Current Server Time: 23:53 (Central Europe)

#apertus IRC Channel Logs

2021/05/22

Timezone: UTC


01:07
tpw_rules
vup: what do i need to do to get started the gsoc project? i figured i would attend the meeting monday.
01:11
tpw_rules
will it be hosted here or over on libera?
02:08
vup
tpw_rules: Attending the meeting on monday sounds good. Furthermore (if not already done), getting the fpga development environment setup is probably a good idea (ie installing vivado, its a huge download...).
02:09
vup
Additionally Bertl will give a introduction on how to get remote access to test hardware soon
02:09
vup
not sure where we will do the meeting, things are still a bit influx...
02:53
tpw_rules
vup: which fpga model do you use? i assume i need to install specific support packages like i do with quartus?
02:54
vup
its a zynq 7020, but IIRC all the zynq 7 series stuff (or maybe even all the 7 series stuff) is bundled together
04:23
Bertl_oO
yes, Monday is a good choice for a remote access session before or after the meeting (or both ;)
04:23
Bertl_oO
off to bed now ... have a good one everyone!
04:23
Bertl_oO
changed nick to: Bertl_zZ
09:30
mithro
left the channel
09:30
aombk
left the channel
09:30
felix_
left the channel
09:31
mithro
joined the channel
09:31
aombk
joined the channel
09:31
felix_
joined the channel
09:33
NausiyanNyan[m]
left the channel
09:34
davidak[m]
left the channel
09:34
bluez_[m]
left the channel
09:34
v1landa[m]
left the channel
09:35
aleb
left the channel
09:35
metal_dent[m]
left the channel
09:35
anu3jn
left the channel
09:35
elkos
left the channel
09:36
eppisai[m]
left the channel
09:36
promach3
left the channel
09:36
karl[m]
left the channel
09:37
RizwanMemon[m]
left the channel
09:39
ServerStatsDisco
left the channel
09:44
NausiyanNyan[m]
joined the channel
09:48
elkos
joined the channel
09:49
davidak[m]
joined the channel
09:49
bluez_[m]
joined the channel
09:49
metal_dent[m]
joined the channel
09:49
anu3jn
joined the channel
09:50
v1landa[m]
joined the channel
10:23
aleb
joined the channel
10:23
promach3
joined the channel
10:24
RizwanMemon[m]
joined the channel
10:29
eppisai[m]
joined the channel
10:31
karl[m]
joined the channel
10:32
se6ast1an
good day
10:33
eppisai[m]
good day! :)
10:50
ServerStatsDisco
joined the channel
11:35
anuejn
tpw_rules: maybe also build some bitstream from the nmigen-gateware codebase and run the testsuite
11:38
dcz
joined the channel
11:39
dcz
hi folks! I'm dealing with v4l2 power management right now, and I thought maybe you have a good idea to implement it
11:40
dcz
the question I have is: how to give the application enough time to do things which take time before the acapture starts: like setting the mode or focus
11:41
dcz
I'm in a pickle, because if the camera is powered on when the application starts the capture, then useful data will be delayed by the init time
11:42
dcz
on the other hand, if the camera is powered on when the application sets mode or focus, and then never starts the capture, then we're wasting power
11:43
dcz
I hope someone here knows the proper way to deal with it
14:33
tpw_rules
wouldn't the user want to see a preview of the camera image once they set the mode and focus?
14:53
Bertl_zZ
changed nick to: Bertl
14:53
Bertl
morning folks!
15:06
dcz
tpw_rules: usually yes, I don't think this is so clear cut. Maybe the shutter is released by a timed event
15:28
dcz
on a related note, the preview doesn't need to be the same mode, and turning the camera off while switching modes is causing actual bugs :)
17:30
tpw_rules
okay, i don't understand how the camera architecture works very much. that is a good thing to bring up at the meeting
17:32
vup
tpw_rules: usually the meeting is more for "quick" updates, but if you have time, we can do some general introduction to the camera after the meeting
17:34
vup
dcz: what is the target application of the camera you are building the v4l2 interface for?
17:34
tpw_rules
i assume that would be good for other gsoc participants too?
17:34
vup
sure
17:35
vup
but everybody needs slightly different details
17:35
tpw_rules
who else got accepted?
17:35
dcz
vup: immediately, it's point-and-shoot, but I want to make it as free of surprises as possible for anyone who wants to program it
17:35
aombk2
joined the channel
17:38
vup
tpw_rules: anuejn, eppisai[m], vnksnkr, and two more I don't see in this channel currently, you can see more details here: https://summerofcode.withgoogle.com/organizations/5025744687726592/
17:39
vup
dcz: well how long is the init time?
17:39
aombk
left the channel
17:40
dcz
I haven't measured yet, thinking there must be a solution independent of init times
17:40
vup
Well that depends on your definition of solution I think. To me this seems like a problem of UX design, which usually does not have a clear cut solution?
17:41
dcz
I see it as a problem of API design... which arguably is UX, I'm just not super familiar with v4l2
17:42
vup
tbh I am not really familiar with v4l2 either :)
17:42
vup
what does powered down mean for you? Is the camera completely shut down (ie would save the parameters set via v4l2 and then apply them once you power the camera on) or just the capture being turned off?
17:43
dcz
the camera is completely cut of from power, including the mechanical parts
17:43
vup
well then I would say you need to power it on before being able to set any parameters
17:44
dcz
I'd agree, but there doesn't seem to be any explicit call for that in v4l2
17:44
vup
dcz: how is the camera connected to the host computer where the v4l2 interface will be used?
17:44
dcz
mipi-csi2
17:45
vup
Hmm I see, not sure how the mipi-csi2 infrastructure looks like for linux. Is it very static, or can you connect and disconnect devices at runtime?
17:46
dcz
It seems so, as the power management calls do indeed reset the focus mechanism without fail
17:46
vup
well is it possible to only make the v4l2 device available when the camera is powered then?
17:47
dcz
I've been thinking of keeping the camera file descriptor open as a "hold power" condition, but that would make using terminal tools annoying
17:48
dcz
not really, the idea is that the device is available for opening all the time, and doing that supplies the power
17:48
dcz
I guess that's my solution
17:48
vup
I mean yeah, that then seems like your solution :o
17:49
vup
s,:o,:),
17:49
dcz
thanks :)
18:56
Bertl
off for now ... bbl
18:56
Bertl
changed nick to: Bertl_oO
19:04
anuejn
dcz: not really an answer to your question but you might be better helped in #libcamera
19:04
anuejn
and you might want to look at their software infrastructure :)
19:04
dcz
oh I didn
19:04
anuejn
(if not already done)
19:05
dcz
nott realize they have a channel
19:06
intrac
left the channel
19:18
comradekingu
left the channel
20:38
intrac
joined the channel
20:59
dcz
left the channel
21:19
dos1
joined the channel
21:47
Spirit532
left the channel
00:52
rektide
left the channel