebin.com/SD204gnD is probably something like my last patch was |
| [19:43:41] | willcooke (willcooke!~will@willcooke.plus.com) has joined #mythtv | |
| [19:43:41] | willcooke (willcooke!~will@willcooke.plus.com) has quit (Changing host) | |
| [19:43:41] | willcooke (willcooke!~will@canonical/willcooke) has joined #mythtv | |
| [19:56:06] | andreax (andreax!~andreaz@p54BF1B1F.dip.t-dialin.net) has joined #mythtv | |
| [19:56:55] | ocrateb (ocrateb!~ocrateb@s83-177-186-55.cust.tele2.se) has joined #mythtv | |
| [20:14:59] | Captain_Murdoch: | markk, iamlindoro, yeah, I left b frames out because of the apple recommendation and the side effects it can cause with segmentation especially related to starting in the middle of a stream or seeking to a new segment in the middle. I don't know if it is possible to work around that in the transcoder by buffering frames, etc., but that would be a lot more logic than is currently there where we just grab a frame, call the encode |
| [20:14:59] | Captain_Murdoch: | r, and write the resulting encoded frame out if one was returned. |
| [20:40:07] | willcooke (willcooke!~will@canonical/willcooke) has quit (Ping timeout: 240 seconds) | |
| [20:40:30] | willcooke (willcooke!~will@canonical/willcooke) has joined #mythtv | |
| [20:46:39] | markk: | I'm obviously missing something here. the b-frames setting made a huge difference – without it my transcoded h264 material is very poor quality (even at higher resolutions). I've adjust for 4mv, level etc and no real difference. |
| [20:46:59] | markk: | (not using vdpau) |
| [21:03:13] | stichnot_ (stichnot_!~chatzilla@192.55.55.39) has joined #mythtv | |
| [21:05:54] | stichnot (stichnot!~chatzilla@nat/intel/x-zdtdjdamdlsvsobk) has quit (Ping timeout: 252 seconds) | |
| [21:05:55] | stichnot_ is now known as stichnot | |
| [21:07:16] | jeff999 (jeff999!~jeff@203-217-51-4.dyn.iinet.net.au) has quit (Ping timeout: 240 seconds) | |
| [21:16:36] | willcooke (willcooke!~will@canonical/willcooke) has quit (Ping timeout: 240 seconds) | |
| [21:18:08] | willcooke (willcooke!~will@willcooke.plus.com) has joined #mythtv | |
| [21:18:08] | willcooke (willcooke!~will@willcooke.plus.com) has quit (Changing host) | |
| [21:18:08] | willcooke (willcooke!~will@canonical/willcooke) has joined #mythtv | |
| [21:19:03] | jpabq (jpabq!~jpabq@mythtv/developer/jpabq) has joined #mythtv | |
| [21:20:43] | jeff999 (jeff999!~jeff@203-173-12-152.dyn.iinet.net.au) has joined #mythtv | |
| [21:22:12] | j-rod|afk (j-rod|afk!~jarod@static-72-93-233-2.bstnma.fios.verizon.net) has joined #mythtv | |
| [21:22:33] | j-rod|afk (j-rod|afk!~jarod@static-72-93-233-2.bstnma.fios.verizon.net) has quit (Client Quit) | |
| [21:23:47] | jeff999 (jeff999!~jeff@203-173-12-152.dyn.iinet.net.au) has quit (Remote host closed the connection) | |
| [21:26:02] | stuarta: | Beirdo: i've quite a few mpeg2's that won't losslessly transcode |
| [21:27:28] | stuarta: | if that's the sort of thing you are after? |
| [21:30:27] | Beirdo: | yeah, not sure what's causing the issues stuartm is seeing, but any known borked file would likely help get to the bottom of it |
| [21:30:57] | stuartm: | stuarta: afaik he's after stuff that doesn't transcode correctly after the qt changes but did before them, that said if he's willing to look at the deadlock and other transcoding failures then great :) |
| [21:31:48] | Beirdo: | Yeah, it would be nice to clean it ALL up, but in particular, any changes that happened in the de-Qt3ing would be good to get behind us |
| [21:31:48] | stuartm: | fwiw, what I'm experiencing is lossless transcodes working but producing artefacts around the cut points when they didn't before the changes |
| [21:32:09] | stuarta: | right, not seen that personally |
| [21:32:12] | Beirdo: | I'm betting it's an off-by-one buried somewhere in there |
| [21:32:49] | Beirdo: | those are insiduous little bitches to find |
| [21:32:55] | stuarta: | heh |
| [21:33:42] | Jordack (Jordack!~jordack@h69-131-44-221.plmomi.dedicated.static.tds.net) has quit () | |
| [21:41:46] | kth (kth!~kth@unaffiliated/kth) has quit (Quit: Leaving.) | |
| [21:42:28] | stuartm: | to tell the truth I can't be 100% certain that it's a bug caused by those changes, it could be a change in the encoding options with the source video ... on balance though a new bug is more probable than a combination of encoding options that no-one has ever encountered before |
| [21:44:15] | stuartm: | over christmas I might be able to run the same file through master and 0.24.1 to compare |
| [21:45:08] | Chutt (Chutt!~ijr@cpe-24-29-225-175.neo.res.rr.com) has quit (Ping timeout: 240 seconds) | |
| [21:46:42] | Beirdo: | K. Well, depending on results, I don't mind digging into it more if we have something to work with :) |
| [21:47:38] | stuartm: | that email to the -dev list from the RadioTimes folk is promising |
| [21:48:15] | stuartm: | well, assuming the quality of the data doesn't change with the out-sourcing |
| [21:48:28] | Beirdo: | yeah, that looks like a potentially good thing |
| [21:48:51] | stuartm: | "with the exception of film ratings" << Ahh, bugger |
| [21:49:04] | stuarta: | i don't believe them anyway |
| [21:49:13] | dekarl: | has anybody looked at http://atlas.metabroadcast.com/ looks interesting for combining our (.de) lot of public VoD services into one ;) |
| [21:50:47] | willcooke (willcooke!~will@canonical/willcooke) has quit (Ping timeout: 240 seconds) | |
| [21:52:37] | stuarta: | i doubt it |
| [21:52:58] | jpabq (jpabq!~jpabq@mythtv/developer/jpabq) has quit (Quit: jpabq) | |
| [21:53:03] | willcooke (willcooke!~will@willcooke.plus.com) has joined #mythtv | |
| [21:53:03] | willcooke (willcooke!~will@canonical/willcooke) has joined #mythtv | |
| [21:53:03] | willcooke (willcooke!~will@willcooke.plus.com) has quit (Changing host) | |
| [21:54:19] | stuartm: | stuarta: do you know something we don't, or is this just normal pessimism? :) |
| [21:54:54] | allesmueller (allesmueller!~allesmuel@unixboard/users/allesmueller) has joined #mythtv | |
| [21:55:08] | stuarta: | pessimism :) |
| [21:55:38] | stuarta: | and there is more likely stuff you know that i don't given i have nearly 5000 mythtv emails to read! |
| [21:55:48] | jmartens (jmartens!~jmartens@s5597ca60.adsl.wanadoo.nl) has quit (Quit: Leaving.) | |
| [21:57:25] | stuartm: | ok, good, I'm not entirely buying into metabroadcast's promises yet but given the alternatives and the fact that it's Christmas I'll remain just about optimistic for now :) |
| [21:58:41] | ** stuarta ponders a bah humbug hat or more pretty lights ** | |
| [21:59:19] | stuartm: | lights are on offer, 3 for 2 at B&Q |
| [21:59:55] | ** stuarta remembers wifey has bought 6 different lights for him to install in the house ** | |
| [22:00:11] | stuarta: | bathroom, utility, lounge x4 |
| [22:09:55] | Chutt (Chutt!~ijr@cpe-24-29-225-175.neo.res.rr.com) has joined #mythtv | |
| [22:14:17] | stuartm: | Beirdo: I'm not sure if it's related to the MythHDD changes you committed (written by Lawrence) but I can reliably crash the frontend by plugging in a USB mass storage device – backtrace coming up |
| [22:15:35] | stuartm: | http://pastebin.com/BjemBqzY – QString segfault |
| [22:15:48] | zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Remote host closed the connection) | |
| [22:17:38] | stuartm: | http://pastebin.com/zjAy8xE6 – Full backtrace including the last hundred lines from the logs |
| [22:18:15] | stuartm: | it's retrying to mount constantly as long as it'd connected |
| [22:20:56] | stuartm: | it's not obvious to me what the problem is exactly otherwise I'd have produced a fix |
| [22:22:45] | stuartm: | agh, it's worse than no ratings info, they've dropped the excellant RT writeups for films |
| [22:23:04] | stuarta: | you mean the ones that gave away the entire plot? |
| [22:25:53] | allesmueller (allesmueller!~allesmuel@unixboard/users/allesmueller) has quit (Ping timeout: 245 seconds) | |
| [22:26:59] | Beirdo: | OMG, nasty! |
| [22:27:28] | Beirdo: | that thing really needs a massive overhaul |
| [22:27:52] | Beirdo: | starting with rm .... |
| [22:27:53] | Beirdo: | :) |
| [22:29:19] | Beirdo: | I guess we need a ticket to track that puppy. I'll see if I can reproduce it at home (sounds like it should be easy to do) |
| [22:31:54] | willcooke (willcooke!~will@willcooke.plus.com) has joined #mythtv | |
| [22:31:54] | willcooke (willcooke!~will@willcooke.plus.com) has quit (Changing host) | |
| [22:31:54] | willcooke (willcooke!~will@canonical/willcooke) has joined #mythtv | |
| [22:32:27] | willcooke (willcooke!~will@canonical/willcooke) has quit (Client Quit) | |
| [22:32:29] | ocrateb (ocrateb!~ocrateb@s83-177-186-55.cust.tele2.se) has quit (Quit: Leaving) | |
| [22:36:15] | j-rod|afk (j-rod|afk!~jarod@static-72-93-233-2.bstnma.fios.verizon.net) has joined #mythtv | |
| [22:39:42] | allesmueller (allesmueller!~allesmuel@mk093111020237.a1.net) has joined #mythtv | |
| [22:39:42] | allesmueller (allesmueller!~allesmuel@unixboard/users/allesmueller) has joined #mythtv | |
| [22:39:42] | allesmueller (allesmueller!~allesmuel@mk093111020237.a1.net) has quit (Changing host) | |
| [22:42:20] | LTHorn (LTHorn!~Android@15.sub-174-253-27.myvzw.com) has joined #mythtv | |
| [22:42:37] | j-rod|afk (j-rod|afk!~jarod@static-72-93-233-2.bstnma.fios.verizon.net) has quit (Quit: Coyote finally caught me) | |
| [22:43:02] | j-rod|afk (j-rod|afk!~jarod@static-72-93-233-2.bstnma.fios.verizon.net) has joined #mythtv | |
| [22:55:43] | stuartm: | stuarta: I never thought that they did give away the plot, just enough commentary to say whether it was worth a couple of hours of your life |
| [22:56:04] | stuartm: | Beirdo: I'll open a ticket :) |
| [22:57:45] | stuartm: | stuarta: you'd be happy with what we've got now, it's more or less what you get from EIT but 14 days vs 7 |
| [22:57:59] | stuarta: | :) |
| [22:58:12] | stuartm: | at least, I hope it's still 14 days ... |
| [23:07:42] | stichnot (stichnot!~chatzilla@192.55.55.39) has quit (Ping timeout: 255 seconds) | |
| [23:13:38] | andreax (andreax!~andreaz@p54BF1B1F.dip.t-dialin.net) has quit (Read error: Connection reset by peer) | |
| [23:18:41] | stuartm: | I'm thinking we need the ability to lookup film and programme reviews and ratings reviews from the guide, there must be at least one good source and API we can use, then either allow the review to be pulled up from the guide or better yet, add it to the information shown on the details screen |
| [23:19:20] | kormoc: | stuartm, Rotten Tomatoes is good for film |
| [23:19:50] | stuartm: | kormoc: I'll take a look, thanks |
| [23:19:59] | kormoc: | This atlas.metabroadcast.com is close to what I wanted to do with MediaGlue, so I might just stuff to it |
| [23:20:04] | kormoc: | (or use it to power MG) |
| [23:53:27] | dblain (dblain!~dblain@mythtv/developer/dblain) has quit (Ping timeout: 240 seconds) | |
| [23:53:46] | dblain (dblain!~dblain@c-76-127-227-175.hsd1.ma.comcast.net) has joined #mythtv | |
| [23:53:46] | dblain (dblain!~dblain@c-76-127-227-175.hsd1.ma.comcast.net) has quit (Changing host) | |
| [23:53:46] | dblain (dblain!~dblain@mythtv/developer/dblain) has joined #mythtv | |
IRC Logs collected by
BeirdoBot.
Please use the above link to report any bugs.