Current Server Time: 11:52 (Central Europe)

#apertus IRC Channel Logs

2019/02/17

Timezone: UTC


23:06
seaman
joined the channel
23:16
seaman
left the channel
00:21
Bertl_oO
off to bed now ... have a good one everyone!
00:21
Bertl_oO
changed nick to: Bertl_zZ
01:54
RexOrCine
changed nick to: RexOrCine|away
01:56
RexOrCine|away
changed nick to: RexOrCine
02:08
RexOrCine
changed nick to: RexOrCine|away
02:09
RexOrCine|away
changed nick to: RexOrCine
05:58
Cyna_
joined the channel
06:53
Bertl_zZ
changed nick to: Bertl
06:53
Bertl
morning folks!
07:01
siddhantsahu
Morning Bertl
07:02
siddhantsahu
Just wandering it is a good idea to build a open source version of CalMan software by SpetraCal?
07:03
siddhantsahu
https://calman.spectracal.com
07:04
siddhantsahu
Understand AXIOM is primary focus but still wandering.
07:23
niemand
joined the channel
07:32
Bertl
not sure what the CalMan software does ... please elaborate
07:35
siddhantsahu
Help to calibrate any display so that there output matches with that of international standards (Rec. 709, DCI P3, sRGB, AdobeRGB etc) but generating 1D, 3D LUTs and ICC profiles.
07:36
siddhantsahu
by* generating
07:36
siddhantsahu
Very crucial tools for filmmakers.
07:37
siddhantsahu
The CalMan software requires a hardware colorimeter to measure the luminance and chromaticity of the display's output and compares it with industry standards.
07:39
Bertl
https://iorodeo.com/pages/colorimeter-project
07:39
siddhantsahu
and generates 1D, 3D LUT and ICC profiles which can be applied to any application like (OpenCine) or can be loaded to monitor directly if it supports 3D LUT
07:39
Bertl
https://displaycal.net/
07:40
Bertl
https://www.linuxlinks.com/monitorcalibration/
07:40
Bertl
https://hughski.com/
07:42
siddhantsahu
Ohh didn't know so many open source alternatives are available
07:43
BAndiT1983|away
changed nick to: BAndiT1983
07:44
Bertl
search engines are a great source of information nowadays :)
07:54
se6astian|away
changed nick to: se6astian
08:12
se6astian
good day
08:18
supragya_
joined the channel
08:28
BAndiT1983
changed nick to: BAndiT1983|away
08:42
johnathan
joined the channel
08:45
Alpha
joined the channel
08:46
Alpha
Hi!
08:46
Alpha
Thanks a lot BAndiT1983
08:46
Alpha
I have one more question
08:47
Alpha
In the Challenge it was told "Output the intensity values of the first 5x5 (square tile)"
08:47
Alpha
What do you guys mean by tile
08:47
Alpha
tile= 1byte? tile=1.5 byte?
08:48
Alpha
Or I misunderstood something.
08:48
se6astian
5x5 pixels
08:50
Alpha
@se6astion, Good morning Sir. Thanks a lot for clearing my doubt.
08:51
Alpha
@supragya_ @johnathan, Good Morning Sir.
08:51
Alpha
@BAndiT1983 Thanks a lot to you too Sir.
08:53
Alpha
left the channel
08:55
Bertl
hey johnathan! how's it going?
08:57
supragya_
hi Alpha
09:07
BAndiT1983|away
changed nick to: BAndiT1983
09:19
johnathan
hey actually very well, was just working on stuff
09:21
Bertl
glad to hear! I read that you have/had some register specific questions?
09:22
supragya_
left the channel
09:36
johnathan
yes I have them figured out via mostly trial and error, but wondering if you could shed some light on what the various registers in gen_init / rmem_conf?
09:38
johnathan
left the channel
09:43
johnathan
joined the channel
09:47
RexOrCine
changed nick to: RexOrCine|away
09:50
johnathan
in any event I have it all working, I'm able to display my structured light patterns via HDMI @ 120 hz. https://vgy.me/pXWRyh.jpg I'm now working on sync the patterns with the actual capture of the images.
09:52
johnathan
here are the settings I ended up coming up with for my HDMI output device: https://pastebin.com/JJ32heEz
09:55
johnathan
I tweaked a lot of them manually, it was somewhat of an analog process and I've used some of the mimg code to be able to write my patterns to memory. I noticed if I display pure white patterns 0xfff there's a lot of flickering (maybe some kind of clipping going on?) but it works fine with 0xccc
09:55
Bertl
interesting, your output is 1:1.1 or so?
09:56
johnathan
"912x1140" 146.000 912 920 984 1040 1140 1150 1156 1170 +hsync +vsync is the modeline
09:56
Bertl
you need to be aware that the output is taken from several buffers in a round robin way (i.e. think circular buffer)
09:57
Bertl
so you want/need to fill all buffers with the same image for testing, otherwise you get a quick sequence of 3 or 4 images
09:57
Bertl
this is especially important for non-image data like the overlay
09:57
johnathan
I hacked the mimg code (where its generating test patterns), I assume that is doing that? in writeline and such?
09:58
johnathan
basically the pattern you see above is done via mimg
09:58
Bertl
mimg has an -a option to fill all buffers at once
09:59
johnathan
yep that's what I'm using, I removed the mask stuff though
09:59
Bertl
well, the memory layout is a little bit tricky and you might end up with a blinking overlay :)
10:00
Bertl
i.e. in each 64bit word you have 4x12bit of color information and 16bit 'overlay'
10:00
Bertl
and the overlay has a bunch of special values which end up with blinking output etc
10:00
Bertl
so you normally want to keep the overlay at zero for normal image output (hence the masking code)
10:00
johnathan
what is the 4th channel in the color? alpha?
10:02
Bertl
the color is generated from bayer pattern so the four values are RG/GB where RG comes from the first (even) line and GB from the second (odd) line
10:03
Bertl
those four values are then mixed together with a color matrix resulting in the output RGB pixel with 8 bit depth
10:03
johnathan
ah gotcha
10:04
johnathan
I noticed when using cmv_snap it outputs the image to the HDMI, is there any way to avoid that? i.e. its controlled via software right?
10:04
Bertl
the two pipelines are basically independent except for a buffer synchronization
10:05
Bertl
the reason why the image ends up being displayed is because the buffers are currently configured to be the same
10:06
johnathan
ok cool, what I basically need to do is sync exposure/captures and different patterns, but haven't dug into it yet
10:06
Bertl
i.e. cmv_snap fills one of the buffers with the acquired image, the synchronization code sends a signal to the output pipeline which switches to the next buffer
10:35
johnathan
left the channel
10:42
Bertl
off for now ... bbl
10:42
Bertl
changed nick to: Bertl_oO
10:43
Y_G
joined the channel
10:43
RexOrCine|away
changed nick to: RexOrCine
10:51
BAndiT1983
changed nick to: BAndiT1983|away
11:06
Y_G
Hi, could anyone check if this is the image we are aiming for after bilinear debayering,wrt T872?
11:06
Y_G
https://paste.pics/9befa517a30cc972840cd0157267a36c
11:08
anuejn
yup, seems to be correct :)
11:09
anuejn
(see anuejn.github.io/batic )
11:13
Y_G
left the channel
11:28
Y_G
joined the channel
11:29
Y_G
@anuejn Thanks
11:32
RexOrCine
changed nick to: RexOrCine|away
11:45
se6astian
Y_G: yes that looks good
13:36
Y_G
Hi se6astian
13:37
Y_G
Could you help me with what should be the resolution of the extracted channel images,should it be (2048*1536) or (4096*3072) ?
13:42
Y_G
wrt T872^
13:51
se6astian
each of the 4 color channels has only a quarter of the total pixels in the image if that is what you mean
14:02
Y_G
So when I save each channel as an image (pgm),should I just save the channel directly(4096*3072/4) or expand the image (4096*3072)
14:02
Y_G
Sorry ,if not clear
14:08
Dev_
joined the channel
14:11
Dev_
Hello to all , se6astian : even i am getting weird output when i am storing r , g1 , b ,g2 as .ppm file . I am getting three images instead of one
14:12
Dev_
This is the output -> https://paste.pics/3291be0b865e1019c9dd1affcd28c336
14:13
Dev_
Of red channel which i have extracted , i remember when i did with lodepng , it was giving single image ,
14:14
Dev_
I have set 2048 * 1537 as height and width for .ppm file , every time change the parameter , it gives either more images or less
14:15
Dev_
Any suggestions please
14:16
Bertl_oO
correct the code you are using?
14:18
se6astian
Y_G: no interpolation of the channels is required, original resolution is fine
14:19
Dev_
left the channel
14:20
Dev_
joined the channel
14:27
Dev_
Bertl_oO : change what ?? , Height and width , can u elaborate please
14:27
Dev_
left the channel
14:28
Bertl_oO
I presume you have written some code to split up the raw image into the channels ...
14:29
Dev_
joined the channel
14:29
Bertl_oO
each channel is 2048x1536 in size, so if a PPM image with this size shows more than one image, there is an error in the code splitting up the raw image
14:30
Dev_
Yes but i giesg that is fine as when i used lodepng library , it was giving fine output
14:30
Dev_
Guess*
14:31
Bertl_oO
so what is the difference now?
14:31
Bertl_oO
i.e. what changed?
14:36
Dev_
First i have stored the each channel in an array butfer , when i used lodepng by passing buffer and 2048 * 1537 as parameters it was giving single image
14:37
Bertl_oO
sounds good, and now?
14:37
Dev_
This time i am writing a .ppm from scratch , by setting same dimensions but it is giving three images
14:37
Dev_
Please see -> https://paste.pics/3291be0b865e1019c9dd1affcd28c336
14:38
Bertl_oO
'from scratch' means?
14:39
Dev_
When i writing it like this -> P6( space) height( space) width( space ) intensity (255) then buffer data
14:40
Dev_
I am *
14:40
Bertl_oO
so you are using a library for that as well, right?
14:40
Dev_
No , just writing in binary file
14:41
Dev_
No external library as specified in task t782
14:43
Bertl_oO
and the buffer is the data for one channel, yes?
14:45
Dev_
Yes
14:45
Bertl_oO
well, there is your problem/error
14:45
Bertl_oO
PPM is a color image format and thus contains RGB values not just a single channel
14:46
Bertl_oO
which means that you get 'three' small images instead of one
14:46
Y_G
@Dev try using pgm format
14:46
Bertl_oO
write a PGM image instead and you should be fine
14:49
Dev_
Okay , I got it , thanks bertl_oO an Y_G :) , I will try it with pgm
14:51
Y_G
@Se6astian Thanks,Could anyone check if this is the extracted channel image we are looking for wrt T872
14:51
Y_G
https://paste.pics/70bfd4a6006b22c12ff2acc7cdfc54fa
14:54
se6astian
yes that looks like one of the channels indeed
14:57
Bertl_oO
well, hard to tell as the image is scaled down ...
14:57
Bertl_oO
but it looks a lot better
14:58
Dev_
left the channel
15:04
Y_G
Thanks @Se6astian and @Bertl_o0 for your help,I'll try other improving it :)
15:04
Bertl_oO
you're welcome!
15:05
Bertl_oO
as a general piece of advise here: as it seems there are a number of people interested in working on OC, but we only reserve a single slot for it this year
15:06
Bertl_oO
you might really want to provide an excellent challenge task solution with lots of extra
15:11
Y_G
I'll try for sure :)
15:14
ASHU
joined the channel
15:18
Dev_
joined the channel
15:22
Dev_
yes , i understand bert_oO :)
15:22
Bertl_oO
you might also want to get a proper IRC client instead of using the web interface :)
15:22
Dev_
Bertl_oO
15:23
Dev_
Yes yes sure , :)
15:23
Dev_
left the channel
15:28
ASHU
left the channel
15:31
sketch_
joined the channel
15:31
sketch_
hallo
15:32
Bertl_oO
hello sketch_!
15:40
sketch_
habe heute nach einer kleinen raw video camera, für kürze sessions gesucht, fergeblich... dann gab ich in die suchmaschine opensource camera ein und entdeckte euch :D
15:44
se6astian
wunderbar :)
15:45
se6astian
was meinst du mit "für kürze sessions" ?
15:50
sketch_
für highspeed run sassions, auf bergen wäldern, hochhäusern und so weiter.
15:52
Y_G
left the channel
15:53
sketch_
am besten zum selber bauen
15:54
se6astian
hört sich spannend an aber kann ich mir noch nicht ganz vorstellen, wir haben keine "action cam" in dem sinn..
15:55
comradekingu
left the channel
15:56
sketch_
ich weiss... aber mit 16mm sensor camera könnte man vielleicht schon was anfangen
16:06
BAndiT1983|away
changed nick to: BAndiT1983
16:12
sketch_
left the channel
17:35
BAndiT1983
changed nick to: BAndiT1983|away
19:18
comradekingu
joined the channel
19:23
BAndiT1983|away
changed nick to: BAndiT1983
20:11
Cyna_
left the channel
20:25
niemand
left the channel
20:27
BAndiT1983
changed nick to: BAndiT1983|away
20:38
BAndiT1983|away
changed nick to: BAndiT1983
21:09
se6astian
changed nick to: se6astian|away
21:24
BAndiT1983
changed nick to: BAndiT1983|away
21:24
BAndiT1983|away
changed nick to: BAndiT1983
22:47
Bertl_oO
off to bed now ... have a good one everyone!
22:47
Bertl_oO
changed nick to: Bertl_zZ