--- Log opened Fri Apr 22 00:00:54 2016 | ||
-!- amiller_ [~socrates1@li175-104.members.linode.com] has quit [Ping timeout: 276 seconds] | 00:04 | |
-!- juri_ [~juri@vpn166.sdf.org] has joined ##hplusroadmap | 00:08 | |
-!- amiller [~socrates1@li175-104.members.linode.com] has joined ##hplusroadmap | 00:12 | |
-!- amiller is now known as Guest84297 | 00:13 | |
-!- helleshin [~talinck@66-161-138-110.ubr1.dyn.lebanon-oh.fuse.net] has quit [Read error: Connection reset by peer] | 00:15 | |
-!- JayDugger1 [~jwdugger@108.19.186.58] has joined ##hplusroadmap | 00:16 | |
-!- JayDugger [~jwdugger@108.19.186.58] has quit [Ping timeout: 276 seconds] | 00:19 | |
-!- Guest84297 [~socrates1@li175-104.members.linode.com] has quit [Changing host] | 00:35 | |
-!- Guest84297 [~socrates1@unaffiliated/socrates1024] has joined ##hplusroadmap | 00:35 | |
-!- Guest84297 is now known as amiller | 00:35 | |
-!- Gurkenglas_ [Gurkenglas@dslb-178-000-178-156.178.000.pools.vodafone-ip.de] has quit [Ping timeout: 240 seconds] | 00:39 | |
-!- JayDugger1 [~jwdugger@108.19.186.58] has quit [Quit: Leaving.] | 00:40 | |
-!- JayDugger [~jwdugger@108.19.186.58] has joined ##hplusroadmap | 00:40 | |
nmz787_i | xentrac_: are you in South America? | 00:44 |
---|---|---|
-!- JayDugger [~jwdugger@108.19.186.58] has quit [Ping timeout: 260 seconds] | 00:52 | |
-!- jcluck [~cluckj@pool-108-16-231-242.phlapa.fios.verizon.net] has joined ##hplusroadmap | 00:58 | |
-!- cluckj [~cluckj@pool-108-16-231-242.phlapa.fios.verizon.net] has quit [Ping timeout: 252 seconds] | 01:01 | |
-!- JayDugger [~jwdugger@108.19.186.58] has joined ##hplusroadmap | 01:05 | |
-!- nmz787_i [~ntmccork@134.134.139.82] has quit [Quit: Leaving.] | 01:14 | |
-!- Urchin [~urchin@unaffiliated/urchin] has joined ##hplusroadmap | 02:38 | |
-!- dbolser [~dbolser@193.62.194.251] has quit [Changing host] | 03:04 | |
-!- dbolser [~dbolser@unaffiliated/faceface] has joined ##hplusroadmap | 03:04 | |
-!- amiller [~socrates1@unaffiliated/socrates1024] has quit [Ping timeout: 240 seconds] | 03:07 | |
-!- amiller_ [~socrates1@li175-104.members.linode.com] has joined ##hplusroadmap | 03:25 | |
-!- amiller_ [~socrates1@li175-104.members.linode.com] has quit [Quit: ZNC - http://znc.sourceforge.net] | 03:29 | |
-!- jtimon [~quassel@18.29.134.37.dynamic.jazztel.es] has joined ##hplusroadmap | 03:31 | |
-!- amiller_ [~socrates1@li175-104.members.linode.com] has joined ##hplusroadmap | 03:33 | |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap | 04:01 | |
-!- helleshin [~talinck@66-161-138-110.ubr1.dyn.lebanon-oh.fuse.net] has joined ##hplusroadmap | 04:20 | |
-!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-bnfdxlhvovbubgau] has joined ##hplusroadmap | 04:31 | |
streety | drethelin: Their product seems to be at http://www.elysiumhealth.com/basis My initial impression is that the marketing is heavy and light on facts. For example, knowing the quantities of the active ingredients should be easier than it is. Based on that I was expecting their prices to be some multiple of competitors for the same compounds but that doesn't appear to be the case. | 05:03 |
streety | for example http://www.lifeextension.com/Vitamins-Supplements/item02031/Optimized-Resveratrol-with-Nicotinamide-Riboside#panelSupplements would work out to be roughly the same price | 05:03 |
-!- Gurkenglas_ [Gurkenglas@dslb-178-000-178-156.178.000.pools.vodafone-ip.de] has joined ##hplusroadmap | 05:26 | |
-!- justanotheruser [~Justan@unaffiliated/justanotheruser] has quit [Ping timeout: 252 seconds] | 05:26 | |
-!- c0rw1n [~c0rw1n@139.117-67-87.adsl-dyn.isp.belgacom.be] has quit [] | 05:43 | |
-!- andares [~andares@unaffiliated/jacco] has quit [Ping timeout: 276 seconds] | 06:16 | |
-!- c0rw1n [~c0rw1n@139.117-67-87.adsl-dyn.isp.belgacom.be] has joined ##hplusroadmap | 06:19 | |
-!- CheckDavid [uid14990@gateway/web/irccloud.com/x-etpvtrsfomkzgxaf] has joined ##hplusroadmap | 06:21 | |
-!- DroneFox [~DroneFox@95.128.68.231] has quit [] | 06:25 | |
-!- wbraun [~wbraun@dhcp-18-189-39-222.dyn.mit.edu] has quit [Quit: wbraun] | 06:32 | |
catern | oho | 06:41 |
catern | JayDugger: with the belt clip, the camera of the phone is facing outward from my waist | 06:41 |
catern | i could use that to record, too! | 06:41 |
-!- wbraun [~wbraun@31-33-141.wireless.csail.mit.edu] has joined ##hplusroadmap | 06:44 | |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Leaving] | 06:46 | |
-!- wbraun [~wbraun@31-33-141.wireless.csail.mit.edu] has quit [Quit: wbraun] | 06:55 | |
kanzure | recent sleeptyping results show a character called "spider monk" because monks at a monastery over a few hundred years selected for smart spiders in a spider breeding program | 07:02 |
catern | genius | 07:09 |
-!- Jp_ [5f8095ec@gateway/web/freenode/ip.95.128.149.236] has joined ##hplusroadmap | 07:17 | |
-!- Jp__ [5f8095ec@gateway/web/freenode/ip.95.128.149.236] has joined ##hplusroadmap | 07:18 | |
-!- augur [~augur@c-73-46-94-9.hsd1.fl.comcast.net] has quit [Remote host closed the connection] | 07:20 | |
-!- Jp_ [5f8095ec@gateway/web/freenode/ip.95.128.149.236] has quit [Ping timeout: 250 seconds] | 07:22 | |
-!- wbraun [~wbraun@dhcp-18-111-124-141.dyn.MIT.EDU] has joined ##hplusroadmap | 07:26 | |
-!- augur [~augur@c-73-46-94-9.hsd1.fl.comcast.net] has joined ##hplusroadmap | 07:29 | |
-!- jtimon [~quassel@18.29.134.37.dynamic.jazztel.es] has quit [Remote host closed the connection] | 07:30 | |
-!- jaboja [~jaboja@vps.jaboja.pl] has joined ##hplusroadmap | 07:30 | |
-!- Jp__ [5f8095ec@gateway/web/freenode/ip.95.128.149.236] has quit [Ping timeout: 250 seconds] | 07:31 | |
JayDugger | Mornng, catern. I am happy to hear of your success. Remember to find a good way to catalog all the recordings you accumulate. | 07:40 |
JayDugger | Spider monk, spider monk. Does what ever a monk can do. Proofs a book, any size. Catches typos, just like flies. Look out! Here comes the spider monk. | 07:42 |
kanzure | you forgot about his vow of silence | 07:44 |
JayDugger | Heh. | 07:44 |
JayDugger | That would be in the second or third verse, right after the bit about Gregorian chant. | 07:45 |
-!- jaboja [~jaboja@vps.jaboja.pl] has quit [Ping timeout: 246 seconds] | 07:46 | |
JayDugger | And illuminated inks for hemolymph. | 07:58 |
-!- jaboja [~jaboja@vps.jaboja.pl] has joined ##hplusroadmap | 08:04 | |
-!- CheckDavid [uid14990@gateway/web/irccloud.com/x-etpvtrsfomkzgxaf] has quit [Quit: Connection closed for inactivity] | 08:22 | |
-!- nildicit [~nildicit@gateway/vpn/privateinternetaccess/nildicit] has joined ##hplusroadmap | 08:23 | |
maaku | kanzure: zappos (which amazon bought) found their optimal inventory management was custom built wheeled robots... | 08:25 |
-!- CheckDavid [uid14990@gateway/web/irccloud.com/x-xebphyqsnxctasti] has joined ##hplusroadmap | 08:39 | |
kanzure | wait, leary was signed up for alcor and cryocare? | 08:42 |
JayDugger | No, he ended up cremated and shot into space in 1997 with Rodenberry and O'Neill by Celestia. | 08:57 |
JayDugger | I vaguely remember some dispute about cryonics versus cremation, but the details aren't in memory. | 08:58 |
-!- nmz787_i [~ntmccork@134.134.139.82] has joined ##hplusroadmap | 09:14 | |
kanzure | looks like he was signed up at one point. weird. | 09:46 |
-!- nmz787_i [~ntmccork@134.134.139.82] has quit [Quit: Leaving.] | 09:48 | |
catern | how fucking useless | 09:50 |
kanzure | "no, no. it's the thought that counts." | 09:51 |
catern | i asked these people http://kaptureaudio.com/ if their thing could handle recording 24/7 | 09:51 |
catern | and they said "well that's not what we're trying to do!!" | 09:51 |
-!- jcluck is now known as cluckj | 09:51 | |
-!- Orpheon [~Orpheon@42.90.62.81.dynamic.wline.res.cust.swisscom.ch] has joined ##hplusroadmap | 09:52 | |
catern | lol there's this I guess http://www.pimall.com/nais/recordwatch.html | 09:53 |
catern | this stuff | 09:53 |
catern | probability that it's garbage: high | 09:54 |
-!- wbraun [~wbraun@dhcp-18-111-124-141.dyn.MIT.EDU] has quit [Quit: wbraun] | 09:57 | |
maaku | catern: I'll send them an email as well | 10:05 |
maaku | maybe if they get more than one inquiry about this they might be more helpful... | 10:05 |
catern | actually, it looks like there are a lot of products dedicated to this purpose | 10:05 |
maaku | there's a fucking huge opportunity here | 10:05 |
catern | http://www.amazon.com/s/ref=nb_sb_noss_1?url=search-alias%3Delectronics&field-keywords=watch+voice+recorder | 10:05 |
kanzure | hm a wristband.. | 10:06 |
maaku | incidentally someone mentioned the cocktail party problem in the scrollback. solution is to put one on each wrist | 10:06 |
maaku | ideally i'd have a smartwatch on one wrist, and a high quality mic (kapture?) on the other | 10:07 |
kanzure | er, my expectation was that the solution would be putting microphones in each ceiling tile | 10:07 |
maaku | kanzure: for a conference, yes, but I'm talking about a mobile record-everything-everywhere-i-go solution | 10:08 |
catern | hmm | 10:09 |
catern | i don't know anything about audio capture | 10:09 |
maaku | i want something looking like a IRC chat log - date, time, nick (voice identification), text (transcribed), and a little play button to hear the original audio | 10:09 |
catern | is it possible to reconstruct a high-quality audio stream from the the two recording wristbands | 10:09 |
catern | is there some off-the-shelf software I can use do that | 10:11 |
catern | i literally don't know | 10:11 |
maaku | catern: sure, that's essentially what high quality microphones do internally. they have multiple mics for noise cancellation and dopler isolation | 10:11 |
maaku | there's plenty of software. | 10:11 |
catern | so you assert, but do you know of a specific example? | 10:12 |
maaku | but what'd be unique and require something custom is that the mics (your wrists) would be in presumed constant motion, both absolutely and relative to each other | 10:12 |
-!- sivoais [~zaki@unaffiliated/sivoais] has quit [Ping timeout: 244 seconds] | 10:13 | |
maaku | catern: because it's literally an entry-level DSP problem, e.g. something that's assigned as homework for a signal processing 101 class | 10:14 |
kanzure | nah just throw in some accelerometers and uh.. something.. | 10:14 |
catern | maaku: oh, ok | 10:14 |
-!- Aurelius_Home [~cpopell@c-73-200-185-48.hsd1.va.comcast.net] has quit [Ping timeout: 276 seconds] | 10:14 | |
maaku | catern: as to the question "which software should I use", idk | 10:14 |
catern | yeah, i get it now | 10:15 |
maaku | kanzure: good point the accelerometer (which is already in the smartphone anyway) should make the problem easier | 10:15 |
-!- sivoais [~zaki@unaffiliated/sivoais] has joined ##hplusroadmap | 10:15 | |
catern | but you don't have an accelerometer in these voice recording watches I assume | 10:16 |
-!- c0rw1n [~c0rw1n@139.117-67-87.adsl-dyn.isp.belgacom.be] has quit [Read error: Connection reset by peer] | 10:17 | |
-!- Burn_ [~Burn@pool-108-31-204-167.washdc.fios.verizon.net] has quit [Ping timeout: 260 seconds] | 10:17 | |
-!- c0rw1n [~c0rw1n@139.117-67-87.adsl-dyn.isp.belgacom.be] has joined ##hplusroadmap | 10:17 | |
maaku | sure but when one is available it is one more piece of data to constrain the problem | 10:18 |
maaku | i don't think it's strictly necessary, but it should help reduce noise | 10:18 |
-!- Burninate [~Burn@pool-108-31-204-167.washdc.fios.verizon.net] has joined ##hplusroadmap | 10:18 | |
maaku | doing this properly is actually a really interesting machine learning problem | 10:19 |
maaku | i can't emphasize enough how much I want this though .. like seriously i'm going to drop all my other tinkering until I have this | 10:19 |
catern | ^5 | 10:20 |
-!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-bnfdxlhvovbubgau] has quit [Quit: Connection closed for inactivity] | 10:20 | |
-!- Aurelius_Home [~cpopell@c-73-200-185-48.hsd1.va.comcast.net] has joined ##hplusroadmap | 10:21 | |
catern | that is similar to how i feel. but before I couldn't think of decent hardware for this | 10:21 |
catern | but the belt-clip-phone would work great | 10:22 |
maaku | I have a personal problem that while I'm highly organized at work with a system to keep track of commitments, it's hard to keep that up at home.. I'd love to be able to quickly review the conversation with my wife to remember what I promised to do, or pull up the chat I had with the teacher or school counselor | 10:22 |
-!- nmz787_i [~ntmccork@134.134.139.82] has joined ##hplusroadmap | 10:23 | |
catern | indeed | 10:24 |
catern | hmm | 10:31 |
catern | if someone has a smartwatch on their wrist (good quality), and a smartphone in their pocket (poor quality), i wonder if you could use both of them with enough DSP magic and get a clear enough signal out to feed into machine learning | 10:33 |
catern | so you could maybe do this without any special hardware/behavior changes | 10:34 |
catern | oh! and phones have multiple microphones, too | 10:35 |
-!- nmz787_i1 [~ntmccork@134.134.137.75] has joined ##hplusroadmap | 10:40 | |
-!- nmz787_i [~ntmccork@134.134.139.82] has quit [Remote host closed the connection] | 10:43 | |
maaku | catern: sent | 10:43 |
maaku | hopefully I can get a better response from them | 10:44 |
maaku | catern: that's the idea .. the larger the separation in space, the better | 10:45 |
maaku | sound moves about 7mm/sample at 44.1kHz | 10:46 |
maaku | so a wristband and pocket phone should be 10's of cm apart, which would be enough to do doppler isolation which not only separates voice samples, but also eliminates background noise | 10:47 |
maaku | that'd probably do more to improve quality than any kind of reconstruction combining the sources | 10:47 |
kanzure | 35106 conversations, 182424 tags, 33600 unique tags | 10:49 |
maaku | so I was imagining sampling both sound feeds on the phone, doing doppler analysis to separate sources and drop line noise, then uploading the separate, compressed feeds for further processing | 10:49 |
maaku | the biggest optimization is just detecting when nothing interesting is going on, because otherwise you have 500MB/day of audio ;) | 10:50 |
catern | 500MB a day = 182.5GB a year | 10:52 |
catern | practically nothing | 10:52 |
maaku | catern: not nothing when you're pushing it through you 4G LTE metered plan ;) | 10:53 |
-!- Filosofem [~Jawmare@unaffiliated/jawmare] has quit [Ping timeout: 246 seconds] | 10:53 | |
maaku | although if you can wait for wifi connectivity i suppose it's a non-issue | 10:53 |
catern | sure, that's why you only push it out on wifi | 10:53 |
cluckj | check on the legality of it too before you start recording everyone :) | 10:54 |
catern | everyone has the right to perfect memory, cluckj | 10:54 |
cluckj | lol | 10:54 |
catern | are you a transhumanist or not huh!?! :) | 10:54 |
cluckj | nope | 10:54 |
cluckj | well...practically yes, philosophically no | 10:55 |
cluckj | http://www.rcfp.org/reporters-recording-guide/state-state-guide | 10:57 |
drethelin | that's a big problem with vidoe surveillance too | 10:59 |
drethelin | You want sensitivity high enough to get small but important things recorded | 10:59 |
drethelin | but not so high you have to sort through gigs of video to find anything | 10:59 |
maaku | http://www.dmlp.org/legal-guide/recording-phone-calls-and-conversations | 11:02 |
-!- nildicit_ [~nildicit@gateway/vpn/privateinternetaccess/nildicit] has joined ##hplusroadmap | 11:02 | |
-!- ArturSha1 [~ArturShai@212.97.31.129] has quit [Ping timeout: 240 seconds] | 11:02 | |
-!- nmz787_i1 [~ntmccork@134.134.137.75] has quit [Quit: Leaving.] | 11:03 | |
-!- wbraun [~wbraun@dhcp-18-111-52-45.dyn.MIT.EDU] has joined ##hplusroadmap | 11:04 | |
kanzure | longest tags: | 11:05 |
kanzure | "dna synthesis as shortcut for molecular manufacturing without figuring out tooltips" | 11:05 |
-!- nmz787_i [~ntmccork@134.134.139.82] has joined ##hplusroadmap | 11:05 | |
kanzure | "Singularity Institute for Making 'Optimization' a Less Ominous Prospect" | 11:05 |
kanzure | Singularity Institute for Ominously-Characterized 'Optimization' | 11:05 |
kanzure | "slower-decaying radioisotope tagged neurotransmitter precursors" | 11:05 |
kanzure | "zen-of-byzantine-agreement-and-problems-of-distributed-systems" | 11:06 |
kanzure | "trip out from the lsd released each time i crack my knuckles" | 11:06 |
-!- nildicit [~nildicit@gateway/vpn/privateinternetaccess/nildicit] has quit [Ping timeout: 260 seconds] | 11:06 | |
kanzure | "likely consequence of the general quality thereof" | 11:06 |
-!- wbraun [~wbraun@dhcp-18-111-52-45.dyn.MIT.EDU] has quit [Client Quit] | 11:08 | |
kanzure | mentioned 196 cities | 11:12 |
-!- Gurkenglas_ [Gurkenglas@dslb-178-000-178-156.178.000.pools.vodafone-ip.de] has quit [Ping timeout: 244 seconds] | 11:14 | |
kanzure | mentioned 2518 people.... yikes. | 11:14 |
nmz787_i | who what where? | 11:17 |
kanzure | just looking over some analytics data from my "write down all conversations ever" file | 11:19 |
drethelin | you should start giving the conversations points | 11:20 |
drethelin | in a scoring system | 11:20 |
drethelin | so you can relive the best ones | 11:20 |
kanzure | well one of the intended purposes was to do tag recommendation, e.g. which things to speak about with which people, such as in the event of neglecting to mention to maaku something $dangerously_important | 11:22 |
-!- amiller_ [~socrates1@li175-104.members.linode.com] has quit [Changing host] | 11:31 | |
-!- amiller_ [~socrates1@unaffiliated/socrates1024] has joined ##hplusroadmap | 11:31 | |
-!- amiller_ is now known as amiller | 11:31 | |
-!- CheckDavid [uid14990@gateway/web/irccloud.com/x-xebphyqsnxctasti] has quit [Quit: Connection closed for inactivity] | 11:32 | |
-!- nildicit__ [~nildicit@gateway/vpn/privateinternetaccess/nildicit] has joined ##hplusroadmap | 11:36 | |
-!- nildicit_ [~nildicit@gateway/vpn/privateinternetaccess/nildicit] has quit [Ping timeout: 252 seconds] | 11:39 | |
maaku | I'd be curious to hear what pasky thinks he can extract from analysis of recorded conversations | 11:51 |
-!- jaboja [~jaboja@vps.jaboja.pl] has quit [Ping timeout: 246 seconds] | 11:53 | |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap | 11:57 | |
-!- wbraun [~wbraun@dhcp-18-111-52-45.dyn.mit.edu] has joined ##hplusroadmap | 12:02 | |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Leaving] | 12:10 | |
pasky | text or audio? | 12:11 |
pasky | hmm i'll talk later | 12:11 |
-!- wbraun [~wbraun@dhcp-18-111-52-45.dyn.mit.edu] has quit [Quit: wbraun] | 12:13 | |
maaku | pasky: both | 12:19 |
maaku | base input would be audio, but with supervised transcription would be assumed | 12:19 |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap | 12:30 | |
-!- jaboja [~jaboja@vps.jaboja.pl] has joined ##hplusroadmap | 12:49 | |
-!- Guest38821 [~lake@23-118-33-136.lightspeed.moblal.sbcglobal.net] has joined ##hplusroadmap | 13:16 | |
-!- Guest38821 [~lake@23-118-33-136.lightspeed.moblal.sbcglobal.net] has quit [Client Quit] | 13:17 | |
nmz787_i | pfft, every day is earth day, except when you're extra-terrestrial | 13:26 |
chris_99 | heh | 13:27 |
chris_99 | nmz787_i, am i right in thinking as the order increases in the diffraction grating formulae, the amplitude gets dimmer? | 13:29 |
-!- jaboja [~jaboja@vps.jaboja.pl] has quit [Ping timeout: 260 seconds] | 13:29 | |
-!- sandeepkr_ [~sandeep@111.235.64.4] has quit [Ping timeout: 240 seconds] | 13:30 | |
-!- Orpheon [~Orpheon@42.90.62.81.dynamic.wline.res.cust.swisscom.ch] has quit [Read error: Connection reset by peer] | 13:45 | |
nmz787_i | yeah I think so... since they're like harmnoics or something | 13:50 |
nmz787_i | i know some spectrometers collect multiple orders and add/average them to increase SnR | 13:50 |
chris_99 | ta, see i was confused, how you'd end up with both 100nm @ 9th order, and 900nm @ 1st order, in the same position | 13:53 |
chris_99 | how you'd distinguish between them | 13:53 |
chris_99 | but i guess the 900nm would just have the highest amplitude | 13:54 |
-!- jaboja [~jaboja@vps.jaboja.pl] has joined ##hplusroadmap | 13:58 | |
nmz787_i | well I think if they overlapped then you're screwed | 14:05 |
chris_99 | but say 100nm @ 9th order, and 900nm @ 1st order: arcsin((1 * 900e-9) / 2e-6)) and arcsin((9 * 100e-9) / 2e-6)) both give the same radians, so theoretically i guess they could end up in the same position | 14:07 |
maaku | catern: kapture response was useless. no plans for API, no desire to support our use case | 14:09 |
maaku | here's some examples of their audio quality at least: | 14:09 |
maaku | http://kapturelive.com/klip/553cf203e13b4c3b03dae6e3 (inside a vehicle) | 14:09 |
maaku | http://kapturelive.com/klip/559c5884e13b4c0c446fef02 (pretty cute) | 14:09 |
maaku | http://kapturelive.com/klip/55cbf255e13b4c026227f25d (funny) | 14:09 |
nmz787_i | chris_99: yeah they definitely can end up like that, I'm just saying you would have to have a good reason to /want/ them to overlap... separating the signal strength of each individual frequency would probably involve some assumptions about your signal source or something | 14:11 |
maaku | catern: he did point me at this though : http://sasquatchbioacoustic.blogspot.com/2016/01/long-duration-recorder-ldr-based-on.html | 14:11 |
xentrac_ | nmz787_i: yes, I'm in Buenos Aires! | 14:11 |
chris_99 | nmz787_i, oh i don't want them to overlap, i'm just confused what happens when they do | 14:11 |
nmz787_i | it would just be like shining a red and a purple laser on the same spot | 14:12 |
nmz787_i | colors would mix apparently to our eyes | 14:12 |
chris_99 | mmm, theres not much you can do then i guess if they do | 14:13 |
chris_99 | so you could get incorrect results i guess | 14:13 |
nmz787_i | reseparate would be possible | 14:21 |
nmz787_i | xentrac_: ever been to Columbia? | 14:22 |
nmz787_i | Colombia* | 14:22 |
xentrac_ | no, when my then-wife and I were trying to figure out where to live, we figured Colombia was too scary for our families to want to visit | 14:30 |
chris_99 | http://www.symmetrymagazine.org/article/lhc-data-at-your-fingertips that's really cool | 14:47 |
-!- Jawmare [~Jawmare@unaffiliated/jawmare] has joined ##hplusroadmap | 14:47 | |
catern | maaku: disappointing. but that is cool. (and amusing that the useful information is coming from a blog about sasquatch bioacoustics) | 14:50 |
maaku | heh | 14:50 |
maaku | catern: I think I'm getting ahead of myself. step 1 is just to setup a background 'record everything' app | 14:51 |
maaku | standard in-ear bluetooth can be used instead of a fancy watch or whatnot | 14:51 |
catern | oho, interesting | 14:51 |
catern | of course in-ear bluetooth should work | 14:52 |
-!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Leaving] | 14:52 | |
maaku | it's meant to pick up the person wearing it though, not the environment | 14:53 |
maaku | so we'll see... | 14:53 |
pasky | maaku: never worked with raw audio up to now, not sure what's hard and what's easy | 15:04 |
maaku | pasky: i guess i was asking at a higher level .. what kind of interesting augmented-human projects you could do with such info | 15:05 |
maaku | e.g. identifying people by voice, automatically picking out commitments (making task entries) and appointments (making calendar entries) | 15:06 |
-!- wbraun [~wbraun@dhcp-18-111-52-45.dyn.mit.edu] has joined ##hplusroadmap | 15:07 | |
pasky | might change over the summer, i'm looking for a student intern that'd try to train a text-to-speech deep network on harry potter ebooks + audiobooks to learn stephen fry's voice, but with a chance to also change the tone of the voice based on what's going on, dialogue etc., i have some good reasons that there is some space for a breakthrough... | 15:08 |
pasky | maaku: hmm, well, i'm better at coming up ideas to augment computers than humans :/ | 15:08 |
pasky | because my own life is really boring | 15:08 |
-!- wbraun [~wbraun@dhcp-18-111-52-45.dyn.mit.edu] has quit [Quit: wbraun] | 15:45 | |
xentrac_ | nmz787_i: what's your interest in Colombia? | 15:48 |
-!- Gurkenglas_ [Gurkenglas@dslb-178-000-178-156.178.000.pools.vodafone-ip.de] has joined ##hplusroadmap | 15:48 | |
nmz787_i | xentrac_: heading there next month for a wedding | 15:49 |
nmz787_i | got a cheap cell phone and cheap laptop in case of robberies... but actually thinking I probably don't need a laptop for 5 days | 15:50 |
nmz787_i | was thinking I might leave the laptop with some child-in-need | 15:50 |
nmz787_i | some me-as-a-kid type kid... but I have no idea how to find someone like that | 15:50 |
-!- jaboja [~jaboja@vps.jaboja.pl] has quit [Remote host closed the connection] | 15:54 | |
-!- HEx1 [~HEx@212.159.112.196] has quit [Remote host closed the connection] | 15:56 | |
xentrac_ | I found them in Venezuela at a free software conference | 15:57 |
-!- HEx1 [~HEx@hexwab.plus.com] has joined ##hplusroadmap | 15:58 | |
nmz787_i | I feel like Google Translate with the offline spanish pack might really come in handy | 15:58 |
nmz787_i | or make me look like a douche | 15:58 |
-!- HEx1 [~HEx@hexwab.plus.com] has quit [Remote host closed the connection] | 15:59 | |
archels_ | there are offline packs for Google Translate?? | 16:02 |
nmz787_i | for a few larger langs | 16:03 |
nmz787_i | I think spanish, maybe arabic | 16:04 |
archels_ | any idea how complete the coverage is of a pack? | 16:05 |
nmz787_i | nope | 16:14 |
nmz787_i | I can only hope all the words I might need to have a effective, if not nice, conversation! | 16:15 |
-!- HEx1 [~HEx@hexwab.plus.com] has joined ##hplusroadmap | 16:23 | |
-!- justanotheruser [~Justan@unaffiliated/justanotheruser] has joined ##hplusroadmap | 16:24 | |
-!- nmz787_i [~ntmccork@134.134.139.82] has quit [Ping timeout: 276 seconds] | 16:29 | |
-!- HEx1 [~HEx@hexwab.plus.com] has quit [Remote host closed the connection] | 16:31 | |
-!- HEx1 [~HEx@hexwab.plus.com] has joined ##hplusroadmap | 16:33 | |
fenn | i wonder if you could get around wiretapping laws by doing the speech-to-text on the phone and throwing away the audio data | 16:57 |
-!- Gurkenglas_ [Gurkenglas@dslb-178-000-178-156.178.000.pools.vodafone-ip.de] has quit [Ping timeout: 268 seconds] | 17:11 | |
fenn | hmm it may be illegal in california to save your emails without informing all parties to the email | 17:12 |
fenn | so i'd want to have a fixed array of microphones with raw uncompressed recording of each one, then dump all that data each night for offline processing, where you do source separation and noise reduction, then compress each source individually along with metadata about its location and amplitude | 17:14 |
fenn | this can then be reconstructed as a 3d audio scene which you can fly around in or simply follow the original path of the microphone array | 17:15 |
fenn | i think having a non-fixed baseline like one microphone on each wrist is just making things difficult for yourself | 17:16 |
fenn | on the other hand it does give you a little bit more parallax to triangulate relatively constant noise sources, because you are moving around more and thus providing more information about the noise | 17:17 |
maaku | fenn: that's where I'm headed, except I don't expect the array to be fixed (e.g. wristbands, phone in pocket that sometimes gets taken out; relative positions will change) | 17:50 |
maaku | but it seems like recording the audio to the phone only doing detection of dead sound (no voices), and bulk uploading when you get wifi is best | 17:51 |
maaku | fenn: the goal (for me) is to capture conversations throughout my life, not just in a conference setting | 17:51 |
maaku | so fixed mic arrays isn't really practical | 17:51 |
maaku | I emailed kaptureaudio again to see if I can get a datasheet/low level API just for my own personal use | 17:54 |
-!- yashgaroth [~yashgarot@2602:306:35fa:d500:f5e0:f867:a11d:8d52] has joined ##hplusroadmap | 17:54 | |
-!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-ycqtuhoddxnhmhoo] has joined ##hplusroadmap | 18:05 | |
-!- nmz787_i [~ntmccork@134.134.139.82] has joined ##hplusroadmap | 18:39 | |
nmz787_i | maaku: fenn: fixed arrays that can move around (phone, other widgets with >1 mic soldered down) seem like a good thing | 19:03 |
nmz787_i | .title http://onesolver.com/ | 19:04 |
yoleaux | OneSolver | 19:04 |
nmz787_i | "OneSolver is a online platform for solving STEM problems (science, technology, engineering, mathematics). The framework can quickly create rich, interactive calculators, with very little underlying code. Solutions are optimized for search engines, but feel free to browse them manually through the Design menu. Our features page contains the latest details about additions to the platform. Please report software bugs via issue tickets. If you want | 19:04 |
nmz787_i | to be notified about major releases or are interested in becoming and contributor join our members list." | 19:04 |
nmz787_i | fenn: I got a zeolite dehumidifier... any chance you think I could tweak it to get water in the desert with it? | 19:06 |
nmz787_i | (or just in summer here, I guess) | 19:06 |
nmz787_i | even with electricity to power it for starters | 19:06 |
-!- Filosofem [~Jawmare@unaffiliated/jawmare] has joined ##hplusroadmap | 19:12 | |
-!- Jawmare [~Jawmare@unaffiliated/jawmare] has quit [Ping timeout: 276 seconds] | 19:13 | |
fenn | by "fixed array" i just mean multiple microphones on a single device | 19:22 |
fenn | like a phone with 16 microphones or whatever | 19:23 |
maaku | fenn: I think there is significant benefit to be had from having large separations between the mics to do doppler separation | 19:23 |
fenn | i don't know what doppler separation is; but yes a larger baseline is better | 19:24 |
maaku | sound moves about 7mm per sample at sea level for 44.1kHz | 19:24 |
maaku | fenn: cocktail party problem | 19:24 |
maaku | https://en.wikipedia.org/wiki/Cocktail_party_effect | 19:25 |
fenn | is doppler separation similar to how those gunshot detection algorithms work? simple trilateration? | 19:25 |
maaku | I've heard it called doppler separation, althoug I'm not really sure why | 19:25 |
fenn | the wikipedia article is about attention and biological stuff | 19:26 |
maaku | oh sorry https://en.wikipedia.org/wiki/Source_separation | 19:26 |
fenn | so are you saying doppler separation = source separation? | 19:27 |
maaku | yes i've heard both terms | 19:27 |
maaku | there's many ways to do it. e.g. you can use the fact that sound arrives at different times at different positions, or that sound drops off at 1/r^2, or echo delays indoors | 19:28 |
fenn | i'm seeing a paper about monitoring train bearings by "doppler separation" but this is a special case where the different train cars have different relative velocities wrt the microphone | 19:29 |
fenn | so each bearing will have an individual frequency shift | 19:29 |
fenn | i keep trying to learn how kalman filters work but it never seems to stick | 19:31 |
-!- Aurelius_Laptop [~cpopell@c-73-129-20-70.hsd1.md.comcast.net] has joined ##hplusroadmap | 19:50 | |
kanzure | https://archive.stsci.edu/mug/mug_2016/PS1_MUG_2016jan14.pdf | 19:54 |
kanzure | "PanSTARRS' imaging survey will make 2 PB available online" | 19:55 |
kanzure | "You [can] buy a 180TB 4U backblaze storage pod assembled for about $10,500. For $21,000 you can buy two and have 60TB to spare. $8,500/ $17,000 if you want to DIY." | 19:56 |
kanzure | ( http://www.backuppods.com/ ) | 19:56 |
xentrac_ | fenn: I have the impression that kalman filters are actually a straightforward application of Bayes' theorem | 20:45 |
xentrac_ | under the assumption that the probability distribution of the thing you're trying to determine is a multidimensional Gaussian | 20:46 |
-!- ArturSha1 [~ArturShai@212.97.31.129] has joined ##hplusroadmap | 20:48 | |
xentrac_ | apparently if both your prior and the conditional probability you're using to update it are Gaussians then the posterior probability is also a multidimensional Gaussian? | 20:48 |
xentrac_ | or maybe it isn't and you just pretend that it is, I don't remember | 20:48 |
xentrac_ | but I think that's enough of the idea to derive Kalman filtering from | 20:49 |
xentrac_ | I think that Kalman filtering may not be a great fit for audio source localization though | 20:50 |
xentrac_ | but maybe that was a potentially unrelated comment | 20:51 |
xentrac_ | nmz787_i: I am sure that you can use a zeolite dehumidifier to get water in the desert | 20:52 |
xentrac_ | I mean you have to use hot air to get the water out of the zeolite, and hot air will keep getting water out of the zeolite until it's saturated | 20:56 |
xentrac_ | which is to say that its dewpoint is the temperature of the hot air | 20:58 |
nmz787_i | yeah I use it in my garage now | 21:00 |
nmz787_i | it works pretty good, constant dehumidification across temperatures, though the data they provide stops around 90F | 21:01 |
nmz787_i | (it works great at i.e. 0C where a refrigerant based wouldn't be able to) | 21:01 |
nmz787_i | (or at least not very well) | 21:02 |
xentrac_ | yeah, I think you can make frost refrigeratively, but I'm sure the process slows quite a bit as the frost begins to insulate your refrigerant coils | 21:12 |
xentrac_ | I mean I'm quite sure you can; I've done it | 21:12 |
-!- drewbot [~cinch@ec2-54-196-245-91.compute-1.amazonaws.com] has quit [Remote host closed the connection] | 21:13 | |
-!- drewbot [~cinch@ec2-54-221-61-242.compute-1.amazonaws.com] has joined ##hplusroadmap | 21:14 | |
maaku | kanzure: or you can build the same thing for around $2k | 21:17 |
xentrac_ | I haven't measured its performance though! | 21:17 |
kanzure | maaku: how are you getting 180 TB for $2k? | 21:18 |
maaku | kanzure: I'm pretty sure "fully assembled" does not include drives | 21:20 |
kanzure | language owe | 21:22 |
xentrac_ | there's some in-depth discussion ofOpenCASCADE at https://news.ycombinator.com/item?id=11533435 | 21:22 |
xentrac_ | also of FreeCAD | 21:22 |
nmz787 | meh, seems like genehacker replied right the top, what kanzure will just tell you soon enough | 21:32 |
nmz787 | that it crashes and isn't reliable, and then try brlcad | 21:32 |
kanzure | gene-h in that thread is a regular of these parts | 21:33 |
kanzure | glad to hear they tried opencascade after i left their group but bummed that everything sucks | 21:33 |
xentrac_ | also the freecad release notes paeg mentions that they now have an alpha CAM module that can export G-code for "a variety of CNC machines" | 21:44 |
xentrac_ | SolveSpace also shows up in the comment thread on HN | 21:46 |
-!- sivoais [~zaki@unaffiliated/sivoais] has quit [Quit: leaving] | 22:00 | |
-!- sivoais [~zaki@unaffiliated/sivoais] has joined ##hplusroadmap | 22:01 | |
nmz787 | xentrac_: have you seen flatcam? | 22:04 |
nmz787 | I was trying to use that to export from gerbers to g-code for my little laser etcher | 22:04 |
-!- nildicit__ is now known as nildicit | 22:07 | |
nmz787_i | this is somewhat interesting http://optalysys.com/technology/watch-videos/ | 22:09 |
abetusk | nmz787, nmz787_i, you need a gerber to gcode exporter? | 22:18 |
Aurelius_Laptop | Bleh. | 22:20 |
nmz787_i | abetusk: yeah, i wanna try etching some boards | 22:34 |
nmz787_i | abetusk: does meowcad have that? | 22:35 |
-!- nmz787_i [~ntmccork@134.134.139.82] has quit [Quit: Leaving.] | 22:37 | |
-!- yashgaroth [~yashgarot@2602:306:35fa:d500:f5e0:f867:a11d:8d52] has quit [Quit: Leaving] | 23:03 | |
-!- nmz787_i [~ntmccork@134.134.139.82] has joined ##hplusroadmap | 23:19 | |
-!- nmz787_i [~ntmccork@134.134.139.82] has quit [Client Quit] | 23:21 | |
-!- Quashie [~boingredd@45.42.8.151] has joined ##hplusroadmap | 23:29 | |
-!- Aurelius_Laptop [~cpopell@c-73-129-20-70.hsd1.md.comcast.net] has quit [Ping timeout: 252 seconds] | 23:41 | |
-!- Aurelius_Laptop [~cpopell@c-73-129-20-70.hsd1.md.comcast.net] has joined ##hplusroadmap | 23:47 | |
--- Log closed Sat Apr 23 00:00:55 2016 |
Generated by irclog2html.py 2.15.0.dev0 by Marius Gedminas - find it at mg.pov.lt!