Current Server Time: 00:32 (Central Europe)

#apertus IRC Channel Logs

2017/02/08

Timezone: UTC


23:35
intracube
left the channel
23:40
intracube
joined the channel
00:21
RexOrCine
left the channel
01:30
davidak
left the channel
02:56
Bertl_oO
off to bed now ... have a good one everyone!
02:56
Bertl_oO
changed nick to: Bertl_zZ
05:52
Spirit532
joined the channel
06:53
LordVan
joined the channel
06:55
LordVan
morning
08:05
dimaursu16
joined the channel
08:05
dimaursu16
left the channel
08:05
dimaursu16
joined the channel
08:12
dimaursu16
left the channel
08:45
morrigan
left the channel
08:45
Bertl_zZ
left the channel
08:51
ItsMeLenny
joined the channel
08:51
morrigan_
joined the channel
08:52
Bertl_zZ
joined the channel
09:04
se6astian|away
changed nick to: se6astian
09:14
ItsMeLenny
left the channel
09:25
Bertl_zZ
changed nick to: Bertl
09:25
Bertl
morning folks!
09:50
ItsMeLenny
joined the channel
10:03
se6astian
changed nick to: se6astian|away
11:00
pusle
left the channel
11:02
Bertl
off for now ... bbs
11:02
pusle
joined the channel
11:02
Bertl
changed nick to: Bertl_oO
11:10
ItsMeLenny
left the channel
13:19
LordVan
left the channel
13:34
RexOrCine
joined the channel
13:52
jucar
left the channel
14:02
se6astian|away
changed nick to: se6astian
14:27
se6astian
good evening
15:08
arpu
joined the channel
15:13
Spirit532
Apertus might be interested in Alexima's latest sensor - LUX160
15:13
Spirit532
16MP @ 330fps, 4K @ 500fps
15:14
Spirit532
No NDA as far as I'm aware, but no public quotes.
15:25
Bertl_oO
it would be great if you could get/link the datasheet then
15:25
Bertl_oO
(so we can check if it would work on the Beta)
15:34
se6astian
good find, I will contact them
15:50
dimaursu16
joined the channel
15:50
dimaursu16
left the channel
15:50
dimaursu16
joined the channel
16:08
Spirit532
Bertl_oO, datasheets aren't public, gotta email them
16:09
Spirit532
it probably can work on the Beta, but not anywhere near the full speed
16:57
davidak
joined the channel
16:59
davidak
good evening everyone
17:00
intracube
hi
17:00
intracube
meeting is imminent? :)
17:00
se6astian
indeed
17:00
davidak
yes
17:01
intracube
pours a coffee
17:01
se6astian
lets go through the flyer text first
17:02
se6astian
at the end of the google doc
17:02
intracube
ok
17:02
se6astian
"Designed by a worldwide community of film-makers, all device schematics have been made publicly available in conjunction with the FOSS/OH ethos."
17:02
se6astian
while I think its good to have the term "filmmakers" in there somewhere
17:02
se6astian
here I think its missguiding/wrong
17:03
davidak
does a cinema camera buyer know what a "FOSS/OH ethos" is?
17:04
se6astian
probably not
17:04
se6astian
lets phrase that out as well
17:04
davidak
also you are right, we are not only filmmakers
17:04
intracube
se6astian: "worldwide community of camera users and specialists"
17:04
intracube
?
17:05
se6astian
well it sounds strange that "users" created something
17:05
davidak
yes
17:06
intracube
I thought about 'enthusiasts', but that sounds too amateur
17:06
se6astian
"Designed by a worldwide community of experts from different kinds of professions"
17:06
se6astian
maybe?
17:06
intracube
it doesn't seem so strange that users are part of the design process. that's a defining part of Open Src
17:06
davidak
what professions?
17:06
davidak
;)
17:07
intracube
"different kinds of professions" - sounds a bit wordy/vague
17:07
intracube
and users are getting involved in the design process (feedback etc)
17:08
davidak
users and makers?
17:08
intracube
so "camera users and specialists" seems quite accurate
17:08
intracube
users and designers?
17:08
davidak
whould you consider calling yourself specialist se6astian?
17:09
davidak
intracube: designer sound like graphic
17:09
se6astian
well everyone is special is his own kind of way :D
17:09
davidak
:D
17:09
se6astian
"experts"?
17:09
intracube
davidak: hm, yeah I guess
17:10
intracube
experts... or specialists :)
17:10
intracube
and it seems like there should be at least one extra comma in the first sentence
17:10
intracube
there isn't much of a natural pause considering the number of words
17:11
se6astian
feel free to add it :)
17:11
davidak
"camera entusiasts and makers"?
17:11
se6astian
"Designed by a worldwide community of film-makers, camera experts and technical specialists" ?
17:11
se6astian
"makers" is good
17:12
davidak
a normal videographer who goes to a store and picks one camera wouldn't consider the axiom beta
17:12
se6astian
but we have the makers in "film-makers" already
17:13
davidak
do we have much people with a free software / open hardware / free culture background?
17:13
intracube
I think the first couple of sentences flows better
17:13
davidak
i would consider myself one :)
17:14
se6astian
we hand out such flyers at maker faires or linux events
17:14
intracube
I think the majority of the people involved have a background in open src
17:14
se6astian
so we can assume that many people reading this will be FOSS aware
17:14
davidak
"Designed by a worldwide community of film-makers, camera experts and open source entusiasts"?
17:14
intracube
(it'd be interesting to know how many people in IRC and elsewhere use linux)
17:14
se6astian
sounds good
17:14
intracube
I'm guessing it's significant
17:14
davidak
that would include Linux people and programmers
17:15
se6astian
changed
17:15
intracube
although that's now several mentions of open source
17:15
intracube
repetition?
17:15
davidak
intracube: just throw in a poll later
17:15
intracube
yep
17:16
se6astian
"All device schematics have been made publicly available in conjunction with the Free Open Source Software/Open Hardware ethos. " is just to explain what open source means
17:16
se6astian
we mention open source two times before that
17:16
se6astian
we need to also tender this flyer to be for non OS aware people....
17:16
davidak
1. it is open source 2. becouse that is important to US
17:16
davidak
so it fits
17:17
se6astian
"All device schematics and software source codes have been made publicly available." remove: "in conjunction with the Free Open Source Software/Open Hardware ethos." ?
17:18
intracube
se6astian: yep, that would be better imo
17:18
davidak
yes
17:18
davidak
it is too long
17:18
se6astian
do you say "software source codes" ?
17:18
se6astian
or "software source code" ?
17:18
intracube
software source code
17:18
davidak
code
17:19
se6astian
fixed
17:19
se6astian
thanks
17:19
se6astian
is the "sustainable" discussion resolved?
17:19
se6astian
or are there any concerns about the word still?
17:20
intracube
it still feels like it's to do with the environment, where I suspect the main meaning here is upgradable
17:20
davidak
i think it was resolved
17:21
se6astian
its good to also include a bit of the environmental aspect
17:21
se6astian
upgrade instead of replace also means less waste, etc.
17:21
davidak
yes, absolutely
17:22
se6astian
of course we cant claim we will save the planet with a few cameras but its the message that counts
17:22
davidak
it is a small step
17:22
davidak
in the right direction
17:23
se6astian
ok next point, the "photographic situation" ?
17:23
se6astian
how can we word that to also include videographers, vision people?
17:24
davidak
with "photographic" i associate a photo and not moving images
17:24
se6astian
agreed
17:24
intracube
"master any project or situation"
17:24
intracube
?
17:24
davidak
yes
17:24
se6astian
right
17:24
davidak
sounds good
17:24
davidak
"provides the building blocks"
17:25
davidak
are it multiple blocks?
17:25
davidak
or one device?
17:25
intracube
it's going to be a modular design
17:25
se6astian
well the modules are each blocks to put together
17:25
intracube
the Gamma, at least
17:25
se6astian
the beta as well
17:25
davidak
ok
17:26
intracube
oh, I thought the Beta was a closed unit
17:26
intracube
hasn't kept up to date with developments :)
17:26
intracube
is there going to be an image nearby that shows the modular aspect of the camera?
17:27
davidak
https://www.apertus.org/axiom-beta-status
17:27
davidak
here are the modules that are currently in development at the bottom
17:27
davidak
and in this concept you see a slot https://www.apertus.org/sites/default/files/axiom-beta-concept-enlosure-03.png
17:28
intracube
is that a render or a photoshop/gimp job?
17:28
intracube
gets side-tracked
17:28
davidak
3D rendering i think
17:29
intracube
looks really good
17:29
davidak
the lens and sockets might be a photograph
17:29
se6astian
its actually a photoshop drawing :)
17:30
davidak
i hope it get's thinner
17:30
intracube
nice
17:30
intracube
I always liked this concept render: https://www.apertus.org/sites/default/files/project/AXIOM_GAMMA-website-banner-01.jpg
17:30
intracube
shows of the modularity well
17:30
se6astian
agreed
17:30
davidak
yes, that looks good and very professional
17:30
se6astian
next point: Super35/APS-C format
17:30
se6astian
any objections if we keep both?
17:31
intracube
might be worth putting the latter in brackets?
17:31
davidak
i think we can attract most attention if we have both
17:31
se6astian
fine for me
17:31
intracube
/ might mean 'or'
17:31
davidak
so photo and video people recognize it
17:31
intracube
but they're almost exactly the same format
17:32
intracube
at least the width dimension
17:32
dimaursu16
left the channel
17:32
intracube
actually, if APS-C is 3:2, it is a bit different
17:32
davidak
Super35 (APS-C)
17:32
intracube
maybe better left as is
17:32
intracube
doesn't matter one way or the other, imo
17:33
se6astian
right, we are through the text!
17:33
intracube
although Super35 would be the more important, since APS-C is based in still photo land
17:33
intracube
yep :)
17:34
davidak
so Super35 (APS-C) would fit
17:34
intracube
yep
17:35
se6astian
changed
17:35
se6astian
now we have the logos
17:35
intracube
although Super35 has a slightly greater height than APS-C
17:35
_florent_
left the channel
17:36
se6astian
we need to differenciate those logos that have added text
17:36
se6astian
and the logos without text
17:36
intracube
differentiate?
17:36
se6astian
yes ;)
17:37
intracube
do you mean spacing/layout or colour?
17:37
se6astian
yes
17:37
se6astian
I was thinking of making the logos plus text into 3 coloumns
17:37
intracube
or the logos by themselves could be horizontal and centred?
17:37
se6astian
but I fear it will look strange that we have so much text for the first item in the first coloumn
17:37
se6astian
and much less in the others
17:37
BAndiT1983
joined the channel
17:38
intracube
and the logos + text could be as they are now (stacked), but also centreed?
17:38
se6astian
let me check if we can do live layout collaboration somewhere like draw.io
17:38
intracube
are there any better CMV12000, FPGA, ARM images? they're not that consistent in style right now
17:39
davidak
we will create icons like that
17:39
davidak
it was only to test how it looks
17:39
intracube
ah ok
17:40
davidak
to have icons for that informations in general
17:40
davidak
first this technical details was also in the text
17:40
davidak
but that would be way too much information
17:42
intracube
currently looks superfluous. 'sensor modules in development' would mean the same
17:42
intracube
remove?
17:42
se6astian
does this work: https://sketchboard.me/zAkPzBjQfPyn#/ ?
17:43
intracube
I'll need to enable JavaScript
17:43
intracube
or just switch to another browser
17:43
intracube
yep. working here
17:44
se6astian
looks good
17:44
se6astian
its really realtime
17:44
se6astian
I see every stroke being drawn
17:45
xfxf
left the channel
17:45
davidak
everyone is going crazy :D
17:46
intracube
I can't figure out how to delete
17:46
davidak
you have to select every line
17:46
davidak
and press Entf.
17:46
davidak
or Del.
17:46
davidak
i don't know how select multiple elements
17:47
BAndiT1983
shift and drag for box selection
17:47
davidak
ah, thanks
17:48
intracube
it also matters what draw mode you're in
17:48
intracube
I couldn't select multiple elements when drawing primitive shapes
17:49
intracube
is there enough room to put all the icon + text elements horizontally?
17:50
se6astian
its a challenge
17:50
se6astian
both vertical and horizontally
17:51
BAndiT1983
do you want it to be static or something like a flowpanel / carousel with slight rotation?
17:53
davidak
i would use static
17:53
davidak
but i'm not a designer
17:54
se6astian
what does a flowplanel look like?
17:54
BAndiT1983
i have to look up an example
17:55
davidak
do we have a newer PCB stack image?
17:55
se6astian
maybe like this now?
17:55
BAndiT1983
something like that, but much less round -> http://www.picz.ge/img/s1/1010/31/6/6e8a7852abce.jpg
17:56
davidak
we don't have hundesds of icons
17:56
BAndiT1983
http://bashooka.com/coding/responsive-jquery-carousel-plugins/
17:56
BAndiT1983
it's not about amount, they can just oscillate slightly, if there is less space available
17:57
se6astian
BAndiT1983: its a print flyer :)
17:57
davidak
:D
17:57
BAndiT1983
ah, :D you see i missed the beginning
17:57
se6astian
:)
17:57
Bertl_oO
changed nick to: Bertl
17:57
BAndiT1983
too much UI development at work ;)
17:57
Bertl
*phew* finally here ...
17:58
davidak
hi Bertl, we are still discussing the Flyer
17:58
Bertl
yeah, no problem ...
18:00
Bertl
what's the 'current' text?
18:00
davidak
https://docs.google.com/document/d/1TPMb4SYdlWopCSHlJm4GcKz_RScZ5h3annAwawtn8vI/edit
18:00
se6astian
added layout draft
18:01
se6astian
lets see what ray can come up with this
18:01
se6astian
but I think the text is really good now
18:01
davidak
yes, i also really like it
18:01
Bertl
is my input appreciated? :)
18:01
davidak
always :)
18:02
Bertl
I'm missing the 'never obsolete' part for the hardware
18:02
se6astian
had to tune the doc sharing settings
18:02
se6astian
once posted here its public....
18:03
se6astian
'never obsolete' added
18:03
Bertl
excellent!
18:03
davidak
">Dual ARM Core and Linux inside."
18:04
davidak
i think "Dual Core ARM" would be more accurate
18:04
Bertl
maybe at the end of the first paragraph something like:
18:04
RexOrCine
Present
18:04
davidak
since it is 1 processor with 2 cores
18:04
Spirit532
se6astian, when you get a reply from Alexima/Luxima(same founder/employees), please PM me with the prices/whatever they give you for the LUX160
18:04
Spirit532
I might be interested in getting one myself.
18:04
Bertl
project or situation and the open interfaces allow to extend those as needed.
18:05
Bertl
(or something along the lines of 'extensions are easy to add')
18:05
Bertl
maybe extend and complement those
18:05
davidak
Bertl: that will make it much longer
18:06
Bertl
okay, length is an issue? or do we just want to keep it as short as possible?
18:06
BAndiT1983
it should have 3 to 5 sentences to make the point
18:06
BAndiT1983
otherwise it overloads people with infos
18:06
Bertl
well, strictly speaking that would be the same (last) sentence :)
18:06
se6astian
"and extendable plug-in interfaces (up to 12 Gbps bandwidth)"
18:07
se6astian
just one word added
18:07
davidak
good
18:07
Bertl
works for me
18:07
se6astian
great
18:08
comradekingu
I asked for permission to see the draft
18:08
davidak
what do you thing about the ARM sentence?
18:08
BAndiT1983
reminds me of intel inside :)
18:08
Bertl
yep Dual Core Arm is better
18:09
davidak
i changed it
18:09
Bertl
the question is, if we should mention the arm separately
18:09
intracube
is it worth adding CPU?
18:09
Bertl
because it is part of the Xilinx ZYNQ
18:09
intracube
I guess that can be inferred from 'dual core'
18:09
davidak
the beta is not special because it uses a CPU or ARM, so i think not
18:10
Bertl
so personally I'd swap that around a little bit
18:10
Bertl
we have other interesting stuff on the Beta
18:10
davidak
Linux already tells people if they know Linux they can easily extend it
18:10
Bertl
maybe have:
18:10
Bertl
(a Zynq Picture) with XIlinx Zynq SoC, Dual Amr Core, ...
18:11
Bertl
then mention the Lattice routing fabric and the Microchip PICs
18:11
davidak
what :D
18:11
Bertl
and finally add a sentence for IMU or so?
18:12
davidak
i don't understand anything and i think someone looking for a camera also
18:12
se6astian
we are already short on space now...
18:12
Bertl
then maybe skip the arm and make it a penguin
18:12
Bertl
and just write Dual Core ARM Linux inside
18:13
BAndiT1983
left the channel
18:13
davidak
"Firmware based on GNU/Linux"
18:13
davidak
i would like that phrase :)
18:13
Bertl
actually Arch/Linux :)
18:13
se6astian
tux logo plus "Dual Core ARM Linux inside" should work I think
18:13
Bertl
but yes, I agree, GNU/Linux is very Richard!
18:14
Bertl
(RMS)
18:15
davidak
i have a picture of him on the wall
18:15
se6astian
lol
18:15
davidak
holding a sign: "With free software you have freedom"
18:15
davidak
you may know that
18:16
Bertl
yep
18:17
se6astian
this http://www.opimedia.be/DS/une/_jpg/Richard-Stallman-on-the-road.jpg
18:17
se6astian
?
18:17
davidak
exactly
18:17
se6astian
very nice
18:17
se6astian
ok can we conclude the flyer for today?
18:17
davidak
yes
18:17
se6astian
anything left you do not like, would like changed?
18:18
se6astian
otherwise we let ray take a go
18:18
se6astian
and discuss what he comes up with then
18:18
davidak
we made good progress the last days :)
18:18
Bertl
the FPGA logo looks ugly
18:18
BAndiT1983
joined the channel
18:19
davidak
Bertl: what would you use?
18:19
se6astian
text above icon says "Icons are just drafts for now, should have simplified design that is the same style:"
18:19
Bertl
I would probably use the xilinx and the zynq and the lattice logos
18:19
intracube
se6astian: the paragraph for the sensor spec is a bit long compared to the other two
18:20
se6astian
true
18:20
intracube
words like pixel could be removed?
18:20
davidak
it's hard to read
18:21
intracube
you mean with pixel removed?
18:21
davidak
yes
18:21
se6astian
removing pixels is fine
18:21
se6astian
I turned "fps" into "frames per second"
18:21
se6astian
but we can go back to fps
18:21
intracube
could bullet points be put between each part of the spec?
18:21
se6astian
just not sure if people will understand
18:22
intracube
I think anyone looking at this level of camera system will understand fps :)
18:23
se6astian
ok
18:23
se6astian
google docs has so many symbols
18:23
se6astian
bathtub!
18:23
intracube
global shutter · 4096 x 3072 · Super35 (APS-C) format · up to 300 fps at 4K · Alternative image sensor modules are in development
18:24
intracube
^ something like that
18:24
BAndiT1983
looks fine with points
18:24
davidak
yes
18:24
Bertl
last comment to the logo section: I would use the manufacturer logos and drop the names from the text
18:24
Bertl
i.e. I would put the MicroZed, Xilinx, Lattice, Microchip and Cmosis logo there
18:25
davidak
so have only CMV12000 in the text?
18:25
Bertl
and just have the text use the specifica, like Zynq, PIC, CMV12k
18:26
se6astian
noted
18:26
Bertl
those technically interested will know what it means
18:26
BAndiT1983
is it "real time" or "realtime", google gives many possible cases, also real-time
18:26
Bertl
and those not interested might recognise the company logos
18:26
davidak
http://www.dict.cc/?s=realtime
18:26
davidak
also both
18:26
BAndiT1983
was there, also not conclusive ;)
18:27
davidak
i would say realtime
18:27
BAndiT1983
searched for "Internal real time video processing" on google to find similar things
18:28
Bertl
okay, move on to the GSoC project brainstorming?
18:28
se6astian
comradekingu: stop changing the text completely!
18:28
davidak
from the translation, i think "real time" means something else
18:28
BAndiT1983
yes, haven't considiered the translation
18:28
Rex0r
joined the channel
18:29
comradekingu
se6astian: i just corrected some errors
18:29
comradekingu
Made comments for the ones that were hardcoded into images
18:29
intracube
as word pairings become popular, they usually end up joined or hyphenated
18:29
comradekingu
You can give suggestion rights if you dont want errors fixed
18:30
intracube
so I think realtime or real-time is more common now
18:31
comradekingu
se6astian: Try reading that sentence now, it makes no sense
18:32
se6astian
you removed half if it
18:32
comradekingu
The part i removed did nothing
18:32
se6astian
unfortunately revision history is hard to read to revert it....
18:32
comradekingu
I could make suggestions
18:33
comradekingu
This is the sentence "The worlds first, fully open source and open hardware motion picture camera system. has been devised to be modular and repairable, designed by a worldwide community of film-makers, camera experts and free software enthusiasts."
18:33
se6astian
we just concluded the meeting that deals with writing this text
18:33
comradekingu
The worlds first comma. provides no info
18:33
davidak
we discussed that several time
18:33
_florent_
joined the channel
18:34
Bertl
hey _florent_!
18:34
intracube
this was the original first para text:
18:34
intracube
"The world’s first, fully open source and open hardware motion picture camera system. Devised to be completely modular and repairable and designed by a worldwide community of film-makers, camera experts and open source enthusiasts. All device schematics and software source code have been made publicly available. Not just a camera, the AXIOM Beta is a constantly evolving, sustainable hardware and software ecosystem. With Interchangeable lens
18:34
intracube
mounts (e-mount as standard), swappable image sensors (Full-frame, High Speed, Four Thirds, 2/3", etc.) and plug-in interfaces (up to 12 Gbps bandwidth), the AXIOM Beta provides the building blocks to master any project or situation."
18:34
comradekingu
Well, my argument is what it is, not how it came to be that way
18:35
intracube
in case it got lost in Google's revision system...
18:35
se6astian
thanks intracube
18:35
se6astian
replacing
18:36
comradekingu
completely, fully, deviced and enthusiasts can be done away with
18:36
comradekingu
and world's
18:36
se6astian
ok flyer meeting concluded
18:36
se6astian
comradekingu: feel free to add comments in the google doc
18:37
comradekingu
ok
18:37
comradekingu
extenable not a word btw
18:37
se6astian
so GSoC is on!
18:37
se6astian
we are VERY late
18:37
davidak
yay
18:37
Spirit532
though it's not going to be very open if it's still in eagle
18:37
intracube
:)
18:37
Spirit532
because autodesk
18:37
se6astian
and I have to admit I have doubts if we can still finish everything in the next 22hours
18:37
davidak
true
18:37
intracube
this was my rough layout idea: http://i.imgur.com/kGVtBkk.png
18:38
comradekingu
I am the green snowy pine trees
18:38
Bertl
yeah, well, autodesk will help to speed up the kicad transition
18:38
se6astian
but lets try!
18:38
BAndiT1983
what the hell, i have missed the point in time when autodesk bought it
18:38
davidak
se6astian: YES
18:38
se6astian
so the application text is pretty much finished
18:39
Spirit532
BAndiT1983, it's also subscription only now.
18:39
se6astian
the biggest thing google reviews is the idea page
18:39
Spirit532
autodesk killed eagle
18:39
davidak
se6astian: i have read it, it's ok
18:39
se6astian
just look at KDE: https://community.kde.org/GSoC/2016/Ideas
18:39
Spirit532
it was "an okay" software to use before
18:39
Spirit532
now it's obsolete
18:39
BAndiT1983
i used target3001 before, was never big friend of eagle user experience
18:39
Bertl
okay, I've got a list of 17 (seventeen) potential tasks/projects for GSoC
18:39
se6astian
great!
18:40
se6astian
how to best organize them?
18:40
se6astian
google doc?
18:40
Bertl
but I would prefer to start with a brainstorming, where everybody says what comes to his/her mind
18:40
se6astian
sure
18:40
se6astian
I ll collect them in a google doc
18:40
se6astian
shoot
18:40
davidak
Bertl: how much students do we want
18:40
se6astian
I d say 3 max
18:40
davidak
and how much tasks do we need so
18:40
Bertl
as many as we get and we can handle
18:40
se6astian
6-9 would make sense
18:40
se6astian
more = better
18:41
Bertl
if we make a list of 20 tasks, I think that is fine as well
18:41
BAndiT1983
which students can apply, just from austria and surrounding or from evrywhere?
18:41
se6astian
right
18:41
Bertl
it doesn't mean that we have to put them all on GSoC :)
18:41
se6astian
worldwide
18:41
intracube
changed nick to: intracube_afk
18:42
BAndiT1983
oha, time difference will be interesting thing
18:42
se6astian
so I added the 4K HDMI fpga code as idea in the lab workboard already
18:42
se6astian
plus a nodejs API
18:42
se6astian
for future interactive webinterfaces, etc.
18:42
davidak
shouldn't it be a Python API some time ago?
18:42
Bertl
isn't nodejs that thing which didn't work?
18:43
Bertl
(i.e. didn't want to install/work/etc)
18:43
BAndiT1983
is there already something like a remote app for android?
18:43
davidak
yes, i think so
18:43
se6astian
hackmd yes
18:43
se6astian
but saying it the way you did it its like saying "isnt that the C++ thing that didnt work" :)
18:43
Bertl
so I don't think java is a good idea _on_ the camera
18:43
BAndiT1983
nooooooo
18:43
davidak
:d
18:43
Bertl
it will make the camera die a horrible death :)
18:43
se6astian
nodejs is javascript
18:44
BAndiT1983
have to develop in java at work, still happy to be c++ developer at home
18:44
davidak
it would need lot's of RAM :D
18:44
BAndiT1983
java is not javascript ;)
18:44
Bertl
okay, so brainstorming!!!
18:44
se6astian
shoot people!
18:44
Bertl
what would you want somebody to develop for the Beta
18:44
Spirit532
nothing containing the word "java" should ever exist on embedded hardware
18:44
Spirit532
ever
18:44
Bertl
(software only, unless you can come up with the hardware)
18:44
BAndiT1983
and still android is one of the best selling plattforms :D
18:45
Bertl
agrees with Spirit532
18:45
davidak
the remote software
18:45
se6astian
define "remote software"
18:45
Spirit532
Qt, maybe
18:45
Spirit532
(if we're talking PC)
18:45
davidak
the software running on the remote
18:45
Spirit532
that's still embedded
18:45
BAndiT1983
like the one with which i can control my canon DSLR from my smartphone
18:45
BAndiT1983
you have live preview and can set shutter, exposure etc.
18:46
Spirit532
well, smartphone apps have to be made with their own SDKs
18:46
Spirit532
so that's a given
18:46
davidak
yes
18:46
se6astian
-) moving the FPN calibration routine plus validation into the camera and all into one script
18:46
davidak
but maybe the remote uses the API?
18:46
BAndiT1983
Qt would be not a big problem, and there can be a module for OC for such things like remote control or similar
18:46
davidak
so we need that first
18:47
Bertl
brainstorming, not arguing or thinking! :)
18:47
se6astian
slower than realtime camera internal hdmi overlay text and primitive drawing framework
18:47
Spirit532
the remote could use the same interface as apps
18:47
Spirit532
(minus live view)
18:47
Spirit532
bluetooth perhaps
18:48
se6astian
"ESP32 bluetooth communication demo" added
18:48
davidak
implement the remote software -> http://files.apertus.org/controller-simulator/
18:48
Spirit532
ESP32 wouldn't be fast enough for live view though, would it?
18:48
Spirit532
davidak, that looks like something I can do in a day
18:48
se6astian
collecting ideas from brainstorming here: https://docs.google.com/document/d/1n7BxN4UFok-JP47zsJ2w6oAErJ1aFpHtssCNx4dB9P0/edit
18:49
BAndiT1983
does Beta has wifi or just BT?
18:49
se6astian
davidak: no hardware yet
18:49
se6astian
wifi and bt
18:49
davidak
ok
18:49
BAndiT1983
canon uses wifi for live preview
18:49
BAndiT1983
it could be useful for static recording, when you don't need very fast preview
18:49
BAndiT1983
there is also not much delay, as i can tell from my experience
18:49
Bertl
BAndiT1983: the Beta itself doesn't have either, but we have an ESP32 shield in the works which does both and we can use USB dongles for Wifi/BT
18:49
se6astian
make that MJPEG encoder fpga code
18:51
se6astian
added
18:51
Bertl
okay, unfortunately not much storming here ... so let's discuss the list I madeL
18:51
Spirit532
I've got some debayering code
18:51
BAndiT1983
Sebastian, what about the new algorithm from Stephan?
18:52
Spirit532
it's not very good though
18:52
Spirit532
just nearest neighbor
18:52
BAndiT1983
i've added simple linear debayering and was on to adding SHOODAK, but priorities changed half a year ago
18:52
BAndiT1983
but nearest is also helpful for draft preview
18:52
Spirit532
I don't really contribute to hw projects I don't own
18:53
Spirit532
and I'd rather drop $5k on a good lathe/mill combo right now
18:53
se6astian
herbert can we control the fan speed on the microzed yet in software?
18:53
davidak
maybe there is already good open source code out there somewhere?
18:53
xfxf
joined the channel
18:53
Bertl
the fan is on a port pin, so PWM could be implemented
18:54
Bertl
but we don't have an implementation yet
18:54
davidak
some python dev to work on Elmyra
18:54
Bertl
it is controlled by one of the routing fabrics
18:55
davidak
active lens mount for aperture control
18:55
davidak
and focus
18:55
davidak
autofocus
18:56
se6astian
is there
18:56
se6astian
Lens Control (various Protocols: Canon, Nikon, Sony, etc.)
18:56
BAndiT1983
there are some command-line tools in the repo, should the get some UI?
18:56
BAndiT1983
*they
18:56
se6astian
good idea
18:57
Bertl
maybe we should simply do a GET/POST interface for camera control?
18:57
comradekingu
Made a comment with a suggestion for the text
18:57
Bertl
with a well defined API
18:57
BAndiT1983
REST API?
18:57
se6astian
added
18:57
davidak
Bertl: would that be called REST-API?
18:58
Bertl
yep, we could do REST
18:58
Bertl
then all kind of client side apps can use that
18:59
davidak
should that api be high performance and written in native language like C++, Go or Rust?
18:59
Bertl
regardless of what underlying language or presentation
18:59
davidak
so a picture is taken immediately when you hit the trigger?
18:59
Bertl
my preference would actually be C
18:59
BAndiT1983
yeah, C on the embedded side for performance
18:59
davidak
C has some problems you see in current OSS like openssl
19:00
BAndiT1983
it has to be evaluated how much performance C++ consumes on chip
19:01
Bertl
davidak: what problems do you expect for a REST api?
19:01
BAndiT1983
https://github.com/babelouest/ulfius
19:01
davidak
someone hacks your camera because of buffer overflows and stuff
19:02
davidak
not sure if that is important here
19:02
Bertl
well, they don't need a buffer overflow to hack it? :)
19:02
BAndiT1983
you should only allow specific things, to overcome that problems or even prevent them
19:02
alexML_
I like the idea of Rust, just that it doesn't seem the easiest thing to learn
19:02
BAndiT1983
hack an open-source camera :D this has funny side to it
19:04
se6astian
enough brainstorming?
19:04
BAndiT1983
the list is long enough, i think
19:04
Bertl
yeah, I don't have a problem with Rust either ... I'm just worried that we will not find a Rust hacker :)
19:04
davidak
or someone to maintain it later
19:04
BAndiT1983
we should use languages which have established over years
19:05
niemand
joined the channel
19:05
niemand
left the channel
19:05
niemand
joined the channel
19:06
se6astian
so now we need to break it down
19:06
se6astian
each idea needs a mentor
19:06
Bertl
okay, so I propose we create 'new' tasks for each of them in the Lab with a common structure
19:06
se6astian
I would do that after we made a selection
19:06
Bertl
and incorporate old tasks with similar/related stuff
19:07
se6astian
each idea for google needs
19:07
Bertl
well, we don't want to lose any of those ideas, no?
19:07
se6astian
for ourselves, sure
19:07
Bertl
we can still cherry pick those for GSoC
19:07
se6astian
ok
19:08
se6astian
please make a note in the google doc if you can think of mentoring any of these projects
19:08
Bertl
alexML_: the problem with overlays is that they come in different flavors
19:08
Bertl
zebra or peaking or similar is best done in the FPGA in realtime
19:09
Bertl
basically when the data passes by
19:09
alexML_
yes; I meant zebras and peaking are very similar in the backend, so they can be grouped in a single project
19:10
Spirit532
are they though?
19:10
Bertl
stuff like frames or grids too, they do not even need input data
19:10
Spirit532
I mean, they're both LPF/HPF
19:10
comradekingu
Could add some shadows from the camera onto the backdrop?
19:10
Bertl
but user elements, info, graphs, etc should be handled by a framebuffer overlay
19:10
niemand
That's music to my ears... :)
19:13
se6astian
Andrej are there any other tasks/subprojects for opencine that you can think of?
19:14
BAndiT1983
i also could supervise the REST API
19:14
se6astian
great
19:14
BAndiT1983
just don't know at the moment where it should be put, as i haven't looked in the camera code yet
19:14
Bertl
a few tasks require that we provide hardware to work on/with
19:15
se6astian
BAndiT1983: dont worry about that
19:15
Bertl
we should color code them or similar
19:15
se6astian
agreed
19:15
BAndiT1983
can it be surpassed by emulation?
19:16
Bertl
some parts maybe, others hardly
19:16
Bertl
but it shouldn't be a problem to provide the required hardware from our side
19:16
comradekingu
Btw, the oshwa cert is no longer self-certification, and it costs money. Im pretty sure they own the blue gear logo and use it for that
19:16
BAndiT1983
depends on the location of the student
19:17
BAndiT1983
or do you want to carry it over to himalayas ;)
19:17
davidak
comradekingu: we are officially certified :)
19:17
comradekingu
oh ok
19:17
davidak
AT01 i think
19:17
comradekingu
How about including some test shot imagery, there was a nice video of some trees yesterday, could use a still from that
19:18
BAndiT1983
Sebastian, you can include me for the API tasks, is just software after all
19:18
comradekingu
It is a bit "concept art" CGI heavy atm
19:18
Spirit532
I'd love to contribute a lot, but I can't afford a beta right now
19:19
Bertl
se6astian: which reminds me that the OSHWA cert logo should be on the flyer as well, probably also OSHpark
19:19
Spirit532
though most of it doesn't require hardware
19:20
davidak
Bertl: we have the open source hardware logo, maybe replace that?
19:20
davidak
else it would be redundant
19:20
Bertl
I would mark the typical FPGA parts separately (not as requires hardware)
19:20
se6astian
Bertl: noted
19:20
Bertl
because FPGA/HDL designs can and should be tested with simulations
19:21
se6astian
davidak: noted
19:21
Bertl
of course, the final code will be integrated and tested on real hardware as well
19:21
davidak
Bertl: for that you could provide ssh access to betas you have
19:21
Bertl
yes, that's the idea
19:22
davidak
if some studen't don't needs it all the time
19:22
comradekingu
All the logos are for the initiated. It would be good to provide some starter info, cheif among which is the copyleft logo, which is instantly recognisable as not standard copyright
19:22
Bertl
we have had a working setup for a remote Beta some time ago
19:22
BAndiT1983
time for jenkins CI i think
19:22
Bertl
so that seems to be doable and reasonable
19:22
se6astian
Bertl: please mark the tasks that are FPGA related but do not require special hardware
19:23
niemand
BAndiT1983, for the FPGA part or other code?
19:23
Bertl
with the symbol I do not see except for a binary code :)
19:23
davidak
BAndiT1983: i have used Drone CI the last days, it's really simple like Travis
19:23
davidak
(i don't like Jenkins because it's too complicated)
19:23
BAndiT1983
jenkins 2.0 is really good
19:24
se6astian
Bertl: yes that one :)
19:24
BAndiT1983
CI would be good to gather the code and to test automatically on hardware, like nightly build with flashing and testing
19:24
comradekingu
se6astian: Can it be RYF licensed?
19:25
Bertl
I don't see why not?
19:26
Bertl
we don't do any DRM and we do not actively support it in the hardware/software AFAIK
19:26
se6astian
moving ideas/projects out now who do no have a mentor
19:27
davidak
should we scrap node/Python API in favor of native language?
19:27
davidak
or do you mean a client for the API?
19:27
Bertl
I think a python interface would still be a good idea
19:28
Bertl
but it can either be a wrapper of a common C library
19:28
Bertl
or already a REST client
19:28
BAndiT1983
both would be good
19:28
comradekingu
I really like this folding orange thing on the logo https://www.apertus.org/sites/default/files/logo-history.jpg
19:28
davidak
when we have REST we can connect with every language
19:28
BAndiT1983
what about performance?
19:30
BAndiT1983
i would suggest REST and some "native" API, e.g. Python
19:30
davidak
what about Elmyra? could Simon mentor some tasks of it?
19:32
Bertl
some of the tasks can also be formulated in different levels of complexity
19:32
Bertl
for example, the sensor simulation for me has three different levels
19:32
Bertl
easy: create a simple test data stream like from the cmv* sensor
19:33
Bertl
medium: simulate the data and timing behaviour of the sensor including data related registers
19:33
Bertl
hard: create a full fledged simulation which can be put into a separate FPGA
19:34
Bertl
similar is true for most of the FPGA related parts
19:35
se6astian
gotta leave now for dinner....
19:35
davidak
students could work as a team
19:35
BAndiT1983
have a nice meal
19:35
se6astian
I gave you edit access
19:35
se6astian
to continue the list
19:35
Bertl
okay, let's take a short break ... I could use something to eat as well
19:35
davidak
same
19:35
se6astian
next step is adding one lab task per idea with brief description and a list of prerequisites, description of programming skills needed and estimation of difficulty level.
19:36
Bertl
can we continue in half an hour or so?
19:36
davidak
ok
19:36
BAndiT1983
by the way node.js and REST go together, if i interpret my google searches correctly
19:36
BAndiT1983
alright
19:37
davidak
BAndiT1983: of course you CAN do that but is it good for our use case in embedded hardware?
19:38
BAndiT1983
depends on what do you want to achieve through it
19:38
davidak
take a picture just now?
19:39
davidak
like tethered shooting
19:39
BAndiT1983
which MCU is used at the moment?
19:39
slikdigit
joined the channel
19:39
davidak
what is MCU?
19:39
BAndiT1983
microcontroller unit
19:40
BAndiT1983
sorry, use this abbreviation since my student years
19:41
davidak
what microcontroller do you mean?
19:41
BAndiT1983
i think the Zynq®-7000
19:41
davidak
Zynq 7020
19:41
BAndiT1983
have looked up the tech specs of beta
19:42
BAndiT1983
usually people use like node.js or python to get REST on their embedded hardware
19:42
BAndiT1983
node.js seems to be faster
19:42
Spirit532
javascript on embedded hardware = death
19:42
Spirit532
or anything involving java
19:43
BAndiT1983
here is also node.js as example for some Zync present -> https://docs.resin.io/installing/gettingStarted-Zynq-ZC702/
19:43
BAndiT1983
???
19:43
BAndiT1983
you know that javascript and java are 2 different things? node.js is based in Google V8 and is driven by C++
19:43
Spirit532
it has the letters in it
19:44
BAndiT1983
we should forbid the letter -> J, V and A as they are evil :D
19:44
BAndiT1983
*leters
19:44
BAndiT1983
*letters
19:45
niemand
running javascript on a mcu seems awkward
19:46
BAndiT1983
on camera-side we could implement in C++, client-side doesn't really matter then
19:46
BAndiT1983
just tell it your smartphone with Android ;)
19:47
niemand
Keep the number of software which needs to be updated and maintained low, to reduce the effort needed
19:47
dimaursu16
joined the channel
19:47
dimaursu16
left the channel
19:47
dimaursu16
joined the channel
19:49
davidak
also remember we need maintainers
19:49
BAndiT1983
it's always a balance act between usability and maintenance, we should decide what's important for the remote control
19:50
davidak
if we get a student that likes Rust and provide amazing code, we still need someone to maintain it in the future
19:50
davidak
i'm not sure how much people are good C programmers
19:50
niemand
and someone who reviews the code in the first place :)
19:50
Spirit532
C is universal
19:51
BAndiT1983
i can review the code if necessary, it's almost my daily task in my job
19:51
Spirit532
and when you aren't here?
19:51
niemand
yes, for rust that's not so easy I think (tough I agree that would be great actually)
19:51
BAndiT1983
what's so great about rust?
19:52
Bertl
a systems programming language that runs blazingly fast, prevents segfaults, and guarantees thread safety.
19:52
BAndiT1983
but there has to be some trade-offs, or not?
19:52
Bertl
well, it's not trivial to read/write :)
19:53
BAndiT1983
the syntax is not to my liking, but my area is object oriented, so i'm biased
19:53
davidak
i think it's just a modern language and C is very old…
19:53
niemand
which is the same for c, but we are all used to c because we have been raised with it
19:53
BAndiT1983
:D instead of milk?
19:53
niemand
I (tried to) answer to Bertl^^
19:54
BAndiT1983
don't underestimate C, it's still very good
19:55
BAndiT1983
and i'm looking forward to C++17
19:56
Spirit532
left the channel
19:56
Bertl
yeah, well, I'm not against Rust per se, but C and gcc is already installed, and Rust seems to come with its own install/update tool (that seems to be modern as well :)
19:57
BAndiT1983
do we have C++ support?
19:57
BAndiT1983
just asking in case we would use some REST lib
19:57
GrahamM
joined the channel
19:58
Bertl
personally I don't like C++ very much, if you want to go object oriented, use a truly object oriented language not a hack :)
19:58
davidak
one point is also that you can code faster in Ruby, Python and i think Go than C or C++
19:58
BAndiT1983
i disagree with that sentence ;)
19:58
Bertl
or do it like the Linux Kernel and code object oriented without language support
19:58
niemand
That cargo bullshit in rust is really disgusting
19:59
BAndiT1983
faster execution depends on many things, especially first execution can take much longer
19:59
niemand
before I start ranting, i'm off
19:59
davidak
(I'm not really a programmer. just know python and perl ;))
20:00
BAndiT1983
i fell outnumbered by hardware guys somehow ;)
20:00
davidak
what i know is that Go software is amazing since the programmers are very entusiastic about it :)
20:00
BAndiT1983
maybe another hipster thing
20:00
davidak
like Syncthing, Grafana, ...
20:01
davidak
BAndiT1983: when it result's in good software i think it is good :)
20:02
BAndiT1983
it also depends on the usage case, imho
20:02
davidak
and i see no one want's to take care of old openssl code
20:02
BAndiT1983
is it bleeding again?
20:02
davidak
not yet
20:02
intracube_afk
changed nick to: intracube
20:02
BAndiT1983
what's the matter with it?
20:03
davidak
but who knows what bugs there are
20:03
BAndiT1983
every software has a lot of bugs, if you search thouroughly
20:03
BAndiT1983
i see that openssl is progressing and there was new release a couple of days ago
20:03
davidak
but there are always the same problems with C code that you don't have with Rust or Go (i think)
20:04
BAndiT1983
that's a complicated topic, the language does not guarantee that thing
20:05
davidak
i'm not sure if you could call C still sustainable
20:05
BAndiT1983
i've seen people writing in C#, like they were used to in C, and it was very awkward and slow
20:05
niemand
left the channel
20:06
BAndiT1983
C is still number one, at least according to TIOBE rating
20:07
davidak
that is not the point
20:07
davidak
Java is also used very often
20:07
davidak
i'm interested what language results in good software
20:07
BAndiT1983
yes, but i meant it more in hardware development related way
20:08
BAndiT1983
language is just a tool for that, the developer creates good software, at least one with enough experience
20:09
davidak
when you have to write more code in C for the same task than you can have more bugs in it
20:09
BAndiT1983
what do you think, how many times i'm writing ranting mails to remote engineers about bad programming
20:09
arpu
left the channel
20:09
davidak
and a language can prevent bugs, like Bertl said Rust prevents segfaults and is thread save
20:09
BAndiT1983
but the trade-off is performacne for example
20:10
davidak
so there are some advantages in some languages
20:10
BAndiT1983
yes and also disadvantages too
20:10
BAndiT1983
i could get C# to leak memory, although it has garbage collector, if one uses the language wrong, then it will result in mess
20:10
Bertl
okay, so let's continue with the task at hand (pun intented)
20:10
davidak
is there a performance issue with Rust or Go?
20:11
Bertl
let's break the task list down and get some descriptions and classifications done
20:12
BAndiT1983
just to finalize the discussion for now -> http://benchmarksgame.alioth.debian.org/u64q/rust.html
20:13
BAndiT1983
you can also select comparison to Go at the bottom
20:15
BAndiT1983
we should create the tasks, respectively sub-tasks for each difficulty, in Lab
20:15
Bertl
so a quick grep/wc (not so easy with a google doc) gives meL
20:16
Bertl
7 Alex, 5 Andrej, 14 Herbert and 1 Sebastian
20:16
Bertl
(note that there are overlaps)
20:17
BAndiT1983
overlaps are welcome, as there could be emergency cases and one mentor is away for the moment
20:17
Bertl
which tells me, that we probably have to drop a bunch of 'my' tasks to stay realistic
20:17
Bertl
yes, I think we definitely take all tasks which have overlaps
20:17
Bertl
i.e. where there is more than one mentor
20:18
se6astian
back
20:19
BAndiT1983
hm, not many overlaps
20:19
Bertl
so those three would be the ones then
20:20
BAndiT1983
not good we need more
20:20
BAndiT1983
i could overlook also the firmware tasks, have my experience from MCU programming with assembler and C
20:20
Bertl
I think the kernel driver tasks are probably very appealing for GSoC folks
20:21
BAndiT1983
it would be interesting for me also
20:21
Bertl
to do or to mentor?
20:21
BAndiT1983
both ;), but its gsoc, so mentor first
20:21
Bertl
so maybe we should simply vote for the tasks?
20:22
Bertl
and please feel free to add yourself as mentor
20:22
davidak
(is the time also payd for the mentor?)
20:22
Bertl
not from google, AFAIK
20:22
se6astian
google also has some recommendations about the projects/ideas
20:22
se6astian
Low-hanging fruit: These projects require minimal familiarity with the codebase and basic technical knowledge. They are relatively short, with clear goals.
20:23
se6astian
google pays 500$ to the org for every mentored student
20:23
se6astian
I would propose we forward that money to the mentors
20:23
Bertl
really
20:23
Bertl
so that is not too bad then
20:23
BAndiT1983
we have to do some preparation, that was just an underestimation, a lot of preparation
20:23
se6astian
well its 3 months of work for 500$ :)
20:23
davidak
:D
20:23
se6astian
Risky/Exploratory: These projects push the scope boundaries of your development effort. They might require expertise in an area not covered by your current development team. They might take advantage of a new technology. There is a reasonable chance that the project might be less successful, but the potential rewards make it worth the attempt.
20:24
se6astian
Fun/Peripheral: These projects might not be related to the current core development focus, but create new innovations and new perspective for your project.
20:24
Bertl
we will live like kings!! :)
20:24
se6astian
Core development: These projects derive from the ongoing work from the core of your development team. The list of features and bugs is never-ending, and help is always welcome.
20:24
se6astian
Infrastructure/Automation: These projects are the code that your organization uses to get its development work done; for example, projects that improve the automation of releases, regression tests and automated builds. This is a category in which a GSoC student can be really helpful, doing work that the development team has been putting off while they focus on core development.
20:24
se6astian
The project you propose will define the tone and scope of your organization's participation in GSoC. It is a key part of your organization's application. Further, it may be the first impression made on a potential student applicant.
20:24
BAndiT1983
i want to see how Bertl tries to pay his rent from GsoC :D
20:25
se6astian
so we should have a bit of everything
20:25
Bertl
I think we have
20:25
se6astian
so lets start with the lab tasks
20:26
se6astian
https://lab.apertus.org/T722
20:27
se6astian
prerequisites, description of programming skills needed and estimation of difficulty level.
20:27
BAndiT1983
will we base it on Stephans new algorithm also or is it not public currently?
20:27
se6astian
we can discuss this with stephan
20:28
Bertl
let's push out a few tasks with different mentors at a time
20:28
se6astian
but there are many things to try with debayering
20:28
se6astian
Bertl, alexML_ please edit the QEMU emulation for the Beta hardware (e.g. boot the custom Arch Linux, take a “fake” picture with cmv_snap) - hard (low-level) (Mentor: Alex, but also need Herbert’s advice)
20:28
se6astian
task
20:29
se6astian
BAndiT1983: please edit the debayering task description
20:29
Bertl
which task is the QEMU one?
20:29
BAndiT1983
i'm on to it at the moment, have to reflect a bit
20:30
se6astian
great
20:30
se6astian
Bertl: could you imagine being backup mentor for the Lens Control task?
20:30
Bertl
why not
20:31
alexML_
Bertl: QEMU is T724
20:31
Bertl
se6astian: but you should talk to the Sebastian guy, he doesn't want to mentor much :)
20:31
Bertl
alexML_: TX
20:34
Bertl
alexML_: so what prerequisites does that task have?
20:39
alexML_
I'd say C, ARM assembly, ability to find one's way around a huge C codebase (QEMU)
20:39
BAndiT1983
very advanced task
20:39
alexML_
yes
20:40
BAndiT1983
but you would point in the right direction, so the difficulty could be reduced
20:41
alexML_
correct; here's where we are with emulating Canon firmwares in QEMU: http://www.magiclantern.fm/forum/index.php?topic=2864.100
20:42
se6astian
alexML_: please review: https://lab.apertus.org/T726
20:42
alexML_
basically enough to boot the GUI on some models and also emulate image capture (enough to get raw image data in a buffer from a DNG)
20:42
BAndiT1983
i'm still waiting for eos760d ;) but i don't have time to help to develop it currently
20:42
se6astian
from RAM perspective would the darkframe processing work inside the camera?
20:42
se6astian
in raw2dng
20:43
alexML_
darkframe creation?
20:44
alexML_
yes, it uses an accumulator to average the frames
20:44
alexML_
could be even accelerated by FPGA
20:44
Bertl
se6astian: is that sufficient for QSoC? https://lab.apertus.org/T724
20:45
se6astian
alexML_: great
20:46
se6astian
Bertl: looks good, google lists Prerequisites and programming skills separately but its a gray area I guess
20:47
Bertl
okay, then let's move on to the next one, we can still refine it lateron
20:47
BAndiT1983
i've adjusted the T722
20:47
davidak
can we add a "Your Own Idea" task like https://community.kde.org/GSoC/2016/Ideas#Project:_Something_that_you.27re_totally_excited_about
20:48
davidak
this way we could also get interesting ideas :)
20:48
davidak
to implement ourselves later
20:49
se6astian
good idea
20:49
Bertl
"surprise us!"
20:50
davidak
finally implement that helicopter with the fan
20:50
se6astian
added
20:50
davidak
then we have a drone
20:50
Bertl
alexML_: I'll do a quick structure for T725 now
20:54
se6astian
we will receive students applications and mentors will pick their students from it
20:54
se6astian
if they want a drone the chance they will be accepted by anyone of us is rather small I guess
20:54
se6astian
the ideas/projects they pick are part of the application
20:56
Bertl
alexML_: please extend T725
20:57
BAndiT1983
jsut PWM the fan high enough for that task :D
20:57
se6astian
its a great low hanging fruit
20:58
davidak
it was a great feature of the alpha
20:58
se6astian
T275: we will need a working off the shelf camera/lens assembly to eavesdrop on
21:00
BAndiT1983
for such info? -> https://nikonhacker.com/wiki/Lens_Serial_Interface
21:01
se6astian
yes lets collect references as well in the task description
21:01
se6astian
there is lots of such information already on the internet
21:01
se6astian
about different lenses/systems
21:01
BAndiT1983
yes, on MagicLantern forum :D
21:01
se6astian
I am sure the entire task could be solved by just collecting google search information :)
21:02
BAndiT1983
https://www.magiclantern.fm/forum/index.php?topic=3796.0
21:05
alexML_
https://www.magiclantern.fm/forum/index.php?topic=18308 (very good docs)
21:06
Bertl
I've created T728, please somebody check
21:10
se6astian
made headers bold and added beta software project tag
21:11
davidak
should we improve the organisation of code and stuff before students search for them?
21:12
se6astian
you mean github?
21:12
davidak
and make sure the latest changes are pushed
21:12
davidak
for example
21:12
se6astian
wouldnt hurt of course
21:12
davidak
i think much is in the wiki or somewhere else...
21:14
BAndiT1983
yeah, we have to prepare separate branches for gsoc, imho
21:14
davidak
if we use GitHub the students can create Pull Requests
21:14
davidak
from their own branch
21:15
BAndiT1983
so they should fork?
21:15
se6astian
yes
21:15
se6astian
thats the standard method to commit own changes
21:15
se6astian
fork, push, merge
21:15
davidak
at least on GitHub
21:15
se6astian
yes
21:15
se6astian
ok I gotta leave for today
21:15
se6astian
I think we are progressing very well
21:15
se6astian
deadline tomorrow 18:00
21:15
davidak
yes
21:16
se6astian
but I leave office at 17:00
21:16
se6astian
then I will have to submit the application
21:16
BAndiT1983
alright, i will try to be before that time at home
21:16
se6astian
of course we can still make changes to our lab
21:16
se6astian
but we do not know if google pulls a static html page from the application process
21:16
se6astian
or if a physical person accesses it at a later point
21:16
se6astian
so we better have everything ready tomorrow
21:17
Bertl
what if a virtual person accesses it at a later point? :)
21:17
BAndiT1983
this would be the year 2035
21:17
se6astian
and what if a vertical person does that?!
21:18
Bertl
we have to add some colors ...
21:19
davidak
also take a moment and think how a students firtst impressiom would be with what we have
21:20
davidak
don't assume they know stuff, link to more information about our project etc
21:20
davidak
it is also a way to promote our vision
21:26
Bertl
definitely
21:29
RexOrCine
Whats the project assignment list called?
21:29
Bertl
hmm?
21:30
RexOrCine
The ideas list, whats it referred to inside GSoC?
21:30
RexOrCine
referred to as*
21:30
Bertl
no idea
21:30
RexOrCine
It was mentioned by someone at one point.
21:31
BAndiT1983
https://docs.google.com/document/d/1n7BxN4UFok-JP47zsJ2w6oAErJ1aFpHtssCNx4dB9P0/edit#
21:31
BAndiT1983
this one?
21:32
RexOrCine
Yeah. se6astian might know. I think the term might have been laid out........
21:33
RexOrCine
Ah, workboard?
21:33
RexOrCine
Must be.
21:34
Bertl
I think that might be the term phabricator uses
21:34
RexOrCine
Ah.
21:34
RexOrCine
It'll be in the application stipulations.
21:36
BAndiT1983
left the channel
21:42
comradekingu
Do get the "obselete" changed to "obsolete"
21:57
Bertl
I'm off to bed now ... have a good one everyone! thanks for the QSoC help!
21:57
Bertl
changed nick to: Bertl_zZ
21:58
davidak
good night Bertl
22:17
Rex0r
comradekingu - Where's this?
22:17
Rex0r
Flyer?
22:22
Rex0r
Ah. I see. The Nofilmschool quote.
22:29
comradekingu
Rex0r: It is in the quote on a few of the images
22:29
comradekingu
yesyes
22:36
se6astian
off to bed as well
22:36
se6astian
good night
22:36
se6astian
changed nick to: se6astian|away
22:55
jucar
joined the channel
22:57
dimaursu16
left the channel