AXIOM Team Talk Volume 12.2



Small Hardware Changes/Upgrades here and there

In this team talk, Sebastian gives a brief rundown of some recent hardware changes to the AXIOM Beta development kits. Changes include; different chips to reduce the Lattice FPGA programming time, a new connector to attach the main cooling fan (which is now held in place by noise reducing nylon screws), some schematic changes relating to the JTAG interface FTDI chips (to reduce power feedback), a new onboard sound buzzer. Also, to reduce mechanical stress on the electronics, we now have new precision manufactured nylon spacers to align the PCB stack at exactly the right distance.



AXIOM Beta Assembly timelapse

This timelapse (yes, it actually is sped up - don't believe the alternative facts shown in the video itself! :) ) shows the entire build process from the initial sensor cleaning, to attaching the PCB boards to each other and fixing all the mechanical components in place with the threaded rods and new nylon spacers. We finally attach the plugin modules, which are held in place by some small custom 3d printed clips.


AXIOM Beta FPN compensation

Camera FPN Calibration Demo

We show you the entire process of capturing 64 dark frames with an AXIOM Beta developer kit and how to process and apply the Fixed Pattern Noise or Row/Coloumn Noise compensation, which is applied inside the camera by the FPGA in real time. Details about the process can be found in the wiki. We also show that after the FPN noise is removed, if we contrast enhance the image, "dynamic row noise" becomes visible. This is a CMOSIS global shutter specific phenomenon. We can already compensate for this noise type for raw still images by utilizing the so called "black columns" along the edge of the sensor - but that correction process is not part of this video and doesn't work in real time yet.


New Team Talk Naming Scheme

Note that from now on, Team Talks shot on the same day will be released under the same volume name (like Volume 12 now), but episodes released with a decimal number added to the end, e.g. Volume 12.2. The next release would be Volume 12.3, etc. until we go to the studio on a different day to shoot Volume 13.

Project: 

19 Comments

1 month ago
David

I like the FPN correction you did. I worked with some cameras from big manufacturers where this could have been a livesaver (or at least spare my nerves).

One thing though: if you adress complex hardware changes, either boil them down to something simple so people who got the camera won't worry or go more into depth. Explain what the problem actually was (so we can really understand), how it is fixed, what it affects and what it won't affect. I am a audio hardwaredesigner myself and had a hard time following you. I dunno how people with little or no electronics background would feel about watching this : )

4 weeks ago
Sebastian

Thanks for the feedback. This team talk episode was indeed more technically indepth than any other before but this was one of the things requested in the recent survey we ran. We are aware that its not for everyone - we will occasionally throw in these tech episodes. Maybe we should label it more clearly though...

1 month ago
Brian

Damn Sebastian, your hand eye coordination is insane. you da boss ;)

P.S. I always get excited when I see updates from you guys. Eagerly awaiting a finished product.

1 month ago
Kimbray

How close are you to having betas with assembly and all boards complete for later beta users? When are you going to have a Canon EF lens mount as well? Thanks!

4 weeks ago
Sebastian

With the electronics manufacturing challenge now getting closer and closer to being solved: https://www.apertus.org/picking-and-placing-hardware-production-progress... the main things left for having a more end user friendly camera is the full enclosure manufacturing and the software. Its hard to say when exactly this will be "good enough" but we hope that we can shift our focus from pick and place robots and assembly processes towards these things now soon.

Canon lens mount is already possible currently with emount to EF adapter. We use several Canon mount lenses with the Beta already.

2 weeks ago
Kimbray

So what is the ETA (roughly the month) when you will have a full enclosure, usable as a daily video camera? Thanks!

4 weeks ago
Jose

Much better video, I love the real time timelapse, but some things:

In the first part of the video a fan noise is heard all the time, for me is disturbing. Is not that hard to remove it with just one extra microphone.

Making yourself a teleprompter and actually looking at the camera is not that hard, people emotionally don't like you looking to a side, it is what someone lying badly to you will do, so it makes you untrustworthy:
https://www.youtube.com/watch?v=uqo_W6tqnf4
https://www.youtube.com/watch?v=7wcVI2B6kXI

There is a club called toastmasters all of you could go in lots of cities in Europe where you could improve your speaking skills by orders of magnitude in a short amount of time. As open source entrepreneurs you seriously need this thing, you need to get your message across to people and Apertus is suffering for this. Every single time you interact with people, from convincing them to give you money to explaining concepts or getting a loan from the bank you need this skill that currently you lack.

With technical explanations that are hard to understand it is a good idea to show a diagram so people could understand and see the forest before explaining the threes, people do this using powerpoint, it takes five minutes. You can learn from a DIY show in German TV for example how they start with a summary because you know what you are saying but people in the audience not. Most people have never done what is shown in the program because they had never done it but they can get the idea from the summary and if they are interested even do it following the technical process step by step. Most of the time will only understand deeply only when they do it but they will only replicate what they are interested in.

Every person in your audience is interested in just a small part of an open source camera. For example I need to have a camera I could use as a scientific sensor without artificial restrictions. Even in cinema there is Advertising industry, film and documentaries with different needs.

For example in your video this is the way I would do it:
[SUMMARY, forest]Now we are going to explain how we correct our cameras from the slight error that every manufactured sensor has.
-[INTRO or OPENING]As you probably know every single pixel in the camera has a different deviation in sensibility to light because of the manufacturing process not being perfect(as no manufacturing process is ever perfect) and the thermal noise of the electronics.
-[BODY]We are going to correct this deviation measuring this noise when the camera light passage is blocked. Taking multiple pictures we can average this error.
[Now you do what you have explained on the screen, this way even if people are not technical enough to understand the details they got the intro and the body explanation, and they got the basic idea(forest), and they are satisfied, but if you just show the technical details upfront they can get confused].

The same thing with the explanation of the electronics, or the manufacturing process: simple summary and explanation before the hard stuff.

Again, this is storytelling 101, but it is important.

3 weeks ago
Sebastian

Hi Jose, many thanks for the feedback!

We deliberately decided against the teleprompter (we have one actually) as result from our community survey: https://www.apertus.org/axiom-team-talk-volume-12.1-article-december-2016

We are aware that being spontenious will not yield perfect results but thats part of the deal, we improvise :)
Team Talks are not meant to be overprodcued TV shows, its a snapshot of whats currently happening in the project shared by the people working on these things - rather authentic maybe not 100% cooherent though. We need to find the right balance between actually working on the project and reporting about it - we are a small team after all.

The structured approach is a good idea, we might give it a try with TT13.

4 weeks ago
Anonymous

I think ALL of us wanted more frequent updates and the technically minded wanted more technical updates. Both of these wants have been addressed in the last few weeks (thankyou and please keep doing so). Jose's comments are valid I suppose, I am native English speaking and technically minded so I have no idea how hard the language barrier may be and the less technically minded may like a simpler explanation first before a technical one, but I do not want you to be discouraged from your other recent improvements. I have no problems comprehending the information in your videos and feel Jose's concerns (particularly of the fan noise) are secondary to having something in the video thats worth while watching in the first place. Please keep up the excellent work.

4 weeks ago
Jose

@anonymous: I am a technical person myself, I understand everything in the videos myself, but I have experience as a speaker too and much more important I was a very bad speaker at first, you can improve.

If your intended audience is cinematographers, well, most of them do not understand deep technical issues. This is not secondary. You have to put yourself in the shoes of your audience, not you. In the same way some people could tolerate noises like their partner snoring and other people just get crazy at them. If I don't care about noises but my audience does, It is no good.

In my experience this is one of the biggest problems of technical people, putting in the shoes of other people, empathy and so on. But this can be learned like anything else.

Me too do not want Sebastian and company to be discouraged by my messages. On the other way I want them to see they can improve a lot in the way they communicate, because the can.

It is always dangerous to use written messages like mine because it can be misunderstood as the emotional content of face to face communication is lost in the sterile words.

It is not pleasurable to criticize the work that someone has done. But it is also one of the biggest problems with startups, to have honest opinions as your family and friends do not want to tell you something that could affect you.

So Sebastian and company if you are reading this remember to take my messages with a grain of salt. Take of them whatever you could find useful and ignore the rest.

3 weeks ago
Sebastian

The thing is that we do not believe there is a prime target audience. There are filmmakers, photographers, scientists, DIY, makers, industrial vision, automation, inspections, etc. professions all coming together here and it goes from pure end users to high tech developers from in both software and hardware and there are people seeing these videos who know nothing about the project yet and others who are following everything we release for years. So we try to cover different aspects of the projects in different depths - clearly its impossible to have something that makes everyone happy all the time but we try to at least have different content for different people :)

We are happy to improve and advance, its the core of open technology and I find this discussion already very fruitful.
One idea would be to invite external directors for each TT episode to allow different styles and different "outside" perspectives and approaches with us "actors" in front of the camera. But its a logistic and financial challenge we first have to find consent for.

1 week ago
Jose

Hi Sebastian
Well, that is the reason that I recommend you to go to a good Toastmasters group, because it will teach you basic stuff in the way to communicate with others as human beings. Scientists, filmakers, photographers... all of them are human beings you need to communicate with.

Those people have a basic program called "competent communicator" in which you have to give ten speeches and you will get to know a ton about things you currently do not know about. You get very useful feedback after each speech from your audience, and your feedback is always positive in a way you can handle it.

You don't need to invite external directors for this, extremely easy and it cost over 50-60 euros per year per person, and you could go and watch a two or three sessions as a guest without paying anything. It is a very inexpensive way to get feedback from 80 or 100 people on your audience at the same time. Over time you learn patterns and develop a sixth sense about what people is thinking in your audience as you could compare your impressions with the written feedback.

When I asked Brook, the creator of a 3d printer that sells in the US, printrbot, what was the most important skill he valued in the process of succeeding in his kickstarter, he told me without a doubt the fact that he was a religious pastor in the past, so he learned to talk in public.

Do not over complicate things, it is a simple thing. It is not about a style. you can give a great speech at 20 different styles at least. It is about learning how to communicate in any style you want.

I mean if you don't have this ability everything will become 10 times harder as if you are not able to communicate there is an obstacle between you and everybody out there.

3 weeks ago
Walter

Hey everyone :)

thank you for the update and thanks for all the imput and feedback.
It is nice to see that people are so interested in improving this project
and donating the best they can give.

In our shot lived time, giving time is the most precious thing one can give.
Thank you for that, to everyone who shared.

See you soon :)
Walter

2 days ago
Adam

Any update on a 12.3 teamtalk ?!?! Please please keep them regular, even if it means they are shorter. It stops me getting nervous.

Add new comment

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
By submitting this form, you accept the Mollom privacy policy.
go back up