[00:03] aaaaaaaaand updated it again [00:03] proper optimization, bugs gone [00:08] so i'm up to 9260 on g4tv.com [00:09] also the number of videos is only 1058 [01:32] does anyone know of a way to fix a broken flv file [01:32] on linux [01:32] i tryed ffmpeg and mencoder and there not working [01:33] do you know how it's brokne? [01:33] broken? [01:33] it stops like in the middle [01:36] ah, truncated [01:36] ? [01:36] you mean the file is truncated [01:36] shorter than it was supposed to be [01:37] yes [01:37] not much you can do to fix it then, but you might try rebuilding the index [01:37] -idx I think, at least for mplayer [01:37] not sure about mencoder; it's been a while [01:38] barring that you might use mencoder to copy just the first N seconds from the file [01:38] that should give you a new file that isn't corrupt, and has very nearly the entire content of the file [01:39] but i want to fix it [01:39] not just part of the file [01:39] some files only play like half a second [01:42] right [01:42] the problem with a truncated file is twofold [01:43] there is an index at the beginning that tells the player where the I-frames are [01:43] the key frames [01:43] with a truncated file, a lot of those don't actually exist [01:43] second, the truncation rarely happens to line up with the frame boundaries in the stream [01:44] nor does it usually line up with the interleave boundaries, where video and audio data are woven together [01:45] and you can't just make up information to fill out the end of the file [01:45] so keep the original, but copy out as much of the video as you can with mencoder to produce a file that plays for just as long (less a tenth of a second perhaps), but isn't corrupt [01:47] balrog_ thinks its bad muxed [01:47] may be, I'd have to look at it [01:48] i hope so [01:48] ah, if that's the case then mencoder won't be able to copy anything out anyway :) [01:49] then how do you fix it if mencoder and ffmpeg will not? [01:58] well, this is certainly interesting [01:59] I just fixed a bug in the core of beautifulsoup [01:59] inadvertently [02:00] i did find video file that just skips ahead by 3mins [02:01] but its only a problem when your trying to skim the video [02:01] godane: fun [02:01] the full video is there [02:01] yeah lots of videos have seeking issues [02:01] ah, that's probably just a bad index. should be easy to fix, since mplayer/mencoder can just scan for I-frames [02:14] how do you scan for i-frames [02:16] some of the videos just stop working at 20secs [02:16] *20secs before it ends [02:18] does it end abruptly as far as the content of the video, I've seen video files before where the length is detected wrong so they play fine, they're just shorter than the reported length [02:19] some of the files are alot bigger then there should be [02:20] if it was wrong length then it should have been smaller [02:20] for example the crazyadam clip was around 18mb [02:20] when trying to get the video out with ffmpeg i only got 8.9mb [02:23] i have found 17 bad videos [02:24] probably old flv [02:24] just like old realmedia - very messy [02:26] then how do you fix the old flv? [02:26] ill have to look at it :/ [02:28] godane: mplayer -idx filename [02:29] that doesn't fix it [03:04] some videos get pink spots when they crash [03:13] colored spots (blocks) just means that there was data missing and the player filled it in [03:18] i think the time jumps [03:19] it will say like it 3:00 then when the problem happens it will stop and got 72:00:00 or something [03:33] this is a first [03:33] i found one video that doesn't skip anywhere near the end but it will still play all of it [03:42] you can't seek near the end? [03:42] the index is busted, or it just doesn't have any i-frames there [03:42] db48x: any news on how to fix the files? [03:43] i think just seeking is not possible with some [03:43] that's quite possible [03:43] some player software can only seek to an i-frame [03:43] basically, an i-frame is the data for a single frame of video, with no reliance on previous frames [03:44] but some just will not get past a point [03:44] b-frames and p-frames contain information about what has changed since the last frame [03:44] godane: even a single bit error can cause a player to bail out and not play the rest of the file [03:44] ok [03:45] or it can cause the player to start hunting through the rest of the file for a place to pick back up [03:45] I had one a few weeks ago that skipped about 15 minutes of plot thickening and jumped right into the denoument [03:50] but if I very carefully seeked to just after the spot where it jumped, then it played those 15 minutes just fun [04:02] found a episode of arena and will wheaton is in it [04:03] heh [04:03] nemesis! [04:03] WHEATON!!!! [04:44] :) [04:45] gn [05:50] i got doctor who magazine off of my drive [05:50] i have burned it a few months back [06:05] godane: if mplayer -forceidx makes it seek properly, then there's your problem - bad index [08:09] chronomex: most of my problems is the files have errors in them [08:11] 0x0089633f: error E12051: previous tag size should be 1389, -302199066 found instead [08:11] 0x0089633f: error E30013: unknown tag type 86 [08:12] 0x00896340: fatal F20016: tag body length (9683438 bytes) exceeds file size [08:12] 3 error(s), 0 warning(s) [08:12] thats the out put with flvmeta -C xp7030crazyadam_flv.flv [08:21] oh [08:31] Wil had a major falling out with Arena [08:31] Well worth keeping that [08:57] how much has someone downloaded before they was warn of going over the cap? [08:58] also i'm running the terminal download parts about 400 to 500 kbytes [08:58] *3 terminals [08:58] trying to find the big videos too [09:03] Just keep going man :) [09:03] Someone said they got calls for using terabytes of bandwidth [09:35] ok [09:35] i'm only have about 163gb [09:41] i think i will need to by another hard drive next month at this rate [09:43] I just had disk-full today :( [09:44] i had to move some 1940s scifi magazines [09:44] i didn't have space for doctor who magazines but i have that burned so i deleted it [09:45] *burned it to bluray [10:31] new telegeography submarine international fiber map is out... wish I could afford to blow the $250- for the wall map. nice. interactive version: http://submarine-cable-map-2013.telegeography.com/ [10:32] turnkit: grab the images and stitch it yourself :) http://b.tiles.telegeography.com/maps/submarine-cable-map-2013/6/33/23.png [10:33] and take it to fedex/Kinko's for printing? hmm... [10:33] :) [10:33] i printed some 60x60 cm poster once with a map from http://www.davidrumsey.com/ i downloaded that way [10:34] what's the easy way to get all the pieces? [10:35] i would think a loop to a index.txt [10:35] using seq like 1 to 30 or something [10:36] yeah, seq [10:36] top corner is 0,0 bottom is 64,64 or so [10:36] no, use proper modern bash! [10:36] wget http://b.tiles.telegeography.com/maps/submarine-cable-map-2013/6/{0..64}/{0..64}.png or so [10:37] one line shell code? hmm... what would that be... [10:37] ah [10:37] will that really work? [10:37] just like that [10:37] ? [10:37] yes [10:37] bash will expand it out for you [10:37] not sure about other shells. [10:37] that only downloads the snippets though, stiching is easy with image/graphicsmagick [10:37] and how about stitching... any suggestions [10:38] ah -- answers before I ask them. I like that [10:38] ;) [10:38] I'm skimming.... http://linuxandfriends.com/2009/07/26/how-to-stitch-pictures-using-imagemagick/ [10:40] okay, I have to actually think here... using the "-append" and "+append" w/ bash... maybe can do in two passes... ? group all horizontal pieces in one pass, then all those in the second [10:40] ? [10:40] sounds valid [10:48] oh that bash line was missing wget putting the images into subdirs :) [10:52] lol... I don't have a linux set up.. (I know, right?) tried to use "bash for windows" ... didn't expand properly. [10:52] bash$ wget http://b.tiles.telegeography.com/maps/submarine-cable-map-2013/6/{0..64}/{0..64}.png --04:50:06-- http://b.tiles.telegeography.com/maps/submarine-cable-map-2013/6/%7B0..64%7D/%7B0..64%7D.png => `@7B0..64@7D.png' [10:53] Guess I need to set up a little Mint box or something... Should do that anyway I guess. [10:53] thanks for the direction. [10:53] to all [10:56] 'tiss a really nice map though, right? I like on of the scales at the bottom showing worldwide bandwidth in thickness... many countries apparently w/ 5000 GB+. [10:58] Well, that's quite enough Cleveland Brown Show [11:10] SketchCow: do you know how to fix bad flv files? [11:11] i some of the flvs with errors and just will stop short of playing the full then [11:11] *thing [11:11] sometimes it will not play at all [11:12] example: vids.g4tv.com/videoDB/015/507/video15507/xp7030crazyadam_flv.flv [11:12] it stops at 1:18 in to it [11:12] No, sadly. Not my thing. [11:13] will your getting all of it [11:13] even if turns out to be garage file in some case [11:29] godane -- if you can automatically detect the bad files perhaps losslessly remux to fix -- this might not stop the corruption from being heard/seen but could allow playthrough still, despite the apparent glitching [11:29] see http://stream-recorder.com/forum/fix-repair-flv-files-make-flv-files-t5751.html?s=0380ac4d809bf9346ed5d907a36ffae6& [11:46] some lost big thinkers video: https://archive.org/details/g4tv.com-video754 [11:47] i say lost cause 1 there is only the new 2.0 big thinkers full episodes i can find [11:47] 2 is cause anything in the double or triple digit is just gone but this i think [11:48] Fuck, I just get so annoyed when people nag about IA + robots.txt shit [11:48] * ersi shakes fist at jwz [14:29] dammit imgur We're really sorry, but uploading from the Tor network has been disabled. Please disable Tor and try uploading again. [14:35] :< [14:36] i wanted to share my fabulously failed stitching attempt :) [14:39] ersi: Wanna know an easy way to make people stop complaining? [14:41] ah, its because find does not sort numerically [14:43] DoubleJ: Kill'em? [14:43] Even easier than murder :P [14:44] threw scantools onto github: https://github.com/joepie91/scantools [14:51] success, 16384x16384 png at 94MB [14:52] that is some sloppy cartography :) [14:58] * joepie91 spams github and twitter with code [14:59] Hah, and I just did the opposite... I grabbed a code from github :) [14:59] and i contemplated uploading some but thought meh [15:08] just added https://github.com/joepie91/anonnews2 [15:10] joepie91: What does it do? [15:11] norbert79: it's the source for http://anonnews.org/ [15:11] :P [15:11] Nice [15:12] Nice work [15:13] joepie91: mine isn't something this extraordinary, i just did a clean install of phpaste and did some minor modifications: https://telehack.tk/pastebin [15:13] norbert79: :) [15:14] also, I am really annoyed that github only seems to count things on the master branch [15:14] for the account overview page [15:14] as if my commits on the develop branch are somehow not relevant... [15:14] Yeah, makes more sense [15:15] ut sometimes I think I have wasted way too much time with DOS based games or gaming at all... Duke 3D BUILD was fun, but now I wish I have coded more :) [15:15] *But [15:15] Though modifying the engine a bit through the CON files and adding new things was also nice too [15:18] something I've noticed is that the WTFPL is starting to be used more and more recently [15:19] oh, right, yeah, I prefer GPL CC or BSD [15:19] but WTFPL is a good idea too [15:23] hmm, why BSD over WTFPL? :P [16:00] joepie91: once there's a merge from the branch into master, those commits will start counting [16:00] db48x: why [16:00] how does it make sense to not count them straight away [16:00] commits are commits [16:07] http://i.imgur.com/Q3lKIr1.jpg [16:15] joepie91: it's just an optimization [16:15] schbirid: great find [16:15] db48x: optimizations that negatively affect practical use, are evil [16:20] heh [16:21] the argument is that unless it's shipped it doesn't count, and everyone ships from the master branch [16:21] that's bullshit [16:21] well, yes [16:21] especially when it affects your 'spree' count [16:21] although it is true that code doesn't count unless it's shipped [16:22] because who commits to a master branch every day? [16:22] branches just don't correspond to shipping [16:22] which thereby defeats the point of counting a spree [16:22] .. [16:22] db48x: the problem is that they are trying to implement two different paradigms [16:22] that horribly clash with each other [16:22] on the one hand, they try to make you keep track of every tiny commit you make - hence the "spree" counting [16:22] on the other hand, they say code doesn't count until it's in master branch and therefore it's not counter [16:22] counted * [16:23] either code counts when committed or it doesn't, can't have both [16:23] yep [16:23] and again [16:23] who the hell commits to master daily [16:23] so, is this part of github open source? [16:23] doubt it [16:34] norbert79: neat, that's my phpaste version ^_^ [16:37] more updates http://www.newyorker.com/online/blogs/newsdesk/2013/01/the-libraries-of-timbuktu.html [16:49] Schbirid -- please do share your stitches! [16:50] where? it's a 90MB png [16:51] actually, sorry no. this is logged and public and i do not want to have my name associated with sharing that stuff. it's a fun scripting exercise [16:52] k [16:53] fair point [16:53] it's 0-63 for both x and y [17:00] I need to build a Linux box. It'll take me a while to get the free time. Did you use imagemagick? Can I cheat and ask, if you did stitched in two lines shell lines, what you used? [17:04] i used graphicsmagick but the usage is the same [17:04] brute force?: imagemagick convert -append 1/{0..64}.png [17:04] i wrote a few more lines but the resulting gm line was one [17:04] nope, check out the "montage" tool [17:04] well starting with zero not one [17:04] thx -- will check it out -- http://www.imagemagick.org/script/montage.php [17:05] you will need to give it the all filenames in the correct order, it wants them ordered by column not by line [17:06] then it is just "gm montage +frame +shadow +label -tile 64x64 -geometry 256x256+0+0 $filenames stitchedtiles.png" [17:06] thanks! [17:07] you will be able to use it to stitch google maps, openstreetmap etc too btw. same setup really [17:08] for $filenames can I use "{0..64}/{0..64}.png " -? or do I just ls > filelist.txt [17:08] I know these are nooby q's [17:09] dont worry about that, i am a noob myself [17:09] haha [17:09] first of all, it is 0 to 63, not 64! (64 images in total) [17:10] I know it's basic but how does $filenames get filled [17:10] you will need to give it all the {0..63}/0.png , then the {0..63}/1.png etc [17:10] that will be your task! [17:10] I think in Linux and DOS if I type 'set' I get a list [17:11] can I just ls > filelist.txt [17:11] and then gm < filelist.txt ? [17:11] in (recent?) bash you can simply do filenames="$filenames $somethingnew" to append [17:11] nope, that would fail horribly [17:11] haha [17:11] maybe ls -1 [17:12] gotta get back to work, sorry. but feel free to ask (and also tell me to stop being a jerk and give you the script). [17:12] haha [17:12] yeah -- man, make it easy on me [17:13] ah -- "single-column -1," [17:13] " -1 list one file per line" [17:13] Hey, any chance, you might want to, you know, just give me the script? [17:14] back to work myself [17:14] thx [20:02] ersi: Right, and it works quite well :)) [20:02] ersi: Thanks for that [21:08] http://rbelmont.mameworld.info/?p=800 [21:21] DFJustin: That is awesome. [21:22] I've been wondering how practical it is to run Atari ST MIDI software in MESS. [21:23] the st is still not really up to snuff [21:23] Aw. [21:27] st software does insane things with the 68000 chip that we can't handle despite running almost everything else in creation that uses one, olivier galibert is currently rewriting the cpu core based on die photos [21:48] heh