#qi-hardware IRC log for Tuesday, 2011-04-19

rjeffriesI suspect the bot that copies this channel to the log file is not working.06:23
kyakddd/t06:26
kyakwhat makes you think that?06:26
wpwrakwolfspraul: hmm, tuxbrain is very silent lately. do you have any life signs from him ?10:59
wolfspraulno10:59
wolfspraulbut... tuxbrain nearly disappeared for 3+ months last year11:00
wolfsprauldiving11:00
wpwrakurgh11:00
wolfsprauland then he reappeared11:00
wolfspraulwell the reality is sometimes with our kind of stuff you cannot hold the same pace all the time11:00
wolfspraulI am also quite a lot underground lately, or am I not? :-)11:00
wolfspraulso he reappeared last time (and stronger than before), he will surely reappear this time11:01
wpwrakyes, indeed. found some "daytime" work ?11:01
wolfspraulno way11:01
wpwrakall mm1-v3 ?11:01
wolfspraulI rather go run 10h / day thinking how to move forward11:01
wpwrak(disappear/reappear) yeah. i'm just wondering because we have the wpan production moving along. so there's an external pace for a while.11:02
wolfspraulyes mostly rc311:02
wolfspraulfrom our forum: "Is it Possible to use nanonote as an oscilloscope (using it's inbuilt ADC using the 3.5mm jack as input) using software like xoscope?"13:55
wpwrakit ought to be, with all the constraints such a solution has (no DC-coupling, only AC-coupling; very limited bandwidth; probably not very linear either; ...)13:57
wpwrakregarding software, it may be easier to write something from scratch than porting an application, particularly if non-X is a requirement13:59
wpwrak(linear) that can of course be partially compensated by software14:00
jesperjhi guys14:02
wolfspraulhi14:04
jesperjIf I recall correctly, when checking out eye tracker designs in the past, I have seen quite a lot of them using analog cameras compared to for example ordinary "webcams". If this is the case, is there any reason why an analog camera would be prefered?14:10
jesperjor could it just be the case that good enough digital cameras were not available when the designs with analog cameras were compiled?14:11
wpwrakbetter optics than cheap webcams perhaps ?14:11
wpwrakyes, that would be my guess14:11
wpwraksome of the early webcams were quite dreadful14:12
jesperjIf I were to play around with making my own eyetracker today, then I would want a very small camera to make it inobtrusive, and then what comes to mind are cameras from cell phones14:12
wpwrakjesperj: yes, why not. phone cams have reasonably good resolution. and they (usually ?) have a fast parallel bus, so you can get the images out quickly enough.14:16
Action: jesperj nods14:21
jesperjI just had to check if I had missed something there, some reason why not to use digital camera14:21
rohnot reaööy14:22
rohirgh. really14:22
rohjesperj: 'no driver support yet'14:22
jesperjhttp://embeddedsystemnews.com/toshiba-introduces-new-8-megapixel-cmos-image-sensor-and-2mp-cscm-in-15-inch-optical-format.html < Cool. Besides CCM cameras there are also CSCM "Chip Scale Camera Module"14:23
jesperjroh: What do you mean with "no driver support yet"?14:23
rohdigital camera -> digital interface.14:26
rohmeans usb or similar.14:26
rohmm1 doesnt have drivers for such afaik14:26
wolfspraulroh: hey there. I have some mechanical m1 feedback.14:26
wolfspraulfirst - Adam will see whether the DMX TX connector is available in a version that has the metal 'push' button on the left side, rather than on top14:26
wolfspraulthat would make it easier to press, because otherwise the top acrylic is in the way a little when trying to press it14:27
wolfspraulif we do that, we would need to move the 'DMX' label to the left, otherwise it would be partially behind the 'push' button14:27
wolfsprauldo you know what I mean?14:27
rohsure14:28
wolfspraulis that OK with you? (we haven't decided whether we want this change, I just try to get the pieces together so we can make a yes/no decision)14:28
rohof course i can do that14:28
wolfspraulit's a small thing, but seems like an easy change if such a connector exists...14:28
wolfsprauldo you have any opposition against this idea?14:29
rohi would just like to remember that the current connector is neutrik. dont use anything cheaper or less qualitative.14:29
wolfspraulanother option would be to cut out some piece from the top acrylic, but that's a bigger change and would look ugly imo14:29
wolfspraulno worries14:29
wolfspraulthe 'push' is still pushable, but for someone with a bit bigger fingers, or even glove or so, it's probably hard to push it14:30
rohi was more worried that the depth to press the button isnt deep enough with 3mm acryllic14:30
jesperjroh: Pardon my ignorance but what is "mm1"?14:30
wolfsprauljesperj: thanks for asking, curiosity rules here :-)14:31
rohjesperj: milkymist one14:31
jesperjoh I am way too curious for my own good :)14:31
wolfspraulanonymous curiosics14:31
rohjesperj: i thought you were referring to the cam input and the discussions about adding one to the kit14:31
wolfspraulwe all suffer from that one14:31
jesperjbut it feels good if that attitude is common in this channel14:32
wolfspraulroh: I'm just testing. I cannot tell what the minimum is that 'push' needs to be pushed in, but there is no problem with that14:32
wolfspraulthe push button definitely releases the cable14:32
jesperjroh: Oh. No I am thinking about pros and cons for usage in constructing an eye tracker.14:32
wolfspraulbut it's hard to push down being squeezed in between the top acrylic and the DMX cable14:32
rohbtw.. there are no analog cams.. or digital ones. there are cams with analog or digital video interfaces. the chips themselved (ccd/cmos) are not really analog anyhow14:32
wolfspraulroh: ok, so please note this. We _may_ move the DMX label to the left for the next order of 80 cases from you, it's not decided yet.14:32
wolfspraulI will let you know asap.14:33
wpwrakroh: sure there are analog cams ;-)14:33
rohwolfspraul: the only thing i can do anyhow is drill the metal sheets from gemmel which arrived yesterday and then start unmounting the cnc mill14:33
wolfspraulroh: then I have another one. THe reception of the internal microphone is quite bad inside the case. How do you feel about adding a hole on the side, between VGA connector and line-in14:33
rohmy electronics workbench is already unmounted and moved (not reassembled yet)14:34
wpwrakjekhor: btw, mm1 refers to this: http://www.milkymist.org/14:34
rohhm. i dont think a simple hole would make it better so much14:34
wolfspraulAdam would then need to see how he can move the microphone closer to the acrylic, or make it point towards the acrylic14:34
wolfspraulno we can make the microphone reach into the hole or so. only have to be careful that it's not becoming too fragile for assembly/disassembly14:35
rohto be fair.. i wonder why you added such a electret mike at all.14:35
wolfspraulyou are saying you would remove the microphone?14:35
rohi wouldnt have added an internal one.14:36
wolfspraulI think an integrated microphone that 'just works' is nice. but inside the case the reception is just not very good, only if it is loud outside (which Sebastien says it is in a VJ environment)14:36
wolfspraulok, but we have it and I think it's nice14:36
rohseems like a leftover from a develboard where somebody wanted to use all the pins on his codec (makes sense on a develboard)14:36
wolfspraulso the question is whether we can improve reception14:36
wolfspraulit's fun that you can clap into your hands and the vga output reacts :-)14:37
rohi am not critisising.. just wondering. dont vj's always get proper line audio from the dj booth?14:37
wolfspraulok let's just be very specific about the upcoming run of 80, rc314:37
wpwrakwolfspraul: how does sound currently reach the mic ? are there holes for it in the acrylic ?14:37
rohwpwrak: resonance mostly14:37
wolfspraulI think it's clear we do not want to move the vga connector or line in/out14:37
wolfspraulso the only question is whether we can improve the microphone reception14:38
wolfspraulwe will not remove the mic in rc314:38
wpwrakroh: so .. negative hi-fi ?14:38
wolfsprauladam can move it closer to the acrylic, and make it point to the acrylic14:38
wolfspraulI think if there would be a hole there, things could be much better14:38
rohwpwrak: its a low-end electrec capsue.. so.. regardless of making holes.. ye. low-fi14:38
wolfspraulwpwrak: well. no.14:38
wpwrakwolfspraul: one hole or a number of holes ?14:38
rohand my fingers are still sleeping14:39
wolfspraulthe gaps in the case somewhere :-)14:39
wolfspraulI was thinking about one hole, potentially large enough so that the mic can reach into it14:39
wpwrakdoes the mic have a surface suitable for exposure ?14:39
rohwpwrak: nope.14:40
wolfspraulhuh? checking...14:40
rohwolfspraul: i think i can imagine.. but it wouldn look pretty14:40
wpwrakso that would suggests many small holes something like <= 1mm14:40
wolfspraulok, then we can also do something with a few small holes, and the mic behind those14:40
wolfspraulthat's also more robust for assembly/disassembly14:40
wolfspraulwhy 'many small holes'? I would make them less and bigger14:41
wolfspraul3-514:41
rohand what keeps the mic there?14:41
wolfspraulit keeps itself there, it's on some feet right now (wait checking...)14:41
wpwrakdrawback: may give an access for spilled liquids. does the mic face upward ?14:41
wolfsprauljust need to find out that points to the side, or simply bend it14:41
rohmine has weird bent pins... , 214:42
rohwpwrak: ack14:42
wolfspraulok mine is sitting flat on the pcb14:42
rohwolfspraul: add some 3$ mic for external connect, remove the internal one ;)14:43
wolfsprauldiameter seems to be 5-6mm (guessed)14:43
rohoutsource the issue11!114:43
wpwrakoh course, most DJs probably have a spare gibbet stowed away for anyone who spills their drink on the equipment, but still ...14:43
wolfspraulmore clutter, I really don't like it14:43
wolfspraulalso more expensive14:43
wolfspraulintegrated mic is nice14:43
rohmaybe you should test the internal mic14:44
rohvia line out.14:44
wolfspraulso... it seems we have a lot of enthousiasm and reached consensus :-)14:44
rohand some headphones or so14:44
wolfspraul(joking)14:45
rohto 'hear wat the mm hears'14:45
wolfspraulmaybe you are right and there are more problems14:45
wolfspraulbut I just want to make a simple improvement now, or understand what simple improvement is possible14:45
wolfspraulit is pretty clear that the mic is really locked away right now14:45
rohcheck how sensitive it is without case and if wood or acryllic makes much difference etc14:45
wolfspraulwell, that's easy14:46
wolfspraulopen the top cover, run some patch, speak into the microphone14:46
rohno. not a patch14:46
wolfspraulthen close the cover, tighten the screws, try the same thing again14:46
wolfspraul:-)14:46
rohi meant AUDIO14:46
wpwrakfor the little holes, probably a center hole plus a ring of peripheral holes would work well. maybe with a ring diameter of ~5 mm. (hole center would be on the ring)14:46
rohnot something which works on it. use headphones.14:46
wolfspraulI am pretty sure that the way we lock it into the acrylic makes it mostly 'deaf'14:47
wolfspraulof course in software we can increase the volume again, but why not add some holes and move the mic close to those holes?14:47
wpwraksomething like 6 holes on the ring14:47
wolfspraulSebastien made the best point - in a typical VJ environment it is so loud that even if it's locked away, it will still work14:48
wpwrakwolfspraul: the case probably also has frequency-dependent attenuation. so you'd have to equalize as well, not just amplify.14:49
wolfspraulbut I think that's a quite extreme argument, even if valid, especially that is no argument why it shouldn't be better in the first place14:49
wolfspraulroh: let's do 2 things:14:49
wolfspraul1) I ask Adam how easy it is to move the mic right next to the side, and make it point to the side (90 degree)14:49
wolfspraul2) same as with 'moving the DMX label', you know that we are looking into this and maybe we want to add some holes there. if so I will try to send you a mechanical sample upfront.14:50
wolfspraulor just some parts so you can try/imagine14:50
wolfspraulyou haven't made a point yet why such holes would be bad, you only say the whole internal mic idea is bad14:51
wpwrakyou could also combine "many small holes" with "lateral placement". that would remove the access for liquids.14:51
wolfspraulwpwrak: you mean just leave it sitting flat on the pcb?14:51
wolfspraulif we add a few holes, we probably also want to add a new label 'MIC'14:52
wpwrakwolfspraul: leaving it flat may not be optimal when the sound comes from the side. but if you can tilt it and make small holes on the side, you'd get reasonably simple assembly plus direct access for soundwaves, yet no new upward-facing hole(s)14:52
wolfspraulupward facing holes?14:53
rjeffrieswpwrak is your email working these days?14:53
wolfspraulI thought roh and I so far talked about holes on the side14:54
wpwrak(hole pattern) with a 1+6 pattern, the distance between holes on the ring would be roughly the same as the distance between holes and the ring and the center hole14:54
wpwrakrjeffries: aye14:54
rjeffriesthx14:54
wolfspraulroh: technically adding holes there and a 'MIC' label is OK for you?14:54
wpwrakrjeffries: at least a dozen bayes-cheating spammers have told me so today :-(14:54
rohwolfspraul: could do that.. dont think it will look nice yet14:55
rohyou mean the side, right?14:55
wolfspraulactually now that I think about it, most 'mic' designs just have 1 small hole14:55
wolfspraullike the Ben :-)14:55
wpwrakwolfspraul: (holes) ah, thought that plan A was big mounting hole on the side, plan B many small holes wherever the mic is facing14:55
rohdoes only work with directional sounds... like headphones.. or telephones14:56
wolfspraulthe mic is facing up, but the distance between top of mic and top acrylic is quite big, and I wouldn't want holes in the top side14:56
rohand the mic is placed directly behind such one-hole-things14:56
wolfspraulyes but we can do that here14:56
wolfspraulone hole, say 2mm diameter? or 1mm? then the mic right behind it?14:56
rohanyhow... i'll need to run now..  hope i can get the drilling for the metal done today... then the machines will be disassembled14:57
wolfspraulok14:57
wolfspraulroh: please remember those 2 potential additions14:57
wolfspraulDMX label move, MIC label & hole14:57
wolfspraul'potential' right now14:57
rohwolfspraul: i'll ask again for final thoughts before giving the big laser order.14:57
wolfspraulthanks14:58
rohcurrently i am just stocking on parts and ramping everything to the 'go'14:58
rohwhile moving ;)14:58
rohbbl15:01
rjeffriesis everyone asleep?15:30
wpwrakor working :)15:30
rjeffrieswpwrak that is a decent alibi. ;)16:53
GNUtoo<rjeffries> wpwrak that is a decent alibi. ;)18:15
GNUtooalibi of that:18:15
GNUtoogit commit crime18:15
GNUtoo?18:16
wpwrakalibi for being silent yet not asleep :)18:19
wpwrakhmm, does anyone know if there's a way to do the equivalent of passing a -geometry argument with SDL ? what i'm looking for is a way to make an SDL window appear at a specific position under X.19:27
wpwrakah, found it.  export SDL_VIDEO_WINDOW_POS=x,y    :)19:28
phretorhi, will this IDE2USB (NEC D720130GC) http://dl.dropbox.com/u/215810/ide2usb.jpg need separate power or it can take it from the USB bus? It comes from a 3.5'' HD enclosure.19:51
Jay7wrong channel but you may just try19:55
phretorJay7: any other suggestion besides #hardware20:06
Jay7sorry, I don't know any..20:07
rjeffrieswpwrak interesting set of slides around TinyOS and wireless PANs. http://www.inf.ethz.ch/personal/mharvan/talks/wsn.pdf20:30
wpwrakheh, yet another C variant ;-)20:45
wpwrakthe problem: global optimization is hard because object files don't carry enough information. attempted solution: dumb down the language a but and beef up the object files a bit, until you can somewhat do global optimization. fact not considered: open source exists.20:51
wpwrakan analogy:20:51
wpwrakthe problem: this knot is too difficult to disentangle. attempted solution: get the knotmaker to make it a little less complicated and use an oiled rope, until you can wiggle it apart. fact not considered: you're carrying a sword.20:52
wpwrakhow hard can it be to put 2 and 2 together and arrive at the correct result ? :)20:53
rjeffrieswpwrak what do you kn ow about the microIP protocol they mention wher eit looks like21:54
rjeffriesthey are able to have a very small IP and TCP/IP that (I think) they tunnel through some lower level protocol21:55
rjeffriesthe aim being to allow a VERY low poere WPAN node be able to conne=ct (eventually, through a gateway)21:56
rjeffriesto the "real internet)21:56
wpwrakrjeffries: isn't uIP a stack of 6lowpan, not a new protocol ?22:01
wpwrakrjeffries: how you connect to the "real" internet is another question. sometimes you do indeed want a layer 7 gateway.22:03
wpwrakrjeffries: non-standard layer <= 3 protocols don't make sense anyway. not being able to fit 6lowpan into some microcontroller is hardly an excuse :)22:05
rjeffrieswpwrak I am out of date re your status with 6lowpan.22:07
wpwrakrjeffries: nothing new for the last few weeks. i'm now working on the production testing.22:10
--- Wed Apr 20 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!