#apertus IRC Channel Logs

good morning

morning folks!

time for some supermarket shopping

if you buy one, please take one for me as well :)

back

I have the supermarket home with me now, should bing it over sometime soon? :P

writing reply to steffany now

that was the plan :D

time for bed

hey

what up?

what's up on your end?

In a nutshell, I was trying to wrap my head about the idea of investing time in this project and also being able to have access to a camera...

good to hear that, btw I wrote the first reply :)

I am a bit confused

I know, but are you refering to this morning email?

no, I didn't reply after that, because I wanted to avoid even more confusion :)

(probably a language problem, I'm obviously not a native speaker)

So am I

excellent! so where do we start?

and what are the short terms goal?

yes, well, let me give you a quick history about the axiom and then I'll get to what we are doing right now, yes?

btw, I speak french too, although I am not as efficient in french as I am in English even as a second language

German, I supposed

sebastian was working with elphel back then (the google cameras)

and so they decided to team up and work on some kind of open source/hardware camera

k

at some point, the elphel cameras turned out to be insufficient (low resolution, etc) for the goal

that was when 'they' decided to actually design a new camera from scratch

that he is looking for somebody who can do FPGA development

incidentially that was the same time when I decided to start with FPGAs

cool

good! that's how things happen

sweet, was the HDMI out a 4K output or a 2K? (HD)

and we had to use that, because all the other pins were already used up by the sensor interface

but we are only able to output full HD with the zedboard

but it was a proof of concept and we actually surpassed our initial goals with that one

It is pretty amazing considering that you never touched FPGA before

I am sure it was... I did the same at some point...

which turned out to be a nightmare to deal with

and when was that?

okay, did you convert a lot of people?

we realized very soon that we need to build a camera development kit (as I call it)

because folks started to rebuild the Alpha, which wasn't designed for that

k

i.e. we wanted something which is small, simple to recreate, and very powerful

the Beta is based on the microzed, which is roughly 5 by 10cm

a note here: we are trying to build a platform not just a camera, so it is important not to target only a minority

what is a platform and what is a minority in that sentence?

Neferstefany: meet danieel, our 'competition' :)

i see very limited possibility to build upon such choice

somebody else having a running kac12040 :)

the KAC is not that difficult to design with

well, my design choices are made more into future, and I can use "secret" devices, while axiom tries to avoid NDA parts

as well as process choice is not "build in your kitchen"

You know, I for one just want to have a camera to shoot movie with the best sensor I can find at a reasonable price and because I am technical and I have a very good base right now, I can certainly reach that goal... If I can help others along the way, well so be it

which sensor?

yeah, rolling shutter sucks

even the high end cameras now build work arounds for the rolling shutter

depends how fast you roll :) sometimes it can be accepted, it still got less noise

I just want to have a global shutter before and now, I could not care less, I want pixels, Color Depth and a lot of Dynamic range...

the 2000 sucked as much as the 12000 do

Interesting to hear that

cmosis is raw cmos :)

it is a tricky sensor to adjust

and lot of post processing to get a decent image

so what is the point to pay for an over priced sensor and to have pay extra for FPGA power... that's my question right now

that cant be true bert :)

So, there is more offering now, Altasens, Trusense, Aptina and CMOSIS, all of them have their kwarks but there is hope to get a decent sensor

we are limiting ourselves to sensors where documentation doesn't require any NDA

@Bert1, I understand, in your case, it is more difficult

Gopro suck, but at some point everybody hated it and now, they are fucking making feature film with it...

it always the same story

YES! ;)

hehe, yeah, so it seems ... so, back to axiom, we are currently designing and testing various parts for the axiom beta, which will soon get a crowdfunding campain to fund the development

for example we are testing the power supply as well as options for clocking and various I/O parts

we have created a list of different perks for the beta, starting from small to large

I haven't done any SDI yet

the SDI and hdmi in 4k

hdmi should be already covered, at least the encoding

4:2:0 in 60p is somehow not meeting my standards

the real hdmi 2.0

the rest is just timing and I/O lanes

even if we could output it by gtp's, you cant view nor record it

sony? :)

Neferstefany: that is good to know :)

28" UHD, 60P Display Port

DP is a smarter choice

and so SDI

what do you mean?

That's a problem... But I thought the whole platform was a module approach, so my thoughts were to have a module with a Artix with Serdes that could do all that

That's was my idea...

In a way, I am competitor too? I do have my own camera design

fight! :)

where are you at then?

but I had a Eng Sample of the CMV12000 which is crap and I wanted changed it but a lot of things happened in my life so I had to shelve it

So, now I want to go back on it, change the sensor either to a 4K or 3K and finish the recorder so I can record in full 12Bits raw no compression and get my PCI express interface working...

Well, I have a day job... so that takes care of that. In the near future, I would like to design a KAC board for my camera that would support the 12M and 6M sensor they have now... and start looking at your project... again, my idea would be to make a module with a small Artix the 35 I guess and do the sDI and HDMI

and some kind of healthy information exchange regarding the sensor

I/O Shield?

it doesn't have a name yet, but yes, here is roughly how it will look

Do you have a larger view of the concept (a picture) I am not sure If I understand it

this image is actually LR flipped

is this a platform for development?

you are probably referring to the AXIOM Gamma design

yes

(at which point we should have completed the beta)

well, you can make it larger if you don't mind the 'camera' getting larger

Okay, I will look it this week... I am in the middle or re equip myself with decent equipment, when I left Los Angeles last year I had to sell most of my stuff including my super computer... so for almost a year I have been using a Core 2 duo (8 years old computer) computer

It makes for very long compilation time with Xilinx!

In the near future I am looking for at least to get an i7 (overclocked)...

it is said to be faster

what Vivado?

we are currently using vivado 2014.1/2

We are using the same at the office... but since I still dealing with the big Artix, I am still using ISE

Bertl: would be interesting to run that on the server to produce night builds and error reports

danieel: yeah, we will probably do something like that soon (for the beta)

cya

luckily your part is supported in free edition

(for a server build)

so how is your camera project going? anything new there?

thinking to go for a larger sensor

(personally i do not care)

not sure about that... i would rather go for panavision