Current Server Time: 09:48 (Central Europe)

#apertus IRC Channel Logs

2013/11/06

Timezone: UTC


01:35
[1]Sasha
joined the channel
03:16
troy_s_
joined the channel
04:15
Bertl
off to bed now ... have a good one everyone!
05:06
[1]Sasha
left the channel
06:48
PhilippeJ
joined the channel
08:04
dmj_nova1
wow that meeting is early
08:23
Sasha_C
joined the channel
09:41
PhilippeJ
Bertl, Hello !, I got a message form Gabe, he will send me what he curently has in term of dng conversion code (python script)
10:47
FergusL
cool ! could I see it ? if he's ok
11:19
Bertl
morning everyone!
11:19
Sasha_C
Morning Bertl
11:40
troy_s_
changed nick to: troy_s
11:42
troy_s
FergusL: Answering white balance... if you think of a device referred image as a volume of color points floating in space, the transform (either XYZ scaling, Von Kries, or Bradford) transforms the original point to the new point - stretching / bending the volume such that the values represent the color under the new white.
11:44
troy_s
FergusL: It isn't quite this simple when we get into the psychophysical domain as phenomenon happen. Lower luminance levels for example, hand off the color receptors to the monochromatic ones (so your eyes see less to no color at lower luminance) and
11:44
troy_s
FergusL: ...metamerism phenomenon where your brain makes up the color seen.
11:44
troy_s
etc.
11:45
PhilippeJ
left the channel
11:45
troy_s
eg FergusL - http://wiki.blender.org/index.php/File:Orange.gif
11:45
troy_s
Local contrast makes your brain interpret identical values differently
11:46
troy_s
(tied to how your brain interprets the scene)
11:47
Bertl
well, isn't that more like this one: http://www.businessballs.com/images/shadow-illusion.jpg
11:47
troy_s
and this http://wiki.blender.org/uploads/b/b2/Grey-cube-color-squares.jpg
11:47
troy_s
where the blue squares on left and yellow squares on right are actually achromatic values
11:48
troy_s
http://wiki.blender.org/uploads/thumb/2/2b/Grey-cube-squares.jpg/800px-Grey-cube-squares.jpg
11:48
troy_s
Bertl: Yes indeed. Different phenomenon however I believe
11:49
troy_s
Bertl: and culture influences perception as well (see Herskavits et al)
11:49
troy_s
I believe Mark Fairchild lists all the psychophysical phenomenon in his CIECAM02 paper.
11:53
Bertl
in the color squares, the funny part is that the yellow and blue squares are actually the _same_ gray :)
11:53
PhilippeJ
joined the channel
12:00
FergusL
http://wiki.blender.org/index.php/User:Sobotka/Color_Management the cyan/green "confusion" is impressive there
12:03
troy_s
left the channel
12:04
troy_s
joined the channel
12:05
FergusL
troy_s: "device referred image" == still bound to the limits of the device space ?
12:08
FergusL
I think I get it about white balance it's again a transform
12:58
troy_s
FergusL: Display / Device / Output referred - 0 to max (eg sRGB display)
12:59
troy_s
FergusL: Scene referred = 0 to theoretical infinity, linear relation.
12:59
FergusL
yes
13:00
FergusL
shall I think about white balance in an image as : One white reference point (moved around in XYZ with transforms) and all other values in the image relative to this reference ?
13:31
troy_s
FergusL: More or less a good summary from the data side.
13:32
troy_s
(and remember, not all transforms for white point are equal - Bradford etc.)
13:39
FergusL
troy_s: in what aren't they equal ? quality ? or simply the result ?
13:40
PhilippeJ
left the channel
13:55
troy_s
FergusL: 'quality' of the transform.
13:55
FergusL
ok
13:56
troy_s
Bertl: And to answer your earlier question: I believe profiling the sensor is a prudent choice to at least get a sense of the color gamut and latitude. I wouldn't publish the results, but it is important stuffs for a camera aimed at cinema crafting.
13:57
FergusL
why wouldn't you publish the results ?
13:57
troy_s
Bertl: The folks that have a hope of being interested in such a device will be deadly interested in the color gamut and latitude. (Namely director of photography archetypes.)
13:58
troy_s
FergusL: Because it is too early and likely not 100% optimal / controlled. It would be a great point of reference for development however.
13:58
Bertl
well, I don't think there is any reason to hide or hold back information in this regard
13:59
FergusL
troy_s: but releasing double checked info with the proper tool later is a good thing, right ? that's what I would do
13:59
Bertl
we are not producing/selling the sensor, and if we find out that this specific sensor is not suitable, we simply switch to a different one
13:59
FergusL
^ good to hear
13:59
troy_s
Bertl: I wouldn't suggest that either. Merely provide ample caveats and not offer an "official" publish of the results. A simple "additional information".
13:59
Bertl
and if somebody misinterprets development results as 'that's how the final image will look like' then they should stay away from development :)
14:00
troy_s
And exactly - it would be one more piece of information to see if the sensor is suitable for imaging.
14:00
Bertl
we are currently testing with a 'known broken' sensor :)
14:01
Bertl
still your 'answer' didn't really answer my question
14:01
Bertl
(which basically was: is it worth to do stills of the IT8 chart with known insufficient lighting)
14:03
Bertl
we will, for sure, make those images once the prototype is mobile and can be used by sebastian and friends for actual testing
14:04
troy_s
Bertl: I think a consistent tungsten image at a few white balance points (tickle registers for this?) would be a magnitude better than where it currently is.
14:05
Bertl
regarding publishing findings: I don't think there is any point for us to fund and build a camera which doesn't do what we want :)
14:06
Bertl
so if the sensor is insufficient or has bad color rendition, it helps to know that early, because we can save the time and spend it on something more productive
14:07
Sasha_C
left the channel
14:17
Bertl
off for now ... bbl
15:04
philippej
joined the channel
15:16
se6astian
joined the channel
15:18
philippej
left the channel
15:19
se6astian
good evening
15:23
FergusL
hello se6astian
15:38
philippej
joined the channel
15:45
se6astian
I just sent the purchase order for the sensor to Cmosis
16:01
se6astian
I scheduled a team meeting for Monday the 11th at 12 PM GMT here on IRC, please join us at the time
16:02
se6astian
the association general assembly will start 2 hours later at 2PM
16:03
se6astian
sorry the meeting is at 12 AM GMT (noon)
16:03
philippej
left the channel
16:20
FergusL
se6astian: a working one for Bertl ?
16:23
se6astian
yes
17:01
rexbron
joined the channel
17:02
rexbron
sup troy_s
17:06
rexbron
You guys got some coverage over on redshark
17:11
se6astian
hi rexbron, yes three articles in the last month :)
17:49
troy_s
rexbron: !
17:50
troy_s
Bertl: Exactly.
18:02
Bertl
back now ...
18:26
troy_s
left the channel
18:26
rexbron
left the channel
19:24
Bertl
off for a nap ... bbl
19:47
aombk
joined the channel
20:31
aombk
left the channel
21:15
se6astian
time for bed
21:15
se6astian
goog night!
21:15
se6astian
left the channel
21:45
Sasha_C
joined the channel