Current Server Time: 02:54 (Central Europe)

#apertus IRC Channel Logs

2018/01/29

Timezone: UTC


00:26
seku
left the channel
01:30
Bertl_oO
off to bed.
01:31
Bertl_oO
changed nick to: Bertl_zZ
02:06
rton
joined the channel
02:24
rton
left the channel
06:43
pusle
left the channel
06:50
pusle
joined the channel
07:35
LordVan
joined the channel
07:48
seku_
joined the channel
08:27
Bertl_zZ
changed nick to: Bertl
08:27
Bertl
morning folks!
09:00
rton
joined the channel
10:10
se6astian|away
changed nick to: se6astian
10:51
Guillaume_Chooks
joined the channel
11:15
Bertl
changed nick to: Bertl_oO
11:41
joy_
joined the channel
11:44
joy_
hello, I will be doing the 3rd task of VHDL challenge that is high speed link...just to make sure what to do: I need to make a data serialiser and deserialiser and check by PRNGs
11:45
joy_
Can anyone please explain what is link training and how to implement them.
11:46
joy__
joined the channel
11:56
joy_
left the channel
11:59
joy_
joined the channel
12:02
joy__
left the channel
12:08
joy_
left the channel
14:49
seku_
left the channel
14:52
seku_
joined the channel
15:01
seku_
left the channel
15:29
LordVan
left the channel
15:40
seku_
joined the channel
15:49
anuejn
i build a small tool for playing with pixelshaders for debayering beta footage: https://anuejn.github.io/batic/ feel free to hack :)
15:51
anuejn
just hack on the code. sliders are created automatically and the preview gets rerendered on the fly
15:51
Rex0r
Link leads to a blank page.
15:55
anuejn
it just loads a big (24M) raw16 page
15:55
anuejn
*image. that takes quite some time
15:56
Bertl_oO
works here
15:57
Bertl_oO
very interesting moire effects when you zoom in/out though
15:58
anuejn
yeah thats related to the way, the browser scales the canvas. wasnt able to fix this yet
15:59
Bertl_oO
maybe only allow specific magnifications?
16:00
anuejn
mmh. meh.. maybe
16:07
Bertl_oO
hmm, why do I also get a pattern when I zoom in to the pixel level?
16:11
anuejn
i dont no either
16:11
anuejn
but thats really bad
16:11
Bertl_oO
I think there is a problem with the order of operations
16:12
Bertl_oO
i.e. the scaling for the zoom should apply after a fixed size debayering
16:12
Bertl_oO
but actually it seems to be the other way round, i.e. the scaling controls which bayer 'sensel' are sampled
16:13
Bertl_oO
note that the scaling/zoom is not something you need in a real world application except for specific transformations
16:14
Bertl_oO
e.g. going from 2160p 'raw' to 1080p 'debayered'
16:14
anuejn
the scaling is done after the debayering
16:14
anuejn
everython is rendered to a fixed size opengl canvas and than this is scaled by the browser
16:15
seku_
left the channel
16:15
Bertl_oO
okay, then the debayering algorithm is really bad :)
16:16
anuejn
you are right but for i think, that it is quite handy for experementing to be able to look at details
16:16
anuejn
of course
16:16
Bertl_oO
yes, I agree that you want to 'zoom in' on the result
16:16
anuejn
the algorithm is the most naive approach i could come up with
16:16
anuejn
but its okay for demoing the tool and to start with
16:17
Bertl_oO
fair enough
16:18
Bertl_oO
so how does it work? get_color_value() is called with a position for the 'to be calculated' pixel, right?
16:18
Bertl_oO
the pos value can be assumed to be integers in what range?
16:19
anuejn
0 to 4096 or 3072 for the example image
16:20
Bertl_oO
the sensel data can be retrieved with get_intensity() I presume, which also takes a two integer pos argument, yes?
16:20
anuejn
yes
16:22
Guillaume_Chooks
left the channel
16:23
Bertl_oO
why is it 0 to 4096 or 3072 and not 0 to 4095/3071 ?
16:26
Bertl_oO
and what is the float range of get_intensity()?
16:26
anuejn
oh i failed of course you are right
16:33
Bertl_oO
hmm, I don't think I get how this works ... here is my trivial code I tested:
16:33
Bertl_oO
https://pastebin.com/raw/ZdLiyWe8
16:34
Bertl_oO
for whatever reason this gives me a dark-ish gray image with colorful patterns when you zoom in
16:35
Bertl_oO
ah, sorry, my fault, I forgot to multiply the coordinates by two again
16:36
Bertl_oO
okay, that seems to work and gets rid of the patterns
16:36
Bertl_oO
int x = (pos.x/2)*2;
16:36
Bertl_oO
int y = (pos.y/2)*2;
16:44
Bertl_oO
or probably more efficient:
16:44
Bertl_oO
int x = pos.x & ~1;
16:44
Bertl_oO
int y = pos.y & ~1;
17:27
Bertl_oO
here a 'high resolution' version:
17:27
Bertl_oO
https://pastebin.com/raw/MCaKp563
17:27
Bertl_oO
btw, the error output does not disappear when everything is fine, which is a little confusing :)
17:50
se6astian
IRC meeting in 10 minutes
17:53
se6astian
anuejn: very cool work with https://anuejn.github.io/batic/
17:54
anuejn
thx :)
17:54
davidak[m]
hello everyone
17:54
BAndiT1983|away
changed nick to: BAndiT1983
17:56
anuejn
btw. what is the license of the images in http://files.apertus.org/AXIOM-Beta/snapshots ?
18:00
se6astian
hi davidak[m]
18:01
se6astian
anuejn: good point, we do not have a license file in there to specify it
18:01
se6astian
but I would say its creative commons
18:01
se6astian
meeting time!
18:02
anuejn
whom is the author?
18:02
se6astian
I think I took that picture
18:02
comradekingu
"creative commons" needs to be specified by the licences on offer from them
18:03
se6astian
for the IRC meeting please pm me now if you have a topic to report/discuss
18:03
se6astian
and I will put the topics in order
18:03
se6astian
www.apertus.org says "Website content available under CC-BY 4.0 International license"
18:04
comradekingu
The default one for Apertus media is https://creativecommons.org/licenses/by-sa/4.0/
18:06
se6astian
anyone with topics for todays meeting?
18:07
se6astian
I will report in the meantime
18:07
se6astian
1. GSoC org application has been handed in successfully last week
18:07
se6astian
as we already participated last year there is a good chance we will be accepted again
18:07
se6astian
but we can only be 100% sure once we get approved
18:08
se6astian
that will be released on Feb 12th
18:09
se6astian
timeline: https://developers.google.com/open-source/gsoc/timeline
18:09
se6astian
2. the AXIOM hardware assembly office has moved
18:09
se6astian
we are now at an industrial manufacturing/fab lab/coworking space at : https://factoryhub.at/
18:10
se6astian
today is our first day "in business" here
18:10
se6astian
I am writing from our new office desks actually :)
18:10
se6astian
moving all the machines and equipment was quite a challenge in the last week
18:10
se6astian
here are some impressions: https://cloud.gerade.org/index.php/apps/gallery/s/PGhonMwtS6m9IbE#20180129_111050.jpg
18:10
se6astian
sorry for potato cellphone image quality
18:11
se6astian
we have all machines set up and operational again at the new space and hope to resume hardware production next week
18:11
Rex0r
Present.
18:12
Bertl_oO
nice
18:13
se6astian
the focus is more on sharing equipment, machines, collaborating
18:13
se6astian
so very much in line with our spirit
18:13
se6astian
but I am sure it will bring some new challenges to solve
18:14
se6astian
3. the website has seen some sections being reworked
18:14
se6astian
https://www.apertus.org/ has a new footer
18:14
se6astian
https://www.apertus.org/mission-statement is new
18:15
se6astian
and contains a summarized but very essential overview of our projects "maniphesto"
18:15
comradekingu
I think "ethos" sounds better than "mission-statement"
18:15
se6astian
new home/landing page is in the works plus some other subpages
18:16
se6astian
comradekingu: noted
18:17
se6astian
4. https://lab.apertus.org/ is seeing increased spam attacks recently
18:17
se6astian
but as we have an all new phabricator already set up and ready at http://lab2.apertus.org/
18:18
se6astian
we will soon make a full switch rather than implement spam countermeasures on the old system
18:18
BAndiT1983
are there new measures in the newer one?
18:18
comradekingu
is a new login needed?
18:18
se6astian
last invitation to play around with http://lab2.apertus.org/ and try to push the system and discover bugs before we switch over
18:18
davidak[m]
is there a ticket for the new footer? i have broken layout in chromium 64
18:19
se6astian
BAndiT1983: Its a much newer release so if we still see spam accounts we can deal with them on the new system rather than the outdated one
18:19
se6astian
comradekingu: accounts and content were already migrated
18:19
se6astian
but will be overwritten once again from lab.apertus.org once we make the production switch
18:19
BAndiT1983
github login for lab2 is not working yet, so i will wait till the switch
18:19
se6astian
davidak[m]: please create a lab ticket in the old phabricator
18:20
se6astian
BAndiT1983: yes the github auth is unfortunately broken with missmatching urls
18:21
se6astian
ok thats it from my side from the top of my head
18:21
Rex0r
We do use the word ethos frequently in other areas of the site/Wiki, mission statement tends to be a universally understood heading where website structure or introduction materials are concerned. For instance we have a page that details the 'open source ethos' and in this sense 'ethos' is referred to in documentation in this sense. In other words, deviating away from the term mission statement may cause confusion in some of
18:21
Rex0r
the project's copy.
18:21
se6astian
anything else to share/discuss while we are all together, now is the time
18:24
comradekingu
Is the red thing fixed?
18:25
se6astian
right, then I conclude todays IRC meeting and schedule the next one for Monday in two weeks
18:25
se6astian
I will add the meeting minutes to the wiki in the coming days
18:25
se6astian
comradekingu: define "red thing"
18:26
comradekingu
The red artifact on the videofeed of the CCC protptype
18:26
se6astian
heard no reports of that so far
18:27
comradekingu
Ok. If you upgrade to Nextcloud 13 (or 14 soon) we can use Talk for videochat meetings
18:28
comradekingu
A fitting way to display the uses of apertus video
18:29
comradekingu
I have found it to tidy up my desk immensely
18:29
davidak[m]
comradekingu: felix_ knows the limitations of the SDI module
18:29
davidak[m]
and will work on that
18:34
davidak[m]
se6astian: https://lab.apertus.org/T918
18:35
BAndiT1983
davidak[m], is the page zoomed somehow? what is your screen resolution?
18:36
davidak[m]
no zoom
18:36
davidak[m]
1366x768
18:37
illwieckz
left the channel
18:37
BAndiT1983
pretty much what i've expected, it seems that we need to adjust the media queries, so the elements are moved/wrapped earlier
18:38
davidak[m]
i think there is enough space around the links
18:40
BAndiT1983
the prototype worked more or less for mobiles and hd ready screens, but current implementation is not doing the right switch
18:40
davidak[m]
when i make the window smaller it get's even worse
18:40
BAndiT1983
tried it already, the links disapper, have to investigate
18:40
BAndiT1983
*disappear
18:41
BAndiT1983
but i remember that se6astian has added some workaround to bypass drupal layouting, maybe it's breaking the switch
18:45
BAndiT1983
se6astian, we need to move the footer out of the main container
18:45
BAndiT1983
then the wrap is working correctly, also when shrinking
18:51
illwieckz
joined the channel
19:01
davidak[m]
se6astian: here are some improvements for the mission statement https://hackmd.io/KYdgzAjArARmAsBaMBOGAGR8Ac1EwCZ4A2RYAEyigggENhQEg===
19:03
davidak[m]
should i create a ticket or do you want to update it directly?
19:07
davidak[m]
i really like that the picture on the page is made with a beta :)
19:09
seku
joined the channel
19:10
Rex0r
Not sure about the empowering. Expectations can't be true or false, but they can be raised or lowered. So the appropriate form for that message would be "raise expectations unduly"... but then, the whole concept of raising expectations is open to a lot of interpretation, so, not sure.
19:11
davidak[m]
you may be right
19:12
Rex0r
And like you say there are time frames for development involved, so...
19:12
davidak[m]
but you know some marketing is very aggresive with raising expectations and we shouldn't do that and i think we didnÄt in the past
19:14
davidak[m]
>We will not exploit the users of the tools we develop.
19:14
davidak[m]
this probaply includes not adding limitations to products to sell a more expensive version
19:15
davidak[m]
but we could add that becouse it's often the case with cameras
19:18
Rex0r
This is the thing with the word marketing. It's really loaded, and what springs to people's minds are things like the Bill Hicks sentiments. But he was talking about the 1980's, he never met people like us. So marketing is like a gun, and it's the consciousness behind the gun that pulls the trigger. There's not much we can do about this but unfortunately the marketplace and those operating inside it give the concept of
19:18
Rex0r
'marketing' a bad name I guess.
19:21
BAndiT1983
changed nick to: BAndiT1983|away
19:21
Rex0r
Luckily the community/team is pretty conscientious on all fronts. We don't need to foster a false sense of security by coming up with a slogan like "Don't be evil". This project is the real deal.
19:23
davidak[m]
sure, but we could differentiate "bad marketing" with our marketing so a new visitor understand what we stand for
19:23
davidak[m]
it's a big thing that cameras have technical limitations
19:24
davidak[m]
like point and shoot and DSLR can't record more than 30 minute videos
19:24
davidak[m]
and point and shoot can't capture RAW
19:24
davidak[m]
...
19:25
davidak[m]
i think that is a good example people would understand
19:41
mithro
Bertl_oO: ping?
19:41
Rex0r
You might be onto something. eg. why mention marketing in the first place. Have improved that section.
19:48
se6astian
off to office now
19:48
se6astian
changed nick to: se6astian|away
19:48
mithro
hey se6astian -- I'll try and get back to you regarding the invoice after wednesday
19:48
mithro
Just got back to the US today
19:52
Bertl_oO
mithro: pong!
20:08
mithro
Cool -> https://hackaday.com/2018/01/24/printed-it-rubber-band-pcb-vice/
20:40
davidak[m]
Rex0r: yes, good solution!
21:29
XDjackieXD
left the channel
21:32
XDjackieXD
joined the channel
22:01
BAndiT1983|away
changed nick to: BAndiT1983
22:11
BAndiT1983
changed nick to: BAndiT1983|away
22:23
XDjackieXD
left the channel
22:25
XDjackieXD
joined the channel
23:45
parasew[m]
left the channel
23:51
parasew[m]
joined the channel