Current Server Time: 21:57 (Central Europe)

#apertus IRC Channel Logs

2015/12/30

Timezone: UTC


23:22
troy_s
Okie. Thanks Bertl_oO .
23:23
troy_s
philippej: ^^. I will also create a task.
23:28
pozitrono
left the channel
23:29
troy_s
Apparently I have progress!
23:30
Bertl_oO
don't let it get too far :)
23:36
troy_s
It didn't. I mashed buttons and got three email verifications, but now when I add my original email it stalls out again.
23:36
troy_s
Ugh.
23:37
Bertl_oO
you sure your "original" email is not filtering the messages out as spam or something like that?
23:41
ar
3r/37
23:41
ar
sorry
23:41
Bertl_oO
no problem, nice to hear from you ar!
23:41
ar
n/p
23:42
ar
the one thing i'm actually interested now is a full-frame still-photo camera
23:42
ar
but i'm not sure if such sensors are available and suitable for your project
23:42
Bertl_oO
the CMV20k is a full frame sensor
23:46
ar
Bertl_oO: it's not listed at http://apertus.org/axiom-beta-imagesensors, right?
23:47
Bertl_oO
correct, but we have been investigating it and a sensor frontend would be possible
23:48
Bertl_oO
http://www.cmosis.com/assets/generate-single-pdfs.php?products=cmv20000
23:51
ar
specs look fine, as long as my totally lame eye is concerned
23:52
ar
i'll need to get a second look when not drunk (i'm at the warsaw hackerspace assembly at 32c3 now and i just finished another long island)
00:13
morethink
joined the channel
00:13
morethink
left the channel
00:16
morethink
joined the channel
00:16
Bertl_oO
off to bed now ... have a good one everyone!
00:16
Bertl_oO
changed nick to: Bertl_zZ
00:53
intracube
the cmv20000 is 4:3 format which makes me think super35 dimensions
00:54
intracube
so similar to the cmv12000 but higher pixel count?
00:54
intracube
and not FF 35mm still format (3:2 aspect)
00:55
intracube
odd that Cmosis don't just list the width/height on the spec sheet
01:33
morethink
left the channel
01:51
troy_s
intracube: Square aspect
01:51
troy_s
intracube: On the sensel sites I suspect, so you should be able to deduce the dimensions
02:22
morethink
joined the channel
02:49
morethink
left the channel
03:24
morethink
joined the channel
03:52
morethink
left the channel
03:58
pozitron
joined the channel
04:02
morethink
joined the channel
04:17
morethink
left the channel
04:27
morethink
joined the channel
04:32
morethink
left the channel
05:52
Bertl_zZ
changed nick to: Bertl
05:52
Bertl
morning folks!
06:10
alexML
morning
06:18
troy_s
Morning all.
06:21
troy_s
Ffffffff
06:23
Bertl
2^32-1
07:35
troy_s
Bertl: Smart ass.
07:36
troy_s
OK. One earthquake per year is enough. Good night.
07:42
Bertl
nn
07:49
Guest18375
left the channel
08:29
pozitron
left the channel
08:41
LordVan
joined the channel
09:07
se6astian|away
changed nick to: se6astian
09:08
se6astian
good morning
09:08
Bertl
morning se6astian!
09:29
aombk2
oh great, the couch video is out
09:30
se6astian
yes :)
09:32
Bertl
good that it wasn't delayed much :)
09:32
se6astian
they claimed they were just waiting for the perfect moment to release it :P
09:35
aombk2
heh
09:56
aombk2
left the channel
10:15
Diggi
joined the channel
10:30
danieel
ar: 12bit and 41dB SNR is very poor for stills camera, the architecture of cmosis is not optimized for that (GS only, 12 bit max)
10:30
Diggi
left the channel
10:37
Bertl
off for now ... bbl
10:37
Bertl
changed nick to: Bertl_oO
10:48
pozitrono
joined the channel
11:19
ar
danieel: thanks
13:34
mimi
joined the channel
13:34
mimi
dear all from apertus, is someone from you still at 32c3 congress?
13:36
se6astian
I think matthias is still there
13:36
se6astian
pmed you his email address
13:36
mimi
until when? If I find time between the talks I'd like to visit you
13:36
mimi
thanks a lot
13:37
mimi
I'd just like to see the camera in real :)
13:44
se6astian
great, just email him, I have no idea how lon he will stay
13:44
se6astian
changed nick to: se6astian|away
13:54
mimi
left the channel
14:01
DoomRainer
joined the channel
14:39
parasew
joined the channel
14:42
parasew
hi i am still at the congress today, in case somebody wants to meet up best would be at 17:45 or 20:00 at the crypto bar (in front of saal g)
14:44
se6astian|away
changed nick to: se6astian
14:46
parasew
left the channel
15:19
LordVan
left the channel
15:20
Mined
joined the channel
15:21
Mined
Looking at one of your speeches on YouTube
15:21
Mined
Awesome project
15:34
pozitrono
left the channel
15:43
arpu
joined the channel
15:47
se6astian
Mined: thanks, let us know if you have any questions
15:51
aombk2
joined the channel
15:53
aombk2
se6astian, i was worrying you may come out as hostile but you did good on that couch interview. nice :)
15:57
tyrone
joined the channel
15:57
tyrone
hi all...
15:59
aombk2
hi tyrone
16:02
aombk2
changed nick to: aombk
16:03
aombk
changed nick to: aombk2
16:03
tyrone
is sebastian active? or are all in party mode :-)
16:03
DoomRainer
left the channel
16:03
tyrone
it's quiet....
16:04
Bertl_oO
is that so? :)
16:08
_tyrone
joined the channel
16:09
_tyrone
ohh the dev master is active :-)
16:10
tyrone
left the channel
16:17
se6astian
we are always having a party in here
16:17
se6astian
aombk2: thanks, I wasnt sure how exactly it would look from the outside as well :)
16:26
aombk2
Bertl_oO, when will we see you in a debate?
16:29
Bertl_oO
whenever there is something technical to discuss :)
16:32
_tyrone
nice to hear :-)
17:24
John_K
@se6astian @Bertl_oO: is there an easy way to trace where the different voltage domains (NW/NN/NE) go to on the different boards?
17:25
John_K
trying to make sure I put the right pots on the right LDO
17:53
Bertl_oO
yes, the schematic for the sensor board explains which voltage is what
17:54
Bertl_oO
(for the sensor) including the proper ranges
17:54
Bertl_oO
http://vserver.13thfloor.at/Stuff/AXIOM/BETA/axiom_beta_sensor_cmv12000_zif_v0.18_r1.2.pdf
18:14
gunnar
joined the channel
18:30
Bertl_oO
hello gunnar!
19:08
gunnar
Hi, good evening, Bertl
19:10
gunnar
(Sorry for the delay, I was watching your 32C3 presentation, and then went astray on your website)
19:12
Bertl_oO
no problem, IRC is perfectly fine with long delays
19:14
pozitron
joined the channel
19:15
gunnar
Would it be technically feasible to use alternating, different exposure times with the Axiom Beta? (I'm thinking of an exposure stack needed for HDR)
19:15
Bertl_oO
sure, why not?
19:16
Bertl_oO
the CMV12k also supports two HDR modes natively
19:17
gunnar
*searching for CMV12k*
19:18
gunnar
Is the CMV12k the sensor used in the Beta?
19:18
Bertl_oO
yep, the k is for thousand, i.e. the sensor is called CMV12000
19:19
Bertl_oO
https://github.com/apertus-open-source-cinema/alpha-hardware/blob/master/Datasheets/datasheet_CMV12000%20v1.12.pdf
19:19
gunnar
So I should RTFM, (and thanks for the link, you beat me...)
19:19
Bertl_oO
it's a slightly older datasheet, but it has the relevant parts
19:20
Bertl_oO
you can get a newer one from cmosis without any problems (i.e. no NDA)
19:21
gunnar
That's rare (the _without NDA_ part) , and very positive. Will look into it right now
19:22
Bertl_oO
it is one of the reasons we are using the cmosis sensors
19:22
Bertl_oO
no point in doing FOSS/OH with an undocumented sensor :)
19:25
Bertl_oO
chapter 5.3 are the HDR modes
19:25
gunnar
page 32, just reading it....
19:26
se6astian
latest datasheet is actually here: https://github.com/apertus-open-source-cinema/beta-hardware/tree/master/Datasheets
19:26
gunnar
CMV12k even supports linewise interleaved exposure times, sweet...
19:26
se6astian
CMV12000 has version 1 (150 FPS) and version 2 (300 FPS) other than that the datasheets should be similar/identical
19:27
aombk2
left the channel
19:27
gunnar
Do you use 8, 10 or 12 bit output, or do you plan to make this user selectable?
19:30
se6astian
12bit currently
19:30
se6astian
user selectable would be prefered at some point in the future
19:32
gunnar
12 bit is probably the best choice for now, I agree, and probably one of the main difficulties is to level feature requests versus implementation manpower....
19:33
gunnar
So, did you do a signal / noise calculation on the 12 bits? How many bits "hold", how much frame-by-frame noise is there, roughly?
19:37
se6astian
http://www.magiclantern.fm/forum/index.php?topic=11787.msg129672#msg129672
19:43
gunnar
You guys are amazingly helpful, thanks, will be reading a while...
19:48
se6astian
you are welcome :)
20:26
tyrone
joined the channel
20:28
troy_s
How goes it se6astian ?
20:28
troy_s
irieger: You have been MIA for a bit. Awaiting your return.
20:29
se6astian
good good, thanks :)
20:29
_tyrone
left the channel
20:45
pozitron
left the channel
21:00
ar
se6astian: 12bit is all available color space, or 12bit/channel?
21:02
se6astian
bayer pattern
21:02
se6astian
12bit per channel
21:12
troy_s
Bit depth is not a colour space.
21:15
baldand
left the channel
21:19
baldand
joined the channel
21:19
slikdigit
joined the channel
21:20
slikdigit
left the channel
21:20
slikdigit
joined the channel
21:23
baldand
left the channel
21:24
baldand
joined the channel
21:24
gunnar
Read the 11-page thread, and A1ex's improvements to FPN, and it seems that there is still room for improvement. Just a silly thought: Since FPN might also be affected by Vcc filtering, and I seem to spot a FPGA directly on the "Main Board", are you sure you are filtering Vcc and keeping Vcc of the Sensor and Vcc of the FPGA as separated as possible?
21:25
gunnar
To be honest, I just drew my assumptions from a picture of the "Main Board", and did not look yet at any PCB layout files.
21:26
gunnar
And I hop I do not insult you, Bertl....
21:26
gunnar
*hope
21:47
gunnar
left the channel
21:51
baldand
left the channel
21:52
baldand
joined the channel
21:54
aombk
joined the channel
21:54
aombk
changed nick to: aombk2
22:28
se6astian
going to bed
22:28
se6astian
I am sure Bertl will be able to answer your question gunnar
22:30
se6astian
changed nick to: se6astian|away
22:33
Bertl_oO
@gunnar no problem with constructive criticism here ...
22:34
Bertl_oO
regarding the FPN part, an insufficient filtering or noise on VCC would result in dynamic noise on the sensor
22:34
Bertl_oO
FPN is something systematic which is tied to the sensor acquisition in some way
22:35
Bertl_oO
regarding the FPGA, we have two on the mainboard actually, but they use completely different supply rails than the sensor, so the crosstalk should be minimal
22:36
aombk2
left the channel
22:46
aombk
joined the channel
22:47
aombk
changed nick to: aombk2
22:47
aombk2
hey Bertl_oO are you there?
22:48
Bertl_oO
kind of ...
22:49
aombk2
interesting
22:50
aombk2
so i wanted to ask, is there a possibility after all is complete to make any module for a lower end tiny sensor?
22:50
Bertl_oO
sure, you "just" have to replace the sensor frontend PCB by your own design
22:51
Bertl_oO
we will probably give a template to start from, but for now you can simply use the CMV12k one and get rid of everything CMV12k specific you don't need
22:52
aombk2
you dont plan designing more PCBs in the future
22:52
aombk2
?
22:52
Bertl_oO
we plan to do PCBs for the CMV2/4k and the KAC
22:53
aombk2
yes i know about that
22:53
Bertl_oO
if there will be "interesting" and well documented sensors available, we will do frontends for those as well
22:55
aombk2
great. i think at some point a small sensor for macro photography / videography would be useful