Current Server Time: 23:31 (Central Europe)

#apertus IRC Channel Logs

2017/03/30

Timezone: UTC


01:28
xeechou
changed nick to: xeechou|afk
02:13
arpu
left the channel
02:29
arpu
joined the channel
02:38
Spirit532_
left the channel
03:32
jucar
left the channel
03:33
jucar
joined the channel
03:46
intracube
changed nick to: intracube_afk
03:58
jucar
left the channel
04:52
Bertl_oO
off to bed now ... have a good one everyone!
04:52
Bertl_oO
changed nick to: Bertl_zZ
05:11
RexOrCine
joined the channel
05:15
RL
left the channel
06:22
Spirit532
joined the channel
06:25
tusle
left the channel
06:54
se6astian|away
changed nick to: se6astian
07:10
usmankhan
joined the channel
07:13
se6astian
good morning
07:13
usmankhan
left the channel
07:47
pusle
joined the channel
08:02
Bertl_zZ
changed nick to: Bertl
08:03
Bertl
morning folks!
08:08
niemand
joined the channel
08:08
niemand
left the channel
08:08
niemand
joined the channel
08:59
mithro
se6astian: morning
08:59
mithro
hey Bertl
09:05
se6astian
hi!
09:14
Bertl
hey mithro! how's going?
09:16
niemand
left the channel
09:24
derWalter
left the channel
09:24
se6astian[m]
left the channel
09:24
MatrixTraveler[m
left the channel
09:24
RexOrCine[m]
left the channel
09:26
derWalter
joined the channel
09:28
RexOrCine[m]
joined the channel
09:41
MatrixTraveler[m
joined the channel
09:46
se6astian[m]
joined the channel
09:55
niemand
joined the channel
10:56
niemand
left the channel
11:39
niemand
joined the channel
11:55
niemand
left the channel
12:19
sagnikbasu95
joined the channel
12:20
sagnikbasu95
Hi Bertl..
12:21
Bertl
hey sagnikbasu95! how's going?
12:25
sagnikbasu95
everything's gr8..just finished our mid-term exams.. so for the sobel filter project...how will it be if there is a top module where all the other module will be instantiated ? this top module will be like a FSM-D structure i.e. based on the input signals received from the user or fpga and the current state..it wll take the appropriate actions
12:26
sagnikbasu95
Like how many rows and coloumns the kernel will have
12:26
Bertl
for testing and similar your testbench will be the top module
12:27
Bertl
for actual use of the filter, the filter will be instantiated in the top module of the camera firmware
12:32
sagnikbasu95
ok..so during the GSoC project..will I be allowed to use to make changes in the camera firmware according to the HDL modules..or I will be only restricted to HDL coding only?
12:33
Bertl
the camera firmware is open source as well, so for sure you can use that and make changes accordingly
12:34
Bertl
we will also provide at least remote access to an AXIOM Beta so you can test the filter on real hardware
13:54
sagnikbasu95
left the channel
14:16
niemand
joined the channel
14:16
niemand
left the channel
14:16
niemand
joined the channel
14:19
jucar
joined the channel
14:26
Bertl
off for a nap ... bbl
14:26
Bertl
changed nick to: Bertl_zZ
14:37
se6astian
gotta go
14:38
se6astian
changed nick to: se6astian|away
14:42
niemand
left the channel
14:50
jucar
left the channel
15:27
se6astian|away
changed nick to: se6astian
16:13
Tanos
joined the channel
16:15
niemand
joined the channel
16:15
niemand
left the channel
16:15
niemand
joined the channel
16:43
Tanos
left the channel
17:09
MichaelH
joined the channel
17:18
intracube_afk
changed nick to: intracube
17:42
Roopal08
joined the channel
17:43
Roopal08
Hey <BAndiT1983>, need to confirm where i m interpreting things in a correct manner or not . I will tell you what i tried .
17:44
Roopal08
In file ProcessingPresenter.cpp , when i commented SHOODAKDebayer* debayer = new SHOODAKDebayer(*_image.get()); debayer->Process(); OC_LOG_INFO("Demosaicing finished");
17:46
Roopal08
and try to view the image , https://drive.google.com/file/d/0B0Xb3WJpS5dFcWhQX2YxM2d1dnc/view?usp=sharing I get this image.
17:47
Roopal08
I should have green and blue components too , which are not coming. So i guess there is some issue in provider->load
17:48
Roopal08
even if i dont comment shoodakdebayer and then try some combinations in BilinearDebayer.h
17:49
Roopal08
in lines redChannel = static_cast<unsigned short*>(image.RedChannel()); greenChannel = static_cast<unsigned short*>(image.GreenChannel()); blueChannel = static_cast<unsigned short*>(image.BlueChannel());
17:52
Roopal08
so by sending redChannel = static_cast<unsigned short*>(image.GreenChannel()); greenChannel = static_cast<unsigned short*>(image.GreenChannel()); blueChannel = static_cast<unsigned short*>(image.GreenChannel());
17:52
Roopal08
I am getting https://drive.google.com/file/d/0B0Xb3WJpS5dFcWhQX2YxM2d1dnc/view?usp=sharing
17:53
Roopal08
this image . And when I try redChannel = static_cast<unsigned short*>(image.BlueChannel()); greenChannel = static_cast<unsigned short*>(image.GreenChannel()); blueChannel = static_cast<unsigned short*>(image.RedChannel());
17:54
Roopal08
I dont get any imge displayed . I should get something right, R and B should behave same way.
17:55
BAndiT1983
joined the channel
17:56
BAndiT1983
Roopal08: have you checked while debugging how the memory at the location of green and blue looks like?
17:56
Roopal08
I have checked what pattern is coming - its 21122112
17:56
Roopal08
its GBRG
17:57
BAndiT1983
usually you can see what the loaders sets, all the infos are there, also pattern
17:58
Roopal08
I tried to see that info on matlab
17:58
BAndiT1983
try to press the button,s there was some small bug so the combinations were not alway working
17:58
BAndiT1983
why matlab?
17:59
BAndiT1983
the loader works as i'M using the same thinng in ocbackup, no need to verify something that is working
17:59
Roopal08
I am very much comfortable on it , so was trying to see basic info on it
17:59
BAndiT1983
https://lab.apertus.org/F1755
17:59
mithro
left the channel
17:59
BAndiT1983
this is actual use of the modules
18:00
BAndiT1983
so the loading works, but maybe there is something deactivated in processingtest, didn't have time for that, my daily job requires more time currently
18:00
Roopal08
I will do that , if i get a rough idea where to look into
18:01
BAndiT1983
just stop the debugger after arrays were filled, line 64 in processingpresenter
18:01
BAndiT1983
then take a look at the memory
18:02
BAndiT1983
i can check it tomorrow as other aperus things need my attention today
18:02
Roopal08
and i have one more doubt . In ProcessingPresenter.cpp .
18:02
Roopal08
Instead of interleavedArray[i * 3] = (redArray[i] >> 4) * 1.0 ; interleavedArray[(i * 3) + 1] = (greenArray[i] >> 4) * 1.0; interleavedArray[(i * 3) + 2] = (blueArray[i] >> 4) * 1.0;
18:02
Roopal08
I use interleavedArray[i * 3] = 0; interleavedArray[(i * 3) + 1] = 0; interleavedArray[(i * 3) + 2] = 0;
18:03
Roopal08
there is no effect on displayed image .
18:03
Roopal08
which means it is displaying before _view->SetThumbnail(_image->Width(), _image->Height(), interleavedArray);
18:04
BAndiT1983
this is processing image for Qt control, image should be shown 2 times, opengl and qt one, so you have to take a look on the setframe, line 52
18:05
Roopal08
Okay .
18:05
BAndiT1983
it should look similar to this -> https://lab.apertus.org/F1511
18:06
BAndiT1983
other image is broken somehow, this one should work -> https://lab.apertus.org/F1511
18:06
BAndiT1983
sorry, wrong link, now -> https://lab.apertus.org/F1512
18:07
BAndiT1983
top one shown through opengl shader with 12 or 16bit per channel if i rememebr correctly, bottom one converted to 8 bit for qt control
18:08
Roopal08
my exceutable file looks like https://drive.google.com/file/d/0B0Xb3WJpS5dFbmE1cDZRT2l0eGs/view?usp=sharing
18:08
Roopal08
and the area below one is for vedio right ?
18:09
BAndiT1983
have you tried to press buttons at the bottom several times and different combinations?
18:09
Roopal08
yeah . I tried those RGB buttons
18:10
Roopal08
and play puse button worked for vedio
18:10
BAndiT1983
they usually work, as the image is showing then take a look in processingview.cpp where it'S prepared for displaying
18:10
BAndiT1983
video is not included in processingtest, as it was added much later
18:11
BAndiT1983
playback controls were only adde for theme tests, actually working ones are in ocbackup
18:11
Roopal08
I gave a .mov file . added a path for it where it was needed and i could see the vedio running properly
18:11
Roopal08
and buttons also worked for it
18:11
BAndiT1983
in processingtest?
18:13
Roopal08
No , while excecuting it was saying it is not able to get media . and it was searching /media/andi/... i have a folder in mine to and added .mov there
18:14
BAndiT1983
of course it said that, as i have created an ISO file with video files for testing and the path was hardcoded for now
18:14
Roopal08
Yeah .
18:15
BAndiT1983
was jsut wondering about buttons, but i think that it works as i added them to the control in the OCui module, but i'M wondering about the video
18:15
Roopal08
play , pause , go to extreme start were working for vedio
18:15
Roopal08
those i checked properly
18:16
se6astian
changed nick to: se6astian|away
18:16
BAndiT1983
alright, will check it tomorrow, as i was only working on ocbackup lately
18:16
Roopal08
okay . So where should i start debugging now ?
18:17
BAndiT1983
at setframe, line 52 in processing presenter
18:17
BAndiT1983
data should be a available at that point
18:17
BAndiT1983
setframe tells view which image to display
18:19
BAndiT1983
in processingview is paintevent() used for drawing the frame
18:20
BAndiT1983
which graphics card do you have?
18:21
MichaelH
left the channel
18:26
Roopal08
I have Intel Corporation Haswell-ULT Integrated Graphics Controller (rev 0b)
18:28
BAndiT1983
hm, you should ensure which opengl version is supported
18:28
BAndiT1983
try glxinfo | grep OpenGL
18:29
BAndiT1983
it could be that other channels are not loaded because of problems with driver or unsupported extensions
18:30
Roopal08
OpenGL vendor string: Intel Open Source Technology Center OpenGL renderer string: Mesa DRI Intel(R) Haswell Mobile OpenGL core profile version string: 3.3 (Core Profile) Mesa 12.0.6 OpenGL core profile shading language version string: 3.30 OpenGL core profile context flags: (none) OpenGL core profile profile mask: core profile OpenGL core profile extensions: OpenGL version string: 3.0 Mesa 12.0.6 OpenGL shading language version stri
18:30
BAndiT1983
another location is ProcessingTest/Controls/PreviewPane.cpp
18:30
BAndiT1983
opengl 3.3 should be sufficient
18:31
BAndiT1983
preview pane initializes opengl and setups textures and so on, not very clean, but for tests it was sufficient
18:32
Roopal08
yeah , Saw that .
18:46
Roopal08
I was even working on my proposal for GSOC . So while making my timeline , I have few doubts .
18:47
Roopal08
As in what exactly is expected . I have to implement SHOODAK , and other algorithms on de-Bayering
18:48
BAndiT1983
generally yes, it's also about acceleration through gpu shader or opencl or arrayfire
18:49
Roopal08
demosaicing and possibly color-grading - multi-threaded solution for it too on OpenMP
18:50
BAndiT1983
color-grading is not important at the moment, but if there is some time left it could be a sub-task
18:50
Roopal08
Okay.
18:51
Roopal08
Will finish up the inital draft of proposal by tomorrow , and share it with you for reviewing.
18:55
BAndiT1983
okay, will look at it on weekend
19:00
Roopal08
cool
19:16
Roopal08
left the channel
19:17
BAndiT1983
left the channel
19:18
se6astian|away
changed nick to: se6astian
19:55
mithro
joined the channel
20:06
Bertl_zZ
changed nick to: Bertl
20:06
Bertl
back now ...
21:01
niemand
left the channel
21:19
se6astian
off to bed
21:19
se6astian
good night
21:19
se6astian
changed nick to: se6astian|away
21:19
Bertl
nn