00:02:25 -!- pikhq_ has changed nick to pikhq. 00:03:34 -!- nooga has quit (Success). 00:05:53 -!- olsner has quit ("Leaving"). 00:08:38 -!- Pthing has quit ("Leaving"). 00:08:55 -!- Pthing has joined. 00:22:41 -!- jix_ has quit ("leaving"). 00:22:54 -!- jix has joined. 00:28:24 -!- BeholdMyGlory has quit (Remote closed the connection). 00:34:10 -!- FireFly has quit ("Later"). 00:42:39 -!- amca has joined. 00:43:40 -!- GregorR-L has quit (Remote closed the connection). 00:43:45 -!- GregorR-L has joined. 00:52:40 -!- coppro has quit (Read error: 54 (Connection reset by peer)). 01:10:56 -!- Asztal has quit ("."). 01:47:26 -!- GregorR-L_ has joined. 01:48:51 -!- GregorR-L has quit (Read error: 104 (Connection reset by peer)). 01:51:33 -!- coppro has joined. 01:56:07 http://codu.org/pics/main.php?cmd=imageview&var1=Assorted%2FGregorInPink.jpg I'm so classy 01:56:22 homo 01:56:31 Don't you wish 01:56:44 no 01:56:47 too femme 01:57:29 I see, you're more into bears then :P 01:58:13 more just 01:58:17 not all pink 01:59:40 Well, all pink isn't my USUAL thing, in fact I only have one pink shirt, one pink tie and one pink hat. I just decided to wear them all together today :P 02:00:08 came over a little queer 02:00:15 we're all queer here. 02:02:20 GregorR-L_: Out of curiosity, do any "god hates fags" people come to your university? 02:02:30 What, Purdue? 02:02:42 I haven't seen any, but at the same time it's hardly PSU. 02:02:51 -!- GregorR-L_ has changed nick to GregorR-L. 02:03:09 (I mean, it is Indiana :P ) 02:03:21 They've come to mine. 02:03:30 What, Westboro? 02:03:35 Have gay sex in front of them. 02:03:37 Purdue is the only big-ten university without a "queer resource center". 02:03:39 ehird: Not them specifically. 02:03:43 Other, smaller groups. 02:03:46 Do it anyway. 02:04:01 There's more gay guys at my school than there are girls at all... 02:04:14 (this, though funny-sounding, is not hard to pull off) 02:04:30 pikhq: Then in terms of maximum utilization of social resources, your choice is clear. 02:12:32 GregorR-L: Hahah. 02:12:53 Gay rationality! 02:12:59 Gaytionality. 02:16:13 pikhq: Which school is this, btw? I forget. 02:16:38 GregorR-L: Missouri University of Science and Technology. 02:16:51 It's not so much that it's a gay haven as it is a not-female haven. 02:16:55 Right. 02:17:17 That being said, smart girls are HOT. 02:17:26 Agreed. 02:56:27 -!- amca has quit ("Farewell"). 03:00:21 -!- coppro has quit ("The only thing I know is that I know nothing"). 03:12:01 -!- coppro has joined. 03:12:11 -!- coppro has quit (Client Quit). 03:14:06 -!- ehird has quit. 03:51:32 -!- GregorR-L has quit ("Leaving"). 03:57:21 Finally, Playstation emulation that doesn't irritate me. 03:59:06 Helps having a Playstation controller. 05:10:45 lol, gamer 06:12:51 -!- Asztal has joined. 07:17:52 -!- Asztal has quit (Read error: 110 (Connection timed out)). 07:19:49 -!- Pthing has quit (Remote closed the connection). 07:27:33 -!- olsner has joined. 07:38:33 -!- asiekierka has joined. 07:38:35 hi 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 08:14:01 -!- kar8nga has joined. 08:15:20 -!- FireFly has joined. 08:27:58 -!- waga has joined. 08:37:14 -!- kar8nga has quit (Remote closed the connection). 08:47:59 -!- waga has left (?). 09:00:02 -!- MigoMipo has joined. 09:04:46 i'm working on my obfuscated Pascal BF interpreter 09:09:49 agh i'm stuck lol 09:10:48 lol 09:10:50 i've done it 09:11:43 Or at least 09:11:47 I've thought I've done so 09:25:41 []++++++++++[>>+>+>++++++[<<+<+++>>>-]<<+<+++>>>-]<<+<+++>>>-]<<+<+++>>>-]<<+<+++>>>-]<<+<+++>>>-]<<<<-][]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]] 09:25:43 ...wtf 09:25:51 that's what my interpreter shows while running this code 09:25:59 []++++++++++[>>+>+>++++++[<<+<+++>>>-]<<<<-] 09:25:59 "A*$";?@![#>>+<<]>[>>]<<<<[>++<[-]]>.>. 09:27:14 yay, i've fixed it 09:27:28 BUT it outputs an I 09:27:30 and not an H 09:28:52 also quine.b gives me weird trash at the end 10:40:56 -!- asiekierka has quit. 10:53:28 -!- nooga has joined. 10:56:13 -!- oerjan has joined. 11:01:53 Woot, I understood almost half of the introduction of fizzie's thesis presentation 11:03:28 well if it was in finnish, then you still beat most of us 11:04:01 It was in English 11:04:17 THEN YOU ARE JUST STUPID 11:05:21 Meh 11:05:44 Well, at least I know what fizzie looks like now 11:05:54 I can stalk him around the campus 11:06:01 yay 11:14:37 where? 11:14:53 i(s the thesis) 11:15:12 Possibly nowhere 11:15:33 indeed, it is quite conceivable that he didn't manage to finish it ;D 11:15:36 4what is it about? 11:15:44 "Methods for Spectral Envelope Estimation in Noise Robust Speech Recognition" 11:15:53 oh sh 11:16:02 i hate signals 11:17:11 no longer will the computer be able to say "I can't hear you, la la la la" 11:24:41 longer computer will say you be able to hear the la la la la, no i can't 11:29:47 ahh 11:29:57 regexps look so eso 11:30:26 text.gsub!(/\/\/.*$/," ") 11:30:26 text.gsub!(/\/\*.*?\*\//m," ") 11:30:27 text.gsub!(/^\s*(#\s*(.*\\\n)*.*$)/," ") 11:30:29 cla = text.scan(/^\s*(class|struct)\s+(\w+)(?:\s*:\s*(protected|public|private)?\s*(\w+))?/) 11:48:41 -!- BeholdMyGlory has joined. 11:52:08 -!- Deewiant has quit (Remote closed the connection). 11:55:48 -!- oerjan has quit ("Gah internet molasses!"). 13:09:22 I still don't know what Deewiant looks like, though I guess I can sort of eliminate, given that I knew most of the other people there. 13:09:29 On the other hand I'm terrible at remembering faces. 13:09:50 They should stamp ID numbers over the faces of everyone I ought to know. 13:20:38 -!- coppro has joined. 13:20:41 -!- Sgeo has joined. 13:22:41 -!- Deewiant has joined. 13:24:38 root is so rude, always rebooting machines when you least expect it 13:25:06 fizzie: I was next to the guy next to you 13:26:28 Okay, that's what I thought. I think. I probably won't recognize you the next time I see you, so maybe it doesn't matter that much. 13:30:25 heh 13:33:01 I also went and wrote that "kypsyysnäyte" thing. 13:33:22 Lit. translated, "ripeness sample". 13:38:06 ilold 13:39:14 guys chat on the same # for years, they go to the same univeristy and they don't bother to meet irl 13:39:30 Oh, and the thesis is nowhere yet, though I think we commonly put those to the web after they've been graded and accepted. 13:39:49 What's "we" here 13:39:59 The authors, right? 13:40:07 I.e. the school doesn't do that 13:40:39 The school doesn't, yes. I just think at least some of the people in the lab have theirs on their personal pages. 13:40:59 Yeah, people do that. 13:41:13 Currently the thesis is in the hands of the professor for final comments before the bind-to-book part. 13:41:35 Or "books". Should figure out how many copies I need, actually. 13:42:02 "Methods for Spectral Envelope Estimation in Noise Robust Speech Recognition" <<< Deewiant: did you understand the spectral envelope estimation half or the noise robust speech recognition half? 13:42:19 Neither 13:42:42 I understood half of the "speech recognition" fifth 13:42:54 But it was maybe about half of the introduction 13:44:02 and here i thought you knew everything 13:44:15 The presentation wasn't really very polished, I just took the conference thing (which didn't really have much of a speech recognition introduction; being the Nth presentation in a speech recognition conference, I thought people might be a bit bored of it) and quickly slapped something on, without practicing how to utter it. 13:44:47 I guess I could do a more followable speech recognition presentation, but that thesis-lecture is officially just 20 minutes and I had to fit the whole thesis in it too. 13:45:26 The slide-PDF had 28 pages already, that's not in the suggested range for a 20 minute show. 13:45:31 Yes, that's of course not the point 13:45:39 what's spectral envelope estimation? something like guessing what spectum is enough for getting all the relevant frequencies? 13:45:41 I think you spent a bit more than 20 minutes anyway 13:45:51 Yes, I hear most people do. 13:46:14 They're not very strict about it, of course 13:47:39 oklopol: http://www.cis.hut.fi/htkallas/speechspec.png <-- that's a speech spectrum and an envelope estimate. 13:48:02 well let me give it an understanding 13:48:09 What's an envelope 13:48:17 Aside from something you put paper in 13:48:44 so basically you the envelop the fft 13:48:50 "b. Electr. Engineering. A curve formed by joining the successive peaks of a graph of an oscillation, esp. a modulated wave." 13:49:01 That's close to the sense. 13:49:12 Yeah, well that's pretty obvious from the pic 13:49:21 I guess I meant to ask what it's good for 13:50:36 Well, for speech that comb-like peaky structure comes from the excitation source, which looks like a pulse train, while the shape of the envelope comes from the resonances in the vocal tract. 13:51:31 And since what distinguishes the meaningful sounds is the vocal tract part (the comb structure is pretty much just about the speaking pitch and prosody and so on) it makes sense to use a nice envelope estimate instead of the spectrum itself. 13:52:05 Okay, cool 13:52:29 alright, but what exactly is the excitation source? 13:53:04 If you want to talk physics, it's the periodic airflow through the glottis. 13:54:19 so basically what happens @ vocal chords 13:56:33 Yes. 13:56:59 People have been modeling the glottal pulse shape, too, it's a sort of a half-sine hump. 13:57:39 would a half-sine hump result in a pulsey fft? 13:59:10 If it's a periodic sequence of humps, yes. Though the peaks aren't then impulses, like they would be for the idealized impulse-train-excitation case. 14:04:16 "the peaks aren't then impulses"? 14:05:09 do you mean they are not just single points, but wider? 14:09:04 Yes. In the very horribly much idealized model the excitation source is an impulse train, the vocal tract is an all-pole filter, and the resulting spectrum is a pure line spectrum, with just single points at the harmonic frequencies, with the envelope shape determined by the vocal tract filter. 14:09:11 Of course that's not at all how it works. 14:10:33 One glottal pulse approximation I've seen looks like g[n] = { 1/2*(1-cos(pi*n/N1)), 0 <= n < N1; cos(pi*(n-N1)/(2*N2)), N1 <= n <= N1+N2; 0 elsewhere }. I guess I could FFT a couple of those to see what the spectrum looks like. 14:19:36 http://www.cis.hut.fi/htkallas/train.png -> http://www.cis.hut.fi/htkallas/spec.png (Please don't ask me to explain the slope.) 14:21:15 train is that g[n] thing right? 14:21:38 Yes. 14:21:47 just making sure 14:22:05 spec is the log-scale (dB in the Y axis, actually) magnitude spectrum. 14:23:34 what does the idealized impulse train look like, unfft'd? 14:26:13 wait wait i think i confused a few terms in what you said earlier 14:26:28 i'm quite slow at getting into things 14:27:05 An impulse train looks like, well, the same except for impulses. Something like g[n] = { 1 if n = 0 (mod T0); 0 elsewhere } with a suitable period T0 for the fundamental frequency you want. 14:27:33 Okay, so that g[n] I gave was just the shape of one of the humps. 14:28:27 was there some name for the two realms, fft'd and unfft'd sound kinda stupid. 14:28:51 Time and spectral domains. 14:29:19 i'm only familiar with the technical details of dft, i'm assuming the terms wouldn't quite mix 14:29:21 alright 14:29:22 Or the other way around. 14:29:35 well yes, ofc 14:30:54 -!- AnMaster has joined. 14:31:08 anyway i guess i was asking what the ideal impulse train looks like in spectral domain, you may have told me already 14:32:29 or is that even well-defined 14:32:57 It should look like an impulse train in the spectral domain too, though with the T0 distance replaced with the corresponding F0. At least that's how it's been talked about. 14:32:58 i guess it might be the limit of some infinite spectrum, but maybe i'll wait for the answer 14:33:08 oh 14:33:18 is there a simple relation? 14:33:24 like F0 = 1/T0 14:33:29 Yes. Like that. 14:33:37 interesting. 14:33:51 I'm not sure if the peaks have a defined value, actually. 14:34:07 should that obvious? 14:34:09 i mean 14:34:12 F0 = 1/T0 14:34:31 Defined height, I mean. 14:34:43 oh? 14:34:47 For a really periodic, infinite sequence, I guess the height could be infinite. But speech people are horrible mathematicians and don't care about that sort of stuff. 14:35:05 yes, i can imagine 14:35:36 I'll check the Rabiner book, he's usually quite formal. 14:36:02 you can be precise without being formal 14:36:23 The single good thing about thesis work is that I have (temporarily, from the library) a pile of the five classical speech textbooks for looking up "trivial" common-knowledge sort of things from. 14:36:26 although i'm not sure what the point of saying that was 14:37:31 you don't read speech literature 14:37:32 ? 14:37:57 I haven't justifiticated buying the books myself, no. 14:39:14 Yes, the DTFT of the signal does not theoretically speaking exist; the mathematical treatment of the book uses the "engineering DTFT". 14:40:55 I mean, obviously it doesn't exist, since you have to have $\sum_{-\infty}^\infty \left| x(n) \right| < \infty$ if you want $x(n)$ to have a DTFT. 14:42:21 i don't know what left and right do in tex 14:42:32 just some kinda alignment? 14:42:55 "\left| x(n) \right|" is just a way to say "| x(n) |" so that TeX knows the |s pair up around x(n), and can spacify them sensibly. 14:43:11 also you can't dtft an infinitely occurring pulse? 14:43:17 ohh 14:44:03 i completely ignored | as a random tex character, but i now realize i've finally reverse-engineered the basic syntax and that doesn't fit it 14:44:30 err 14:44:45 actually i should probably ask what a dtft is :P 14:45:54 X(w) = infinite-sum-with-n of x(n) e^{-jwn}, where w wants to be the omega character when it grows up, j's the good old imaginary unit, and e's e. 14:46:54 That's the DTFT; I guess it doesn't converge very often. 14:48:18 Oh, and x(n) is continuous and infinite and all that fluff. 14:49:42 oh, mister omega? in dft omega is a primitive n'th root of unity, but i don't know what it is here 14:52:14 dft being discrete fourier transform, not sure it's obvious you know i mean tht 14:52:17 *that 14:52:55 Yes; the DFT is much nicer for an engineer, since it's just finite-length sequences and everything's so discrete. 14:54:28 X(w) = infinite-sum-with-n of x(n) e^{-jwn} <<< ah misread this as X(n), i now realize that doesn't even make sense 14:55:40 oh it's discrete-time fourier transform; so this is discrete already 14:55:54 The n time step is discrete. 14:56:03 The values of the signal are continuous, though. 14:56:03 ah 14:57:11 okay i think i see how that works now 14:59:48 sorry, phone 14:59:50 Anyway, since the DTFT tends to not exist, they use this "engineering DTFT" built out of Fourier series coefficients and Dirac delta functions, so you get infinite energy at the harmonic frequencies but the integral over the spectrum is still finite. 15:00:03 I'm not sure if that's a common approach or just a peculiarity of this book. 15:00:12 Signal processing is not my strongest subjects, ironically. 15:00:38 i was gonna say something, but i kinda lost my train of thought 15:01:31 drummer called to say he can't come to play metal with us, because he has to attend church 15:02:33 i'm not even going to ask for details 15:02:46 He is going to BURN the church, maybe. 15:03:18 i'd love to know signal processing, but that doesn't really fit my current degree 15:03:34 maybe 15:04:05 actually i'm not sure it's so much church, iltahartaus, however that translates, or whatever that means, at the army 15:04:14 -!- GregorR-L has joined. 15:04:55 atheists don't have to attend, but they don't have the evening free either 15:05:01 because that would be unfair 15:05:17 wtf conversation did I just step into? 15:05:35 it's about signal processing 15:05:48 Ngraargh. I have this tracking number for this mail package, and it seems they've decided to Screw Up(tm) with it. 15:06:31 i'm tlking about it with fizzie, as you can clearly see 15:06:36 The tracking system says "An irregularity has been observed in the processing of the package, delivery may be delayed". 15:08:36 -!- BeholdMyGlory has quit (Remote closed the connection). 15:16:06 i wonder if there's a programming system that runs ON iphone 15:16:30 I'll bet making it would violate the AUP in some stupid way :P 15:16:40 probably yes 15:17:26 i had cool spreadsheet with usable lisp on my symbian phone 15:17:41 but it was pain to write something with phone's keyboard 15:17:47 * GregorR-L wonders why a spreadsheet would have lisp built in ... 15:19:10 The spreadsheet program is a mode in emacs? 15:19:45 i don't know but it was nice 15:20:02 formulas were written in lisp 15:21:25 "Python on iPhone actually rather good", Nov 2008, http://mail.python.org/pipermail/python-list/2008-November/686098.html 15:22:04 probably it's python interpreter that runs programs written for iphone 15:22:07 not on iphone 15:23:39 What? No. "Install iPhone/Python (examples, the hello world is a PyObjC call into the iPhone API to load and scroll the contacts list in the phone)" 15:23:40 And so on. 15:24:30 " probably it's python interpreter that runs programs written for iphone" this makes NO SENSE 15:24:50 why? 15:26:24 GregorR-L: he means it only runs programs that are written with the iphone in mind 15:26:40 yup 15:26:51 well i was just joking, that makes no sense 15:26:54 :D 15:27:02 well duh 15:27:04 um 15:27:11 oh duh, well good. 15:28:00 okay 15:28:26 that python is probably a runtime environment for python programs that runs on iphone 15:28:38 what i want is development environment that runs on iphone 15:28:57 ;D 15:29:23 Sure, it "just runs Python programs", but you only need a text editor to write those python programs, and it's a development environment. 15:29:53 editing text on iphone sucks 15:30:19 Surely someone must have written a BF interpreter for the iPhone? 15:30:20 i imagine simpler, perhaps semi-visual editor for building programs 15:30:41 And eight buttons instead of lots means easier to hit them 15:31:02 FireFly: something like that, yep 15:31:18 You wouldn't mind that it's BF? 15:31:24 i would 15:31:35 "ibrainfsck: A native iPhone/iPod touch IDE for the brainfuck programming language." 15:31:48 "IDE" >_O 15:31:50 There, a programming IDE 15:32:17 If a GUI domain is made for PSOX, it would be possible to make a Brainfuck IDE in Brainfuck 15:32:30 wut? 15:32:46 Sgeo: IDE does not imply GUI. 15:33:13 GregorR-L: add bf debugger and voila 15:33:30 Hm, can't one make use of ANSI escapes with BF? 15:33:50 Egg-zactly ... though that just means you'd need a good terminal for iPhone 15:34:02 uhuh 15:34:08 that'd be hard 15:34:19 iphone keyboard SUCKS!!! amagad! 15:34:33 What?! You mean the commercials LIE?! 15:34:50 yes, yes it does 15:35:01 lack of tactile feedback = bad 15:35:19 Isn't there some sort of vibration-feedback thing?-) 15:35:22 `google openpandora 15:35:23 Also we now have a dedicated press relations guy, if you need anything press related please email chip@openpandora.org and he will do his best to make ... \ [13]BLOG - [14]Forums - [15]Developers - [16]Press area 15:35:36 WAY TO NOT GIVE A URL, HACKBOT 15:35:43 If even the iPhone keyboard sucks, I wonder how horrible those non-qwerty Android phones everyone seems to be making are. 15:35:44 ? 15:36:10 Oooh, are they FITALY? 15:36:27 I wouldn't mind a Dvorak keyboarded cellphone 15:36:34 FITALY DAMN IT FITALY 15:37:59 It's just a term they use for touch-screen-only things. 15:38:45 "various strokes (rather than taps) are used for both shifting case and selecting symbols." 15:39:06 Pie menu style? 15:39:29 I wonder if there's a Dasher thing for iPhone. 15:40:37 -!- ehird has joined. 15:41:28 * AnMaster installs Ubuntu in a vm 15:44:15 dasher is even usable 15:44:38 it is 15:44:51 dasher? 15:45:43 http://en.wikipedia.org/wiki/Dasher 15:48:28 nooga: dasher is even usable 15:48:28 ehird: it is 15:48:48 wow, that's probably the first time we agreed 15:48:51 oh right, I remember seeing that 15:50:51 Hm.. I think I've seen some input system like that in some DS homebrew 15:51:20 in what ds 15:51:25 -!- olegfink has quit (Remote closed the connection). 15:51:43 Homebrew for the Nintendo DS 15:51:49 ah 15:52:08 Okay, not quite like that 15:52:28 i saw cool one 15:52:58 table with consonants, and when you click on choosen one - a pie menu with vowels appears 15:55:28 oh, there is a dasher for iphone 15:56:04 "People such as scientist Stephen Hawking wouldn't have a chance in the U.K., where the NHS would say the life of this brilliant man, because of his physical handicaps, is essentially worthless." 15:56:22 ... Correct me if I'm wrong, but... Isn't Stephen Hawking British? 15:56:54 lol 15:56:55 Yes 15:56:57 Yes he is. 15:57:16 Quoted-guy was confused by his computer's accent ;) 15:57:22 pikhq, where is that quote from 15:57:23 but he sits in USA? 15:57:30 nooga: No. 15:57:34 oh 15:57:35 AnMaster: Article on Reddit. 15:57:49 GregorR-L: Foiled by DECtalk again! 15:57:52 then i was confused by his computer's accent 15:58:07 He's at the University of Cambridge. 15:58:45 pikhq, DECtalk? 15:59:04 AnMaster: That's the name of his text-to-speech system. 15:59:21 It was, obviously, designed by DEC. 15:59:30 DEC sounds so retro 15:59:35 for some reason... 15:59:36 ;P 15:59:55 IIRC, it's on a small, battery-powered UNIX system in his wheelchair. 16:00:06 heh 16:00:15 pikhq, how does he interact with it I wonder 16:01:17 AnMaster: Don't recall. 16:02:30 -!- Asztal has joined. 16:02:31 -!- BeholdMyGlory has joined. 16:10:15 -!- nooga has quit (Remote closed the connection). 16:18:37 I misread that as "in some DS hebrew", thought it to be some sort of Hebrew-localized input thing. 16:21:10 -!- Asztal has quit (Connection timed out). 16:21:27 wow: http://www.inference.phy.cam.ac.uk/dasher/movies/EyeDasher.4800.mpg 16:21:43 anything that's non-qwerty can't be that bad 16:22:29 -!- Asztal has joined. 16:30:51 [16:00] AnMaster: pikhq, how does he interact with it I wonder 16:30:52 throat vibrations 16:31:10 wow 16:31:16 must be kind of painful 16:31:24 no, just slow 16:31:29 all his interviews are scripted 16:31:30 well yeah 16:31:36 ehird, heh? 16:31:37 in Q&A sessions he takes >20m to say a line 16:31:49 it's like 16:31:54 question? 16:32:00