| norayr | i have opened the issue https://github.com/maemo-leste/bugtracker/issues/699 and there is a video which illustrates it: https://toobnix.org/w/dhb9VMzyGASkiYxmeoYB7G | 01:26 |
|---|---|---|
| soldan | well, the mail app, modest, works perfectly. profanity works fine | 02:23 |
| soldan | empathy and dino doesnt work | 02:23 |
| soldan | im gona try what web-browser i can use | 02:24 |
| rafael2k | which ISP the N900 uses? | 08:29 |
| rafael2k | btw, I'll try to push for a libcamera GSOC to work on this: https://libcamera.org/open-projects.html (OMAP3 ISP pipeline handler) | 08:32 |
| rafael2k | it would be directly beneficial to us | 08:32 |
| rafael2k | It would be nice if we could send a N900 running Maemo (and with libcamera set) in case any developer there picks the task | 08:33 |
| sicelo | omap3isp | 08:38 |
| rafael2k | right | 08:39 |
| sicelo | Laurent definitely would help | 08:39 |
| rafael2k | Laurent is listed as mentor | 08:40 |
| rafael2k | for the GSOC project | 08:40 |
| sicelo | gawd! it has 350 hours there, compared to <200 for some of the others :p | 08:40 |
| rafael2k | eheheheh | 08:40 |
| rafael2k | students have plenty of time! | 08:40 |
| sicelo | hehe | 08:41 |
| sicelo | if they won't feel that these devices are dinosaurs | 08:41 |
| sicelo | luckily, at least for N900, student wouldn't be starting from scratch. pavel has taken 'green' pics with it ... i think even documented it on our mailing list | 08:42 |
| sicelo | we have broken dts atm for the camera, afaict, but that should be easy (i'll actually do it) | 08:46 |
| rafael2k | indeed, it was working at some point: https://lwn.net/Articles/685308/ | 08:53 |
| rafael2k | not perfect, but almost there certainly | 08:54 |
| freemangordon | n900 camera system is kind of special (as everything with n900 :) ) | 08:54 |
| freemangordon | it uses gpio to control a mux that switches between front and back camera | 08:55 |
| freemangordon | back then there was no kernel support for this | 08:55 |
| rafael2k | pinephone does this too | 08:55 |
| rafael2k | now there is! | 08:55 |
| sicelo | there might be now ... i seem to think pp usees it | 08:55 |
| sicelo | ah yes | 08:55 |
| freemangordon | how is it inplemented? | 08:55 |
| rafael2k | and with libcamera, we can properly expose all the controls | 08:55 |
| freemangordon | *implemented | 08:55 |
| rafael2k | as multiple endpoints of the controller | 08:55 |
| freemangordon | in terms of kernel/v4l that is | 08:56 |
| rafael2k | then you can select the endpoint | 08:56 |
| freemangordon | and which driver does the switching? | 08:56 |
| freemangordon | isp? | 08:56 |
| rafael2k | I think if you just open one camera or another, the driver selects the appropriate endpoint | 08:56 |
| freemangordon | (in terms of n900 that is) | 08:56 |
| rafael2k | take a look here: | 08:56 |
| freemangordon | which driver? :) | 08:56 |
| rafael2k | https://github.com/megous/linux/commit/ed71df3572dd998bcee96568e40b214145529f44 | 08:57 |
| freemangordon | BTW, I have docs with camera specs around, if someone is interested | 08:57 |
| rafael2k | this is for the PP sun6i-csi | 08:57 |
| rafael2k | "Allow to register multiple parallel bus subdevices and switch between | 08:57 |
| rafael2k | them by enabling/disabling media graph links." | 08:57 |
| freemangordon | this is not upstream, no? | 08:57 |
| rafael2k | no but it is supported upstream | 08:58 |
| freemangordon | also, n900 is worse than that ;) | 08:58 |
| rafael2k | (I mean, in other drivers) | 08:58 |
| freemangordon | back camera is parallel and from is serial, IIRC | 08:58 |
| freemangordon | *front | 08:58 |
| sicelo | freemangordon: i'm interested in the docs, please :-) | 08:58 |
| freemangordon | sec | 08:58 |
| rafael2k | anyway, still, it is not rocket science, mostly play with command and control (not that it is easy...) | 08:59 |
| rafael2k | In 350 hours it can be done : ) | 08:59 |
| freemangordon | hehe | 09:00 |
| sicelo | if you already know the basics, C++/C | 09:00 |
| sicelo | so for me, i need 3500 hours :p | 09:00 |
| rafael2k | :P | 09:01 |
| freemangordon | sicelo: http://95.43.220.235/n900/camera/ | 09:02 |
| freemangordon | sicelo: knowing c/c++ is the easiest part | 09:05 |
| freemangordon | c has, how many, 19 reserved words? /me check | 09:05 |
| sicelo | thanks, i already had two of the docs, besides the smia functional specifications. tyvm | 09:06 |
| freemangordon | hmm, 32, this seems to have increased back from k&r days :) | 09:06 |
| sicelo | hehe, yes, the languages themselves are easy ... it's groking them that's the problem :p | 09:06 |
| sicelo | structs and pointers | 09:07 |
| sicelo | and other stuff that you guys seem to magically know. recent example - by looking at the N900 dts, Wizzup realized the modem used PIO. i've looked at it a couple of times, and even today i can't tell how that conclusion was reached | 09:09 |
| * sicelo should get a mentor | 09:09 | |
| sicelo | < freemangordon> sicelo: knowing c/c++ is the easiest part <<== what's the harder part? | 09:16 |
| uvos__ | knowing how everyones elses software, frameworks work/ fit together | 09:20 |
| uvos__ | a (c) libarary/framework is kinda its own language but with 100 to 1000 of keywords instead of just 32 | 09:21 |
| uvos__ | in a way | 09:21 |
| rafael2k | indeed | 09:22 |
| rafael2k | btw, this a front cam screen cap: https://www.abradig.org.br/maemo-crazyness/pinhole-working/screen-cap-2.jpg | 09:22 |
| rafael2k | greenish... need to play with while balance controls in pinhole | 09:22 |
| uvos__ | green istent really a white balance thing | 09:22 |
| uvos__ | its mostly about the sensetivity curve of cmos sensors | 09:23 |
| rafael2k | I'm guessing | 09:23 |
| uvos__ | you have to callibrate it away | 09:23 |
| rafael2k | how? | 09:23 |
| rafael2k | change the gamma curve? | 09:23 |
| rafael2k | or do a red / blue balance? | 09:24 |
| uvos__ | so for raw (DNG) conversion | 09:24 |
| uvos__ | eatch camera has a "base" curve for eatch channel | 09:24 |
| uvos__ | that is applied to convert the cameras lopsided colorspace to (s)RGB | 09:24 |
| uvos__ | idk how this relates/ works in libcamera ofc | 09:25 |
| rafael2k | MegaPixels set some registers in the driver and trigger some controls... | 09:26 |
| rafael2k | it is not the green the image there | 09:26 |
| uvos__ | the software has to do the covnersion | 09:26 |
| uvos__ | megapixels might be doing it | 09:26 |
| uvos__ | and assuming the pp sensor | 09:26 |
| rafael2k | the camera needs appropriate calibration controls to be set | 09:27 |
| rafael2k | and megapixels seem to do it right... I'll take a look | 09:27 |
| rafael2k | uvos__, can be, we can even have an option to save the DNG | 09:33 |
| rafael2k | sicelo, tks! | 09:45 |
| rafael2k | Can I say in libcamera that we can donate a N900 if a student gets a GSOC to do this work? | 09:46 |
| rafael2k | (of the N900 camera support) | 09:46 |
| sicelo | yes :-D | 09:46 |
| rafael2k | ok! | 09:46 |
| uvos__ | we also can also donate a mapphone as a streach goal, its front camera is pretty mutch in the same state as n900 (omap4s isp is in stageing and out of tree drivers for the sensor exist) | 09:50 |
| uvos__ | presumeably most of the work can be shared here | 09:51 |
| uvos__ | note "front" | 09:51 |
| rafael2k | wrote | 10:55 |
| Wizzup | sicelo: re: pio, I just looked at it prior and probably concluded that probably dma just never worked, and probably someone here confirmed that and I just went with it | 11:10 |
| Wizzup | I don't think I deduced it from the dts | 11:10 |
| Wizzup | and yeah we can definitely donate mapphone(s) | 11:11 |
| rafael2k | uvos__, then if someone picks the job, you became a mentor! :P | 11:12 |
| rafael2k | ( :P but seriously ) | 11:13 |
| freemangordon | Wizzup: yes, it was me (pio) | 11:31 |
| Wizzup | right | 11:36 |
| freemangordon | sicelo: no, you got it wrong - 'we' don't know it magically, it is all hard work looking in | 11:38 |
| freemangordon | the source code of others | 11:38 |
| freemangordon | reading through docs, etc. | 11:38 |
| freemangordon | that's it | 11:38 |
| Wizzup | there is some experience factor I think | 11:39 |
| freemangordon | when you do that for years, you got experience and usually know what to look for and where | 11:39 |
| freemangordon | Wizzup: yes | 11:39 |
| freemangordon | so, it is the same as with other things - hard work | 11:40 |
| freemangordon | and dedication | 11:40 |
| freemangordon | sure, some of 'us' are better than others, but that's to be expected | 11:40 |
| Wizzup | mhm | 11:41 |
| freemangordon | Wizzup: speaking of pio, I don;t remember seeing a patch sent upstream, did I miss something | 11:43 |
| Wizzup | sicelo is in charge of this atm I think :) | 11:45 |
| Wizzup | or rather, I'm didn't send anything :D | 11:45 |
| freemangordon | hehe | 11:45 |
| freemangordon | be careful what you wish for :p | 11:46 |
| freemangordon | ttyl | 11:46 |
| sicelo | mmm, i guess it should just be a question/discussion? i suppose we're not at patch stage, since we're just forcing pio on a driver that could possibly be used by something else with dma | 11:50 |
| Wizzup | the question I guess is what changes that made it now take a dma path when it's not supported | 11:51 |
| Wizzup | changed* | 11:51 |
| norayr | having fcamera and blessn900 working on all maemo phones would be amazing. | 12:22 |
| Wizzup | blessn900 isn't foss jus so you know | 12:24 |
| Wizzup | afaik | 12:24 |
| bencoh | fcamera would be cool indeed :) | 12:27 |
| bencoh | but looks like rafael2k already has an app of his own anyway | 12:28 |
| sicelo | yeah, seems good to me | 12:28 |
| bencoh | still, I'm not quite sure how the fcamera kernel driver(s) worked, but having a look at it might ease development on mainline | 12:30 |
| bencoh | iirc they bypassed the ISP in every aspect (?) | 12:30 |
| sicelo | the little i read seems to say - forget about fcamera :-P | 12:31 |
| sicelo | it needed its own special kernel module, and a lot of other hacks | 12:32 |
| bencoh | sicelo: well, fcam aimed at providing frame-accurate metadatas | 12:37 |
| bencoh | nobody did anything close to that on linux back then | 12:38 |
| uvos__ | well besides all the other phoens running linux presumably | 12:39 |
| bencoh | back then? | 12:39 |
| bencoh | and no, frame-accurate wasn't (and still isn't?) a common thing on linux | 12:39 |
| bencoh | v4l2 had no proper metadata handling mechanism back then | 12:39 |
| uvos__ | android probubly, motorola linux (dont remember what it was called, most late motorola featire phones ran linux) | 12:39 |
| bencoh | (it's hardly better nowadays) | 12:40 |
| bencoh | uvos__: in 2009? | 12:40 |
| uvos__ | yes | 12:40 |
| uvos__ | in 2007 even | 12:40 |
| bencoh | samsung implemented there own framework, but afaict it came a bit later | 12:40 |
| uvos__ | anyhow | 12:40 |
| bencoh | their* | 12:40 |
| uvos__ | these devices where locked down linux | 12:40 |
| uvos__ | so probubly terrible hacks | 12:40 |
| uvos__ | (besides android idk how the state was on android that early) | 12:41 |
| uvos__ | android is a good bit older than n900 after all | 12:41 |
| bencoh | only 2y ? | 12:41 |
| uvos__ | a good bit in this space | 12:41 |
| bencoh | not even 2y, it was released in sept 2008 apparently | 12:42 |
| uvos__ | https://en.wikipedia.org/wiki/MotoMagx | 12:42 |
| Wizzup | was it actually used by anyone then? | 12:42 |
| uvos__ | btw | 12:43 |
| uvos__ | Wizzup: sure it was pretty popular | 12:43 |
| uvos__ | really popular really | 12:43 |
| uvos__ | but not in europe | 12:43 |
| sicelo | bencoh: i completely agree with what you said, re-fcam. so yes, it's hard to use it with regular linux because of exactly that same reason | 12:43 |
| uvos__ | i used to use this ^^^ | 12:43 |
| Wizzup | looks like google doubt bought it in 2005 | 12:43 |
| Wizzup | s/doubt// | 12:44 |
| uvos__ | sue i mean post htc dream ofc | 12:44 |
| uvos__ | and post d1 it was the second largest platform | 12:44 |
| uvos__ | this is same time as n900 | 12:44 |
| uvos__ | anyhow | 12:44 |
| uvos__ | this is hardly relevant | 12:44 |
| sicelo | yes, I was still trying to understand how it's relevant. to be clear, fcam that's being discussed is not the stock camera | 12:45 |
| bencoh | what's the current camera state on n900/mainline btw? | 12:47 |
| bencoh | according to elinux it worked on v4.14-rc1 | 12:47 |
| sicelo | drivers and isp work | 12:47 |
| sicelo | but i think something's wrong with some clocks ... probably just a simple dts update (that i will look into) | 12:48 |
| bencoh | ah | 12:48 |
| sicelo | actual capturing is a completely different problem ... but maybe libcamera helps there | 12:48 |
| sicelo | see complex cameras talk on YouTube, by Pavel Machek | 12:49 |
| sicelo | at least he could take green photos of his horses :-p | 12:49 |
| bencoh | haha | 12:51 |
| bencoh | yeah, whitebalance/focus is yet another story, but once you have a way to capture frames and control the sensor/actuator, it's already a great start | 12:52 |
| sicelo | https://lists.dyne.org/lurker/message/20200306.084252.32b5515a.en.html | 12:53 |
| sicelo | i think this was 5.x era | 12:53 |
| bencoh | nice | 12:53 |
| Wizzup | tmlind: fyi I also ordered two mz615 just now, so we have a mz617 with microsd slot | 13:49 |
| Wizzup | probably easier for testing/debugging (at least for me) | 13:49 |
| Wizzup | I can get a third one if you'd want to use one | 13:49 |
| Wizzup | (I'll send one to uvos) | 13:49 |
| sicelo | will Leste be represented at FOSDEM? | 13:50 |
| Wizzup | sicelo: there is no stand, room, talk, or anything | 13:52 |
| Wizzup | but I will be there | 13:52 |
| Wizzup | pavel will be there | 13:52 |
| Wizzup | not sure who else :) | 13:52 |
| sicelo | cool | 13:52 |
| Wizzup | we should probably update our fosdem schedule app | 13:53 |
| Wizzup | updated it | 14:32 |
| Wizzup | oh I need to update the logo | 15:16 |
| soldan | hi there, me agaian | 18:33 |
| Wizzup | \o | 18:33 |
| soldan | which xmpp client do you use in maemo leste with n900? i try empathy but do not open, tried psi and the screen do not fit and i cant log | 18:33 |
| soldan | i tried tkabber and i cant login | 18:33 |
| soldan | the only that works to me was profanity, but i want to try a gui-client | 18:34 |
| soldan | piding works but the screen do not fit well | 18:34 |
| Wizzup | that's right, this is still work in progress | 18:37 |
| Wizzup | we will make our conversations work with it | 18:38 |
| soldan | ohhhh wow, so im gonna wait :D | 18:38 |
| soldan | conversations its the best | 18:38 |
| Wizzup | right, yeah, waiting unfortunately :) | 18:49 |
| soldan | no problem, I can wait :D | 18:52 |
| Wizzup | freemangordon: I'm working with some folks on a user manual, what theme do you think we should use for the screenshots? | 19:13 |
| Wizzup | alpha? | 19:13 |
| Wizzup | or maybe devel? | 19:13 |
| Wizzup | (this has the bonus of hiding some of our images being off center like in the clock/alarm and in ham) | 19:14 |
| freemangordon | Wizzup: beta or matrix :) | 19:37 |
| freemangordon | devel is too broken | 19:38 |
| freemangordon | umm, why off center? adjust the resolution. | 19:38 |
| freemangordon | I guess you use VM, no? | 19:42 |
| Wizzup | freemangordon: I suppose, but he has his d4 | 19:48 |
| Wizzup | matrix doesn't seem a good fit imo :p | 19:49 |
| freemangordon | Wizzup: is it really a good idea to use device when we have VM images? | 20:55 |
| freemangordon | well, matrix is my favourite :p | 20:56 |
| freemangordon | but yeah... | 20:56 |
| freemangordon | maybe beta, it is more soft and eye candy | 20:56 |
| uvos | device vs vm probubly dosent matter | 21:38 |
| uvos | but doing the screenshots at n900 resolution makes sense | 21:39 |
| uvos | untill all the issues are fixed | 21:39 |
| Wizzup | hmmm | 21:52 |
| Wizzup | I don't think he has a n900 | 23:40 |
| Wizzup | I think maybe what makes the most sense rn is to get the text written, screenshots are relatively easy to re-do | 23:40 |
| Wizzup | since I don't see an easy way | 23:40 |
| Wizzup | vm and n900 cannot do some tihngs | 23:40 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
