| Thursday, February 27th, 2014, 00:00 UTC | ||
| [00:00:11] | nameless` (nameless`!~nameless@digi00846.digicube.fr) has quit (Ping timeout: 252 seconds) | |
| [00:18:11] | nameless` (nameless`!~nameless@digi00846.digicube.fr) has joined #mythtv | |
| [00:21:59] | knightr_ (knightr_!~Nicolas@69-165-170-178.dsl.teksavvy.com) has joined #mythtv | |
| [00:22:10] | knightr (knightr!~Nicolas@mythtv/developer/knightr) has quit (Read error: Connection reset by peer) | |
| [00:22:44] | jya: | ohhhhhh... 12 hours run of live TV, not a single interruption, and can rewind 12 hours behind and forward.... first time *ever* |
| [01:25:12] | dmfrey (dmfrey!~dmfrey@65-78-98-83.c3-0.smt-ubr1.atw-smt.pa.cable.rcn.com) has joined #mythtv | |
| [02:09:39] | jya: | stuarta: did you intend to put all your Qt5 changes in fixes/0.27 ? if so, having the cherry-pick information would have been better |
| [02:09:47] | monkeypet: | jya that is awesome! |
| [02:09:56] | monkeypet: | congrats |
| [02:10:37] | jya: | rewind causes stutter for a few seconds from time to time... something isn't quite right when re-initialising the player... need to look into it. |
| [02:11:02] | doev (doev!~doev@p4FD42FEE.dip0.t-ipconnect.de) has quit (Ping timeout: 264 seconds) | |
| [02:11:21] | doev (doev!~doev@p4FD42F96.dip0.t-ipconnect.de) has joined #mythtv | |
| [02:11:23] | jya: | also , if you rewind when you're within 15s of the end of the recording, to jumps to the next one the following second or so. there's a race condition occurring .. can't be bothered fixing it it's rare enough |
| [02:13:44] | monkeypet: | I only cherrypicked 20f7cd2858e79385ceec07b2c19c7e4f84714e18 |
| [02:14:07] | monkeypet: | Should i cherrypick the others to help test? |
| [02:14:22] | jya: | there's a few to cherry-pick (and that I intend to) |
| [02:14:26] | jya: | let me make a list |
| [02:14:48] | monkeypet: | If you give me a list, i'll patch it in and test it overnight. |
| [02:15:12] | jya: | in order: |
| [02:16:28] | jya: | 20f7cd2858e79385ceec07b2c19c7e4f84714e18 |
| [02:16:28] | jya: | fafc658597b547e2f0173032515e9790170ebcbe |
| [02:16:30] | jya: | 443b624ebcc8ea56026cb49ca833edc7d73832c1 |
| [02:16:31] | jya: | 534394068373c52fd712d653e5fea725c564a673 |
| [02:16:31] | jya: | 11daeaadc52214605e204cdfe6f7c001c9c1cc9b |
| [02:16:34] | jya: | those are those I intend to backport |
| [02:16:58] | jya: | both backend/frontend should be patched |
| [02:18:54] | taylorr: | jya: did that patch I sent you the other night work? |
| [02:19:05] | jya: | you sent me a patch? |
| [02:19:50] | jya: | what was it about? |
| [02:19:58] | taylorr: | I sent you a link to patch on my website |
| [02:20:11] | jya: | I didn't see it |
| [02:21:46] | jya: | only thing I want to "fix" is the buffer size used by the ringbuffer from time to time. It leads to spurious log and too long wait or delay. Not point reading in 1MB block when the data being written to the disk is at 3MBit/s and log warning when it takes more than a few ms to read. |
| [02:21:59] | jya: | of course it's going to take a while to read 1MB in this case |
| [02:23:34] | jya: | monkeypet: I will push a new 0.27 branch with the proposed changes devel/027candidates |
| [02:34:28] | monkeypet: | jya: I'll pull from there, easier for me. |
| [02:34:40] | jya: | yeah, there are conflicts cherry-picking too |
| [02:37:23] | jya: | taylorr: can you tell me what the patch was about and where I could find it? (if not I'll check the irc log) |
| [02:38:38] | jya: | monkeypet: done |
| [02:40:28] | ** skd5aner claps for jya's livetv bug smash-a-thon ** | |
| [02:41:40] | skd5aner: | jya: thank you for all the time spent, I know you are just scratching an itch (and trying to make the wife happy again ;-) ) but, I can tell you, my wife will be equally as happy – as will I and I'm sure several other users |
| [02:42:57] | jya: | I love making wife (not just mine) happy |
| [02:43:45] | jya: | in the mean time... mac pro ordered :) |
| [02:43:46] | skd5aner: | I've got to run, but just dropped in briefly and read your updates |
| [02:43:53] | skd5aner: | wanted to say thanks |
| [02:43:54] | skd5aner: | good night |
| [02:44:43] | taylorr: | jya: I sent a PM with a link to a patch that forced the backend to transition every 30 secs for livetv |
| [02:44:56] | taylorr: | you replied... just curious if you got any use out of it |
| [02:45:00] | jya: | taylorr: oh that one.. I did.. |
| [02:45:09] | jya: | I thought it was stichnot sorry |
| [02:45:19] | jya: | it worked great, I changed it to 1 minute |
| [02:45:32] | jya: | wouldn't have been able to play with live TV seeking without it ... thanks ! |
| [02:45:50] | taylorr: | cool, I never tried getting ff and rew to work across livetv transitions |
| [02:46:20] | jya: | yeah.. problem with that one, I never knew it was broken to start with... I always assumed you could rewind across program |
| [02:47:11] | taylorr: | jya: curious, do you do your myth dev on a macbook? |
| [02:47:27] | jya: | taylorr: you're much more familiar than me in regards to the playback code... if you don't mind trying the current code and check why sometimes when you rewind to the beginning of a program, it stutters like crazy and in the log you see that it drops frame to catch up for about 5–6s |
| [02:47:37] | jya: | taylorr: on an imac yes |
| [02:48:15] | jya: | I do use a macbook air when I was testing the backend. as I can sit in my couch in the same room as the tv |
| [02:48:24] | taylorr: | I should have asked if you used os x for development |
| [02:48:32] | jya: | I do too |
| [02:48:44] | jya: | my entire changes to the frontend was done on a mac |
| [02:49:15] | taylorr: | cool, as far as trying your changes you'll be sad to hear that I still run 0.25-fixes... just haven't had time to upgrade |
| [02:49:27] | jya: | working with xcode simplify things greatly, in particular multi-threaded stuff. you have all at a glance |
| [02:49:59] | jya: | I can also have helper so it displays the Qt objects (Qstring etc..) directly in the variables pane |
| [02:50:29] | taylorr: | that's cool, I just got started using Macs mid last year... just ordered the wife an air today |
| [02:50:53] | jya: | I love my air... I have the 2011 model, it started to be a bit slow |
| [02:51:10] | jya: | SSDs are great, but really performance do degrade significantly over time |
| [02:51:28] | jya: | on my iMac, the SSD gives me 110MB/s read/write only now. |
| [02:51:28] | taylorr: | I really like OS X overall... but Finder is disappointing |
| [02:51:36] | jya: | was 3 times that fast new |
| [02:51:50] | wagnerrp: | TRIM should resolve that |
| [02:51:58] | taylorr: | jya: even with trim enabled? are the ssd's getting full? |
| [02:52:15] | jya: | I don't use the finder much... as you say it's very .... poor. And it's 100 times better than it was just a couple years ago. You coulnd't even copy/paste files |
| [02:52:21] | taylorr: | I pretty sure that genuine apple SSD drives enable trim by default |
| [02:52:48] | jya: | trim is enabled on both machine... but it doesn't prevent wear in the end... |
| [02:52:51] | taylorr: | if you installed a thirdparty ssd you'd need to patch a file to get trim working |
| [02:53:06] | jya: | it's just a mechanical/electronic fact that the more it goes, the slower it gets to format a block |
| [02:53:28] | taylorr: | that's not cool, I didn't expect that |
| [02:53:34] | jya: | mine are all apple stuff... not going to put appart an imac or macbook air unless I absolutely must |
| [02:53:44] | wagnerrp: | actually, the more you use it, the faster it can format blocks |
| [02:53:56] | wagnerrp: | the issue is in whether those blocks can retain data over a useful duration |
| [02:54:03] | jya: | a typical MLC flash sector can give you 10,000 format cycle.. the more you do, the slower it becomes. |
| [02:54:07] | taylorr: | jya: I'm running a Hackintosh for my desktop |
| [02:54:22] | jya: | TRIM mediate the issue by spreading the wear across all blocks, so it is better |
| [02:54:29] | jpharvey (jpharvey!~jpharvey@host109-148-236-111.range109-148.btcentralplus.com) has quit (Ping timeout: 240 seconds) | |
| [02:54:35] | jya: | but it's just a matter of time before it becomes slow |
| [02:54:54] | wagnerrp: | wear leveling is independent of TRIM |
| [02:55:00] | jya: | and I've found that compiling and dev, SSD aren't that great. lots of small files, lots of writes |
| [02:55:13] | wagnerrp: | wear leveling will happen regardless. TRIM just allows wear leveling to work more effectively |
| [02:55:16] | taylorr: | jya: do you compile on the air? |
| [02:55:43] | jya: | of course it has an impact on wear... you work on blocks that are free... |
| [02:55:46] | jya: | taylorr: I do yes |
| [02:56:11] | wagnerrp: | yes, and without trim, you can have much the same effect by never touching a big chunk of logical space for wear leveling to use |
| [02:56:25] | wagnerrp: | TRIM just allows you to format the whole thing |
| [02:56:41] | jya: | wagnerrp: I've written sufficient flash controller firmware to know precisely what wear on a ssd is :) |
| [02:57:48] | taylorr: | jya: so are you saying that the more wear that occurs that the reads slow down? |
| [02:58:04] | wagnerrp: | and that statement doesn't make sense |
| [02:58:28] | wagnerrp: | in terms of the firmware, once you fill up the drive, it may take more time to figure out how to move stuff around to free up a new block |
| [02:58:46] | wagnerrp: | but electronically, the more you use a cell, the weaker it should become, and the faster you can flash it |
| [02:59:16] | jya: | no, I mean the more wear you have, the slower the writes... It's pretty rare for a typical dev usage to only have reads |
| [02:59:21] | wagnerrp: | until such time that the cell is so weak that it cannot be trusted to store data |
| [03:00:06] | jya: | that's not what I meant by wear. It becomes physically slower to format (turn all bit to 1) the more you do it on a single block |
| [03:00:54] | wagnerrp: | right, and that's due to the controller not having enough free space to effectively perform garbage collection, because without TRIM, it doesn't actually know where that free space exists |
| [03:00:58] | taylorr: | I was hoping these drives would last 10+ years and still perform like new :( |
| [03:01:12] | jya: | taylorr: they won't |
| [03:01:17] | wagnerrp: | it basically ends up having to garbage collect on every write |
| [03:01:28] | jya: | not with your typical MLC NAND |
| [03:02:17] | jya: | I've formatted the entire drive (flash format, not OS format) several times, it never goes back to the speed you got when new |
| [03:02:26] | clever: | jya: yeah, ive also nuked an SD card just by compiling omxplayer a dozen times on my pi |
| [03:02:47] | clever: | various text files that had never been modified since install started turning up corrupt |
| [03:03:17] | jya: | clever: same here... I used a pi to run a solar monitoring tool I wrote (http://mediaserver.avenard.org/power) |
| [03:03:28] | clever: | ive since moved the source dir to nfs and mounted a tmpfs on /tmp, turns out the compiler dumps a lot of junk in /tmp and goes right back to the same problem |
| [03:03:29] | jya: | that generates a lot of logs |
| [03:03:43] | clever: | the most important thing, is to turn on block dump |
| [03:03:51] | clever: | disable syslog, then echo 1 > /proc/sys/vm/block_dump |
| [03:03:54] | wagnerrp: | that sounds like a deficiency in the firmware programming, rather than a physical issue with NAND flash itself |
| [03:03:55] | jya: | after about 3 monhts, the pi became unusable.... writing to the sd card would literally takes seconds |
| [03:03:58] | clever: | it will print every disk read and write to dmesg |
| [03:04:06] | clever: | grep for writes, and you have the pid and path it wrote to |
| [03:04:13] | clever: | kill everything! |
| [03:04:37] | clever: | also, the virtual hwclock, it writes the system time to the sd card every 30 minutes |
| [03:04:43] | jya: | I'm more disappointed with my imac |
| [03:04:48] | clever: | so if you shutdown improperly, it restores the clock with a 30min accuracy |
| [03:04:51] | jya: | it has a 256gig SSD and a 2TB hitachi hard drive |
| [03:04:59] | clever: | that will also have an impact on your disk lifetime, if you go long term |
| [03:05:07] | jya: | the hitach hard drive now gives me faster benchmarks than the SSD after 3 years |
| [03:05:36] | clever: | i never noticed speed problems with my sd card, it just started corrupting files and fs metadata without warning |
| [03:06:38] | jya: | talking about SD card... |
| [03:06:57] | taylorr: | jya: have you tried creating a partition smaller than the drive size to see if it helps? |
| [03:06:59] | jya: | I sent a broken SD to Sandisk about 2 months ago, I never got its replacement |
| [03:07:15] | jpharvey (jpharvey!~jpharvey@host109-148-236-119.range109-148.btcentralplus.com) has joined #mythtv | |
| [03:07:33] | jya: | taylorr: I haven't... simply because the way the mac application works, it makes it very difficult to spread the data across the disks |
| [03:07:39] | clever: | taylorr: enless you run the special trim stuff, the sd card controller will think the rest of the disk is still 'in use', and wont erase it |
| [03:07:41] | jya: | too much hassle. |
| [03:08:08] | clever: | if you make a partition thats only using the lower half of the disk, the controller has no way to know the upper half is free, and will keep its data safe |
| [03:08:25] | taylorr: | clever: not on an ssd |
| [03:08:35] | wagnerrp: | yes, unless you tell it otherwise using TRIM |
| [03:08:41] | jya: | taylorr: on a SSD, partitions really aren't much relevant. it will still end up writing everywhere on the disk |
| [03:08:45] | clever: | you would need to use the trim command to tell the ssd to un-allocate those blocks entirely |
| [03:08:48] | jya: | that's if the SSD does its job properly |
| [03:09:28] | clever: | jya: hmmm, seems i still havent fixed my ipv6 issues, it takes a minute to open your site up |
| [03:09:43] | jya: | that site only has ipv4 |
| [03:09:49] | clever: | mediaserver.avenard.org. 86261 IN CNAME htpc.avenard.org. |
| [03:09:50] | clever: | htpc.avenard.org. 3460 IN AAAA 2001:470:36:b1d::10:11 |
| [03:09:55] | clever: | dns says otherwise |
| [03:10:00] | jya: | yeah... that's broken |
| [03:10:10] | clever: | ah, then its probly no my end this time |
| [03:10:14] | jya: | htpc.avenard.org is another machine |
| [03:10:27] | jya: | problem is I did a cname |
| [03:10:36] | jya: | but also defined an ipv6 address |
| [03:10:36] | clever: | my tunnel broke a few days ago, and caused similar issues at various random sites |
| [03:10:54] | wagnerrp: | i really aught to get my tunnel working again |
| [03:10:57] | jya: | it's 2001:470:36:b1d::10:14 |
| [03:10:57] | clever: | yeah, cname inherits both, its a pain when i want the A records to match and the AAAA's to differ |
| [03:11:09] | clever: | so i need to manualy fix the ip on 5 different A records every time it changes |
| [03:11:19] | wagnerrp: | or shut down route advertising completely |
| [03:11:26] | jya: | the problem is that the dyndns provider I use, I haven't managed to find a way to set more than two hosts |
| [03:11:54] | clever: | jya: ahhh, ive been using zoneedit, but they migrated the site a few years back and lost the dyn updater |
| [03:11:58] | jya: | so the URL in v6 is: http://[2001:470:36:b1d::10:14]/power/ |
| [03:12:01] | clever: | and the register page seems to have vanished |
| [03:12:23] | clever: | amazon also works great, only 50 cents/month per domain |
| [03:12:57] | clever: | wagnerrp: i also had route advertisement problems a few months back, the 2nd router refused to stop sending them |
| [03:13:11] | clever: | so it advertised a null route, and sucked up all traffic in the house |
| [03:13:26] | jya: | I use hurricane electric tunnel, and their dyndns tool |
| [03:13:47] | clever: | HE also, but not using their dns, they didnt have it when i started |
| [03:13:54] | jya: | but I'm yet to understand their documentation on how to update anything else but the main domain |
| [03:14:39] | wagnerrp: | clever: the issue is IPv6 AAAA records still resolve, and the machines think they have a valid route to access them, so it's frustrating waiting for things like package updates to timeout and drop to IPv4 |
| [03:15:07] | clever: | wagnerrp: yeah, i had that very issue yesterday, the tunnel was mis-configured, so i didnt get any ipv6 replies |
| [03:15:22] | clever: | ipv6 outbound worked, but not inbound |
| [03:15:47] | clever: | jya: do you know anything about using solar power in canada? |
| [03:15:55] | jya: | if someone find the HE URL to use to update a particular hostname |
| [03:16:01] | jya: | I can remove the cname then |
| [03:16:23] | wagnerrp: | i would expect it's something along the lines of "don't bother unless you have no grid access" |
| [03:16:43] | clever: | yeah, off the grid is simpler, but way more expensive |
| [03:16:52] | jya: | clever: no... I know how to use solar period... whatever entitlements or funding you can get in your local area varies. even in Oz, it changes per state *and* per infrastructure supplier |
| [03:17:07] | clever: | ah |
| [03:17:16] | jya: | I have a 6.4kW system on my roof... |
| [03:17:22] | jya: | grid connected |
| [03:18:09] | clever: | ah |
| [03:18:22] | clever: | ive got just plain old 100amp service here (120v) |
| [03:18:44] | wagnerrp: | no split phase? |
| [03:18:46] | jya: | we're in 230V zone here, won't see 100 amps in residential setup |
| [03:18:57] | clever: | wagnerrp: yeah, 2 phases, its probly 50 amps per phase |
| [03:19:05] | wagnerrp: | so 240v |
| [03:19:07] | jya: | typical install is 65A single phase, or 35A per phase |
| [03:19:14] | clever: | wagnerrp: yeah is you measure between the phases |
| [03:19:16] | clever: | if* |
| [03:19:44] | clever: | the garage has 200amp service, because the previous owner did welding |
| [03:19:57] | clever: | and with the house on only 100amp, the lights dim any time a large motor starts up |
| [03:20:39] | jya: | pfff... he.net does show the instructions on how to set dns, but only if you're not logged in |
| [03:20:40] | jya: | https://dns.he.net/?action=logout |
| [03:20:48] | clever: | lol |
| [03:21:07] | clever: | yep, they override the index page if you login |
| [03:21:24] | clever: | the tunnel page is different, it properly adds an authed only link on the site |
| [03:21:30] | jya: | last I tried that (from finding it in their forums) it didn't work |
| [03:22:39] | clever: | looks like you need to flag a record as being dynamic |
| [03:22:43] | clever: | and each record gets its own pw? |
| [03:22:55] | jya: | yes, I did that. |
| [03:23:14] | jya: | but everytime I attempted to update that IP address only, it would give me an error |
| [03:23:36] | clever: | i should try it out, since my current dns hoster seems to be on its death bed |
| [03:24:06] | clever: | https://www.zoneedit.com either that, or i'm blind and cant find a register link anywhere |
| [03:24:30] | jya: | I dont use zoneedit anymore.. they require you to login to your account every few days |
| [03:24:35] | jya: | same with dyndns.org |
| [03:24:44] | clever: | zoneedit doesnt expire things on my like that |
| [03:24:51] | clever: | ive left it untouched for months before |
| [03:25:07] | clever: | but they broke the dyn update when they migrated the site |
| [03:25:35] | jya: | maybe I can go back to them then |
| [03:26:01] | jya: | though, I like he.net; their service is easy to use; allow everything and I could easily delegate my ipv6 net |
| [03:27:14] | wagnerrp: | the problem is they require your ISP to be using sane network practices |
| [03:27:22] | clever: | wagnerrp: ? |
| [03:27:34] | wagnerrp: | which mine apparently stopped doing when they transitioned me to ADSL2 and a POS westell modem |
| [03:27:41] | wagnerrp: | which apparently block IP protocol 41 |
| [03:27:44] | clever: | ah |
| [03:27:56] | jya: | ohhhh...he.net worked for my domain ! |
| [03:27:56] | jya: | cool |
| [03:27:59] | clever: | mine POS still allows protocol 41, but it does even worse stuff |
| [03:28:07] | clever: | the tunnel wont activate if they cant ping your wan ip |
| [03:28:16] | clever: | the POS router blocked pings, while the config said pings where allowed |
| [03:28:23] | clever: | i had to turn pings off, then back on to fix the tunnel |
| [03:28:39] | jya: | just had to use a different URL |
| [03:28:53] | taylorr: | jya: digging a little bit about ssd speed degradation led me to this piece of software... http://diglloydtools.com/disktester.html |
| [03:29:30] | taylorr: | it has a recondition option that supposedly will erase the free area so that it is optimized for writes |
| [03:29:36] | clever: | taylorr: oh yeah, that reminds me, ive seen tools before that would test the speed of normal platter disks, and it reveals that each group of tracks has a different speed |
| [03:29:43] | clever: | the closer you get to the outside of the disk, the faster the platter is moving relative to the head |
| [03:29:48] | clever: | and the higher the data rate |
| [03:30:08] | taylorr: | jya: the recondition is based off this article http://macperformanceguide.com/Storage-SSD-Reconditioning.html |
| [03:30:09] | jya: | taylorr: I'm a bit wary about those tools, because unless they directly write to the controller, they never really format a block as it should |
| [03:30:31] | NightMonkey_ (NightMonkey_!~NightrMon@173-164-139-193-SFBA.hfc.comcastbusiness.net) has joined #mythtv | |
| [03:30:48] | jya: | there's a massive difference between formatting a flash sector and writing to the disk |
| [03:30:50] | taylorr: | jya: doesn't seem any risk to it |
| [03:31:02] | jya: | risk now, but extra wear yes |
| [03:31:36] | clever: | wagnerrp: another major issue ive run into on this POS router, it cant port forward traffic from inside the lan |
| [03:31:41] | clever: | if i open the public ip, i get the router config |
| [03:31:56] | jya: | clever: very few do |
| [03:31:58] | NightMonkey (NightMonkey!~NightrMon@pdpc/supporter/professional/nightmonkey) has quit (Read error: Connection reset by peer) | |
| [03:32:21] | clever: | jya: i was using a custom router based on linux and a 200mhz pentium, which handled the HE tunnel and did everything just fine |
| [03:32:22] | jya: | on mine, I got around by setting different name resolution depending if you're on the WAN or LAN |
| [03:32:28] | clever: | it had no trouble with 600kbyte/sec on dsl |
| [03:32:37] | jya: | mine is a linux based router: asus rt-n66 |
| [03:32:41] | clever: | jya: i tried that, then i ran into a new problem |
| [03:32:48] | jya: | with my own firmware (I'm a dev on the tomato project) |
| [03:32:52] | clever: | jya: if you set the dns to 192.168.2.61, everything on wifi breaks |
| [03:32:59] | clever: | turns out, nothing on the wifi can talk to a wired device |
| [03:33:05] | wagnerrp: | as far as i'm concerned, it should be in bridge mode, and have no need to forward traffic |
| [03:33:19] | wagnerrp: | but i can't seem to get my router to work as a dumb modem |
| [03:33:23] | jya: | clever: why don't you set the wifi in bridge mode? |
| [03:33:38] | NightMonkey_ is now known as NightMonkey | |
| [03:33:39] | clever: | jya: the website for the router says that no such feature exists |
| [03:33:45] | clever: | jya: and there is no option to turn it off anywhere |
| [03:33:46] | jya: | you have to be crazy or have a lot of spare time to do otherwise |
| [03:33:58] | NightMonkey (NightMonkey!~NightrMon@173-164-139-193-SFBA.hfc.comcastbusiness.net) has quit (Client Quit) | |
| [03:34:02] | jya: | clever: I'd be surprised if it was antything but bridge mode |
| [03:34:12] | clever: | jya: the isp 'customized' the firmware |
| [03:34:14] | NightMonkey (NightMonkey!~NightrMon@173.164.139.193) has joined #mythtv | |
| [03:34:14] | NightMonkey (NightMonkey!~NightrMon@pdpc/supporter/professional/nightmonkey) has joined #mythtv | |
| [03:34:14] | NightMonkey (NightMonkey!~NightrMon@173.164.139.193) has quit (Changing host) | |
| [03:34:20] | jya: | actually, only advanced router let you set the wifi network to not be bridged |
| [03:34:23] | clever: | the router and modem are seperate, but the modem isnt standard |
| [03:34:36] | clever: | the modem outputs 802.1q tagged packets |
| [03:34:43] | clever: | no normal router can handle that |
| [03:34:45] | jya: | so you get a different subnet on the wifi than on the wired? |
| [03:34:52] | clever: | jya: worse, the same subnet |
| [03:35:04] | jya: | then it must be bridged |
| [03:35:16] | clever: | yet ping and arp cant make contact with the other side, ever |
| [03:35:26] | clever: | but both sides have internet access |
| [03:36:14] | peper03 (peper03!~peper03@mythtv/developer/peper03) has quit (Ping timeout: 264 seconds) | |
| [03:36:46] | peper03 (peper03!~peper03@mythtv/developer/peper03) has joined #mythtv | |
| [03:36:54] | clever: | as for why i cant just drop in a decent router, the modem outputs 802.1q packets, vlan tagging |
| [03:37:12] | clever: | so the router must have support to untag the packets, and treat each vlan properly |
| [03:37:16] | jya: | that's the dumbest network configuration I've ever seen |
| [03:37:30] | jya: | youre better off turning of the wifi feature, and plug an AP instead |
| [03:37:38] | clever: | thats what i did |
| [03:37:48] | clever: | i hooked up the old dlink router with dhcp disabled |
| [03:37:48] | wagnerrp: | apparently Comcast does something similar |
| [03:38:01] | wagnerrp: | their modems advertise two separate wifi networks |
| [03:38:10] | wagnerrp: | one for internal use, and one for public external use |
| [03:38:15] | clever: | my router has the option to advertise up to 4 |
| [03:38:49] | clever: | but there is no option to turn the isolation off, and the website claims they dont even have the ability to isolate things |
| [03:39:01] | clever: | ive tried twice to upgrade to N band |
| [03:39:06] | clever: | both times failed horribly |
| [03:39:26] | clever: | the other N band router sucked up all ipv6 traffic with its own route advertisement packets, but the wan port went nowhere |
| [03:41:47] | clever: | i have been considering replacing the crap router with the sun system i have setup downstairs |
| [03:43:04] | jya: | clever: allright, no more CNAME thing, should all work fine now http://mediaserver.avenard.org/power |
| [03:43:25] | clever: | 84214 seconds left on the cname TTL |
| [03:43:45] | jya: | I've changed the TTL to 300s |
| [03:44:11] | clever: | yep, once i nuke the dns cache it updates |
| [03:44:39] | clever: | cpu : TI UltraSparc IIIi (Jalapeno) |
| [03:44:50] | clever: | would this thing make a good router? |
| [03:45:18] | clever: | i'm not even sure how fast it is, the kernel isnt showing a bogomips or mhz |
| [03:45:18] | wahrhaft_ (wahrhaft_!~quassel@cpe-24-210-69-143.columbus.res.rr.com) has quit (Read error: Connection reset by peer) | |
| [03:45:46] | wagnerrp: | probably eats far more power than simply picking up a cheap atom, if you want to don't want to flash a consumer router |
| [03:46:13] | clever: | it already has to stay on 24/7 as my 1tb NAS unit |
| [03:46:23] | clever: | so power usage isnt much of an issue |
| [03:46:41] | wahrhaft (wahrhaft!~quassel@cpe-24-210-69-143.columbus.res.rr.com) has joined #mythtv | |
| [03:47:13] | clever: | and it has dual gigabit, so its got the perfect hardware |
| [03:47:32] | wagnerrp: | presuming your uplink is anywhere near that speed |
| [03:47:39] | clever: | fiber |
| [03:48:08] | clever: | DCC received file ____ [850MB] from _____ in 00:02:24 [6047.20kB/s] |
| [03:48:20] | wagnerrp: | so... ~50kbps |
| [03:48:47] | clever: | in 00:01:27 [7645.90kB/s] |
| [03:48:55] | clever: | it goes up and down |
| [03:49:08] | clever: | in 00:00:48 [8214.61kB/s] |
| [03:49:36] | clever: | in 00:00:51 [8940.02kB/s] |
| [03:50:41] | clever: | this kind of bandwidth was enough to overload he old 200mhz pentium i was using before |
| [04:23:10] | dmfrey (dmfrey!~dmfrey@65-78-98-83.c3-0.smt-ubr1.atw-smt.pa.cable.rcn.com) has quit (Quit: Ex-Chat) | |
| [04:27:39] | Warped (Warped!~Warped@108-85-161-113.lightspeed.cicril.sbcglobal.net) has joined #mythtv | |
| [04:30:49] | arescorpio (arescorpio!~arescorpi@56-57-245-190.fibertel.com.ar) has quit (Excess Flood) | |
| [05:36:27] | Goga777 (Goga777!~Goga777@128-71-38-29.broadband.corbina.ru) has joined #mythtv | |
| [05:43:22] | Goga777 (Goga777!~Goga777@128-71-38-29.broadband.corbina.ru) has quit (Quit: Leaving) | |
| [06:16:08] | jya: | right... I think I'm done with liveTV... it's working to my expectation now... |
| [06:16:47] | jya: | if someone with knowledge about the video player now wants to have a look at tell me why sometimes rewinding to the beginning of the recording cause massive stutters for a few seconds... |
| [06:16:56] | jya: | I'd greatly appreciate |
| [06:17:12] | jya: | plus looking at the fix would give me a better understanding on how that mess works |
| [07:06:06] | FabriceMG (FabriceMG!~Thunderbi@217.112.59.207) has joined #mythtv | |
| [07:55:09] | dekarl1 is now known as dekarl | |
| [08:17:14] | SteveGoodey (SteveGoodey!~steve@host217-42-221-59.range217-42.btcentralplus.com) has joined #mythtv | |
| [08:49:34] | stuarta: | jya: yes i did, by cherry-pick info do you mean "picked from <blah>" ? |
| [08:49:41] | jya: | yep |
| [08:49:52] | jya: | cherry-pick -x does that for you |
| [08:50:02] | stuarta: | sigh, i just used cherry-pick |
| [08:50:23] | stuarta: | typical |
| [08:50:28] | jya: | just a bit clearer when you look back into history, so you see what's going on |
| [08:50:41] | ** stuarta makes a mental note of -x flag ** | |
| [09:04:29] | SteveGoodey (SteveGoodey!~steve@host217-42-221-59.range217-42.btcentralplus.com) has quit (Quit: Konversation terminated!) | |
| [09:04:53] | Casper0082 (Casper0082!~Casper@pool-71-185-186-159.phlapa.fios.verizon.net) has quit (Ping timeout: 265 seconds) | |
| [09:06:04] | stuartm: | it's one of those things that git should do by default, but doesn't for some stupid reason |
| [09:07:22] | stuartm: | e.g. --signoff when applying a patch by someone else, or --rebase when doing a git pull |
| [09:07:25] | SteveGoodey (SteveGoodey!~steve@host217-42-221-59.range217-42.btcentralplus.com) has joined #mythtv | |
| [09:08:02] | stuartm: | jya: nice work |
| [09:09:19] | jya: | stuartm: thanks... |
| [09:13:27] | stuarta: | agreed, nice work |
| [09:13:40] | jya: | very selfish really |
| [09:14:29] | stuarta: | not really, most things we fix are because we want them to work for us |
| [09:15:26] | jya: | and now, I do want to buy a flex machine, a super juicer, a ninja vacuum cleaner (been watching TV commercials for the past week) |
| [09:16:36] | jya: | we knew using a juicer, aka a nutrient extractor could make you feel better about yourself and spend better time with your kids |
| [09:17:15] | ** stuarta chuckles ** | |
| [09:18:12] | Casper0082 (Casper0082!~Casper@pool-71-185-186-159.phlapa.fios.verizon.net) has joined #mythtv | |
| [09:41:21] | SteveGoodey (SteveGoodey!~steve@host217-42-221-59.range217-42.btcentralplus.com) has quit (Quit: Konversation terminated!) | |
| [09:48:32] | SteveGoodey (SteveGoodey!~steve@host217-42-221-59.range217-42.btcentralplus.com) has joined #mythtv | |
| [09:52:16] | stuarta: | stuartm: i have a suggestion. the web backend tries to load channel icons, if you don't have any, that results in broken links next to every channel in the program listing. maybe add a default "no icon" icon? |
| [09:52:38] | stuartm: | stuarta: yeah I'll do that :) |
| [09:53:19] | stuarta: | cheers :) |
| [10:08:39] | aberrios (aberrios!~aberrios@195.130.201.200) has joined #mythtv | |
| [10:17:06] | ** stuarta adds background channel icon search to his list ** | |
| [10:26:29] | nyloc (nyloc!~quassel@p57B4FEA6.dip0.t-ipconnect.de) has joined #mythtv | |
| [11:39:53] | amessina (amessina!~amessina@2001:470:c1dc:7779:d6be:d9ff:fe8d:7c1e) has quit (Remote host closed the connection) | |
| [11:44:05] | amessina (amessina!~amessina@2001:470:c1dc:7779:d6be:d9ff:fe8d:7c1e) has joined #mythtv | |
| [11:50:33] | andreaz (andreaz!~Andreaz@tmo-106-181.customers.d1-online.com) has joined #mythtv | |
| [11:50:38] | stuartm: | got a little bogged down trying to improve the HLS experience when I really need to get back to the WebFrontend |
| [11:50:39] | andreaz (andreaz!~Andreaz@tmo-106-181.customers.d1-online.com) has quit (Client Quit) | |
| [11:52:24] | stuarta: | stuartm: have you run a web debugger against what it produces? |
| [11:52:53] | stuarta: | throws quite a few 404's for me, i'm currently trying to get to the bottom of why it does what it does |
| [11:53:05] | stuartm: | think my machine is just too slow to ever give the sort of performance from HLS that I'd like, looking at 40+ seconds to transcode a 10 second of segment of an 1080 H.264 recording down to 720p with deinterlacing |
| [11:53:34] | stuartm: | stuarta: a few times yeah, where are you seeing the 404s? |
| [11:55:06] | stuarta: | some of them relate to recordings which were done by an old backend, i've fiddled the database to tell the backend they were done by it. primarily though, it's when it cycles throught the last 10 recordings |
| [11:55:29] | ** stuarta reloads ** | |
| [11:55:46] | stuarta: | hmm, looks like it's still trying to get stuff from the old backend. |
| [11:55:57] | stuarta: | i see the alternate backend name in the request url |
| [11:56:26] | stuarta: | Request URL:http://ganymede:7544/Content/GetPreviewImage? . . . p;Format=JPG |
| [11:56:34] | stuarta: | Referer:http://localhost:7544/ |
| [11:57:31] | stuarta: | stuartm: ganymede is the name of the box that was the backend when these were recorded. updated the hostname field of the recording to the current backend name, but it still does it |
| [11:59:04] | stuarta: | so i dunno if it's just my setup, or an actual issue |
| [12:01:54] | ** stuarta heads off for lunch ** | |
| [12:03:40] | stuartm: | stuarta: port is 6544 not 7544 |
| [12:04:07] | stuartm: | unless you've changed it? |
| [12:23:05] | sraue (sraue!~stephan@xbmc/staff/sraue) has quit (Ping timeout: 246 seconds) | |
| [12:35:45] | sraue (sraue!~stephan@xbmc/staff/sraue) has joined #mythtv | |
| [13:18:07] | stuartm: | heh, Apple's own docs can't agree on the bitrate to use for various resolutions |
| [13:22:05] | stuarta: | stuartm: i run both prod and dev on the same box, prod = 6544, dev = 7544 :-) |
| [13:22:24] | stuartm: | ah |
| [13:24:18] | stuartm: | not exactly sure what's happening there, will take a look later, maybe be some subtly of the storage group stuff |
| [13:24:32] | stuarta: | yeah, that's what i was thinking |
| [13:36:19] | stuartm: | s/subtly/subtlety/ |
| [13:40:30] | stuarta: | my babelfish translated for me |
| [13:41:12] | stuartm: | :) |
| [13:41:24] | ShapeShifter499 (ShapeShifter499!~ShapeShif@unaffiliated/shapeshifter499) has joined #mythtv | |
| [13:41:30] | ShapeShifter499 (ShapeShifter499!~ShapeShif@unaffiliated/shapeshifter499) has left #mythtv ("Leaving") | |
| [14:28:05] | brfransen (brfransen!~brfransen@64.179.169.226) has quit (Remote host closed the connection) | |
| [14:28:27] | brfransen (brfransen!~brfransen@64.179.169.226) has joined #mythtv | |
| [14:36:12] | zentec (zentec!~zentec@75-128-89-222.dhcp.bycy.mi.charter.com) has quit (Ping timeout: 245 seconds) | |
| [14:44:36] | pvr4me (pvr4me!~pvrme@d24-150-182-175.home.cgocable.net) has joined #mythtv | |
| [14:54:43] | pvr4me (pvr4me!~pvrme@d24-150-182-175.home.cgocable.net) has quit (Quit: Colloquy for iPhone - http://colloquy.mobi) | |
| [15:09:49] | zentec (zentec!~zentec@75-128-89-222.dhcp.bycy.mi.charter.com) has joined #mythtv | |
| [15:16:07] | dekarl1 (dekarl1!~dekarl@p4FCEE5EA.dip0.t-ipconnect.de) has joined #mythtv | |
| [15:21:45] | Casper0082 (Casper0082!~Casper@pool-71-185-186-159.phlapa.fios.verizon.net) has quit (*.net *.split) | |
| [15:21:46] | peper03 (peper03!~peper03@mythtv/developer/peper03) has quit (*.net *.split) | |
| [15:21:46] | dekarl (dekarl!~dekarl@p4FCEE5EA.dip0.t-ipconnect.de) has quit (*.net *.split) | |
| [15:21:48] | monkeypet (monkeypet!~monkeypet@c-24-6-133-78.hsd1.ca.comcast.net) has quit (*.net *.split) | |
| [15:21:48] | Captain_Murdoch (Captain_Murdoch!~cpinkham@c-67-170-1-245.hsd1.wa.comcast.net) has quit (*.net *.split) | |
| [15:29:24] | peper03 (peper03!~peper03@mythtv/developer/peper03) has joined #mythtv | |
| [15:30:12] | Casper0082 (Casper0082!~Casper@pool-71-185-186-159.phlapa.fios.verizon.net) has joined #mythtv | |
| [15:32:27] | Captain_Murdoch (Captain_Murdoch!~cpinkham@c-67-170-1-245.hsd1.wa.comcast.net) has joined #mythtv | |
| [15:32:28] | Captain_Murdoch (Captain_Murdoch!~cpinkham@c-67-170-1-245.hsd1.wa.comcast.net) has quit (Changing host) | |
| [15:32:28] | Captain_Murdoch (Captain_Murdoch!~cpinkham@mythtv/developer/CaptainMurdoch) has joined #mythtv | |
| [15:35:27] | monkeypet (monkeypet!~monkeypet@c-24-6-133-78.hsd1.ca.comcast.net) has joined #mythtv | |
| [15:47:00] | pvr4me (pvr4me!~pvrme@d24-150-182-175.home.cgocable.net) has joined #mythtv | |
| [15:49:39] | pvr4me (pvr4me!~pvrme@d24-150-182-175.home.cgocable.net) has quit (Client Quit) | |
| [16:15:13] | andreaz (andreaz!~Andreaz@tmo-106-181.customers.d1-online.com) has joined #mythtv | |
| [16:45:25] | FabriceMG (FabriceMG!~Thunderbi@217.112.59.207) has quit (Quit: FabriceMG) | |
| [16:54:20] | NightMonkey (NightMonkey!~NightrMon@pdpc/supporter/professional/nightmonkey) has joined #mythtv | |
| [16:58:43] | monkeypet: | jKlaus what's the question? |
| [16:59:24] | monkeypet (monkeypet!~monkeypet@c-24-6-133-78.hsd1.ca.comcast.net) has quit (Quit: BitchX Official IRC Channel -- #BitchX on EFNet) | |
| [16:59:41] | monkeypet (monkeypet!~monkeypet@c-24-6-133-78.hsd1.ca.comcast.net) has joined #mythtv | |
| [17:03:32] | dekarl1 is now known as dekarl | |
| [17:29:33] | pvr4me (pvr4me!~pvrme@d24-150-182-175.home.cgocable.net) has joined #mythtv | |
| [17:29:57] | monkeypet: | I deployed devel/027candidates on my backend/frontend. |
| [17:33:45] | monkeypet: | I'll watch LiveTV today all day and see how it looks. |
| [17:35:40] | pvr4me (pvr4me!~pvrme@d24-150-182-175.home.cgocable.net) has quit (Quit: Colloquy for iPhone - http://colloquy.mobi) | |
| [17:39:13] | pvr4me (pvr4me!~pvrme@24.150.182.175) has joined #mythtv | |
| [17:42:30] | pvr4me (pvr4me!~pvrme@24.150.182.175) has quit (Client Quit) | |
| [17:46:50] | Tobbe5178 (Tobbe5178!~asdf@h104n2-sv-a13.ias.bredband.telia.com) has joined #mythtv | |
| [17:48:31] | andreaz1 (andreaz1!~Andreaz@tmo-109-173.customers.d1-online.com) has joined #mythtv | |
| [17:50:29] | andreaz (andreaz!~Andreaz@tmo-106-181.customers.d1-online.com) has quit (Ping timeout: 252 seconds) | |
| [17:52:15] | rsiebert (rsiebert!~quassel@g231187067.adsl.alicedsl.de) has quit (Quit: No Ping reply in 180 seconds.) | |
| [17:52:35] | rsiebert (rsiebert!~quassel@g231187067.adsl.alicedsl.de) has joined #mythtv | |
| [17:54:54] | Anssi (Anssi!~anssi@mandriva/developer/anssi) has quit (Quit: leaving) | |
| [17:55:17] | Anssi (Anssi!~anssi@mandriva/developer/anssi) has joined #mythtv | |
| [18:00:33] | stuartm: | apparently getting a free upgrade to 50Mbps on my cable, was pretty pleased until I checked the schedule online and discovered it wouldn't be for another year ... |
| [18:02:38] | brfransen (brfransen!~brfransen@64.179.169.226) has quit (Ping timeout: 264 seconds) | |
| [18:05:11] | natanojl (natanojl!~jonatan@mythtv/developer/natanojl) has joined #mythtv | |
| [18:23:31] | andreaz1 (andreaz1!~Andreaz@tmo-109-173.customers.d1-online.com) has quit (Read error: Connection reset by peer) | |
| [18:23:35] | brfransen (brfransen!~brfransen@64.179.169.226) has joined #mythtv | |
| [18:32:18] | joki (joki!~joki@p54862E83.dip0.t-ipconnect.de) has quit (Ping timeout: 240 seconds) | |
| [18:38:52] | joki (joki!~joki@p5486065C.dip0.t-ipconnect.de) has joined #mythtv | |
| [18:44:07] | TheCrasher (TheCrasher!~TheCrashe@p5DCE4DC5.dip0.t-ipconnect.de) has joined #mythtv | |
| [18:46:01] | natanojl (natanojl!~jonatan@mythtv/developer/natanojl) has quit (Ping timeout: 240 seconds) | |
| [18:52:45] | rsiebert_ (rsiebert_!~quassel@g231087064.adsl.alicedsl.de) has joined #mythtv | |
| [18:56:02] | rsiebert (rsiebert!~quassel@g231187067.adsl.alicedsl.de) has quit (Ping timeout: 264 seconds) | |
| [19:00:24] | monkeypet: | comcast calls me to try to get me to sign up for cabletv every few weeks. |
| [19:20:32] | andreaz (andreaz!~andre_000@p5DCA3126.dip0.t-ipconnect.de) has joined #mythtv | |
| [19:21:37] | dekarl: | stuartm, another use case that just came up with W+WAF... A playlist editor for the frontends web frontend would be nice to switch from "currently playing episode xx from a list of kids stuff in random play" to "after the current episode switch to a movie for woman+kid" when playing back recordings. |
| [19:24:39] | wagnerrp (wagnerrp!~wagnerrp_@mythtv/developer/wagnerrp) has quit (Read error: Connection reset by peer) | |
| [19:25:12] | cesman (cesman!~cesman@pool-173-60-115-40.lsanca.fios.verizon.net) has joined #mythtv | |
| [19:25:12] | cesman (cesman!~cesman@pdpc/supporter/professional/cesman) has joined #mythtv | |
| [19:25:12] | cesman (cesman!~cesman@pool-173-60-115-40.lsanca.fios.verizon.net) has quit (Changing host) | |
| [19:25:17] | wagnerrp (wagnerrp!~wagnerrp_@mythtv/developer/wagnerrp) has joined #mythtv | |
| [19:33:34] | Warped (Warped!~Warped@108-85-161-113.lightspeed.cicril.sbcglobal.net) has quit (Quit: ChatZilla 0.9.90.1 [Firefox 27.0.1/20140212131424]) | |
| [19:42:38] | Warped (Warped!~Warped@108-85-161-113.lightspeed.cicril.sbcglobal.net) has joined #mythtv | |
| [19:43:33] | brfransen (brfransen!~brfransen@64.179.169.226) has quit () | |
| [19:52:33] | stuartm: | dekarl: for the frontend playlist, or the streaming video? |
| [19:53:35] | stuartm: | dekarl: mind adding that (and anything else), here – https://code.mythtv.org/trac/wiki/WebFrontendWishlist |
| [20:28:27] | monkeypet: | jya: so far with devel/027candidates, LiveTV is perfect. |
| [20:29:03] | monkeypet: | jya: watched many hours of TV so far. |
| [20:30:32] | monkeypet: | jya: the only glitches were due to my bad OTA reception. |
| [20:43:57] | dekarl: | stuartm: I was thinking of the frontends playlist. |
| [20:45:02] | stuartm: | dekarl: ok, tis a good idea, will take a look |
| [20:45:16] | stuartm: | if it's added to the wishlist I won't forget about it :) |
| [20:46:19] | dekarl: | ok, I wanted to see how you (and others) like the idea first ;) |
| [20:46:36] | stuartm: | it's probably something that should get added to both the backend web interface and the frontend web interface |
| [20:47:55] | dekarl: | we could start with adding a link "play on frontend X" next to the "play via HLS". That's one of the functions I like in mythweb |
| [20:48:00] | intok (intok!~junko@h69-131-100-177.nwblwi.broadband.dynamic.tds.net) has joined #mythtv | |
| [20:48:34] | intok (intok!~junko@h69-131-100-177.nwblwi.broadband.dynamic.tds.net) has left #mythtv () | |
| [20:50:20] | stuartm: | dekarl: yeah, it was actually my plan that the 'Watch' link would bring up a popup menu asking whether it should be played in the browser, or on a frontend |
| [20:54:16] | warpme (warpme!~piotro@89-79-250-31.dynamic.chello.pl) has joined #mythtv | |
| [20:54:56] | JohnBergqvist (JohnBergqvist!~JohnBergq@79-71-116-191.dynamic.dsl.as9105.com) has joined #mythtv | |
| [20:56:23] | warpme: | stuartm: regarding 40s to transcode 10s material via HLS: on what CPU is it? |
| [20:57:19] | JohnBergqvist: | Since there's no-one active on the LIRC support channel and there's more likely to be UK users here, would anyone like to help me with troubleshooting an LIRC problem? |
| [20:57:31] | JohnBergqvist: | Also in other news, i shall be making the jump to satellite on Saturday :D |
| [20:58:03] | stuartm: | warpme: AMD 4850e (2x 2.5Ghz) – not exactly cutting edge :) |
| [21:01:02] | brfransen (brfransen!~brfransen@64.179.169.226) has joined #mythtv | |
| [21:01:23] | warpme: | stuartm: I had this CPU in previous incarnation of my BE. I remember it was able to transcode interlaced 1080i to 360p on realtime. is it also Your experience? |
| [21:02:08] | stuartm: | haven't actually tried 1080i to 360p, just 1080i to 720p |
| [21:03:17] | stuartm: | this is with deinterlacing enabled and the preset at 'veryfast', results when I originally tested ultrafast were terrible, haven't retried since enabling deinterlacing |
| [21:04:50] | JohnBergqvist (JohnBergqvist!~JohnBergq@79-71-116-191.dynamic.dsl.as9105.com) has quit (Quit: Leaving) | |
| [21:08:50] | stuartm: | ok, going to 360p is much faster than realtime |
| [21:08:55] | warpme: | stuartm: ahh. I see. 720p is no go with this CPU according to my experience. I think issue isn't in IPC but rather with absence of AVX extensions. Latest x264 can use latest greatest achievements in CPU extensions and I believe this is why IVB i3 can do 1080i->720p at 130% of realtime while AMD with the same clock offers only 30–50%. How "fresh" x 264 are You using? |
| [21:13:16] | stuartm: | looks like 0.125 |
| [21:13:30] | stuartm: | looks pretty old :/ |
| [21:16:35] | jya: | google is down for me... amazing... |
| [21:19:20] | warpme: | stuartm: interesting things are going to be with QSV (Intel) and AMD VCE. VCE is particularly interesting as it will land into Mesa and can produce directly TS. BTW: latest Intel intel-vaapi lib provides already API to HW encode to H264. Rough benchmarks are: i3 QSV is 3–5x faster that high-end i7 |
| [21:22:02] | jya (jya!~jyavenard@mythtv/developer/jya) has quit (Quit: jya) | |
| [21:24:03] | jya (jya!~jyavenard@mythtv/developer/jya) has joined #mythtv | |
| [21:26:03] | warpme: | stuartm: if You have spare time – pls try upgrade x264 to latest one (http://git.videolan.org/?p=x264.git;a=summary) |
| [21:26:52] | stuartm: | jya: lucky you |
| [21:28:00] | stuartm: | warpme: will give it a go, although I expect realtime transcoding to 720p will remain impossible with this CPU |
| [21:29:00] | stuartm: | means that I may need to defaulting to 720p for the WebFrontend, 360p may be the better all-round default |
| [21:29:10] | warpme: | jya: hats off to Your work about LiveTV. I think this is major thing despite believe that MythTV is PVR. My observations are that even if myth shines as PVR – initial adoption is driven by LiveTV experience... |
| [21:30:55] | stuartm: | I'm actually pretty impressed to discover than transcoding 1920x1080 to 360p is as fast as transcoding 720x576 to the same res |
| [21:34:15] | pvr4me (pvr4me!~pvrme@207.164.79.55) has joined #mythtv | |
| [21:36:31] | warpme: | stuartm: I think this is v.good assumption. Lets look on use cases: I think major use case for webfrontent playback capability is "out-of-lan" scenario. We can safely assume in such case display device will have limited res (phone/tables/small laptop) – so 360p will be OK for 90% of cases. My tests with torc on iPad4 shows 360p is relatively way enough regarding picture quality. I think much higher impact on experience has ability to start |
| [21:36:32] | warpme: | playback on arbitrary point. |
| [21:38:06] | stuartm: | we default to 360p for non-local clients, just thought we might be able to use 720p for local clients, but if realtime transcoding to 720p is a stretch for older hardware then it will backfire |
| [21:39:48] | stuartm: | 360p isn't pretty reasonable on a small screen, even a high res one like my tablet and it's passable on a desktop – certainly no worse than some broadcast SD |
| [21:40:11] | stuartm: | which is a surprise, wasn't expecting it to be OK :) |
| [21:40:34] | pvr4me (pvr4me!~pvrme@207.164.79.55) has quit (Quit: Colloquy for iPhone - http://colloquy.mobi) | |
| [21:42:30] | warpme: | stuartm: right. I would say: mandatory transcode practically excluding local-client use-case as 360p is too low res for HD content and 720p is too demanding – assuming current average HW. Adding multi-user only multiplying problem :-( |
| [21:43:48] | warpme: | stuartm: right. 360p is from information BW perspective equal to 720i which is more than 576i/480i... |
| [21:46:21] | dekarl: | Why is 360p to low res? I'd say sourcing 360p from HD H.264 instead of SD MPEG-2 is better because you have less artefacts to encode, so more bits go towards actualy picture signal instead of encoding noise |
| [21:46:36] | warpme: | warpme: so if codec is good in smoothing quantisation effects and decoder is doing good deblocking – results might be pretty good! |
| [21:47:15] | stuartm: | right, but it's misleading to say that 360p > 576i purely because of a higher framerate |
| [21:47:31] | pvr4me (pvr4me!~pvrme@207.164.79.55) has joined #mythtv | |
| [21:48:29] | dekarl: | oh, do we deinterlace 50i to 25p or 50p in the HLS encoder? |
| [21:49:13] | warpme: | stuartm: right. thats why I prefer to operate on information BW :-) |
| [21:49:27] | stuartm: | dekarl: 25fps, wanted to keep the framerate consistent with the source |
| [21:49:44] | stuartm: | plus lower framerate = better quality at the same bitrate |
| [21:49:54] | pvr4me (pvr4me!~pvrme@207.164.79.55) has quit (Client Quit) | |
| [21:50:36] | stuartm: | 50i = 25fps, 2 fields per frame |
| [21:50:37] | dekarl: | stuartm: sounds sane. I was under the impression that we have some places in our code base where frames/field per second are mixed up ... |
| [21:51:37] | stuartm: | dekarl: quite possible |
| [21:51:41] | dekarl: | not only https://code.mythtv.org/trac/ticket/12047 but other places, too |
| [21:53:45] | stuartm: | in the case of the transcoder, or HLS specifically, we don't use any of our video output deinterlacing but libav's deinterlacing filters, so any bugs in our code would be bypassed |
| [21:57:01] | stuartm: | I think it would be difficult to use our deinterlacers, they are in the video output code and the transcoder never uses that, it just decodes the frame then hands it straight back to ffmpeg to encode again |
| [21:57:42] | warpme: | dekarl: I think deinterlace is reasonable. Keeping 25p allows not to loose any information. How 25p is displayed is another thing (with 50 frames + motion interpolation or 25 without). I believe trade-off here is spatial resolution vs fast moving content vibrancy. I personally prefer higher spatial resolution (25p) over lower vibrancy (50i). But this is of course mather of taste :-) |
| [21:57:44] | stuartm: | it would be nice if the transcoder could use hw decoding and encoding, but well beyond the scope of what I'm doing |
| [21:59:23] | dekarl: | warpme: while I agree that this is something that is relevant for big screens I'm not so sure for phones and tables :D |
| [22:00:00] | stuartm: | btw, 40Kbps audio is awful, made it the default for 360p HLS, but I really should have tested it ... |
| [22:00:07] | stuartm: | back up at 64Kbps now |
| [22:00:35] | dekarl: | stuartm which encoder is that? |
| [22:00:47] | stuartm: | mp3 |
| [22:00:53] | stuartm: | so lame |
| [22:01:02] | stuartm: | might be OK for AAC |
| [22:01:26] | dekarl: | just wondering because ffmpeg has a long list on how to choose the AAC/MP3 encoder to use in order of quality http://trac.ffmpeg.org/wiki/AACEncodingGuide |
| [22:01:28] | stuartm: | but we're using mp3/lame for some reason |
| [22:01:49] | dekarl: | not the link I'm looking for but has the list at the end |
| [22:02:04] | dekarl: | oh, what reason is it? |
| [22:04:04] | stuartm: | dekarl: no idea, Captain_Murdoch may know |
| [22:04:54] | dekarl: | http://trac.ffmpeg.org/wiki/GuidelinesHighQualityAudio says "Even though AAC should generally give better quality than MP3 at the same bitrates, libmp3lame should give better quality output than libfaac." and the only encoder that is better would be "libfdk_aac" whatever that is |
| [22:05:30] | warpme (warpme!~piotro@89-79-250-31.dynamic.chello.pl) has quit (Quit: warpme) | |
| [22:06:15] | dekarl: | but it appears to be at higher bitrates then 64kbit/s |
| [22:07:16] | Captain_Murdoch: | mp3 is lowest common denominator, some devices didn't support AAC audio I believe, so we defaulted to MP3 and let users override that if they wanted AAC. |
| [22:08:01] | Captain_Murdoch: | I haven't tried anything as low as 40 I believe. |
| [22:09:13] | Captain_Murdoch: | there was also the issue of requiring libfaac. at the time, we were steered away from the ffmpeg built-in AAC encoder by the guy who wrote it or was maintaining it. it was marked experimental. the story might be different now, so we could use the built in AAC encoder possibly. |
| [22:10:55] | stuartm: | documentation still describes it as experimental |
| [22:11:50] | Captain_Murdoch: | with the deinterlacing, I'm pretty sure we used to de-interlace by default when transcoding and commercial flagging, then somewhere along the line that changed and they weren't receiving deinterlaced data anymore. |
| [22:12:53] | Captain_Murdoch: | internal AAC worked well in my testing, but he said that it had issues and asked us not to use it by default, so we obliged. |
| [22:15:21] | stuartm: | looks like we'll use AAC if libfaac is compiled in, otherwise mp3 unless lame is missing, in which case we fall back to AAC with the internal encoder |
| [22:15:40] | stuartm: | so faac > lame > internal aac |
| [22:16:23] | Captain_Murdoch: | yeah. by default we don't configure with libfaac, so we fall through to mp3. so if the user does configure with libfaac, it enables faac. then on the odd case that we compiled without lame we fall through to internal. |
| [22:16:27] | stuartm: | we don't let the consumer of the content pick the codec, although it can be forced via a setting |
| [22:16:51] | Captain_Murdoch: | right, it's not in the API, just a setting I think IIRC. I wanted to add that to the API |
| [22:17:01] | stuartm: | it may be better in the API |
| [22:17:07] | Captain_Murdoch: | :) |
| [22:17:09] | Captain_Murdoch: | yep |
| [22:17:55] | stuartm: | similarly we might want to let the H.264 profile to be specified via the API |
| [22:18:53] | stuartm: | and ogg support might be nice :) Although I've no idea if any HLS clients support it with ogg – possibly chrome/android |
| [22:19:18] | Captain_Murdoch: | I agree. I had shifted focus to the on-demand stuff instead of adding enhancements to the existing code. |
| [22:22:44] | stuartm: | scratch that, Chrome doesn't support HLS (outside Android), I though it did, but it looks like they've instead been working with the MediaSource API instead |
| [22:23:52] | stuartm: | ... time for bed |
| [22:25:21] | jya: | stuartm: did you correct the playlist file thing you mentioned yesterday? where the file is erased and re-created after each new segment? |
| [22:26:53] | stuartm: | jya: turns out that wasn't exactly what was happening, combination of two identical transcode jobs trying to write to the same playlist file and me making an incorrect assumption while mis-reading the code |
| [22:27:05] | TheCras2 (TheCras2!~TheCrashe@p5080735D.dip0.t-ipconnect.de) has joined #mythtv | |
| [22:27:18] | jya: | so still a minute or so to start ? |
| [22:28:26] | stuartm: | instead what we do is create the playlist from scratch in a temp file, then copy that into place overwriting the old copy – still not particularly efficient but not what I thought was happening |
| [22:29:39] | stuartm: | jya: seems to depend on both the player and how fast the transcoding proceeds – we don't produce a playable playlist until the first segment has completed transcoding, for some players a single segment seems to be enough, others want more |
| [22:30:07] | stuartm: | and some players are failing to check for updates to the playlist frequently |
| [22:30:26] | TheCrasher (TheCrasher!~TheCrashe@p5DCE4DC5.dip0.t-ipconnect.de) has quit (Ping timeout: 252 seconds) | |
| [22:31:12] | stuartm: | so even if the playlist is updated with a playable segment, the player might take a while before re-requesting it |
| [22:31:58] | jya: | what I'll try to do today if I have the time, is modify the ffmpeg used by AirVideo and see which arguments it's using for transcoding on either iphone or ipad... I've said I was goign to do that years ago.. never got to do it. |
| [22:32:16] | jya: | With AirVideo, playback starts instantly, and the quality is always outstanding |
| [22:32:36] | jya: | he does have some patches applied on ffmpeg, but I have the source of those, so can check |
| [22:32:43] | stuartm: | transcoding to 360p I can get something playable in 6–7 seconds, but with JW Player if I hit play before than time it seems to enter a permanent 'failed' state and I need to refresh the page |
| [22:33:12] | Captain_Murdoch: | is segment size still 10s or is it at 4s now? |
| [22:33:15] | stuartm: | HTML 5 video player on android will actually put up a spinning 'buffering' type animation for that period |
| [22:33:23] | jya: | with AirVideo, I press play, and within 1–2s playback starts... always been the case even when I was running it on a core 2 duo server |
| [22:33:29] | stuartm: | Captain_Murdoch: still 10 |
| [22:33:39] | jya: | Captain_Murdoch: AirVideo is 4s |
| [22:33:45] | Captain_Murdoch: | you could try lowering that, I can't remember if the spec allows lower. |
| [22:33:52] | jya: | I thought you reduced it on mythtv ? |
| [22:33:56] | Captain_Murdoch: | I think we did try, but migth have run into issues with some players. |
| [22:34:03] | Captain_Murdoch: | I cant' recall if it was left or bumped back up. |
| [22:34:05] | jya: | Captain_Murdoch: there's nothing in the spec about how long a segment should be |
| [22:34:15] | jya: | just that the segment shouldn't be more than what's in the playlist |
| [22:34:31] | stuartm: | jya: would be good to know exactly what encoding parameters he's using, resolution, profile, whether that includes deinterlacing and the audio-only stream |
| [22:34:35] | jya: | it can be less , and I've seen it often where the actual segment is shorter than what was in the playlist |
| [22:35:02] | jya: | yeah... having said that, I never tried to play an interlaced video with AirVideo |
| [22:35:18] | jya: | I typically watch my TV series at night, that I downloaded ... some places |
| [22:37:26] | stuartm: | spec only says that "A typical target duration is 10 seconds" |
| [22:37:56] | jya: | the longest I've seen is with France24 at 20s |
| [22:38:00] | jya: | Al Jazeera is 16s |
| [22:38:43] | stuartm: | I'll play with a shorter duration, that would help with the initial start speed, but not much else |
| [22:39:18] | jya: | from what I've seen with AV, thjey use 4s and that works very well |
| [22:40:26] | warpme (warpme!~piotro@89-79-250-31.dynamic.chello.pl) has joined #mythtv | |
| [22:42:15] | stuartm: | the actual speed of the transcoding is still the bottleneck for most systems, possibly some tweaking that can be done there, but I'm not sure what I've not already tried |
| [22:42:29] | doev (doev!~doev@p4FD42F96.dip0.t-ipconnect.de) has quit (Ping timeout: 272 seconds) | |
| [22:45:51] | stuartm: | one thing we might want to do is support cutlists – seems a waste of processing time to transcode the adverts |
| [22:49:35] | stuartm: | noticed some audio 'squelching' at segment boundaries that needs fixing |
| [22:49:50] | warpme: | stuartm: Here in Poland much more annoying is 3–5min of adverts at beg of program. Currently HLS users are forced to show them as there is no possibility to start HLS from other point than beg.of show. This is VERY annoying for myth users who since years have comfort of skipping adverts... |
| [22:50:24] | stuartm: | and the cropping of the 8 lines from 1088 content with swscale doesn't work, going to try the crop filter instead |
| [23:23:40] | warpme (warpme!~piotro@89-79-250-31.dynamic.chello.pl) has quit (Quit: warpme) | |
| [23:29:39] | dekarl1 (dekarl1!~dekarl@p4FCEFF43.dip0.t-ipconnect.de) has joined #mythtv | |
| [23:32:02] | dekarl (dekarl!~dekarl@p4FCEE5EA.dip0.t-ipconnect.de) has quit (Ping timeout: 264 seconds) | |
| [23:43:06] | warpme (warpme!~piotro@89-79-250-31.dynamic.chello.pl) has joined #mythtv | |
| [23:43:26] | aberrios (aberrios!~aberrios@195.130.201.200) has quit (Ping timeout: 264 seconds) | |
| [23:44:57] | jpharvey_ (jpharvey_!~jpharvey@host109-148-236-119.range109-148.btcentralplus.com) has joined #mythtv | |
| [23:45:10] | jpharvey (jpharvey!~jpharvey@host109-148-236-119.range109-148.btcentralplus.com) has quit (Ping timeout: 272 seconds) | |
| [23:45:11] | jpharvey (jpharvey!~jpharvey@host109-148-236-119.range109-148.btcentralplus.com) has joined #mythtv | |
| [23:45:13] | jpharvey (jpharvey!~jpharvey@host109-148-236-119.range109-148.btcentralplus.com) has quit (Ping timeout: 272 seconds) | |
| [23:45:19] | warpme: | jya: one question regarding LiveTV: is watching programs saved in LiveTV group from last LiveTV chains supported?. When I try to play any of such recordings – I'm getting 30sec or so black screen and return to MainMenu. |
| [23:51:14] | aberrios (aberrios!~aberrios@195.130.201.200) has joined #mythtv | |
| [23:51:52] | warpme (warpme!~piotro@89-79-250-31.dynamic.chello.pl) has quit (Quit: warpme) | |
IRC Logs collected by
BeirdoBot.
Please use the above link to report any bugs.