• [00:08:27] * dcramer_ (n=davec@dcdsl.ebox.com) Quit ()
  • [00:11:18] * guillaum1 (n=Guillaum@AMontsouris-153-1-91-205.w90-2.abo.wanadoo.fr) Quit ("Leaving.")
  • [00:11:47] * guillaum1 (n=gl@AMontsouris-153-1-91-205.w90-2.abo.wanadoo.fr) has joined #beagle
  • [00:12:11] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [00:20:08] * rsalveti (n=salveti@pc-a81040.wlan.inet.fi) Quit (Read error: 104 (Connection reset by peer))
  • [00:21:01] <geckosenator> my beagle gets stuck on ntpdate booting now
  • [00:21:43] <Russ> ^C?
  • [00:22:39] * likewise (n=likewise@82-171-51-231.ip.telfort.nl) Quit ()
  • [00:30:24] * brolin (n=brolin@190.157.13.75) has joined #beagle
  • [00:36:48] * robclark (n=robclark@nat/ti/x-917f7718f72f8180) Quit (Remote closed the connection)
  • [00:36:53] * florian (n=fuchs@f049188050.adsl.alicedsl.de) Quit ("Verlassend")
  • [00:37:13] * TAK2004 (n=thomas@dslb-088-072-218-108.pools.arcor-ip.net) Quit ("Verlassend")
  • [00:38:29] * lcuk (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) Quit ("Leaving")
  • [00:42:04] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [00:42:28] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [01:04:25] * bmxr (n=bmxr@S01060018f3b11a22.vf.shawcable.net) has joined #beagle
  • [01:08:47] * robclark (n=robclark@166.128.18.100) has joined #beagle
  • [01:14:31] * bkero wonders how octoread will run on the begale
  • [01:14:32] <bkero> beagle
  • [01:19:04] <fenn> what's octoread?
  • [01:19:21] <fenn> ocropus?
  • [01:19:30] <fenn> are you making an auto-translator thingy?
  • [01:20:37] <bkero> Yea, ocropus
  • [01:20:38] <bkero> My bad
  • [01:21:05] <bkero> Yea, I'm doing on-screen crop that passes that off to imagemagick, does some conversion, then passes it on to ocropus
  • [01:22:58] <fenn> keep me updated
  • [01:23:15] <bkero> Heh, interested?
  • [01:24:00] <bkero> I'm using vga glasses and a webcam with it. So takes webcam output, select a region, webcam takes a pic, crops to that region, and passes it off
  • [01:24:14] <bkero> Some image analysis to make it auto-select regions would be really cool
  • [01:25:27] <fenn> seems like you'd have to do a variety of different threshold parameters for "real life" since it's not just black text on white page
  • [01:25:44] <bkero> Dithering and setting bitdepth seems to help a lot
  • [01:31:14] <Russ> your brain really fools you into thinking that this is a simple problem
  • [01:31:42] <Russ> especially since its never a problem your brain solves, at least in that order
  • [01:31:59] * robclark (n=robclark@166.128.18.100) Quit (Read error: 110 (Connection timed out))
  • [01:33:31] <bkero> I've been doing facial analysis on images for a school project for a while now. I'm not deluded about how difficult the concept is.
  • [01:34:20] <Russ> its just finding the text region in an image with shadows and partial blockage
  • [01:34:51] * motion8 (i=43353c6a@gateway/web/ajax/mibbit.com/x-da90da74c672ba47) Quit ("http://www.mibbit.com ajax IRC Client")
  • [01:36:05] <bkero> Yea, I'm not even bothering to find the text, just doing it manually
  • [01:37:25] <Russ> that sounds like fun
  • [01:37:32] <Russ> which vga classes did you choose?
  • [01:38:23] <bkero> Myvu crystals for the prototype
  • [01:44:52] <Russ> its so awesome what can be put together these days for less than $500
  • [01:50:04] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [01:52:09] * parapete (n=burgrlov@86.130.9.92) Quit ("Leaving")
  • [02:09:08] * brolin (n=brolin@190.157.13.75) Quit (Client Quit)
  • [02:15:29] <fenn> bkero have you seen the 'squirrel tracker' FPGA DSP demo? http://thomaspfeifer.net/fpga_dsp_bildverarbeitung.htm
  • [02:16:04] <fenn> (scroll down)
  • [02:21:19] * dcramer (n=davec@dcdsl.ebox.com) has joined #beagle
  • [02:24:55] <bkero> Ok, that's really impressive
  • [02:26:25] <fenn> seems like you could do one for each kana
  • [02:26:36] <fenn> however the algorithm works.
  • [02:42:34] * uberfry (n=spinl0ck@85.93.208.138) has joined #beagle
  • [02:46:10] * fulgas is now known as FuL|OUT
  • [03:01:16] * JuanG (n=Juan@192.94.94.106) has joined #beagle
  • [03:01:23] * JuanG (n=Juan@192.94.94.106) has left #beagle
  • [03:09:32] * uberfry (n=spinl0ck@85.93.208.138) Quit (Read error: 145 (Connection timed out))
  • [03:18:07] * emeb (n=ericb@ip72-223-90-212.ph.ph.cox.net) has left #beagle
  • [03:31:12] * esuie (n=dennis@222.65.175.96) has joined #beagle
  • [05:06:44] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [05:17:33] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [05:18:34] * dioioib (n=dioioib@d121-173-32.home3.cgocable.net) Quit ()
  • [05:23:50] * vijay (n=vijay@203.199.213.3) Quit (Read error: 110 (Connection timed out))
  • [05:47:17] * bmxr (n=bmxr@S01060018f3b11a22.vf.shawcable.net) Quit (Read error: 54 (Connection reset by peer))
  • [05:58:49] * sakoman_ (n=sakoman@static-74-41-60-154.dsl1.pco.ca.frontiernet.net) has joined #beagle
  • [06:18:45] * sakoman (n=sakoman@static-74-41-60-154.dsl1.pco.ca.frontiernet.net) Quit (Read error: 110 (Connection timed out))
  • [06:32:19] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) has joined #beagle
  • [08:16:33] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) Quit (Read error: 110 (Connection timed out))
  • [08:41:30] * kblin (n=kai@samba/team/kai) has joined #beagle
  • [08:41:39] <kblin> morning folks
  • [08:47:45] * robclark (n=robclark@166.128.18.100) has joined #beagle
  • [09:15:19] <julianoliver> buenas dias a todos
  • [09:15:38] * ldesnogu_ (n=ldesnogu@ven06-2-82-247-86-183.fbx.proxad.net) has joined #beagle
  • [09:34:29] * Russ (i=foobar@ip70-176-253-20.ph.ph.cox.net) Quit (Read error: 113 (No route to host))
  • [09:54:58] * adj_ is now known as adj
  • [09:57:25] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) has joined #beagle
  • [10:05:44] * lcuk (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) has joined #beagle
  • [10:07:10] * Xenion (n=robert@p579FCCCD.dip.t-dialin.net) has joined #beagle
  • [10:17:32] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) Quit (Read error: 110 (Connection timed out))
  • [10:17:49] * Xenion (n=robert@p579FCCCD.dip.t-dialin.net) Quit ("Verlassend")
  • [10:21:28] * mib_s5sulr (i=d04f0f66@gateway/web/ajax/mibbit.com/x-742fb4f7ff5cbc5f) has joined #beagle
  • [10:27:44] <dasnake> morning all
  • [10:27:44] <dasnake> I finally managed to have bb working with r86 hdtv
  • [10:27:44] <dasnake> seems that CVT computed timings were not good for htdv
  • [10:29:32] * magnet (n=magnet@90.27.43.123) has joined #beagle
  • [10:29:41] * Psychiatras (i=1000@212.122.84.232) Quit (Read error: 104 (Connection reset by peer))
  • [10:31:05] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Read error: 104 (Connection reset by peer))
  • [10:31:43] <dasnake> uhmm, in u-boot .. saveenv doesn't want to save optargs, any ideas about this fact?
  • [10:32:19] <mru> I've seen a few reports about u-boot not saving env properly
  • [10:32:28] <mru> never had the problem myself
  • [10:33:29] <dasnake> u-boot I'm using is 2009.01-00013-g52eddcd
  • [10:34:09] * Psychiatras (i=1000@212.122.84.232) has joined #beagle
  • [10:34:30] <dasnake> another problem I'm having is:
  • [10:34:37] <dasnake> I've setup the 720p mode in modedb.c
  • [10:34:48] <dasnake> but when I boot it set as pixel_clock 13468
  • [10:35:04] <dasnake> instead of 74250
  • [10:35:25] <dasnake> then if I manually set it via sysfs it's ok
  • [10:39:25] * mib_s5sulr (i=d04f0f66@gateway/web/ajax/mibbit.com/x-742fb4f7ff5cbc5f) Quit ("http://www.mibbit.com ajax IRC Client")
  • [11:01:50] * julianoliver (n=julian@164.pool85-53-3.dynamic.orange.es) Quit ("Lost terminal")
  • [11:05:14] * mpoullet (n=mpoullet@77-22-121-59-dynip.superkabel.de) has joined #beagle
  • [11:06:08] * abitos (n=nixgibts@91.46.71.169) has joined #beagle
  • [11:11:15] <dasnake> I've troubles compiling alsa in .. which is the right card for bb?
  • [11:19:49] * TAK2004 (n=thomas@dslb-088-074-035-015.pools.arcor-ip.net) has joined #beagle
  • [11:26:32] * AudioFreak|sleep is now known as AudioFreak
  • [11:27:56] <dasnake> SND_OMAP_SOC <- is this the kernel module for sound on bb?
  • [11:29:07] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) has joined #beagle
  • [11:31:00] * spinl0ck (n=spinl0ck@vodsl-10890.vo.lu) has joined #beagle
  • [11:31:31] * ogra (n=ogra@ubuntu/member/ogra) Quit (Read error: 113 (No route to host))
  • [11:35:42] * esuie (n=dennis@222.65.175.96) Quit ("leaving")
  • [11:38:14] * mib_06h7j5 (i=91638602@gateway/web/ajax/mibbit.com/x-76e8ee6eee48d85d) has joined #beagle
  • [11:38:24] * mib_06h7j5 (i=91638602@gateway/web/ajax/mibbit.com/x-76e8ee6eee48d85d) Quit (Client Quit)
  • [11:41:46] <AudioFreak> has anyone recently tried building a kernel with the tidspbridge module?
  • [11:42:03] <AudioFreak> the dsp howto seems outdated
  • [11:42:50] <AudioFreak> the patches mentioned don't apply, and there is a tidspbrdige branch of the linux-omap kernel
  • [11:43:33] <AudioFreak> but building that kernel fails on an implicit declaration of function 'memzero'
  • [11:45:11] <mpoullet> good morning
  • [11:45:34] * mpoullet is now known as mpoullet|home
  • [11:45:49] <AudioFreak> morning
  • [11:46:23] <AudioFreak> well actually afternoon here :P
  • [11:48:35] * ogra (n=ogra@ubuntu/member/ogra) has joined #beagle
  • [11:49:08] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) Quit (Read error: 110 (Connection timed out))
  • [11:52:10] <AudioFreak> no-one able to help me? :(
  • [11:53:15] <koen> AudioFreak: beagle is done by 'catalog' people from TI, so they use dsplink, no dspbridge
  • [11:53:43] <AudioFreak> hmm, how should i use the beagle dsp then?
  • [11:53:44] <koen> if you use angstrom you can install the kernel modules, dsp codecs and gstreamer plugins straight from the package feeds
  • [11:54:11] <AudioFreak> so the dsp howto is really way outdated or something?
  • [11:54:55] <koen> it's targeted at dspbridge, which probably only nokia uses
  • [11:55:07] <koen> and the guy who wrote that howto works at nokia, so....
  • [11:55:56] <AudioFreak> hmm, ok.. strange it is called beagleboard/dsp howto..
  • [11:56:24] * ClaudeQC (n=ClaudeQC@bas1-quebec03-1279635805.dsl.bell.ca) has joined #beagle
  • [11:56:51] <dasnake> the alsa support for beagleboard isn't in the linux-omap tree?
  • [11:57:03] <dasnake> which patchset I should merge to get alsa?
  • [11:57:17] <AudioFreak> keon: how do i install these kernel modules?
  • [11:59:20] <koen> AudioFreak: opkg ti-cmemk-module ti-dsplink-module ti-dmai-module (from the top of my head)
  • [11:59:37] <koen> 'opkg install gstreamer-ti' will pull in everything for gstreamer based stuff
  • [12:02:26] <mpoullet|home> dasnake: I think the best repo for alsa support is Sakoman's one
  • [12:05:21] <dasnake> mpoullet|home: I tried to add that but I've confused my git
  • [12:05:27] <dasnake> and now I'm confused myself
  • [12:05:34] <dasnake> :)
  • [12:05:55] <dasnake> there's also a dss2.2 branch in the sakoman repo
  • [12:06:03] <dasnake> is it similar to latest tomba repo?
  • [12:06:25] * mpoullet|home never tries Sakoman's repo...
  • [12:11:38] <mpoullet|home> does someone know what is the new --no-check-certificate dir in DL_DIR?
  • [12:12:49] <AudioFreak> koen: opkg cannot find these packages? can i download and install them manually?
  • [12:13:33] <AudioFreak> i tried google but it can't tell me much about it
  • [12:16:14] <kblin> is anyone here running angstrom with IPv6? I'm having a hard time getting my debian-armel to respond to neighbour-discovery packets with my davicom usb nic, but on an intel-based box, that NIC's working just fine
  • [12:21:05] * tzhau_ (n=tzhau@dsl-245-92-01.telkomadsl.co.za) Quit (Read error: 110 (Connection timed out))
  • [12:21:16] * tzhau_ (n=tzhau@dsl-245-92-01.telkomadsl.co.za) has joined #beagle
  • [12:21:58] * likewise (n=likewise@82-171-51-231.ip.telfort.nl) has joined #beagle
  • [12:22:12] <AudioFreak> i'll some more try later today, going visit my parents now
  • [12:22:23] * AudioFreak is now known as AudioFreak|Away
  • [12:29:06] * geckosenator (n=sean@71.237.94.78) Quit (Read error: 110 (Connection timed out))
  • [12:39:27] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [12:44:17] * ogra (n=ogra@ubuntu/member/ogra) Quit (Read error: 104 (Connection reset by peer))
  • [12:45:26] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [12:54:56] * likewise (n=likewise@82-171-51-231.ip.telfort.nl) Quit ()
  • [12:56:50] * ogra (n=ogra@ubuntu/member/ogra) has joined #beagle
  • [12:58:34] * lcuk (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) Quit ("Leaving")
  • [13:05:22] * Openfree (n=dennis@222.65.175.96) Quit (Remote closed the connection)
  • [13:08:31] * Openfree (n=dennis@222.65.175.96) has joined #beagle
  • [13:08:48] * FuL|OUT is now known as fulgas
  • [13:10:02] <mpoullet|home> there's an issue with the alsa-state package: http://pastebin.com/d6f86e017
  • [13:11:02] <mpoullet|home> oe_unpack_file uses dest as src
  • [13:11:19] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-df602fa6df7cf92c) Quit ("http://www.mibbit.com ajax IRC Client")
  • [13:25:06] <koen> mpoullet|home: try pinging kergoth in #oe about that
  • [13:25:29] <Crofton|work> kergoth won't be awake for a bit :)
  • [13:25:42] <Crofton|work> hmm
  • [13:25:51] * Crofton|work wonders what TZ difference is
  • [13:25:57] <Crofton|work> since we changed to DST
  • [13:26:08] <Crofton|work> and .az does not
  • [13:26:26] <Crofton|work> bother, this means he is 3 hours behind me
  • [13:39:55] * julianoliver (n=julian@164.pool85-53-3.dynamic.orange.es) has joined #beagle
  • [13:41:22] <julianoliver> hi.
  • [13:42:00] <julianoliver> i'm benchmarking the BB for computer vision applications. i've written a basic first test webcam capture application using Intel's OpenCV.
  • [13:42:07] <julianoliver> would someone with beagleboard, angstrom and a webcam mind compiling this simple application and telling me the camera they have and the number of frames per second they get?
  • [13:42:13] <julianoliver> http://rafb.net/p/52QDyf72.html
  • [13:44:51] <koen> julianoliver: mail it to the list, I might have time for it next week
  • [13:49:05] * julianol1ver (n=julian@164.pool85-53-3.dynamic.orange.es) has joined #beagle
  • [13:49:19] * julianoliver (n=julian@164.pool85-53-3.dynamic.orange.es) Quit (Read error: 104 (Connection reset by peer))
  • [13:53:00] <mpoullet|home> koen: mwester in #oe helps me to fix the alsa issue
  • [13:57:55] * robclark (n=robclark@166.128.18.100) Quit (Read error: 110 (Connection timed out))
  • [13:59:43] <jkridner> good morning all
  • [14:00:50] <Crofton> gm
  • [14:02:21] * urnenfeld (n=vision@209.pool85-58-29.dynamic.orange.es) has joined #beagle
  • [14:02:35] <mpoullet|home> gm
  • [14:02:55] <jkridner> hi Crofton, mpoullet|home
  • [14:04:30] <urnenfeld> Hi people, I have seen some comments comparing the beagleboard with openpandora, or the gp2x, but it isnt right, beagelborad is just a board, no screen no inputdevice, am i right?
  • [14:05:23] <jkridner> you are right.
  • [14:05:37] <jkridner> the similarity between the OpenPandora and the BeagleBoard is that they use the same processor.
  • [14:05:50] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-c9d4032c12d55baa) has joined #beagle
  • [14:06:31] <jkridner> you might be able to buy a case and some game controls to turn a BeagleBoard into a retro gaming platform, but that isn't really in the class as the OpenPandora.
  • [14:06:54] <jkridner> many of the OpenPandora folks are using the BeagleBoard to write software for the OpenPandora before the system ships.
  • [14:07:54] <urnenfeld> ahamm thanks jkridner from http://beagleboard.org/ is possible to purchase and get it shipped to an european country?
  • [14:08:21] <jkridner> yes, Digi-Key ships world-wide.
  • [14:09:36] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-c9d4032c12d55baa) Quit (Client Quit)
  • [14:14:22] <urnenfeld> and... are the specs of the HW enough open to write drivers for it? to port a new OS?
  • [14:16:28] <julianol1ver> urnenfeld: i paid EUR64 to bring 2 beagle boards into Spain.
  • [14:17:47] <julianol1ver> (pay on pick up)
  • [14:18:07] <kblin> julianol1ver: yeah, that seems the average import tax
  • [14:18:19] <urnenfeld> julianol1ver what we call, "a cobro revertido" ?:P
  • [14:18:52] <julianol1ver> jeje si, exactamente.. pero no es caro.
  • [14:19:47] <urnenfeld> 64 included shipping and tax?
  • [14:19:57] <julianol1ver> yep all inclusive..
  • [14:20:42] <kblin> it's a nice little piece of hardware.. now if I could make the software do what I need, I'd be happy :)
  • [14:20:57] <julianol1ver> kblin: agreed..
  • [14:23:18] <urnenfeld> thanks all, I am looking for an ARM platttform enough open to port Haiku OS, or to try it... considered the openpandora and the gp2x... are there more around?
  • [14:26:11] <julianol1ver> urnenfeld: touchpad, TI Zoom..
  • [14:26:21] <julianol1ver> oops.. s/touchpad/touchbook.
  • [14:27:21] * lcuk (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) has joined #beagle
  • [14:27:34] <julianol1ver> would targetting the NEON instruction set benefit frame capture rate from a V4L2 device?
  • [14:29:32] * ali_as (n=ali_as@ambix.plus.com) has joined #beagle
  • [14:33:42] <urnenfeld> wow the Ti zoom it not really cheap...
  • [14:33:48] <urnenfeld> is*
  • [14:33:57] <julianol1ver> nope
  • [14:34:09] <urnenfeld> 900$?
  • [14:37:43] <julianol1ver> i doubt it will ever be a consumer-grade device so much as a reference kit for developers..
  • [14:37:53] <julianol1ver> could well be wrong however
  • [14:39:17] <julianol1ver> koen: can you reccommend any way of speeding up V4L2 video capture? my reference application is about as minimal as it gets: http://rafb.net/p/52QDyf72.html
  • [14:39:40] <julianol1ver> koen: i get about 30-40FPS under Debian on my Intel laptop, just 7FPS on the BB.
  • [14:40:02] <julianol1ver> (BB/Angstrom)
  • [14:40:07] <jkridner> urnenfeld: I believe there is plenty of documentation to introduce new operating systems. Please take a look at the system reference manual and the OMAP3530 and TPS65950 technical reference manuals.
  • [14:40:09] * as_leep (n=ali_as@ambix.plus.com) Quit (Read error: 113 (No route to host))
  • [14:40:26] <jkridner> also, all of the Linux drivers are open source, so you can examine how they were done.
  • [14:40:49] <julianol1ver> koen: i've targetted the NEON by passing -ftree-vectorize -ftree-vectorizer-verbose=6 to g++
  • [14:46:56] * spinl0ck (n=spinl0ck@vodsl-10890.vo.lu) Quit (Read error: 60 (Operation timed out))
  • [14:51:54] <vlad_> julianol1ver: you're dooooomed
  • [14:58:16] <julianol1ver> vlad_: hmm... any suggestions?
  • [14:58:19] * tequilaworm (n=chatzill@212-198-79-134.rev.numericable.fr) has joined #beagle
  • [14:58:47] <vlad_> julianol1ver: no, just doomed if you're trying to speed things up with -ftree-vectorize.. as I understand it, it generates awful code
  • [14:58:55] * BThompson (n=Bernie@host-69-146-39-154.static.bresnan.net) has joined #beagle
  • [14:59:35] <tequilaworm> Does anyone uses Eclipse as a default IDE? I coud not find a recent doc with how to configure Eclipse in elinux systems
  • [15:00:24] <julianol1ver> vlad_: right..
  • [15:01:56] <julianol1ver> admittedly there's little to 'vectorise' in the above code anyway..
  • [15:02:56] <julianol1ver> there has to be a means to speed up video capture on the Beagle else it will be of little use for 'real time' computer vision applications.
  • [15:03:00] * julianol1ver keeps digging
  • [15:03:57] <kulve> julianol1ver: what's there to speed up the capture? Isn't it just copying data from place a to place b?
  • [15:08:23] * likewise (n=likewise@82-171-51-231.ip.telfort.nl) has joined #beagle
  • [15:08:43] <jkridner> raster: the audio was a bit muddy on your Bossa conference video. Just got finished watching most of it and am downloading the code samples now.
  • [15:10:07] * sakoman__ (n=sakoman@static-74-41-60-154.dsl1.pco.ca.frontiernet.net) has joined #beagle
  • [15:10:53] <dasnake> there's something that puzzle me .. why beagle isn't an option of alsa Kconfig?
  • [15:10:54] * zedstar (n=john@fsf/member/zedstar) has joined #beagle
  • [15:11:03] <dasnake> I've omap3beagle.c there, but no way to compile it
  • [15:11:12] <dasnake> maybe I've messed up repos
  • [15:11:52] <kulve> dasnake: at least at some point it got removed from the Kconfig because of a merge mistake. I haven't checked the situation lately
  • [15:12:09] <dasnake> ok
  • [15:12:13] <dasnake> seems removed now
  • [15:12:18] <dasnake> I
  • [15:12:20] <dasnake> err
  • [15:12:30] <dasnake> I'm checking also via web git tree browser
  • [15:12:36] <julianol1ver> kulve: yes.. see my code. it's ~25fps on my Atom netbook and around ~7fps on the BB.
  • [15:13:33] <julianol1ver> kulve: i use Intel's OpenCV, yet profiling doesn't seem to reveal any lower level functions which would benefit from vectorisation.. it's just raw capture, windowing and data copying.
  • [15:13:42] <jkridner> I'm going to be asking myself if there is any way to run this toolkit in a browser like what the Cappuccino and Objective-J. :)
  • [15:15:02] <kulve> julianol1ver: mem bandwidth in BB is much slower than in atom. But I guess plain video capture shouldn't take that much (depends on the resolution of course)
  • [15:16:12] * urnenfeld (n=vision@209.pool85-58-29.dynamic.orange.es) Quit (Read error: 110 (Connection timed out))
  • [15:20:41] <dasnake> what's the max Amperes the BB can draw?
  • [15:21:08] <julianol1ver> kulve: i'm capturing at 320x240 on the BB @ ~7fps.
  • [15:23:00] <kulve> julianol1ver: definitely should be faster even without any neon tricks
  • [15:23:10] <suihkulokki> dasnake: http://www.beyond3d.com/content/news/711 (and the document linked from the page)
  • [15:24:50] * sakoman_ (n=sakoman@static-74-41-60-154.dsl1.pco.ca.frontiernet.net) Quit (Read error: 110 (Connection timed out))
  • [15:25:53] * fulgas (n=fn@85.138.15.70) Quit (Read error: 110 (Connection timed out))
  • [15:29:04] <dasnake> uhmm, an usb hub powered by 5V@1A seems a bit tight then
  • [15:29:06] * brolin (n=brolin@190.157.13.75) has joined #beagle
  • [15:34:22] * geckosenator (n=sean@71.237.94.78) has joined #beagle
  • [15:35:38] * emeb (n=ericb@72.223.90.212) has joined #beagle
  • [15:36:54] * likewise (n=likewise@82-171-51-231.ip.telfort.nl) Quit ()
  • [15:37:20] <dasnake> uhmm, there's a way to disable alsa from kernel boot args?
  • [15:37:43] <dasnake> I finally managed to compile a new kernel with alsa support but it get stuck during boot process
  • [15:37:51] <dasnake> I don't really know if alsa is to blame
  • [15:37:51] <julianol1ver> kulve: i agree..
  • [15:41:53] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) has joined #beagle
  • [15:43:37] * BThompson (n=Bernie@host-69-146-39-154.static.bresnan.net) Quit (Read error: 54 (Connection reset by peer))
  • [15:47:28] * tequilaworm (n=chatzill@212-198-79-134.rev.numericable.fr) Quit (Read error: 110 (Connection timed out))
  • [15:50:43] * ant__ (n=andrea@host48-248-dynamic.15-87-r.retail.telecomitalia.it) has joined #beagle
  • [15:59:05] * julianoliver (n=julian@164.pool85-53-3.dynamic.orange.es) has joined #beagle
  • [16:02:24] * julianol1ver (n=julian@164.pool85-53-3.dynamic.orange.es) Quit (Read error: 113 (No route to host))
  • [16:02:30] <julianoliver> .
  • [16:15:43] * Openfree (n=dennis@222.65.175.96) Quit ("Leaving")
  • [16:16:53] * Openfree (n=dennis@222.65.175.96) has joined #beagle
  • [16:18:51] * FuL|OUT (n=fn@85.138.42.115) has joined #beagle
  • [16:19:19] * FuL|OUT is now known as fulgas
  • [16:20:01] * Batko_Marto (n=Batko_Ma@CPE001346f996d2-CM001e6b1a8d1e.cpe.net.cable.rogers.com) Quit (Remote closed the connection)
  • [16:31:30] * fulgas (n=fn@85.138.42.115) Quit (Read error: 104 (Connection reset by peer))
  • [16:32:43] * fulgas (n=fn@85.138.42.115) has joined #beagle
  • [16:36:15] <dasnake> anyone have experience with a52 audio decoding with the bb?
  • [16:36:32] <dasnake> I've finally managed to compile both dss2 and alsa on the same kernel, but I got :
  • [16:36:36] <dasnake> a52: CRC check failed! 0.104 ct: -0.004 1963/1963 ??% ??% ??,?% 0 0
  • [16:36:36] <dasnake> a52: error at resampling
  • [16:36:41] <dasnake> and video stuck
  • [16:36:51] <dasnake> with -ao null the video play nice
  • [16:37:24] * bmxr (n=bmxr@S01060018f3b11a22.vf.shawcable.net) has joined #beagle
  • [16:38:03] * rsalveti (n=salveti@pc-a81040.wlan.inet.fi) has joined #beagle
  • [16:43:02] * Russ (i=foobar@70.176.253.20) has joined #beagle
  • [16:48:17] <vlad_> hrm, the Overo is a 3503? not a 3530?
  • [16:48:21] * mib_vto769 (i=7aa62e54@gateway/web/ajax/mibbit.com/x-0f075b84e1c44996) Quit ("http://www.mibbit.com ajax IRC Client")
  • [16:48:29] * mib_vto769 (i=7aa62e54@gateway/web/ajax/mibbit.com/x-d35f70dffa31c441) has joined #beagle
  • [16:48:39] <Russ> afaik
  • [16:48:48] <vlad_> that's unfortunate
  • [16:48:49] <Crofton|work> vlad_, yes, although the gumstix roadmap says they willhave a 353-0 version
  • [16:48:53] <vlad_> ah
  • [16:56:26] <jkridner> now that GStreamer-TI is in the Angstrom feeds and the SGX code can be built-in too, I expect they'll release 3530 soon.
  • [16:57:24] * fulgas (n=fn@85.138.42.115) Quit (Read error: 104 (Connection reset by peer))
  • [17:02:51] * Openfree is now known as Openfree|away
  • [17:02:53] * kozak (n=subbu@117.192.13.5) has joined #beagle
  • [17:07:23] * mpoullet|home (n=mpoullet@77-22-121-59-dynip.superkabel.de) Quit (Read error: 113 (No route to host))
  • [17:10:14] * fulgas (n=fn@85.138.20.210) has joined #beagle
  • [17:10:17] <Crofton> we need to tear down the barrier to dsp usage
  • [17:10:36] <Russ> Mr. Crofton, tear down that wall
  • [17:11:59] <Crofton> I have many windmills to tilt at ....
  • [17:18:29] * fulgas is now known as FuL|OUT
  • [17:18:33] * FuL|OUT is now known as fulgas
  • [17:20:35] * mib_5rz2nt (i=63921a39@gateway/web/ajax/mibbit.com/x-68193047a6bbb7a7) has joined #beagle
  • [17:21:05] <mib_5rz2nt> I have a general question about beagle board based product development and I would appreciate the answers
  • [17:21:47] <mib_5rz2nt> Once I develop a product with beagle board, how can I reduce the product cost? Can I remove the unused hardware components?
  • [17:27:03] * brolin (n=brolin@190.157.13.75) Quit (Client Quit)
  • [17:30:15] <mru> you can make your own board based on the beagle board
  • [17:30:41] <mru> if you do that, you can of course omit any components you don't need
  • [17:30:48] <mru> though that won't be many
  • [17:31:28] <mru> can't think of much beyond the dvi framer
  • [17:31:38] <mru> and connectors of course
  • [17:32:16] <mib_5rz2nt> I actually only need a TFT LCD and USB connection for my purpose. Thanks. what can I reduce here?
  • [17:32:29] <mru> the dvi framer
  • [17:32:30] <Russ> mru: anything mmc
  • [17:32:42] <mru> and any unused connectors
  • [17:33:03] <Russ> anything svideo or audio
  • [17:33:21] <mib_5rz2nt> I literally don't need anything else. My application is very custom. I just need webcam, audio and some alrithms for capture
  • [17:33:28] <mru> none of that is more than a few passive components
  • [17:33:53] <Russ> mib_5rz2nt: then maybe a omap3 is beyond what you need/want
  • [17:34:11] <mib_5rz2nt> are there any design services that can help me modify this board at affordable cost
  • [17:34:24] <Russ> heh
  • [17:34:31] <Russ> in that case, just use it as is
  • [17:34:46] <mru> if the form factor is ok, the simplest and cheapest is to just skip the unused components
  • [17:34:48] <Russ> unless you plan to produce hundreds of thousands of units
  • [17:35:12] <mru> a few unused solder pads won't hurt anything
  • [17:35:23] * julianoliver (n=julian@164.pool85-53-3.dynamic.orange.es) Quit (Nick collision from services.)
  • [17:35:50] <mib_5rz2nt> I will need a few hundred units at max. This application is for helping some building inspectors
  • [17:36:34] <Russ> yes, then $20 of engineering time to reduce cost wouldn't help you much
  • [17:36:38] <Russ> er, $20k
  • [17:37:25] <Russ> also, asking someone to make masks for the pcb to make just a few hundred will push you past the digikey per unit cost I think
  • [17:37:36] <Russ> how much does a solder mask cost these days?
  • [17:38:15] <mib_5rz2nt> are there any cheaper alternatives ($50 or below) which can have an TFT and USB OTG which I could afford?
  • [17:38:33] <Russ> a really cheap digital camera?
  • [17:38:34] * julianoliver (n=julian@164.pool85-53-3.dynamic.orange.es) has joined #beagle
  • [17:39:00] <mib_5rz2nt> which digital camera runs linux and is open source?
  • [17:39:19] <julianoliver> digital cameras running linux?
  • [17:39:24] <Russ> I'm just saying, the only things I've seen with that hardware is a cheapo digital camera
  • [17:39:33] <Russ> and those are produced in huge volumes
  • [17:39:42] <Russ> so I think you need to rethink things a bit
  • [17:40:04] <Russ> the TFT alone might blow your budget in small quantities
  • [17:40:15] <Russ> you could try buying old iPAQs in bulk
  • [17:40:32] <Russ> or zaurus
  • [17:40:36] <mib_5rz2nt> what is IPaq?
  • [17:40:52] <julianoliver> a product from Compaq, a PDA.
  • [17:41:01] <Russ> http://www.ipaqlinux.com/
  • [17:41:02] <mru> a cross between an ipod and a compaq
  • [17:41:08] <Russ> nice mru
  • [17:41:48] <Russ> http://en.wikipedia.org/wiki/IPAQ
  • [17:42:07] <mib_5rz2nt> Ipaq seems to be a good idea but I need to check if it supports USB host for webcam
  • [17:42:28] <Russ> what do you need the device to do?
  • [17:43:00] <mib_5rz2nt> I need to connect a USB webcam which will capture images. I will analyze the images and generate some data on display
  • [17:43:08] <julianoliver> mib_5rz2nt: it's not fast enough for computer vision.. i've already looked into this.
  • [17:43:16] <julianoliver> nor is it fast enough for augmented reality.
  • [17:43:29] <mib_5rz2nt> so back to beagle board then?
  • [17:43:32] <Russ> mib_5rz2nt: are you looking to do stuff realtime?
  • [17:43:37] * julianoliver used to run Linux on an iPaq
  • [17:43:42] <mib_5rz2nt> yes, this will be realtime
  • [17:43:51] <julianoliver> mib_5rz2nt: it's probably not fast enough either. i'm working on AR for the beagleboard.
  • [17:43:55] <Russ> and for less than $50 per unit?
  • [17:44:14] <julianoliver> mib_5rz2nt: today's benchmarking puts the BB at 7FPS using OpenCV capturing at 320x240.
  • [17:44:28] <mib_5rz2nt> I was thinking I could buy TFTs etc. from someone in Asia in bulk
  • [17:44:35] <julianoliver> i'm trying to speed it up but as kulve said there's so little memory bandwidth..
  • [17:44:48] <Russ> mib_5rz2nt: a few hundred isn't really "bulk"
  • [17:45:08] <Russ> at least in the electronic component industry
  • [17:45:23] <julianoliver> mib_5rz2nt: i'm looking at the Atom again for computer vision right now, a VIA board.
  • [17:45:44] <Russ> julianoliver: too bad there isn't a mini-pci slot, you could put a fast GPU in there
  • [17:46:24] <julianoliver> mib_5rz2nt: Atom has awful power drain and is far too hot to be fun, but it might be the safest bet without vectorisation of the lower layers of OpenCV and or your tracker of choice.
  • [17:46:27] <mib_5rz2nt> that is why my best shot would be a GPS board or PDA board that have USB OTG connector
  • [17:46:48] <Russ> mib_5rz2nt: your performance requirements are pretty daunting
  • [17:46:52] <julianoliver> Russ: it's not the GPU really, that's fast.
  • [17:47:06] <mib_5rz2nt> actually my performce requirement is not tood bad
  • [17:47:20] <julianoliver> Russ: errm.. i mean OpenGL ES support is the least of the problems with the Beagle.
  • [17:47:20] <Russ> julianoliver: GPU will give you the memory bandwidth you need
  • [17:47:23] <mib_5rz2nt> I am just capturing video frames in < VGA
  • [17:47:28] <Russ> julianoliver: I don't mean a GPU for display
  • [17:47:32] <julianoliver> ahh right..
  • [17:47:33] <Russ> julianoliver: I mean a GPU for computing
  • [17:47:35] <mib_5rz2nt> and check for certain patterms
  • [17:47:42] <julianoliver> yes, i see.
  • [17:47:42] <Russ> mib_5rz2nt:
  • [17:47:52] <julianoliver> mib_5rz2nt: ARToolKit etc?
  • [17:48:01] <Russ> mib_5rz2nt: its the checking for patterns in realtime where you will run into problems
  • [17:48:22] <kulve> julianoliver: sounds like you are doing something wrong. Faster hw doesn't help in that case
  • [17:48:34] <julianoliver> mib_5rz2nt: i have a few projects using fiducial marker tracking i am attempting to port to the BB. i think that should be O.K using ARToolKitPlus..
  • [17:48:46] <julianoliver> kulve: VLC is no better.. what frame rate can you capture at using VLC?
  • [17:49:15] <mib_5rz2nt> generally webcams run for 10 Mbps at full speed. This is must slower than reading a flash drive
  • [17:49:43] <kulve> julianoliver: I don't capture anything, so I can't tell
  • [17:50:12] <julianoliver> kulve: right
  • [17:50:45] <julianoliver> kulve: you're welcome to try with my capture code: http://rafb.net/p/52QDyf72.html
  • [17:52:26] <philv> I've always found OpenCV is hideously slow, period. Perhaps check what insanity OpenCV is up to during the capture.
  • [17:52:45] <julianoliver> philv: i get 30-40fps on an Atom netbook.
  • [17:53:19] <Russ> If you can get 7fps now, I bet with dsp and/or neon integration, you could get realtime
  • [17:53:24] <julianoliver> philv: OpenCV is fast for capture when you follow a few basic rules. it's used in nearly every computer vision and robotic application precisely because it is pretty good code.
  • [17:53:37] <philv> Hrmn, I would have said about the opposite. :-)
  • [17:53:43] * TAK2004 (n=thomas@dslb-088-074-035-015.pools.arcor-ip.net) Quit (Remote closed the connection)
  • [17:53:44] * eFfeM (n=frans@195-241-226-180.ip.telfort.nl) has joined #beagle
  • [17:53:44] <philv> It's research-quality code.
  • [17:54:12] <geckosenator> what does taht mean?
  • [17:54:14] <julianoliver> philv: well i've seen it deployed fairly often..
  • [17:54:45] <philv> I've seen it used in prototypes, and prototypes that have been accidentally deployed as operational systems.
  • [17:55:27] <ldesnogu_> julianoliver, I guess it's faster on your Atom because there are some SSE(2) optimizations in the code
  • [17:55:33] <mib_5rz2nt> how come digital cameras are so small and they show pretty good video these days?
  • [17:56:22] <mru> if there are already sse optimisations, adding neon should be easy
  • [17:56:22] <julianoliver> ldesnogu_: yes, this could very well be the case.. i've done profiling using gprof but can't yet see where..
  • [17:56:46] <julianoliver> mru: Russ i compiled for NEON on the BB itself, with no gain. i suppose it's time to bite the bullet and try DSP.
  • [17:57:04] <Russ> compiling for neon really isn't going to do much for you
  • [17:57:18] <Russ> just like sse and simd isn't magic either
  • [17:57:22] <ldesnogu_> julianoliver, you need to use asm to get any significant speedup with neon
  • [17:57:30] <julianoliver> Russ: yes, it seems so.
  • [17:57:49] <julianoliver> ldesnogu_: right.. admittedly that's an area i haven't touched in years..
  • [17:57:57] <Russ> http://en.wikipedia.org/wiki/ARM_architecture#Advanced_SIMD_.28NEON.29
  • [17:58:43] <ldesnogu_> it'd be great to put arm asm code in opencv which was started by Intel :p
  • [17:58:58] <mru> I've seen about 10% speedup with the armcc vectoriser
  • [17:59:05] <Russ> ldesnogu_: intel had fingers in the ARM pie for quite some time
  • [17:59:06] <ldesnogu_> on opencv?
  • [17:59:12] <mru> on ffmpeg
  • [17:59:15] <ldesnogu_> Russ, yes but no NEON ;)
  • [18:00:04] <julianoliver> regardless, i can see most roads point to a rewrite of the lower layers of Intel's OpenCV: http://www.hbrobotics.org/wiki/index.php5/Beagle_Board#Using_the_BeagleBoard.27s_DSP_for_vision_processing
  • [18:01:37] <ldesnogu_> eeeeek they are using intrinsics for SSE
  • [18:01:55] <mru> eeek indeed
  • [18:01:55] <julianoliver> ldesnogu_: they are?
  • [18:02:02] <ldesnogu_> http://opencvlibrary.svn.sourceforge.net/viewvc/opencvlibrary/trunk/opencv/src/cv/cvfilter.cpp?revision=1649&view=markup
  • [18:02:08] <Russ> mib_5rz2nt: go read that whole page that julian linked
  • [18:02:41] * mpoullet|home (n=mpoullet@77-22-121-59-dynip.superkabel.de) has joined #beagle
  • [18:02:58] <ldesnogu_> that could mean a talented programmer (mru? :) could perhaps make something that'd run circles around opencv on an atom
  • [18:03:43] <mru> do I smell a challenge?
  • [18:03:50] <ldesnogu_> really?
  • [18:03:56] <ldesnogu_> :)
  • [18:04:02] <julianoliver> mru: many would be super greatful, except Intel of course ;)
  • [18:04:13] <mru> sounds perfect to me
  • [18:04:17] <julianoliver> hehe
  • [18:04:21] <mru> what does opencv do?
  • [18:04:33] <Russ> I'm guessing it be too much work to extend gcc to allow embedded of dsp code
  • [18:04:37] <ldesnogu_> computer vision library
  • [18:04:54] <julianoliver> it's a computer vision library used in a heap of robotic and embedded space applications.
  • [18:04:57] <mru> Russ: that's not how the dsp works
  • [18:05:35] <Russ> I realize that, I'm just imagining a very complex system
  • [18:06:14] <Russ> where a bit of object code gets compiled seperately and special bits of code are added to get it loaded into the dsp, etc
  • [18:06:51] <ldesnogu_> Russ, that looks similar to how you program Cell's PS3 under Linux
  • [18:07:06] <julianoliver> mru: it's used heavily for more than image processing: typically face, object and 'natural feature tracking'. has a heap of utils for matrix manipulation and projection also (2d->3d homographies etc).
  • [18:07:22] <Russ> ldesnogu_: and some CUDA stuff too
  • [18:08:12] <ldesnogu_> Russ, is using CUDA wise when OpenCL should be coming?
  • [18:08:22] <Russ> I'm not saying its wise
  • [18:08:34] <ldesnogu_> no, that was a side question :)
  • [18:08:39] <Russ> I'm just saying that a develpment model exists where you can embed code and it...
  • [18:08:40] <Russ> oh
  • [18:08:40] * guillaum1 (n=gl@AMontsouris-153-1-91-205.w90-2.abo.wanadoo.fr) Quit (Read error: 110 (Connection timed out))
  • [18:08:49] <philv> cubins aren't embedded in your code, heh
  • [18:08:53] <Russ> no, especially since its not open afaik
  • [18:10:26] * bmxr (n=bmxr@S01060018f3b11a22.vf.shawcable.net) Quit (Remote closed the connection)
  • [18:10:29] * ldesnogu_ suddenly notices OpenCV is C++ which might make mru angry
  • [18:10:44] * guillaum1 (n=gl@AMontsouris-153-1-55-75.w86-212.abo.wanadoo.fr) has joined #beagle
  • [18:10:49] <Russ> ldesnogu_: didn't someone already mention that its research quality code
  • [18:11:12] <mru> c++ does make me angry
  • [18:11:17] <ldesnogu_> I don't know what "research quality" means
  • [18:11:19] <Russ> I'm pretty sure that means c++ without knowledge of what bogs down c++
  • [18:11:25] <julianoliver> mru: understandably
  • [18:11:28] <mru> ldesnogu_: it means that it doesn't actually work
  • [18:11:49] <julianoliver> mru: OpenCV is used in a heap of robotic products. battle bots especially.
  • [18:11:58] <julianoliver> (unfortunately)
  • [18:11:58] <ldesnogu_> I've seen code written by scientists that was much smarter and brilliant than anything available otherwise
  • [18:12:13] <mru> few programmers have the slightest idea about how computers and compilers actually work
  • [18:12:13] <Russ> like minix?
  • [18:12:31] <mru> such knowledge is absolutely necessary to write efficient c++ code
  • [18:12:41] <philv> mru: wow, someone who finally gets it! :D
  • [18:12:50] <Crofton|work> "research quality code" rofl
  • [18:13:06] <koen> julianoliver: are you designing your own board?
  • [18:13:11] <mru> now those who know these things don't "need" c++ in the first place
  • [18:13:20] <koen> julianoliver: if so, you could use the ISP module to capture the video
  • [18:13:26] <Crofton|work> my friend just started in grad school and is refactoring some "research quality" code into something maintainable
  • [18:13:28] <mru> so we're left with a bunch of c++ programmers that don't know what they're doing
  • [18:13:35] <ldesnogu_> C++ should only be used by people having spent years looking at generated code
  • [18:13:36] <koen> julianoliver: if not, you'll have to profile and see what's the slow bit
  • [18:14:09] <koen> julianoliver: when I tested face recognition on the beagle with opencv the most time was spent in parsing the haarcascade xml file
  • [18:14:10] <mru> it's actually sometimes easier to understand what c++ code does by looking at the assembler output
  • [18:14:11] <Crofton|work> what is the current state of oprofile on cortex-a8?
  • [18:14:20] <julianoliver> koen: thanks, i'm still gprof'ing.
  • [18:14:20] <mru> Crofton|work: working
  • [18:14:22] <ldesnogu_> mru, couldn't agree more...
  • [18:14:29] <Russ> ldesnogu_: I fear any other method would just lead to cargo cult programming
  • [18:14:53] <julianoliver> koen: this is my test code for raw capture.. got to get the frames first! http://rafb.net/p/52QDyf72.html
  • [18:15:17] <mru> ldesnogu_: people who have spent years looking at generated code know better than to use c++
  • [18:15:26] <eFfeM> koen, hi, had a nice weekend ?
  • [18:15:28] <mru> because they know what a mess it makes
  • [18:15:30] <ldesnogu_> mru, that's the point :-)
  • [18:15:58] <ldesnogu_> young programmers around me don't understand why I never use C++
  • [18:16:01] <julianoliver> mru: several examples of what OpenCV is used for can be seen here: http://www.youtube.com/results?search_type=&search_query=opencv
  • [18:17:03] <koen> julianoliver: http://www.flickr.com/photos/koenkooi/3039017696/
  • [18:17:05] <mru> ldesnogu_: how old are you, if you don't mind?
  • [18:17:24] <Russ> I thought i recognized opencv from somewhere
  • [18:17:27] <koen> eFfeM: yep, had a very nice weekend, and you?
  • [18:17:31] <julianoliver> koen: yes i've run the face detection code on the BB but it's awfully slow..
  • [18:17:37] <julianoliver> i think 4 seconds per face.
  • [18:17:42] <eFfeM> koen, me too, ty
  • [18:17:56] <Crofton|work> koen, you are to result oriented
  • [18:18:02] <Russ> my school uses it quite a bit
  • [18:18:08] <julianoliver> koen: admittedly i didn't target the DSP, just stock samples from the Angstrom distribution of OpenCV
  • [18:18:10] <geckosenator> does the face detection run on the dsp?
  • [18:18:11] <ldesnogu_> mru, hum let's say I've been programming for 28 years :-)
  • [18:18:15] * Crofton|work is off to resemble an overweight 50 year old ona bicycle
  • [18:18:20] <eFfeM> koen, apart from my usb problems (http://groups.google.com/group/beagleboard/browse_thread/thread/fb3caeb7ffdcc02f#)
  • [18:18:28] <Crofton|work> ldesnogu, ok, that makes you mid 30's
  • [18:18:38] * eFfeM does not understand why a device works when directly connected and not through a hub
  • [18:18:57] * Crofton|work first wrote code circa 1977
  • [18:19:15] <geckosenator> eFfeM: usb is complicated :-P
  • [18:19:21] <Crofton|work> this does not mean i am very good though
  • [18:19:37] <eFfeM> btw n00b q: if I want to write a do_install that adds some things do I have to start with oe_runmake install or what
  • [18:19:46] <ldesnogu_> Crofton, no I'm early 40 :)
  • [18:20:02] <eFfeM> geckosenator: yes; life is strange, but usb is more strange
  • [18:20:48] * eFfeM wrote his first proggie in 1976 :-) in pascal and a little bit later in 8008 asm :-)
  • [18:20:49] <koen> eFfeM: remember that a hub re-encodes the complete usb bitstream
  • [18:21:10] <jkridner> Pascal in '76?
  • [18:21:14] <koen> eFfeM: so if you have a timing issue, or a buggy hub....
  • [18:21:17] * mru started poking at computers around age 12
  • [18:21:23] * kozak (n=subbu@117.192.13.5) Quit ("Leaving.")
  • [18:21:24] <eFfeM> koen, true but I would expect the code to drive a hub be stable, i tried different hubs
  • [18:21:24] <Crofton|work> :)
  • [18:21:32] * koen wasn't around in the 70s
  • [18:21:36] <eFfeM> but ofc they could be the same chipsets
  • [18:21:38] * abitos (n=nixgibts@91.46.71.169) Quit (Read error: 110 (Connection timed out))
  • [18:21:40] <Russ> eFfeM: does this work on your PC EHCI?
  • [18:21:44] <Crofton|work> koen missed all the fun
  • [18:21:50] <Russ> koen: I was for 2 1/2 months
  • [18:21:55] * mru missed a lot of the fun
  • [18:22:01] <julianoliver> koen: this is face detection ;) http://code.google.com/p/ehci/
  • [18:22:14] <eFfeM> russ, it works on pc directly not sure if I tried the hub on pc, will do so
  • [18:22:15] <geckosenator> eFfeM: also the hub can run at one speed, and the device at another
  • [18:22:17] <ldesnogu_> jkridner, Pascal can be traced back to late 69
  • [18:22:19] <mru> trying to get some of the action on embedded stuff now
  • [18:22:23] <geckosenator> eFfeM: does it run at high speed though?
  • [18:22:24] <Russ> what a poorly named project
  • [18:22:32] <eFfeM> but the same hub worked on nslu2
  • [18:22:37] <geckosenator> it might actually use a different driver depending on if you use a hub or not
  • [18:22:41] * jkridner only remembers BASIC, Fortran, and assembly in '76. I was only 5, so I guess I might have missed that Pascal existed.
  • [18:22:42] * mru never learned pascal
  • [18:22:45] <julianoliver> Russ: hehe agreed..
  • [18:22:56] <eFfeM> well the hub itself works because my usb ethernet dongle works for me
  • [18:23:10] <geckosenator> well I get "Error: selected processor does not support `cpsie i'
  • [18:23:14] <jkridner> I wrote a couple labs in Pascal after learning C. Seemed like a pointless language from where I sat.
  • [18:23:17] <geckosenator> when I try to compile 2.6.26
  • [18:23:26] <koen> julianoliver: I'm still waiting for artoolkit (free) to support GLES :(
  • [18:23:30] <mru> geckosenator: why so old kernel?
  • [18:23:32] <Russ> I want to see if that guy shaves off an eyebrow if that ehci thing wigs out
  • [18:23:45] * ogra (n=ogra@ubuntu/member/ogra) Quit (Read error: 113 (No route to host))
  • [18:23:58] <Russ> eFfeM: device reset after firmware loading is a complex thing
  • [18:24:03] * jkridner thinks eFfeM must have a simple life if USB is more complex.
  • [18:24:20] <eFfeM> jkridner pascal was developed 68/69 and published in 70. then again it could be that my first program was in algol
  • [18:24:24] * eFfeM even did f66
  • [18:24:39] <geckosenator> mru: I'm trying to figure out why my wifi card only works in older kernels
  • [18:24:48] <mru> back to this opencv thing, is it possible to add assembler optimisations without writing any c++?
  • [18:24:59] <mru> geckosenator: that's a valid reason
  • [18:24:59] <eFfeM> russ, true; i have also problems with the webcam, will try to test on pc
  • [18:25:19] <Russ> eFfeM: with the same hub
  • [18:25:30] <geckosenator> mru: eFfeM has the same issues with his wifi card
  • [18:26:09] <ldesnogu_> mru, that won't be trivial, opencv massively uses operator and template
  • [18:26:13] <eFfeM> russ ofc with the same hub, directly it works
  • [18:26:20] <mru> ldesnogu_: idiots
  • [18:26:31] <Russ> ofc?
  • [18:26:31] <eFfeM> and wifi zd1211 stopped loading firmware in .28 used to work in .27
  • [18:26:32] <ldesnogu_> but there are some private structures
  • [18:26:37] <mru> so that's what "research quality" means
  • [18:26:43] <ldesnogu_> mru, look line 1878 of http://opencvlibrary.svn.sourceforge.net/viewvc/opencvlibrary/trunk/opencv/src/cv/cvfilter.cpp?revision=1649&view=markup
  • [18:26:48] <eFfeM> jkridner still thinks life is simple after that :-)
  • [18:26:53] <ldesnogu_> I guess this can be rewritten in asm
  • [18:27:27] <geckosenator> eFfeM: can you compile .27 and get wifi to work?
  • [18:27:30] <geckosenator> eFfeM: I'm having trouble with that
  • [18:28:21] <julianoliver> .
  • [18:28:21] <julianoliver> .
  • [18:28:32] * julianoliver (n=julian@164.pool85-53-3.dynamic.orange.es) Quit (Remote closed the connection)
  • [18:28:44] * julianoliver (n=julian@164.pool85-53-3.dynamic.orange.es) has joined #beagle
  • [18:28:49] <eFfeM> geckosenator: haven't tried to build .27 myself, not sure if it is still on the feed
  • [18:28:52] <mru> ldesnogu_: have these people never heard of function pointers?
  • [18:29:00] <geckosenator> eFfeM: can you try?
  • [18:29:14] <ldesnogu_> mru, that'd make the code readable
  • [18:29:15] <eFfeM> later, trying on the target now
  • [18:29:53] <mru> how do they choose which implementation to use?
  • [18:30:03] <eFfeM> usb 1.1 webcam works fine through this hub on opensuse
  • [18:30:07] <ds2> bahh....who needs function pointers or for that matter any specific ttypes of pointers...make everything a void *
  • [18:30:08] <ldesnogu_> they only have one I guess :-)
  • [18:30:23] <ldesnogu_> hum they probably have plain C++ too in fact
  • [18:30:29] <geckosenator> ds2: void * isnt' compatible with function pointers
  • [18:30:42] <geckosenator> ds2: since a function pointer might not fit in void *
  • [18:30:49] <ds2> geckosenator: they are if you call though with in line assembly ;)
  • [18:30:53] <eFfeM> pvr works also like a charm through the hub on this opensuse system (it is a 2.0 device)
  • [18:30:57] <geckosenator> ds2: not portable :-P
  • [18:31:02] <julianoliver> eFfeM: are you trying to use a webcam on the BB?
  • [18:31:08] <eFfeM> yes
  • [18:31:14] * julianoliver 's router barked and so i lost track.
  • [18:31:16] <mru> geckosenator: void * is compatible with function pointers on all systems that matter
  • [18:31:16] <ds2> unless you are on a platform where size of data ptr == size of code ptr, they will fit just fine
  • [18:31:20] <Russ> fine, cast it
  • [18:31:28] <ldesnogu_> that's crazy, opencv API is C and the internal C++
  • [18:31:28] <ds2> screw potability
  • [18:31:34] <geckosenator> mru: I can think of architectures I care about that they are not
  • [18:31:36] <ds2> silly userland notions :P
  • [18:31:41] <julianoliver> eFfeM: i have one working today, through a powered hub, but only at 7fps @320x240
  • [18:31:49] <eFfeM> julianoliver: on the otg port it more or less works but I get colored stripes in the picture, on the ehci it does not
  • [18:31:55] <mru> geckosenator: can you give an example, I'm curious
  • [18:32:01] <eFfeM> julianoliver: I used DISPLAY=:0.0 mplayer -tv width=320:height=240:driver=v4l2 -fps 15 tv://
  • [18:32:09] <julianoliver> eFfeM: right.. i don't have such problems..
  • [18:32:09] <eFfeM> stupid smileys
  • [18:32:11] <geckosenator> mru: an avr with more than 128k of flash
  • [18:32:13] <ldesnogu_> mru, I bet some DSP :)
  • [18:32:16] <julianoliver> eFfeM: one moment..
  • [18:32:32] <mru> geckosenator: not likely to run opencv though
  • [18:32:35] <geckosenator> mru: it's true that they currently use trampolines.. but I might need to change that for the new avrs which support up to 4 megs
  • [18:32:36] <julianoliver> eFfeM: try: mplayer -cache 128 -tv driver=v4l2:device=/dev/video1:width=640:height=480 -vo xv tv://
  • [18:32:40] <geckosenator> mru: yes fair enough
  • [18:32:42] <julianoliver> eFfeM: (change the capture size)
  • [18:32:43] <Russ> eFfeM: that doen't work on most webcams I don't think
  • [18:32:55] <Russ> and why the video1?
  • [18:33:01] <mru> geckosenator: 8-bit avr?
  • [18:33:03] <geckosenator> mru: right now they just have stub functions that jump to the high address though
  • [18:33:04] <eFfeM> julianoliver: russ, that very same command works perfectly on opensuse
  • [18:33:05] * jkridner|work (n=a0321898@nat/ti/x-c3a8f051972f194c) has left #beagle
  • [18:33:07] <geckosenator> mru: yeah
  • [18:33:11] * HokieTux (n=HokieTux@wireless-5192.wireless.ece.vt.edu) has joined #beagle
  • [18:33:14] <geckosenator> mru: so code pointers would need to be 3 bytes
  • [18:33:14] <mru> avr32 is like any 32-bit system
  • [18:33:23] <geckosenator> mru: this is 8 bit htough
  • [18:33:27] <Russ> "Selected input hasn't got a tuner!"
  • [18:33:28] <julianoliver> eFfeM: you need to look after memory consumption, at least i did with mplayer.
  • [18:33:29] <Russ> thats what I get for my webcam
  • [18:33:47] <eFfeM> and it streams on my beagle too but gets color artefacts
  • [18:33:50] <julianoliver> eFfeM: mplayer -cache 128 -tv driver=v4l2:device=/dev/video1:width=320:height=240 -vo xv tv:// works for me.
  • [18:34:02] <julianoliver> oops /dev/video0 i meant to say
  • [18:34:20] <julianoliver> Russ: try /dev/video0
  • [18:34:31] * Wowbagger_ (n=wowbagge@d154-20-187-250.bchsia.telus.net) has joined #beagle
  • [18:34:35] <ds2> mru: 8051's with SDCC may have issues but I think the compiler tries to fix that with attributes
  • [18:34:37] <julianoliver> vlc simply hangs trying to capture from the v4l2://
  • [18:34:42] <Russ> xv works, but it claims [VO_XV] Could not grab port 57
  • [18:34:52] <mru> ds2: again unlikely to run opencv
  • [18:35:16] <ds2> mru: there are bigger things that will break with SDCC then ptr sizes
  • [18:35:19] <julianoliver> Russ: odd..
  • [18:35:58] <eFfeM> see for an example picture from my webcam: http://mail.google.com/mail/?ui=2&ik=44cf509bc5&view=att&th=11fc9067ba55df51&attid=0.1&disp=attd&realattid=f_frtnai6x0&zw
  • [18:36:13] <julianoliver> mru: what is your preliminary verdict regarding performative OpenCV on a BB?
  • [18:36:22] <julianoliver> (if at all)
  • [18:36:23] <Russ> cheese works great though
  • [18:36:34] <julianoliver> Russ: 'cheese'?
  • [18:36:42] <ds2> 286 protected mode might be another example
  • [18:36:43] <mru> julianoliver: painful
  • [18:36:48] <julianoliver> hmmm
  • [18:36:48] <eFfeM> see also this thread: http://mail.google.com/mail/?ui=2&ik=44cf509bc5&view=att&th=11fc9067ba55df51&attid=0.1&disp=attd&realattid=f_frtnai6x0&zw
  • [18:37:00] <ds2> with a medium or compact model compile
  • [18:37:05] <mru> julianoliver: I certainly won't do it unless I'm paid to
  • [18:37:14] <julianoliver> understandable..
  • [18:37:33] * mpoullet|home (n=mpoullet@77-22-121-59-dynip.superkabel.de) has left #beagle
  • [18:37:40] <julianoliver> mru: it's a little OT but how much work would you envisage such a rewrite taking?
  • [18:38:00] <mru> julianoliver: no idea
  • [18:38:09] <mru> I don't know how much code is involved
  • [18:38:28] <philv> julianoliver: what is your end application, anyways?
  • [18:38:39] <julianoliver> a good thing to do is to gprof the facedetect app in the standard Angstrom distribution of OpenCV.. that gives an idea.
  • [18:38:49] <julianoliver> philv: a hand-held augmented reality application.
  • [18:38:58] <mru> hand-tuning a page of C code in assembler can take anything from an hour to a week
  • [18:39:04] <mru> it depends on what it does
  • [18:39:07] <julianoliver> here's my last Augmented Reality application: http://www.vimeo.com/1320756
  • [18:39:51] <julianoliver> i'm working on another right now, particularly teaching computers to recognise buildings and other urban features.
  • [18:40:11] <eFfeM> geckosenator: now doing bitbake -b linux-omap_2.6.27.bb
  • [18:40:17] <philv> Ah, I did something like the latter, but for satellite imagery
  • [18:40:45] <mru> philv: sure it wasn't a missile guidance system?
  • [18:41:04] * julianoliver coughs
  • [18:41:09] <philv> mru: Client was non-military, hah.
  • [18:41:13] <eFfeM> n00b q: if I want to write a do_install that adds some things do I have to start with oe_runmake install or what
  • [18:41:18] <mru> uh oh... what's that thing coming towards my house?
  • [18:41:22] <julianoliver> philv: it's a fun area to develop within
  • [18:41:50] <philv> Yeah, it's interesting, for sure.
  • [18:45:06] * ogra (n=ogra@ubuntu/member/ogra) has joined #beagle
  • [18:45:29] <eFfeM> anyone an idea on my install q: I tried both oe_runmake install and oe_runmake install DESTDIR=${D}
  • [18:45:30] <eFfeM> both without success
  • [18:48:41] * mib_9p8ets (i=8ddbcc55@gateway/web/ajax/mibbit.com/x-166280f7e464496c) has joined #beagle
  • [18:49:21] <eFfeM> webcam on usb hub on beagle gives pwc: isoc_init() submit_urb 0 failed with error -28
  • [18:49:32] <mib_9p8ets> Does anyone know anything about installing programs like cheese in Angstrom linux on the BB?
  • [18:50:17] <julianoliver> eFfeM: i had this but not with a PWC and only with the un-powered hub..
  • [18:50:47] <eFfeM> this is with a powered hub on ehci, on the otg port I get the picture I pasted before
  • [18:51:06] <julianoliver> odd..
  • [18:51:20] * florian (n=fuchs@f049169204.adsl.alicedsl.de) has joined #beagle
  • [18:53:56] <eFfeM> there is a config option for -28: Improved Transaction Translator scheduling (EXPERIMENTAL) trying that one
  • [19:04:42] <eFfeM> hm the .27 kernel i made does not come beyond uncompressing linux :-(
  • [19:04:46] <eFfeM> no idea why this is
  • [19:04:57] <eFfeM> coffee break ....
  • [19:16:15] * ClaudeQC (n=ClaudeQC@bas1-quebec03-1279635805.dsl.bell.ca) has left #beagle
  • [19:19:06] * atin_ (n=atin@c-71-232-7-71.hsd1.ma.comcast.net) has joined #beagle
  • [19:24:00] <atin_> woof, ogre3d on beagle! way to go koen :)
  • [19:24:35] <atin_> (I haven't been keeping up with beagle since january, just catching up on the emails :) )
  • [19:27:23] * mib_9p8ets (i=8ddbcc55@gateway/web/ajax/mibbit.com/x-166280f7e464496c) Quit ("http://www.mibbit.com ajax IRC Client")
  • [19:28:52] <kblin> atin_: at how many seconds per frame? ;)
  • [19:29:36] <atin_> I haven't tried it yet - just saw the posts on the mailing list :) - I'm still building my world since I haven't updated anything since like january or early feb!
  • [19:30:14] <atin_> (and mostly trying to catch up on musb work since I was working on that too)
  • [19:30:25] <eFfeM> '
  • [19:30:40] <eFfeM> geckosenator: can't get the .27 kernel beyond uncompressing linux
  • [19:31:34] <eFfeM> how far did you get ?
  • [19:34:11] <eFfeM> i build zd1211 with debug my log is here: http://pastebin.com/m377acab8
  • [19:34:30] <julianoliver> eFfeM: did you build cheese yourself?
  • [19:34:48] <julianoliver> (or is there an ipkg around i don't know of..)
  • [19:34:49] <eFfeM> cheese?
  • [19:35:02] * eFfeM does not understand
  • [19:35:04] <julianoliver> oh.. perhaps i have my nicks mixed up.
  • [19:35:10] <julianoliver> eFfeM: a capture application.
  • [19:35:26] <eFfeM> never used it, I played with vidcat, mplayer and motion
  • [19:35:32] <julianoliver> Russ: did you build cheese yourself?
  • [19:35:48] <julianoliver> eFfeM: (sorry, a long weekend is taking its toll..)
  • [19:35:53] <eFfeM> np
  • [19:39:59] <atin_> sakoman__: ping?
  • [19:41:30] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [19:44:19] <eFfeM> anyone an idea why my linux-omap-2.6.27-r14 does not get beyond uncompressing ?
  • [19:47:12] * mib_5rz2nt (i=63921a39@gateway/web/ajax/mibbit.com/x-68193047a6bbb7a7) Quit ("http://www.mibbit.com ajax IRC Client")
  • [19:50:06] * mib_5rz2nt (i=63921a39@gateway/web/ajax/mibbit.com/x-70dd2d881f549c19) has joined #beagle
  • [19:53:05] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-4c2541d13811d9ff) has joined #beagle
  • [19:53:59] * cgs (n=carl@130.120.124.202.static.snap.net.nz) has joined #beagle
  • [19:55:16] <eFfeM> do i need a specific u-boot version that goes with .27 ????
  • [19:55:45] <eFfeM> or is there a differnet address?
  • [20:00:16] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-4c2541d13811d9ff) Quit ("http://www.mibbit.com ajax IRC Client")
  • [20:00:49] * ahto (i=5b9bb29a@gateway/web/ajax/mibbit.com/x-9e6d11963c48d94e) has joined #beagle
  • [20:01:18] * recalcati (i=5d905a4e@gateway/web/ajax/mibbit.com/x-f15dc9a4dd4de9f4) has joined #beagle
  • [20:04:30] <sakoman__> atin_: pong
  • [20:05:04] <Crofton> eFfeM, enable LL DEBUG
  • [20:05:13] <atin_> sakoman__: hey, what is the latest u-boot repository? did the musb stuff all get into our mainline?
  • [20:05:39] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [20:05:43] * fulgas is now known as FuL|OUT
  • [20:06:21] <sakoman__> atin_: same place as always: http://www.sakoman.net/cgi-bin/gitweb.cgi?p=u-boot-omap3.git;a=summary
  • [20:06:34] <sakoman__> the usb stuff has not mde it upstream
  • [20:06:39] <atin_> so did the omap3-usb get merged up?
  • [20:06:43] <atin_> ah
  • [20:06:50] <eFfeM> Crofton, LL DEBUG for what? getting.27 to run? or getting more info on my webcam problem? or on the wifi problem ?
  • [20:06:51] <sakoman__> current state is in omap3-dev-usb
  • [20:07:09] <atin_> I got email from Wolfgang today asking about my changes, I thought maybe that was the usb stuff
  • [20:07:16] <sakoman__> I tried to do a merge with upstream to get it up to date, but there are many merge errors
  • [20:07:24] <atin_> but maybe it was the other non-usb related stuff he was talking about
  • [20:07:28] <sakoman__> I don't have time to deal with them right now
  • [20:07:48] <atin_> ok, I'm just getting back up to speed - so let me see what I can do :)
  • [20:07:51] <sakoman__> atin_: yes, just about 100% of omap3-dev is upstream now
  • [20:07:54] * ahto (i=5b9bb29a@gateway/web/ajax/mibbit.com/x-9e6d11963c48d94e) Quit ("http://www.mibbit.com ajax IRC Client")
  • [20:07:59] <sakoman__> I think there is just one pending patch
  • [20:08:19] <sakoman__> much work is required to get omap3-dev-usb ready
  • [20:08:36] <atin_> ok, I've been getting email from Dirk so I think he has been doing stuff there
  • [20:08:41] <atin_> he might be the only one tho :)
  • [20:09:03] <sakoman__> yes, I think that is so. He is on holiday for a week or so now
  • [20:09:15] <sakoman__> getting his fix of sunshine :-)
  • [20:09:31] <atin_> ok, I'll build dev-usb soon - trying to build all our mainline stuff right now.
  • [20:09:40] <sakoman__> It is a nice day here, so I am on my way outside too!
  • [20:09:45] <atin_> "preferred version git of u-boot not available"
  • [20:09:49] <atin_> what does that mean?
  • [20:10:00] <sakoman__> from OE?
  • [20:10:02] <atin_> yeah
  • [20:10:14] <Crofton> maybe I should check the input voltage range on a linksys switch
  • [20:10:49] <sakoman__> atin_: I think it means that there is a preferred version set for the machine you are building, but no corresponding recipe for that preferred version
  • [20:11:06] <atin_> hmm, let me look at my conf stuff ...
  • [20:11:28] <sakoman__> so you'd have to look in you machine.conf to see what u-boot version is set and y to figure out why there is no recipe
  • [20:12:05] <atin_> MACHINE="beagleboard"
  • [20:13:16] <atin_> huh. beagleboard.conf has PREFERRED_VERSION_u-boot = "git"
  • [20:16:18] <atin_> sakoman__: I'll figure it out .. :) got enjoy the sun
  • [20:16:32] <atin_> sakoman__: its cloudy and cold in Boston
  • [20:17:08] * jkridner_ (n=jason@c-76-31-18-64.hsd1.tx.comcast.net) has joined #beagle
  • [20:18:21] <atin_> guess I should ask koen about oe stuff - the beagleboard.conf is from openembedded itself after I did a git pull.
  • [20:18:27] <eFfeM> well in the netherlands it is atm just dark
  • [20:18:39] <atin_> and it has preferred version for u-boot set to "git"
  • [20:20:02] <atin_> which actually seems correct to me . . .
  • [20:20:29] <atin_> eh, whatever - I'll be building it by hand anyway..
  • [20:24:13] * jkridner (n=jason@c-76-31-18-64.hsd1.tx.comcast.net) Quit (Read error: 110 (Connection timed out))
  • [20:35:29] * mib_5rz2nt (i=63921a39@gateway/web/ajax/mibbit.com/x-70dd2d881f549c19) Quit ("http://www.mibbit.com ajax IRC Client")
  • [20:35:47] <recalcati> hi everybody
  • [20:38:18] <Jiuy> hi
  • [20:39:06] <geckosenator> hi
  • [20:40:16] <recalcati> anybody is able to configure ipk for installing from the pc the packages ?
  • [20:40:48] <recalcati> for example opkg install gnash worked, but not opkg install vlc
  • [20:41:07] * atin_ is now known as atin_AWAY
  • [20:41:20] * FuL|OUT is now known as fulgas
  • [20:47:58] <ds2> sakoman:is all your PSP LCD changes in L-O?
  • [20:52:29] * javaJake (n=javaJake@unaffiliated/javajake) has joined #beagle
  • [20:53:29] <javaJake> I recently reinstalled my Gentoo OS, and now minicom isn't showing the BB output via serial. Is there anything particular that I should configure my system with to get minicom to function correctly?
  • [21:00:47] <javaJake> I've tried all the available serial ports, but nothing is working
  • [21:00:59] <Russ> did you follow the instructions?
  • [21:01:41] <recalcati> minicom without hardware and software flow control
  • [21:01:49] <recalcati> 115200 8n1
  • [21:04:29] <recalcati> ctrl a then z then o then choose Serial port setup and so uncheck if checked using F and G
  • [21:04:59] <Russ> recalcati: I kinda like people to solve things themselves, I hate spoon feeding
  • [21:05:01] <javaJake> recalcati: yea, I've done all that
  • [21:05:19] <recalcati> any permission pb?
  • [21:05:21] <Russ> o
  • [21:05:32] <Russ> minicom would tell you if it had a permissions problem
  • [21:05:50] <javaJake> Yea, running as root
  • [21:05:53] <AudioFreak|Away> good evening everyone
  • [21:05:57] <Russ> I'm assuming that you are plugging it into a hardwired port on your computer?
  • [21:06:03] * AudioFreak|Away is now known as AudioFreak
  • [21:06:17] <recalcati> yes, but you have to read messages. sometimes people don't read messages
  • [21:06:19] <javaJake> Russ: Yea
  • [21:07:15] <javaJake> Oh GOSH
  • [21:07:19] <javaJake> Russ: NO
  • [21:07:20] <javaJake> :|
  • [21:07:22] <javaJake> Dadgum!
  • [21:07:26] <javaJake> It was just a little teensy bit out
  • [21:07:31] <javaJake> ...
  • [21:07:33] <Russ> damned gentoo
  • [21:08:26] <AudioFreak> Russ: remember we tried to get the DSP working yesterday?
  • [21:08:38] <Russ> ya
  • [21:08:44] <AudioFreak> koen told it will not work at all.. what we tried
  • [21:08:54] <AudioFreak> i need dsplink instead of dspbridge...
  • [21:09:11] <AudioFreak> now trying to figure out how to get that working ;)
  • [21:09:30] <Russ> he'd be the one to ask
  • [21:10:06] <Russ> or maybe mru
  • [21:15:04] <Russ> julianoliver: you need to combine this stuff with face tracking so that it makes a regular screen 3d
  • [21:15:46] <julianoliver> Russ: yes i've already written the base code for this! i should be cleaning it up and uploading it before teaching on the topic, mid april.
  • [21:16:01] <Russ> That would be really cool
  • [21:16:13] * mib_fe4ubl (i=c23f8ba3@gateway/web/ajax/mibbit.com/x-6acd7a5b6491516b) has joined #beagle
  • [21:17:27] <julianoliver> Russ: i'll be sure to post the code here when it's cleaned up.
  • [21:18:44] <mib_fe4ubl> Rev C planned to be at Digi-Key by Mar 30
  • [21:18:49] <mib_fe4ubl> whats new with rev C?
  • [21:18:51] * pcgeil (n=steffen@p5B17ED32.dip.t-dialin.net) has joined #beagle
  • [21:19:23] <Russ> 256MB RAM, host EHCI interface, LCD interface
  • [21:19:23] <Russ> but check the wiki
  • [21:21:55] <eFfeM> is mar 30 confirmed; i heard htey wanted to get rid of the rev b's they had in stock first
  • [21:22:33] <javaJake> eFfeM: heheh, likely ;)
  • [21:24:04] <mib_fe4ubl> thats cool
  • [21:24:09] * Crofton|work (n=balister@pool-96-240-173-146.ronkva.east.verizon.net) Quit (Read error: 104 (Connection reset by peer))
  • [21:24:16] <mib_fe4ubl> but i've ordered the one that is on sale now :(
  • [21:26:29] * Crofton|work (n=balister@pool-96-240-173-146.ronkva.east.verizon.net) has joined #beagle
  • [21:29:41] * mib_fe4ubl (i=c23f8ba3@gateway/web/ajax/mibbit.com/x-6acd7a5b6491516b) Quit ("http://www.mibbit.com ajax IRC Client")
  • [21:31:38] * abitos (n=nixgibts@dslb-084-057-146-023.pools.arcor-ip.net) has joined #beagle
  • [21:33:08] * svolpe (n=Gerrath@unaffiliated/gerrath) Quit ("leaving")
  • [21:33:57] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [21:39:33] <AudioFreak> arrr, starting all over, rather frustrating weekend..
  • [21:39:46] <AudioFreak> had no progress at all it seems :(
  • [21:41:54] <Russ> AudioFreak: you learn by not being afraid to make mistakes
  • [21:43:03] <AudioFreak> yeah I know, and at least I learned what options don't work ;)
  • [21:49:55] <AudioFreak> hmm, once you know how to set up your openembedded environment it's done quite fast :P
  • [21:52:29] <AudioFreak> now waiting for bitbake to compile a new image from scratch
  • [21:53:02] <Russ> oe really is pretty cool
  • [21:53:31] <Russ> wonder if there is a way for me to augment my angstrom sources with the packages I've built with oe
  • [21:54:31] * eFfeM (n=frans@195-241-226-180.ip.telfort.nl) Quit ("Leaving.")
  • [21:55:02] <Crofton|work> russ what do you mean?
  • [21:55:25] <Russ> when I build packages with bitbake, I get a bunch of ipk files
  • [21:55:33] <Crofton|work> yeah
  • [21:55:33] <Russ> I want opkg update to see those
  • [21:55:51] <Crofton|work> you need to figure out how this feed stuff works :)
  • [21:56:06] <Crofton|work> are they of interest to the general public?
  • [21:56:21] <Russ> nope
  • [21:56:33] <Crofton|work> figure out how to add your own feed then
  • [21:56:35] <Russ> they all already exists within oe
  • [21:57:06] <Russ> has anyone else noticed how useful the networkmanager connection sharing is when connecting the beagleboard via usb?
  • [21:57:45] <recalcati> I modified /etc/opkg/beagleboard-feed.conf to point to my apache
  • [21:58:13] <recalcati> but I can install some packages, not all the packages present on my pc
  • [21:58:25] <Crofton|work> ok, now I have four serial ports available :)
  • [22:10:08] * may (i=4772ba7e@gateway/web/ajax/mibbit.com/x-d376f817ef17861b) has joined #beagle
  • [22:10:08] <Crofton> ok, OSK (OMAP1) is running again
  • [22:10:17] * zedstar (n=john@fsf/member/zedstar) Quit (Remote closed the connection)
  • [22:12:11] * felipec (n=felipec@a91-153-251-222.elisa-laajakaista.fi) Quit (Read error: 113 (No route to host))
  • [22:13:52] <geckosenator> I'm thinking about hooking up a monochrome lcd
  • [22:13:57] * ant__ (n=andrea@host48-248-dynamic.15-87-r.retail.telecomitalia.it) has left #beagle
  • [22:14:12] <geckosenator> normally they have 8 datalines, and do 8 pixels at a time
  • [22:14:19] <geckosenator> I guess linux framebuffer supports it?
  • [22:15:28] <recalcati> tcpdump -A doesn't want to show me the contents of html pages
  • [22:16:12] <Russ> use wireshark?
  • [22:18:58] * Wowbagger_ (n=wowbagge@d154-20-187-250.bchsia.telus.net) Quit ("Leaving")
  • [22:19:28] <ds2> Crofton: what tree is/was the OSK broken in?
  • [22:20:20] * bkero wonders if Greg KH's SGX code can be altered to work with the beagle.
  • [22:20:28] <recalcati> textual wireshark?
  • [22:20:45] <Crofton|work> ds2, I am trying de break oe suport
  • [22:22:33] <ds2> Oh that stuff
  • [22:22:41] <Russ> bkero: did he actually push out any userland stuff?
  • [22:22:57] <recalcati> Russ: also I can also sniff from loopback interface (tcpdump -i lo) because if I launch only tcpdump I get the followuing message:tcpdump: unsupported data link type USB_LINUX
  • [22:23:22] <Russ> recalcati: I think thats a warning
  • [22:23:33] <Russ> hmmm..I get worse errors
  • [22:23:33] <recalcati> no, it stops
  • [22:23:45] <Russ> tcpdump: Can't open USB bus file /sys/kernel/debug/usbmon/0t: No such file or directory
  • [22:23:48] <bkero> Russ: No, the userland hooks are in his patches, but he's cutting them out and submiting 2d-only
  • [22:23:50] <recalcati> I think it is due to usb to eth interface
  • [22:24:16] <bkero> Russ: My roommate has taken the 5 patches off LKML and got it merged successfully with 2.6.29-rc8. So it's working for him on his Vaio P
  • [22:24:18] <Russ> recalcati: wireshark does work though
  • [22:24:33] <bkero> Apparently the code is endian dependent though, which is the only setback for getting it to work on a beagle.
  • [22:24:33] <Russ> recalcati: er, wait
  • [22:24:37] <Russ> I'm wiresharking usb
  • [22:24:40] <recalcati> I'm compiling it.
  • [22:24:40] <Russ> protocol
  • [22:24:59] <geckosenator> does anyone know where to get the datasheet the describes the musb stuff?
  • [22:25:16] <recalcati> bitbake wireshark, but I need textual commandline
  • [22:25:21] <recalcati> thx
  • [22:26:31] <Russ> why not dump on the host side?
  • [22:27:48] <Russ> recalcati: weird, I can't figure it out either
  • [22:29:14] <recalcati> I need to know which link I'm going to ask to a server.
  • [22:29:39] <recalcati> It could be better to use a bridge tun+tap, but sniffing is easier
  • [22:30:03] <recalcati> wireshark can be textual?
  • [22:30:13] <Russ> recalcati: apparently, tcpdump thinks that all interfaces starting with usb are usbmon, and it tries to dump usb traffic
  • [22:30:36] <recalcati> ahhh, horrible
  • [22:30:42] <Russ> http://www.mail-archive.com/debian-bugs-dist@lists.debian.org/msg626072.html
  • [22:33:07] <recalcati> thx. I had eth0 and usb0, if I launch tcpdump -A only it tries all ifaces and so has the usb pb, but if I launch only tcpdump -i eth0 it works
  • [22:33:36] <ds2> geckosenator: yes.
  • [22:40:49] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [22:43:40] * Abraxas3d1 (n=michelle@cpe-24-94-7-251.san.res.rr.com) has joined #beagle
  • [22:43:52] * Openfree|away is now known as Openfree|home
  • [22:44:05] <ds2> Russ: ifrename them then!
  • [22:44:11] * Abraxas3d1 (n=michelle@cpe-24-94-7-251.san.res.rr.com) has left #beagle
  • [22:52:15] <geckosenator> ds2: where can I find them?
  • [22:54:24] <recalcati> tcpdump -l -n -A -i eth0 port 80 . it works perfectly
  • [22:55:56] <ds2> geckosenator: They are under an NDA. The doc is from Mentor but you need to talk to your chip maker (TI in this case)
  • [22:56:12] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) Quit (Read error: 110 (Connection timed out))
  • [22:56:15] <ds2> geckosenator: you got diffs against DSS2 for the PSP LCD?
  • [22:56:18] <geckosenator> really? that's horrible
  • [22:56:29] <ds2> Horrible? you are telling me :D
  • [22:56:32] <geckosenator> ds2: yes, but I only used it with dss1
  • [22:56:47] <geckosenator> I have a patch that is supposed to work
  • [22:56:51] <geckosenator> with dss2
  • [22:57:01] <ds2> I got stories but the contents of the story is under NDA :(
  • [22:57:06] <geckosenator> or maybe just the latest in tomi's tree already works
  • [22:57:15] <geckosenator> well I don't sign NDA anyway
  • [22:57:29] <ds2> the supposed patch would be a better starting point then just from scratch
  • [22:57:42] <geckosenator> well for dss1 it was super easy
  • [22:57:43] <ds2> someone of us are under them already for other purposes
  • [22:57:48] <geckosenator> I just modified some fields in a struct
  • [22:58:04] <geckosenator> well I'm not under any NDA and I don't plan on it
  • [22:58:19] <ds2> I know how to do it for DSS1.. done it before the SDP... it is the DSS2 I need cuz I want SVideo and other scaling stuff (dynamic switching to DVI)
  • [22:58:27] <geckosenator> oh
  • [22:58:30] <geckosenator> did you get the lcd working?
  • [22:58:49] <ds2> I am setting up software to test it after working through some power issues
  • [22:59:19] <ds2> turns out I need one more MOSFET since I think /DE is used by DVI so I can't just disable it
  • [22:59:46] <geckosenator> oh
  • [23:00:14] <geckosenator> i tried to send you the patch
  • [23:03:29] <ds2> yeah, you are not nat'ed correctly
  • [23:03:51] * brolin (n=brolin@190.157.13.75) has joined #beagle
  • [23:04:08] <ds2> your send has a source of 192.168.0.107 which is not very reachable from my end :D
  • [23:05:42] <ds2> got it, thanks
  • [23:06:04] <geckosenator> ds2: heh, I thought dcc could go through the server if needed
  • [23:06:09] <geckosenator> it would be direct though
  • [23:07:14] <ds2> DCC needs a nat helper to fix it up
  • [23:07:37] <ds2> all DCC does is send a CTCP request with the source address + port; the other end is expected to open a TCP connection to the address+port
  • [23:07:56] <ds2> Linux has a NAT help that will rewrite the request on the fly
  • [23:08:32] <geckosenator> well too late I mailed it
  • [23:12:22] * pcgeil (n=steffen@p5B17ED32.dip.t-dialin.net) Quit ()
  • [23:12:34] <Russ> no wonder the musb driver is in the state that it is in
  • [23:13:02] <Russ> I don't think most companies realize the ease and extent of corporate espionage
  • [23:13:19] <geckosenator> it would be nice to have specs on it
  • [23:13:46] <geckosenator> otherwise I don't want to waste time dealing with it, I'm just going to get the wifi adapter that works in 2.6.28
  • [23:13:55] <geckosenator> since it's $8.99
  • [23:14:19] <Russ> if companies knew, they would understand that their tech specs are peanuts
  • [23:15:07] <dasnake> why is omapfbplay really really slow (11fps) while mplayer goes fine without framedropping? (2.6.29-rc8-omap + dss2 tomba)
  • [23:16:11] <geckosenator> Russ: it's actually worse for them to not distribute it
  • [23:17:33] <ds2> I think I can say that the docs do not help with MUSB
  • [23:18:05] <ds2> I have been hip deep in MUSB and docs do not do much
  • [23:18:41] <geckosenator> well that's good to know
  • [23:19:29] <ds2> I think most of the people working on MUSB see it that way too
  • [23:21:44] <Russ> I wonder how many of the problems are silicon bugs
  • [23:22:41] <geckosenator> yeah
  • [23:22:51] <geckosenator> but maybe you can work around silicon bugs with software
  • [23:23:38] <Russ> usb bugs are notoriously hard to track down, and you generally don't find the silicon ones till the driver is very robust
  • [23:24:44] * abitos (n=nixgibts@dslb-084-057-146-023.pools.arcor-ip.net) Quit (Read error: 113 (No route to host))
  • [23:26:53] * kb5mu (n=kb5mu@cpe-24-94-7-251.san.res.rr.com) has joined #beagle
  • [23:28:37] <sakoman__> ds2: the panel driver for PSP display is already in DSS2
  • [23:28:42] * AudioFreak is now known as AudioFreak|sleep
  • [23:29:13] * recalcati (i=5d905a4e@gateway/web/ajax/mibbit.com/x-f15dc9a4dd4de9f4) Quit ("http://www.mibbit.com ajax IRC Client")
  • [23:29:19] <sakoman__> ds2: drivers/video/omap2/displays/panel-samsung-lte430wq-f0c.c
  • [23:34:30] * florian (n=fuchs@f049169204.adsl.alicedsl.de) Quit (Read error: 110 (Connection timed out))
  • [23:35:50] * ldesnogu_ (n=ldesnogu@ven06-2-82-247-86-183.fbx.proxad.net) Quit ()
  • [23:36:09] <ds2> sakoman: it isn't in my version of it... I am not blowing on my house of cards
  • [23:36:29] <ds2> strictly speaking that isn't the PSP LCD
  • [23:38:30] <geckosenator> it turns out I wasn't applying all the patches I needed for the kernel
  • [23:38:38] <geckosenator> so I"m doing it now
  • [23:42:47] * Abraxas3d1 (n=michelle@cpe-24-94-7-251.san.res.rr.com) has joined #beagle
  • [23:44:10] <geckosenator> I could really use an lcd bigger than the psp one
  • [23:45:04] * brolin (n=brolin@190.157.13.75) Quit ("Lost terminal")
  • [23:46:48] * Abraxas3d1 (n=michelle@cpe-24-94-7-251.san.res.rr.com) has left #beagle
  • [23:46:55] * Abraxas3d1 (n=michelle@cpe-24-94-7-251.san.res.rr.com) has joined #beagle
  • [23:54:01] * Tukachinchilla (i=816ef113@gateway/web/ajax/mibbit.com/x-0535d21f307cb7b1) has joined #beagle
  • [23:55:00] <Tukachinchilla> does anyone know where I can get the footprints for all the components used in the beagle board?
  • [23:56:36] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [23:56:43] * Abraxas3d1 (n=michelle@cpe-24-94-7-251.san.res.rr.com) has left #beagle