#apertus IRC Channel Logs

\o/

mh?

Hey yeah following this closely. Fantastic is all I can say. Always been a fan and run Magic Lantern the pairing of ML with Axiom is perfect.

the only ppl out there, thinking something else are BM, Sony, R3D and Arri :P

Rebelj12a, why "Hmm"?

aha

that's fine

Yeah I'm all in for gamma though. Will there be crowd funding for that model?

But I had to think about it for a while too. I figured, I'm in no rush to upgrade to a completely working retail 4k production camera, so it seemed like a good idea to help support the Axiom campaign/open source movement in the mean time, and I'm sure we'll end up with something awesome in the long run. :)

Oh yeah, someone will. I've been looking at android Dev and other options. DSLRController even though its not technically a DSLR would be great.

Been working on a (at some point) wireless follow focus and lan arduinoc control for the BMCC. I mean I'll open source it and how to make it once its done, the programming itself is quite simple in all honesty. Everything builds when more people have knowledge.

its amazing the campaign funds just go up and up

you could ask indiegogo to extend the campaign for a year

maybe, but we want to make a Beta :)

it would be awesome

right

but my campaign is not going so well

yeah, I checked regularily ... unfortunately it is still in the early stage

Rebelj12a: have you considered working on a follow focus for the Beta yet?

Oh well the follow focus I'd program again is simplistic. It will run off arduino and connect via 15mm rod to any system.

anybody knows about finished campaigns? can i edit them after they are over?

Just the parts. No images yet. Its 1 of a few projects I have going. Still buying the arduino pieces as I can. Touch screen, SD card shield, for prototyping. Saving money by recycling knobs and buttons from old electronics however final product will be similar to the Redrock version, although larger, dual 15mm attachment, since I want to put a gearing system in for different speeds.

care to show us?

Let me find them

Ah bollocks its at the office. Don't know about you guys but its 8:15pm here hah

no problem

Oh wait found it other notebook hah. Oh jeeze late night.

http://imgur.com/wBZMBDb

http://imgur.com/7GurcMr

Started as a personal project to link h4n recording and BMCC recording. Then I was just disappointed the power of arduino would be so underused just for that so.

Ah current is success with activating recording and built in lanc functions using arduino. Friend of mine has a 3d printer, going to fabricated a control module and run it to the arduino so all the buttons are mapped. Wired connectivity and buttons working-then wireless-then follow focus module and programming-wireless between controller and arduino, presets and auto saved settings for specific cameras and lenses. Auto focus end

I see

so this 4k sample dng you posted on the campaign is without olpf?

Either way time is my enemy. I'm not a Dev or manufacturer, gotta make the money first. Plus the Ronin is giving me a devil of a time. Might have to send it back think there's something wrong with the tilt motor.

mostly because I'm interested in learning about those things

it might not be obvious, but I'm somewhat new to all movie camera

Everyone starts somewhere. Hah

O.o Black magic camera update brb

BMCC doesn't have timecode sync. Yeah telling them to both start at the same time.

or is it like telling the BMCC to start recording the delay could be anywhere from 10ms to 300ms sorta thing?

I haven't had a chance to test it. Its a side project. I'm assuming there will be some delay since its over lanc. To reduce delay I was researching directly bit driving the outputs and bypassing the fabricated interface. Haven't done enough research and testing to make sure that will work yet.

sounds like an interesting project. Hopefully since both devices use solid state storage the delay it consistent and thus you program around it

H4n zoom uses SD. Either way that I can always change timing by position of the initial arm and record press

have a good one and cya!

se6astian: Congratulations! Can't wait till April.

hi Paperhat, probably yes, we will thoroughly check prices once we have the total number of cameras

uh oh

(Y)

time to go to wokr

Would you mind looking at my thoughts on heat management/body design? I know that document is way too long - I really endulged myself :)

I see you are using freecad just as I do

nice morning folks!

__anton___: good document, needs some time to work through...

mornings :) felt like rejoining for the last few minutes of the campaign

yes, incredible. would have been enough for another stretch goal *grins*

indeed

whoa, will mechanical amount for that much?

1000 EUR is not the estimated price for mech. parts but you cant get a plastic body for 5 EUR. Even the mold for one are +30k EUR.

There are a couple of very precise parts where we can't save money...

lens and sensor mount i presume

Congrats guys! That's one heck of a finish!

official countdown

Actually I had rather nervous dreams :)

as the last perk was running out....

90 seconds!

20

5

3

1

over!

yipieyeeee

Well, now what? :)

on the bermudas..

or cayman islands...

congrats everyone!

thanks!

the hidden goal was at 175k :-(((

congratulations se6astian and everyone who worked hard to reach this !

Ok, once again congratulations! Coming from the -ahem- last funder of the campaign!

KurtAugust: you waited for the last minute :D

and wasnt able to donate anymore :)

you could extend the campaign "in private" with donations on the apertus page. leave a message on the campaign site...

maybe someone should sell upgrade trip flights to vienna

thanks :)

some stats

€ 36,000

yeah, i went to sleep when it was around 135k

1.5k per hour. impressive!

moin

vienna i suppose.

maybe they are already off celebrating :)

i opened one yesterday. for my birthday, and the best bday present ever :)

:D

for windows users: http://dimio.altervista.org/eng/ the tool Dsyncronize rocks!! (syncronizing two folders in realtime)

daFred_, we'd like to make it fair for those who trusted us and comited -in time- for the campaign. Else in 5 years we'll still be selling axiom beta vouchers:-) Otoh, the donation page is up. Beside, we'll come up with a solution quickly for those who discovered the project too late :-)

i didn't mean to sell anything with this link, just to use the opportunity to support the development further.

please dont

ok..

its just 3h after campaign end

omg... after weeks of using kaspersky on my laptop which handles my internetconnection, it decides that my homenetwork is not trusted anymore and blocks all traffic on it ^^

se6astian: can we build something liike this: http://i00.i.aliimg.com/img/pb/433/887/498/498887433_749.jpg pleeeeeeeeeeeeeeeeeaaaaaaaaaaseeeeeeeeee!

or something like this : https://www.youtube.com/watch?v=cVIr9MyqLqc

I had more something like this in mind: http://i.ytimg.com/vi/nE1fzM6MD14/maxresdefault.jpg

I am eternally sorry, I said I wanted to contribute a symbolic donation but didn't do it... I watched it evolved and thought "oh I'll do it near the end..;" but now the end has passed ! hahaha ! stupid...

did you people believed it, that the campaign would turn out so successful?

aombk: mid campaign it was even hard to believe we will reach the goal at all

i wonder what triggered the popularity increase, apart from the ml support. or is this what always happens to crowd funding campaigns?

not really :)

but maybe its just that people wait till the end to decide

working on it now

moin Bertl

Morning Bertl !

I think what happened is pretty common in crowd funding

se6astian: actually all I really wanted was the t shirt and stickers :)

FergusL, if you suggest something great to visit in paris i can give you my apertus button

If you're there this week end we run open doors days

Hahaha good idea yeah

it wont even have openings for your arms!

yes, the text will be totally unreadable!

Guys, could smb tell me how Nikon lens mount is going to be made? Is it planned to be 2 parts: barrel + ring? How is the barrel going to be made? How is it coated black on the inside?

Congratulations!! I expected it to cross 150K last night but 170K??? WOW!!

habit

Mine is the 2nd batch but financially after May will be much more comfortable for me

regarding soldering a lot: http://www.societyofrobots.com/robotforum/index.php?PHPSESSID=8c46817f27ae58954d0ef2b87848c18e&topic=2770.msg19337#msg19337

for the Beta we will streamline that process a bit :)

Umibuta: sure, you can decide when to get your Beta

but you can always choose to wait

Thanks!! That's good to hear and I understand your concern on economy of scale :)

Coating the inside black? If it's aluminum what about anondizing?

__anton__: sure why not

1. put the bayonet ring at the right distance to the sensor (FFD)

E.g. can it be made on a processing center from a larger brick of aluminium.. wanted to know how expensive that is.. if we are anodising we have to anodise the whole thing right? And it gets more expensive with size too doesnt it?

not sure if anodising is a good coating for reducing light reflections next to the sensor

a number of tricks can be used to make the inside non-reflective

spraying like with car parts for example would work

but please folks, collect all the ideas on the wiki

no but we need to be sure whereever it is made that people are paid fairly and there are health and safety standards that are on paar with inner EU ones

http://www.motip.com/products/motip/automotive/universal-lacquers/lacquer-black/

I had some stuff made in Singapore. Milled and anodised. How much it compares to EU and the shipping/taxes involved might be a concern.

quick front end builder, might be interesting for us : https://github.com/BrentNoorda/slfsrv

I was looking for some of the drawings that I had made. The dimensions are similar to the size of the Beta. I could get a quote for you just for cost comparison purpose if needed. No harm asking.

Also a question/ suggestion. How different are the barrel size for different lenses? If they are not too different, would it be better to have the barrel part of the body?

adds another interface with tolerances

as it gets down to business now, i want to drop two links, which are very important to me and my future. i want to work like and with this in all the projects i engage with: https://www.ecogood.org/ and http://www.mitgruenden.at/ - both links are in german and mostly relevant for businesses/associations/companies and individuals, from/located in germany and austria

derWalter: Thanks. It is interesting to read how community/economy ethics have evolved.

wow, thats a hard one, this needs either to be cast, where you need a casting form and molten alloy or milled from a solid block

alloy is a very very very harmfull ressource :/ better to focus on organis fibers

wouldnt even be that much work

You mean it costs lots of pollution to get aluminium?

ULTRA lots

We must use Norwegian aluminium

hydropowered aluminium

just get sheet alu and bend it, thats cheap and easy, albeit a bit timeconsuming

hrhrhr: http://eramet.no/en/environment-and-climate/

http://www.cbc.ca/news/world/red-mud-toxic-waste-of-aluminum-refining-1.906411

__anton__: the russians have atomic power that makes a bad joke of even the simplest safety measures. And their mining is not very clean

don't get me wrong, I'm all pro eco system but it's not the materials or the production, it is the careless handling or cheap production techniques

be back later!

and its the planned obsolecence

a modular system has a pro and a contra side to this, pro side, old parts can be reused, contra side you will more likely change a parts of the system, more often, as you dont have to switch the whole system anymore. even on the contra side there is the pro that you switch just one part and not the whole, but! the danger is to use cheap materials/building processes so the hardware fails more often, but doesent hurt so much as sin

... but thats a thing on which i get going on for ever, sorry for that... i am just so hurt by the way we are treating our world and the planet we live on ://

derWalter: I am totally with you here, things should last

that is a big hope for me regarding apertus, to build a real qualityproduct, which lasts and like good wine, just gets better over the years :) grows and grows but doesnt break :) a hardcore hq camera.

philippej_: problem with the client?

now should be ok, sorry for all those "has quit"

€174,520 :D

thanks :-)

intracube: thanks!

testing phabricator, it's a very exciting tool :-)

Ah ok, will there be an emulation software possibly. Came across the idea when I was looking at Arri's virtual camera. Something like that but also with a programming emulator similar to how android can test its apps without breaking a physical phone. Might make it safer for more complex programs

it can be as low level as hardware design (fpga code) that is supposed to be emulated, but I don't know if it really works in practice

then you have the linux layer, meaning you could test stuff on a similar linux install on a normal computer

and probably some arduino code (or similar) which is also testable

All my of my favorite codes. Yeah I was just thinking there are people who are going to want to push the limits of the system. I suppose its more complex than android though. Where emulation as proof of concept beforehand is.more difficult

you can grab a microzed dev board and start right now

the most expensive device, the sensor could be emulated as well, but it's not a trivial task

I like that solution, I'm just tossing ideas around so.

Especially if one becomes quite popular.

People charging for themes or templates etc.

I've had a dystopian dream where Linux was no longer open and everyone charged for updates or features... D:

Yeah that's why I was wondering. Then on the other hand with it being Linux people can program in all sorts of licensing schemes, I'll draw my comparisons from working with Minecraft which is not open source but hey its java and lots of people can code it. Where software that works with it can become very popular on its own, however certain greedy administrators take that open source code API and charge for features of it for "e

Hey congrats guys

buying the sensor and plugging it in

thats the concept of open hardware

Got it derWalter...may be I'll try to put my question this way

Hm although wouldn't the Linux kernel need to be updated with drivers for the new module?

well, i guess u have to unplugg the camera from the powersource anyways

Well yes but on a base level in order for Linux to register new hardware it needs to have the kernel drivers modified, especially if the new sensor has different specifications or wodnt you be limited to the previous sensors capabilities since none of the operating system functions have been updated?

One would argue this one has the potential, if not already is. Granted the Alexa has a whole slew of nice features.

level in terms of best sensor available in the market

Its not all about pixel peeping D:, arguably though right now, you could buy the Alexa sensor and put it in the Beta right?

There is no such thing as a "best sensor". There are different things in what it can be best.

and we really get a lot

something like a PC-based raw recording solution... wonderful

Although couldn't this be a discouragement from any major manufacturers to create accessories for the beta. Since the market isn't guaranteed because of just how customizable it is. There's no guarantee of a market for a particular item.

Anyways yeah.

1080p, 2k ... no difference :)

For vfx and 3d compositing, big difference

canon has the edge on low light, for sure :D

BM 4k has c12000 right?

afaik

300gps@4k, thats around 8gb/s

Yes man, it's fun....The philosophy is what attacted me to this not the technology at this moment

Any sensor experts: what percentage of an image's credit can be attributed to the actual sensor and what percentage to tweaking?

I mean the sensor may require more or different processing than the built in one. However all those things can be addressed with the open source programming.

well, what the axiom team has done is really impressive with the alpha. and same for ML with ... mlv and dual-iso. fun times ahead

Rebelj12a, thats the big advantage, processing is reprogrammable... FPGA

you can even push the fps by doing so

afaik there is no fixed processing

alexML uttered some thoughts along those lines, extrapolating something like 800 at 1080p

yes

Rebelj12a, the FPGA is already very powerful

he went down the math path till the theoretical maximum which is something over 1000fps but at very very very low resolution

and the axiom is a linux computer :)

So what CPU will it use?

I am not talking about color grading here, how much percentage/stops of DR, Lowlight perf can be improved with tweaking at the programming level

FPN correction, LUTs, all that stuff

_Raj_: If you can do oversampling, you can get more out of it. But I have no idea how to do that properly with something that is moving.

That is, shoot at say 300 fps, reduce it to 60 fps

Dual I so and such is a great software hack to expand this. That's why the ML announcement was so exciting to me at least.

i think its all in the campaign updates

thats why apertus insisted on choosing sensor with open and freely available datasheets... afaik

im really happy MLV gets on board too. means the postprocessing workflow i use (chmee's raw2cdng) will be working.

those hooks are open and documented on the axiom. no reverse engineering

Yeah exactly

afaik no audio yet. but.. you see, on the remote stretch goal i had seen something that looked like timecode ;)

Ah ok, background I'm more of a hobbyist engineer at this point but as a filmmaker I know there are a few pillars of good film. Audio is one of them and the built in audio on the BMCC leaves much to be desired and is the cause of many complaints.

<-- audio engineer

Ah good I'm glad its handled then.

The h4n and the BMCC the preamp leaves much to be desired. It seems the ticket to good audio will be a decent preamp.

I think with this project there will appear some mini opensource 4K RAW recoding PC-s too.

Rebelj12a: Forget the h4n. Terrible preamps. Yes. I just got a tascam dr-60. Also cheap and much cleaner preamps.

But I admit I love MLV with sound on the MLed 5dIII. So much less work in post.

Yeah I have my h4n connected to a Redrock power distributor.

guys, we should speak about Axiom :)

i don't even need sound on my beta =)

From a filmmakers perspective audio is just one of the things I've seen a lot of complaints about with the BMCC. Which is why I brought it up. Didn't remember reading anything about audio on the axiom

there will be nice modules in the gamma =)

Awesome

PCIe to SSD maybe?

Hopefully something that can take more punishment. Pcie connections can be delicate. Plus most ssd are sata connections

just my take

nice autotune vid :)

Rebelj12a: there are PCIe SSD boards

the problem will probably to make it easy... and cheap.

HDMI feels like the way to go for the axiom. as we have 3 ports already

then a PCIe SSD board could be used for buffer

and after that transfer the data to PC

I'm just thinking loud

youd need that as sustained write.

same here, also thinking out loud :)

yes

if the CPU/whatevs can handle realtime jpeg lossless compression (which i doubt) you could essentially halve the bandwidth

1 DNG is around 24-25 MB as I know

Unless it was processed directly using an algorithm, like what RED does with its processing.

heh

i'm not engineer, what closed ip core mean?

they have license that you can only use with the xilinx zynq fpga's

Dedicated cores for for just PCI express?

Hm

the campaign said it would be a PC recording solution

it's vhdl code that you get with the zynq fpga's and you have license restrictions so it cloud be problematic to publish public and share.... (not 100% shure)

and the axiom page hints at packing RAW files into a HDMI stream

That would be brilliant. What is it hdmi 3 now? Can't remember the revision but it has built in Ethernet capability.

tyron_: i see, but then how to transfer 25 MB x Y frame to the PC?

ah ok :)

opencores.org has a lot ip cores but they have to be portet to the zynq and also testing if it works...

i gather that in/camera raw recording will be on gamma... beta sounds too early

but i will let myself be surprised positively if they manage that

sure, its just that there is still so much to do

HDMI2 is out on TVs

but not on that axiom board i suppose

1.4 at most i guess

surami: it's not impossible obviously, but pci-e would require us to write our own pci-e core, which is a significant amount of work

http://wolfcrow.com/blog/what-is-6g-sdi/

A single 24Bit 4:4:4 (8bit per channel) pixel can actually store the value of two 12bit RAW pixels (2x 12bit = 24bit). So with 1080p60 we could already store 12bit UHD RAW 30 FPS video.

#axiom gamma

well, we will find sooner or later :D

hi yanez

dmjnova: yes I can imagine that it's not simple, I don't have this kind of knowledge, I just started to search the availabe options on the board and think loud

maybe I'm already at Gamma again :D

a simpler protocol is needed probably

yanez, we were hypothesizing a bit

but where to go with all that data :)

4k

12

each HDMI channel can handle 4k30frames at 12bit

and how to put it on an SSD eg.

ok, and how you record the data?

or money..

yanez, yesterday you were proposing some kind of early timecode

(timecode is a lot of beeps that get recorded on one audio channel. which can be transformed by software automatically to a timestamp). that kind of audio can sync up automatically with the video (if the metadata has the right timestamp)

i look forward to the work done on lossless jpeg compression

still 800 gigs per hour *coughs*

http://en.wikipedia.org/wiki/Timecode

anybody has experience with wordpress gallery plugins?

so in theory we can get 12bit 4K 30 fps RAW out from one HDMI

surami, thats what se6astian has postulated, yes

then we need a "frame grabber" what converts that data into DNG somehow?

converting the datastream into DNG, CinemaDNG, MLV

ML Team has already said they will support MLV format for the axiom. and theres converters for that

i know ML

but HDMI to DNG or MLV is how?

there will be a converter... either by axiom, ML or joint effort

i see

cya

ah too bad, i tried to find a good timecode explanation for yanez

Sorry, had to yell.

campaign details and analysis published: https://apertus.org/crowdfunding-campaign-details-analysis-article-2014

whoa, what a spike

nice post!

seku: I'd say the ML spike

most campaigns get a spike, but ML and stretch goals multiplied it a lot

good night for everyone!

Yeah sorry traveling on the road. Need to verify by sasl somehow when on cell data

Blast missed the audio discussion train