Current Server Time: 22:32 (Central Europe)

#apertus IRC Channel Logs

2017/03/04

Timezone: UTC


00:04
dimon_
joined the channel
00:43
dimon_
left the channel
02:05
jucar
joined the channel
02:06
Bertl
off to bed now ... have a good one everyone!
02:07
Bertl
changed nick to: Bertl_zZ
02:22
BAndiT1983
changed nick to: BAndiT1983|away
03:24
karan
joined the channel
03:25
atoav
joined the channel
03:28
d0
left the channel
03:56
ItsMeLenny
joined the channel
05:07
RexOrCine
joined the channel
05:47
atoav
left the channel
06:46
dimon_
joined the channel
06:54
Mehri
joined the channel
06:55
Mehri
hi guys.
06:57
Mehri
left the channel
07:00
jucar
left the channel
07:06
se6astian|away
changed nick to: se6astian
07:09
dimon_
left the channel
07:29
Spirit532
joined the channel
07:53
karan__
joined the channel
07:55
karan
left the channel
07:55
karan__
changed nick to: karan
07:56
Bertl_zZ
changed nick to: Bertl
07:56
Bertl
morning folks!
08:14
jucar
joined the channel
08:15
jucar1
joined the channel
08:19
jucar
left the channel
08:26
jucar1
left the channel
08:37
anuditverma
joined the channel
08:40
anuditverma
left the channel
08:41
anuditverma
joined the channel
08:43
anuditverma
left the channel
08:52
karan
left the channel
09:31
karan
joined the channel
09:52
Bertl
off for now ... bbl
09:52
Bertl
changed nick to: Bertl_oO
09:56
BAndiT1983|away
changed nick to: BAndiT1983
10:09
BAndiT1983
changed nick to: BAndiT1983|away
10:10
BAndiT1983|away
changed nick to: BAndiT1983
10:11
BAndiT1983
changed nick to: BAndiT1983|away
10:11
BAndiT1983|away
changed nick to: BAndiT1983
10:44
arpu
left the channel
10:57
arpu
joined the channel
11:40
karan_
joined the channel
11:42
karan
left the channel
11:43
karan_
changed nick to: karan
11:50
RexOrCine
left the channel
12:29
g33kyaditya
joined the channel
12:45
g33kyaditya
Hey all !
12:45
g33kyaditya
I'm Aditya. Anyone around ?
12:46
BAndiT1983
hi
12:47
g33kyaditya
Could you tell me more about OpenCine ? Specifically about the task involved in this year's GSoC
12:49
BAndiT1983
OpenCine is planned as a toolsuite, which should help the people, on set or in post-production, to backup, manage and color-grade RAW clips
12:50
BAndiT1983
the task focuses on different de-Bayering methods, which should be applied to RAW data, which we will get from axiom beta for example
12:50
BAndiT1983
as many people asked about OC(=OpenCine), i'm thinking about adding new tasks, but haven't had time to think about it
12:51
BAndiT1983
one can refer to https://lab.apertus.org/T727 in that case
12:52
g33kyaditya
Wikipedia tells me there are multiple algorithms for de-Bayering. Is there a library involved to do this sort of thing
12:53
g33kyaditya
Well, ofcourse. OC sounds really cool, especially for some of us movie buffs, Lol ;)
12:53
BAndiT1983
currently i've implemented 2 algorithms to be able to preview (C)DNG files, but we also have got algorithms
12:53
BAndiT1983
from another team member
12:54
BAndiT1983
problem with libs is often the problematic usage, e.g. libraw, often also performance/speed etc.
12:54
BAndiT1983
so sticking to won implementations at the moment, which should be accelerated by uisng OpenCL, GLSL shader or similar
12:55
BAndiT1983
*own
12:55
g33kyaditya
Plus the lib could go out of maintenance
12:56
BAndiT1983
yes, this is also a problem
12:56
g33kyaditya
Nice. I've never really done this sort of thing before but would love to try my hands on it
12:57
BAndiT1983
if you have development experience, then you could try to build it
12:58
BAndiT1983
instructions can be found in repository -> https://github.com/apertus-open-source-cinema/opencine
12:58
g33kyaditya
Sure thing. Thanks. I'm setting up my build system as we speak :)
12:58
BAndiT1983
very good
12:59
g33kyaditya
Also, I wanted to ask, do I need to know a lot of Image Processing already? Or can I pick some stuff specific to the needs
13:01
BAndiT1983
i am also not an expert in image processing, but there is ProcessingTest sub-project in the source, which you can inspect to see some implementations, it's a sandbox to play around with OC things or test new features
13:02
g33kyaditya
Nice. I'll have a look
13:05
intracube_afk
changed nick to: intracube
13:12
jucar
joined the channel
13:36
ItsMeLenny
left the channel
13:42
jucar
left the channel
14:02
g33kyaditya
What is exactly the avformat package ? CMake gives me an error while making the project QtAV
14:02
BAndiT1983
changed nick to: BAndiT1983|away
14:50
RexOrCine
joined the channel
15:28
karan
Bertl_oO :ping
15:28
Bertl_oO
pong
15:30
karan
Bertl_oO :i gone through schematics. where i can get more details like which pmod is used
15:30
Bertl_oO
pmod?
15:31
karan
fan is connected to PMOD module right?
15:31
Bertl_oO
nope
15:31
Bertl_oO
the fan is connected to the main board
15:32
Bertl_oO
you might be looking at the wrong camera, i.e. AXIOM Alpha not Beta
15:32
Bertl_oO
back then, we had the fan connected to the PMOD (via some electronics)
15:33
karan
http://vserver.13thfloor.at/Stuff/AXIOM/BETA/BETA_interface.pdf
15:34
Bertl_oO
ah, yes, that was the original design we presented for the indigogo campaign
15:35
karan
then where is updated version:(
15:35
Bertl_oO
back then, the Beta design consisted of a single board with two shields and a uSD/FAN controller
15:36
Bertl_oO
there is no updated version available yet
15:36
Bertl_oO
check the schematic for the mainboard for details regarding the fan
15:37
Bertl_oO
http://vserver.13thfloor.at/Stuff/AXIOM/BETA/axiom_beta_main_board_v0.36_r1.2.sch
15:37
dimon_
joined the channel
15:40
karan
umm not getting what this has..
15:41
Bertl_oO
first page, bottom right D7/8
15:42
Bertl_oO
shows the speaker and right beside it D6 is the FAN connector
15:42
Bertl_oO
driven by N_IO0
15:43
Bertl_oO
that is connected to the RFE on page 6
15:43
Bertl_oO
pin 1 aka PL2C
15:55
intracube
changed nick to: intracube_afk
16:02
karan
sorry ,there is no page no.
16:05
Bertl_oO
the sheets are numbered
17:13
karan__
joined the channel
17:16
karan
left the channel
17:16
karan__
changed nick to: karan
17:20
karan
Bertl_oO: not able to follow you ,please elaborate
17:21
Bertl_oO
you got the schematic opened, yes?
17:21
karan
ya
17:21
karan
on web
17:22
Bertl_oO
hmm, that won't show you much I guess
17:22
Bertl_oO
I can make you a pdf though, if you don't want to install Eagle
17:22
karan
k
17:23
karan
is egle is needed furthur
17:25
Bertl_oO
not really, it just makes it easier to find stuff, because you can search for pins/points and highlight components
17:29
Bertl_oO
http://vserver.13thfloor.at/Stuff/AXIOM/BETA/axiom_beta_main_board_v0.36_r1.2.schematic.pdf
17:34
Bertl_oO
here you can find the code for the current 'pass through' setup for the MachXO2
17:34
Bertl_oO
http://vserver.13thfloor.at/Stuff/AXIOM/BETA/LATTICE/
17:34
Bertl_oO
note that this can be used to build designs with the Lattice Diamond tools
17:35
Bertl_oO
also note that for the PWM solution, this needs to be adapted to allow to control both, the PWM settings and the control lines
17:35
anuditverma
joined the channel
17:36
anuditverma
left the channel
17:37
Bertl_oO
one solution to that problem might be to use I2C over two wires and map the sensor registers to I2C
17:38
Bertl_oO
another solution could be to use some kind of serial packet protocol and handle both, I2C and SPI over that
17:38
anuditverma
joined the channel
17:41
Bertl_oO
give me a second to update that though, as we are using a newer version now
17:42
karan
left the channel
17:44
karan
joined the channel
17:46
Bertl_oO
okay, should now be all up to date
17:55
RexOrCine
left the channel
17:55
intracube_afk
changed nick to: intracube
18:07
anuditverma
left the channel
18:27
se6astian
changed nick to: se6astian|away
18:28
karan
Bertl_oO :do i need to use lattice diamond tool
18:55
karan
left the channel
19:10
venky18
joined the channel
19:17
venky18
left the channel
19:19
dimon_
left the channel
19:19
RexOrCine
joined the channel
19:24
RexOrCine
left the channel
19:29
anuditverma
joined the channel
19:39
niemand
joined the channel
20:05
Spirit532
left the channel
20:12
dimon_
joined the channel
20:20
anuditverma
Hi BAndiT1983, we were discussing yesterday about something need to discussed with maltefiala, I will be here for quite a while now, meanwhile I am following the updated task description on the idea page.
20:56
BAndiT1983
joined the channel
21:09
niemand
left the channel
21:13
fredy
left the channel
21:14
illwieckz
left the channel
21:14
intracube
left the channel
21:17
fredy
joined the channel
21:17
fredy
changed nick to: Guest60136
21:18
illwieckz
joined the channel
21:18
intracube
joined the channel
21:18
anuditverma
Hi, could anyone please help me out, I want to try  the example calls for changing parameters on the camera internal command line mentioned on the Wiki page, how can I test these calls without the camera h/w ? I want to know if there is a provision to test these out on a virtual setup or an emulator.
21:20
davidak
joined the channel
21:26
BAndiT1983
there are 2 tasks in the lab, which describe how to call QEMU
21:26
BAndiT1983
maltefiala also posted a link to more or less recent image of the camera firmware
21:27
BAndiT1983
we won't use this command line calls in the future, as the daemon, which i'm developing at the moment, will have better performance and access management
21:28
BAndiT1983
just take a look at the recent task for dameon, web interface and similar
21:30
anuditverma
Okay on it, thanks.
21:32
davidak
left the channel
21:35
davidak
joined the channel
22:21
Guest60136
changed nick to: fredy
22:21
fredy
changed nick to: Guest16894
22:33
Bertl_oO
anuditverma: besides the QEMU setup, there is a remote accessible camera setup (might not be available yet) ... check with se6astian|away for details when he is around
22:34
anuditverma
I will see for sure, thanks Bertl_oO
22:34
Bertl_oO
note that parameter changes on the camera basically boil down to one of three actions
22:34
Bertl_oO
1) writing to a register (memory location)
22:35
Bertl_oO
2) accessing an I2C or SPI device on one of the busses
22:35
Bertl_oO
3) modifying the FPGA firmware (by uploading new bits)
22:36
Bertl_oO
the first one can be simulated with any memory address, so no problem there
22:36
Bertl_oO
the second one is trickier in QEMU, it reqires some virtual QEMU devices in the right places (not there yet)
22:37
Bertl_oO
the third one is nothing which will have an effect on the QEMU anyway, so it can mostly be replaced by a dummy call/file
22:38
anuditverma
okay, so one of the three, as in it can handle only one process thread at a time ?
22:38
Bertl_oO
'it' being the daemon I presume?
22:38
anuditverma
yes
22:39
Bertl_oO
for the first step, a single threaded solution would suffice I guess
22:39
Bertl_oO
not sure multi threading (for access) is required
22:40
Bertl_oO
but there will be more than one connection to the daemon
22:40
anuditverma
okay
22:40
Bertl_oO
the basic idea is to collect information centrally and provide a 'buffered' version to clients
22:41
Bertl_oO
actions like changing settings should only be done in a synchronized way
22:48
dimon_
left the channel
22:49
davidak
left the channel
22:50
anuditverma
the camera firmware here that we are talking about, is it on docker ?
22:54
Bertl_oO
not as far as I know
22:57
BAndiT1983
multithreading need s arising automatically, as i've already preplanned it for sockets
22:57
BAndiT1983
testing currently, before i proceed to the protocol
22:58
BAndiT1983
why should it be used in docker? camera has to run as fast as possible without too much overhead
22:58
BAndiT1983
i have a feeling, that people jump nowadays on some hip technologies, but don't evaluate what stands behind it
22:58
dimon_
joined the channel
22:58
Bertl_oO
docker is not inherently slow, just bloated :)
22:59
BAndiT1983
exactly, i know that it's fast, but the bloat is bothering me
23:00
BAndiT1983
same problem with whole java application server thing, it's very annoying to wait for the application to deploy and then to get some error message which does not tell anything helpful
23:00
anuditverma
not too much of a docker fan, but I might take a step back from the discussion, please carry on O:-)
23:00
BAndiT1983
you should not step back, discuss instead, a every input is helpful
23:00
Bertl_oO
anuditverma: what I'd be interested in is, why do you ask?
23:00
anuditverma
BAndiT1983 +1 for java annoyance
23:02
anuditverma
I was curious if I can get hands on playing around with example calls for changing parameters on the camera, so that's why I asked
23:02
Bertl_oO
do you have an ARM system available?
23:03
BAndiT1983
raspberry pi should be sufficient for such playing around, in my opinion
23:03
BAndiT1983
at least to understand the basics
23:04
anuditverma
I do have a Raspberry Pi 2 Model B
23:04
anuditverma
it has ARMv7
23:05
Bertl_oO
yup
23:05
Bertl_oO
that would be pretty similar from the kernel/userspace side
23:05
Bertl_oO
we had a raspian running on the AXIOM Alpha
23:06
Bertl_oO
was replaced by Arch Linux for the Beta
23:06
BAndiT1983
archlinux for raspi is available, so it could be the test base for you, anuditverma
23:06
Bertl_oO
http://pastebin.com/raw/6XHBinDf
23:07
anuditverma
Bertl_oO I want to know why raspian was discarded ? any specific reason?
23:07
BAndiT1983
i don't remember where the a9 is fitting, iirc it is called armv7-a9
23:08
BAndiT1983
as armv8 is awaiting release soon, but i could be mistaken as i don't follo it closely
23:08
Bertl_oO
detailes for the ARM cores can be found when searching for Zynq SoC on the Xilinx site
23:08
BAndiT1983
raspbian was too bloated possibly
23:08
BAndiT1983
arch is much slimmer
23:08
Bertl_oO
anuditverma: we dropped raspian, because it is not a raspberry pi after all
23:08
anuditverma
I recently upgraded my Pi to Jessie Pixel version
23:08
anuditverma
ohh okay
23:09
anuditverma
ok ok cool
23:09
Bertl_oO
the zynq is on a MicroZed and Xilinx also has a 'custom' Linux variant PetaLinux? for it
23:09
BAndiT1983
never used UI on raspi, as ssh was sufficient to get the job done
23:10
BAndiT1983
don't know if petalinux has all the packages we need
23:11
Bertl_oO
but as Arch has worked well, is pretty small and nicely supported on ARM we went for that
23:12
BAndiT1983
so, if i get my raspi and arch to build the daemon, would it also run on microzed?
23:12
Bertl_oO
pretty sure, you might need to adjust the memory locations though
23:13
BAndiT1983
not a big problem, it was just about getting it that way or using cross compiler, but there is also the thing with checking which features are available in the compiler
23:15
BAndiT1983
prior to that the structure has to be there, afterwards i can adjust it, if the build fails because i mixed C and C++ too much
23:31
anuditverma
it's a bit late here, I might come up with more queries tomorrow and try to gasp more knowledge about the tech :) , meanwhile I will read more links/resources mentioned on the idea page and I will see if I can get the latest archlinux running on the pi
23:33
BAndiT1983
have fun :)
23:38
dimon_
left the channel
23:38
anuditverma
left the channel
23:59
BAndiT1983
left the channel