| 00:00.52 | *** join/#edev SpeedEvil1 (~user@tor/regular/SpeedEvil) |
| 00:41.01 | *** join/#edev bsdfox_ (~bsdfox@unaffiliated/bsdfox) |
| 01:33.42 | *** join/#edev roxfan (dunno@91.180.158.57) |
| 01:45.31 | *** join/#edev SpeedEvil (1000@tor/regular/SpeedEvil) |
| 04:42.58 | *** join/#edev mbuf (~user@118.102.130.6) |
| 04:55.43 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 07:54.11 | *** join/#edev amurray (amurray@cpc1-bath4-0-0-cust211.aztw.cable.virginmedia.com) |
| 08:18.47 | *** join/#edev SpeedEvil (~user@tor/regular/SpeedEvil) |
| 08:48.47 | *** join/#edev florian_kc (~fuchs@Maemo/community/contributor/florian) |
| 09:10.05 | *** join/#edev gratz|home (~gratz@cpc3-brig15-2-0-cust237.3-3.cable.virginmedia.com) |
| 09:29.13 | *** join/#edev mike_s (~quassel@82-69-35-33.dsl.in-addr.zen.co.uk) |
| 09:33.57 | *** join/#edev gratz|home (~gratz@cpc3-brig15-2-0-cust237.3-3.cable.virginmedia.com) |
| 09:42.34 | *** join/#edev pirho (debian-tor@gateway/tor-sasl/pirho) |
| 09:47.37 | *** join/#edev mike_s_ (~quassel@82-69-35-33.dsl.in-addr.zen.co.uk) |
| 10:14.56 | *** join/#edev mike_s (~quassel@82-69-35-33.dsl.in-addr.zen.co.uk) |
| 10:50.21 | *** join/#edev SpeedEvil (1000@tor/regular/SpeedEvil) |
| 11:36.47 | *** join/#edev amurray (amurray@cpc1-bath4-0-0-cust211.aztw.cable.virginmedia.com) |
| 12:01.30 | *** join/#edev PierreLuc (~PierreLuc@ip-66-254-44-78.mqdsl.megaquebec.net) |
| 12:03.07 | *** join/#edev gustavoz (~gustavoz@host57.190-30-11.telecom.net.ar) |
| 12:36.00 | *** join/#edev PierreLuc (~PierreLuc@ip-66-254-44-78.mqdsl.megaquebec.net) |
| 13:38.06 | *** join/#edev melchior (~melchior@AStrasbourg-552-1-46-148.w90-26.abo.wanadoo.fr) |
| 13:50.16 | *** join/#edev kergoth (~kergoth@ip24-251-170-95.ph.ph.cox.net) |
| 14:22.54 | *** join/#edev gustavoz (~gustavoz@host57.190-30-11.telecom.net.ar) |
| 14:27.04 | *** join/#edev prpplague (~danders@nat/ti/x-ywklxnjgiokjqbkh) |
| 14:49.30 | *** join/#edev sndcrb (~sndcrb@easy-ff00c200-144.dhcp.inet.fi) |
| 15:16.55 | *** join/#edev kergoth_ (~kergoth@ip24-251-170-95.ph.ph.cox.net) |
| 15:17.53 | kergoth_ | morning |
| 15:53.46 | *** join/#edev gratz|home (~gratz@cpc3-brig15-2-0-cust237.3-3.cable.virginmedia.com) |
| 16:01.52 | chouimat | morning |
| 16:07.11 | dijenerate | prpplague: ping |
| 16:07.17 | prpplague | pong |
| 16:07.27 | dijenerate | prpplague: how goes it? |
| 16:07.31 | SpeedEvil | Anyone done FIR filters? |
| 16:07.31 | prpplague | dijenerate: it goes |
| 16:07.48 | prpplague | dijenerate: be down there on the 9th thru the 14th |
| 16:08.03 | dijenerate | prpplague: sweet, well this time we should have a beer |
| 16:08.14 | prpplague | dijenerate: really need to get someone to help me out with some support issues down there |
| 16:08.49 | dijenerate | prpplague: that won't be a prob, we can talk |
| 16:08.59 | prpplague | dandy |
| 16:09.59 | Mirell | Hmm.. |
| 16:10.09 | Mirell | I are happy. I actually wrote a simple kernel module. |
| 16:14.05 | SpeedEvil | Congratulations! |
| 16:14.55 | Mirell | chille: And actually, it was because I was investigating your question |
| 16:15.16 | Mirell | chille: Turns out, no, there's not, you have to write a bloody kernel module to do it accurately. |
| 16:16.12 | Mirell | Obviously you'd never need to know the HZ of the kernel for a conversion from jiffies to milliseconds in USERSPACE |
| 16:16.37 | Mirell | I think I'll try poking and see what it takes to make a proc file now.. |
| 16:18.55 | SpeedEvil | can't you work it out from /proc/timer_list |
| 16:19.10 | Mirell | There's no HZ listed there. |
| 16:19.26 | Mirell | Sure, the default is 100. On x86. Sometimes. |
| 16:20.11 | SpeedEvil | idle_jiffies divided by the sum of the idle_times? |
| 16:20.46 | SpeedEvil | It's of the right order of magnitude |
| 16:20.52 | SpeedEvil | Not the sums |
| 16:21.09 | Mirell | Hmm. |
| 16:21.11 | Mirell | Perhaps. |
| 16:21.23 | Mirell | But that wouldn't have led me down trying to learn how Linux timers and kernel modules work. |
| 16:21.27 | Mirell | So that's far less interesting. |
| 16:23.09 | SpeedEvil | Oooh! |
| 16:23.17 | SpeedEvil | I have a sort-of-related question. |
| 16:23.57 | SpeedEvil | cat /proc/timer_Stats |
| 16:24.00 | SpeedEvil | Sample period: 2.506 s |
| 16:24.00 | SpeedEvil | <PROTECTED> |
| 16:24.00 | SpeedEvil | <PROTECTED> |
| 16:24.00 | SpeedEvil | <PROTECTED> |
| 16:24.06 | penguinece | SpeedEvil, what about FIR filters? |
| 16:24.08 | SpeedEvil | What sets the process name and PID |
| 16:24.34 | SpeedEvil | If I have a FIR, and I want to make an EQ, can I simply take an array of the EQ set to -20dB cut at 100Hz (say), 20dB boost at 100Hz, and then interpolate between them to make the desired amount of boost or cut? |
| 16:24.34 | SpeedEvil | And can I then simply add the coefficients of all the other bands, and put them in the filter? |
| 16:24.34 | SpeedEvil | <PROTECTED> |
| 16:26.10 | Mirell | Fuzzy logic. |
| 16:26.43 | penguinece | -20db at 100hz and 20db at 100hz? EQ == equalizer? |
| 16:26.47 | SpeedEvil | Specifically, I had a kernel function - in cpufreq actually - coming up as 'awk' repeatedly, and I was unable to find what should be sertting it |
| 16:26.49 | SpeedEvil | penguinece: yes |
| 16:26.54 | SpeedEvil | Sorry. |
| 16:27.05 | penguinece | I'm confused how you have -20db and 20db at 100hz. |
| 16:27.34 | penguinece | SpeedEvil, are you familiar with bode plots? |
| 16:27.38 | SpeedEvil | penguinece: For each frequency band of the EQ, you have two sets of coefficients for the FIR filter - +20dB, and -20dB. |
| 16:28.16 | SpeedEvil | Can you simply linearly interpolate between the two states, to get a sane filter that will boost or cut proportionally. |
| 16:28.35 | penguinece | Ah I see what you're going for. |
| 16:28.54 | SpeedEvil | I think there are issues with running out of headroom. |
| 16:28.59 | SpeedEvil | But other than that. |
| 16:29.52 | penguinece | SpeedEvil, so you want to make an equalizer out of a set of cascaded FIR filters and you're wondering if you can linearly interpolate the coefficient values for an individual stage? |
| 16:30.36 | SpeedEvil | No - I'm wondering if I can use one FIR filter, with a set of 2* equaliser band coefficients. And simply add the filter coefficients for each band into one filter. |
| 16:35.54 | penguinece | I guess I don't understand what you're going for. sorry. |
| 16:36.18 | SpeedEvil | Background. I have a sound chip with a FIR in it. Just one. |
| 16:36.30 | SpeedEvil | I want to make an equaliser. |
| 16:37.11 | SpeedEvil | I know that in principle I can go from arbitrary filtershape to FIR coefficients -> arbitrary filter (with ripple) through maths I don't quitre understand yet. |
| 16:37.59 | *** join/#edev gustavoz (~gustavoz@host57.190-30-11.telecom.net.ar) |
| 16:38.24 | SpeedEvil | I was wondering if I could - in the implementation - skip all of the hairy maths - and simply sum the FIR coefficient from a simple boost or cut on each band. |
| 16:39.33 | SpeedEvil | Sorry - I'm probably being confusing, I should do some reading. |
| 17:05.31 | penguinece | Sorry I got distracted. it is indeed fun to draw semi-random lines and make an FIR filter out of it, I've written matlab scripts to do that. Designing a single stage (one FIR filter) equalizer is fair game, but it's going to be a relatively high order filter (or just a really awful equalizer). The coefficients in your filter transfer function don't correspond to gains/attenuations at different frequencies, so you can't just, say, adjust the 4 |
| 17:05.31 | penguinece | th coefficient to increase the gain at 100hz by 2 db. |
| 17:09.45 | SpeedEvil | Yes, I know. |
| 17:10.12 | SpeedEvil | I'm assuming that you have a table of 5*128*2 numbers - 5 bands - 128 coefficients - boost and cut |
| 17:10.46 | SpeedEvil | you then go through this table simply taking proportionally from each bands coefficient to generate the 128 output coefficients. |
| 17:11.18 | SpeedEvil | So for a 'neutral' filter - you average each bands 'high' and 'low' coefficient values, and add them all together to get the final coefficients |
| 17:23.15 | chille | Mirell: well, i figured out that i could get uptime with 1/100th seconds precision from /proc/uptime, so i set up a /proc filesystem instead :) |
| 17:29.39 | *** join/#edev CosmicPenguin (~nobody@rrcs-67-52-130-30.west.biz.rr.com) |
| 17:29.39 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 17:53.27 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 18:31.25 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 18:32.09 | *** join/#edev ohsix (ohsix@66.220.111.99) |
| 18:43.26 | SpeedEvil | I really want a nice log of all scheduling events. |
| 18:43.44 | SpeedEvil | So I can take a scope to a system, measure the power trace, and actually see stuff happen, and correlate it. |
| 18:43.57 | SpeedEvil | Scheduling and all wakeup events. |
| 18:47.59 | *** join/#edev miltonm (~miltonm@32.97.110.55) |
| 19:38.56 | *** join/#edev kergoth (~kergoth@ip24-251-170-95.ph.ph.cox.net) |
| 20:06.25 | *** join/#edev miltonm (~miltonm@32.97.110.53) |
| 21:02.16 | wmat | anyone picking up one of these Chumby2 boards? |
| 21:03.16 | wmat | http://www.adafruit.com/index.php?main_page=product_info&cPath=46&products_id=278 |
| 21:03.23 | wmat | pretty cheap |
| 21:03.56 | wmat | ARM926EJ-S |
| 21:11.00 | m4t | wmat , nope, i got a chumby one though |
| 21:14.13 | SpeedEvil | That is an interesting price |
| 21:15.05 | prpplague | for that kind of money you could get a didj or explorer |
| 21:16.13 | SpeedEvil | yeah |
| 21:17.29 | calculus | or a hawkboard |
| 21:18.10 | prpplague | or 18 MSP430 boards |
| 21:19.36 | bsdfox_ | has anyone actually gotten a launchpad yet? |
| 21:20.19 | SpeedEvil | I want the RF watch |
| 21:20.25 | SpeedEvil | but I havre no spare cash |
| 21:21.58 | prpplague | save your money for a panda |
| 21:22.17 | florian | panda? |
| 21:22.29 | prpplague | florian: omap4 based dev board |
| 21:22.39 | florian | prpplague: uh cool |
| 21:22.56 | prpplague | similar in nature as the beagle |
| 21:23.27 | calculus | http://www.arm.linux.org.uk/developer/machines/list.php?id=2791 is the most I could find on it |
| 21:23.31 | SpeedEvil | Nokia are spamming me at the moment with free n900s. |
| 21:23.38 | SpeedEvil | I got one for dev work. |
| 21:23.45 | SpeedEvil | And they've just randomly sent me another. |
| 21:23.52 | florian | heh |
| 21:24.01 | SpeedEvil | I suspect I'll have to return it. |
| 21:24.20 | prpplague | invites florian , SpeedEvil and calculus to join #pandaboard |
| 21:25.05 | florian | :-) |
| 21:26.02 | SpeedEvil | Unfortunately I have _no_ spare cash. As in I spent 50 quid on groceries last month. |
| 21:26.26 | prpplague | SpeedEvil: 50 quid? for just yourself? |
| 21:26.36 | bsdfox_ | do they have sperm banks in scotland? |
| 21:27.04 | CosmicPenguin | prpplague: everybody in your family avoid earl? |
| 21:27.11 | prpplague | CosmicPenguin: yea |
| 21:27.36 | prpplague | CosmicPenguin: thanks for asking |
| 21:28.00 | CosmicPenguin | I think about you every time a hurricane heads toward the windward islands |
| 21:28.04 | SpeedEvil | yes. Just me. |
| 21:28.28 | prpplague | SpeedEvil: i hope most of that is beer |
| 21:28.36 | SpeedEvil | Unfortunately not. |
| 21:28.39 | prpplague | CosmicPenguin: yea, barbados is pretty well protected with the trade winds |
| 21:28.49 | SpeedEvil | Probably the largest line item was milk, followed by banannas. |
| 21:28.58 | SpeedEvil | UHT milk ++ |
| 21:29.08 | prpplague | CosmicPenguin: the hurricane needs to form south of the equator around the cost of africa before it heads across to barbados |
| 21:29.48 | prpplague | CosmicPenguin: http://www.nhc.noaa.gov/gtwo_atl.shtml |
| 21:29.58 | prpplague | CosmicPenguin: see the storm at the bottom right next to africa? |
| 21:30.04 | prpplague | CosmicPenguin: that one is of concern |
| 21:30.19 | SpeedEvil | How often does barbados get hurricaned? |
| 21:30.38 | CosmicPenguin | Fiona isn't a concern? |
| 21:30.45 | CosmicPenguin | I guess it will slip northwest |
| 21:30.53 | prpplague | CosmicPenguin: no, its too far north already |
| 21:31.01 | prpplague | SpeedEvil: about once in 10 years |
| 21:31.10 | prpplague | SpeedEvil: and a really bad one about once every 25 years |
| 21:31.12 | SpeedEvil | Ah. |
| 21:31.23 | CosmicPenguin | Lots of tropical waves recently |
| 21:31.31 | SpeedEvil | Scotland is even more infrequent yet. |
| 21:31.45 | SpeedEvil | About the worst possible here is very deep snow. |
| 21:31.57 | CosmicPenguin | Just wait until we turn off the gulf stream |
| 21:32.01 | SpeedEvil | Or worse, internet falling over. |
| 21:32.02 | SpeedEvil | Indeed. |
| 21:32.05 | prpplague | hehe |
| 21:32.15 | SpeedEvil | I'm north of most of the habited parts of canada |
| 21:32.20 | prpplague | SpeedEvil: most of wife's ancestors were scottish and irish |
| 22:06.22 | *** join/#edev Rubix (debian-tor@gateway/tor-sasl/rubix) |
| 23:05.16 | *** join/#edev JoeLlama (~snork@unaffiliated/joellama) |
| 23:10.25 | *** join/#edev miltonm (~miltonm@32.97.110.51) |
| 23:47.10 | *** join/#edev SpeedEvil (~user@tor/regular/SpeedEvil) |