Which image sensor would you get in your Axiom Beta camera?
Submitted by Sebastian on Wed, 04/09/2014 - 18:30
We would really have loved to visit Las Vegas this year in order to demonstrate our AXIOM Alpha prototype in front of an audience outside of Europe - in fact, we even had invitations to do so in two separate booths there.
Unfortunately, we are not in a position where we can afford the cost of flying several people to the United States of America and cover one week's accommodations for them. We simply cannot justify spending that much of the funds we have received as donations so far on a trip like this.
Most of those funds have already been spent on the development of AXIOM Alpha and we have no intentions of stopping there.
But as usual, we do not require a large booth at the NAB show-floor to make a huge announcement. (By the way, there are a number of major companies who could easily afford to rent space and set up a booth there, but have also decided not to attend this trade show)
So, before we show you what's coming up next, just imagine you're walking up to our booth at the NAB Showfloor - it's all there, the cheering crowd, exciting gear and great people all around us ...
That being said, we hereby announce the AXIOM Beta!
It will be smaller and cheaper, exhibiting far more streamlined features that are much better suited for shooting, testing and development purposes than the AXIOM Alpha could ever be.
This time we need your contribution to make it happen, as we will start a crowd funding campaign to cover hardware and software development. We plan to start with an initial batch of approximately 250 cameras once we know that the design actually works, and this first batch will be exclusively reserved for crowd-funding backers. Each of those cameras will have a degree of personalization possible.
As you may know, the AXIOM Alpha is our proof of concept prototype for the future AXIOM platform. While it was a great way to showcase our concept and verify that the chosen technical approach for building a cinema camera like this works, it has also revealed some shortcomings and a few problems with the chosen components - specifically the Zedboard hardware, which has turned out to be unsatisfactory for this task.
We've also had a large number of developers and hackers willing to buy a complete AXIOM Alpha, however the high component cost (mostly due to the CMV12000 image sensor) and the difficulty in replicating our custom designed enclosure and lens mount have prevented this from happening on a large scale.
Finally, there are the form factor and size of the Zedboard resulting in a bulky and problematic enclosure with a rather unfortunate placement of sensor and lens mount.
While there are some companies out there who trick their customers into paying premium prices for a beta product, eventually ruining their entire project, this is not how we work, so we want to make sure that you know exactly what you are going to receive. Actually, we want to make sure that you are aware of what you are getting yourself into right from the beginning! The AXIOM Beta is a BETA product (which can clearly be deduced from its name). There will be bugs and other issues for sure, but we are confident that with your help, your experience and last but not least your input, we will hunt down every single one of them and turn it into a feature.
Of course, as always, there is the small possibility that the AXIOM Beta will work flawlessly from day one, but there is no guarantee that it won't fail miserably first. Nothing can be perfect right away, and our drive towards perfection may require some time and a number of firmware upgrades before all aspects of the Beta's operation are running smoothly.
We're inviting you to be part of this journey, not as a passive bystander but rather as a contributor at the forefront of progress shaping the final product. In return for your assistance, we're offering the hardware at cost, which brings us to the next point:
Backers who support the AXIOM Beta development in the crowd funding campaign with a minimum amount of 350 EUR / 485 USD or more will receive the AXIOM Beta (including all the necessary I/O modules) for manufacturing costs only (i.e. material and assembly), at approximately 550 EUR / $760 USD (without image sensor, see table below).
But rest assured, no matter what hardware options you choose, you will only pay for the materials, assembly and shipping. Naturally, if we manage to attract more backers than initially planned, the cost of production per unit will be reduced and everyone will get to pay less. The main focus of this campaign is to find early adopters and get the hardware into the hands of hackers and developers. Once our product matures, you can expect the AXIOM camera to grow in complexity along with its feature set and price. For this reason, it is unlikely that future AXIOM models will be as affordable as what we are offering in the AXIOM Beta crowd funding campaign.
If you decide to upgrade your AXIOM Beta to a successive product (e.g., say, the AXIOM Gamma, if we continue using the greek alphabet naming scheme), the main (and most expensive) component, the sensor, can very likely be reused. The Microzed is also an independent product and can be utilized for other projects or resold again.
In the beginning, we will create an AXIOM Beta using the Truesense KAC12040, as we would like to gain experience with another image sensor than the one featured in the AXIOM Alpha. Once this is working, we'll start adding additional image sensor options based upon what the majority of people in our community are interested in. This sensor may come from Cmosis, Truesense or a different manufacturer, as long as they provide datasheets and use compatible interfaces. Please cast your vote below to give us an idea of which options would be preferred.
Resolution | 4096 x 3072 |
Optical format | Super35/APS-C |
Frame rate | 300 fps (10 bit) 180 fps (12 bit) |
Dynamic range | 10 F-stops, up to 15 F-stops (HDR) |
Shutter | Global |
Resolution | 4000 x 3000 |
Optical format | 4/3" |
Frame rate | 70 fps |
Dynamic range | 12 F-stops (Rolling Shutter) 9.3 F-stops(Global Shutter) |
Shutter | Global & Rolling |
Resolution | 3360 x 2496 |
Optical format | 4/3" |
Frame rate | 104 fps (10 bit) |
Dynamic range | 10 F-stops, up to 15 F-stops (HDR) |
Shutter | Global |
Which image sensor would you get in your Axiom Beta camera?
Soon, so stay tuned, as we will share further details about the AXIOM Beta once we have finished the EU Horizon 2020 grant application. At present, we need all of our eyes, minds and energy focused on completing the documents for this program.
72 Comments
You should put the "Upgrade
You should put the "Upgrade Path" text before the votes pool... It's very useful to know that we can keep the beta sensor ot further camera models :)
Good idea! Done.
Good idea! Done.
I think you guys should also
I think you guys should also take a look at Aptina sensors. The newest 1" sensor from them, the AR1411, does 4k60p and 1080p120. In fact, it can do higher than 4k60p
Dynamic range of 79db, which would be around 13 stops I believe. Doesn't say if it is global shutter or rolling shutter though, nor the bits per channel, but you can find that out with the datasheet.
https://www.aptina.com/products/dslr/ar1411hs/
Looks like an impressive little sensor.
Thanks Adam for your
Thanks Adam for your suggestion. I would primary buy the Apertus for slow motion video. 1080p@120 sounds interesting, would love to see the pricing compared.
great features thanx for
great features thanx for sharing
The Aptina has a rolling
The Aptina has a rolling shutter and encodes at 12 bits.
There are many choices but a
There are many choices, but a Module (with User programmable Interface) beats a raw Sensor for plug-and-play.
It is like having a Desktop Computer with a PCI Bus, just plug in any Graphics Card and Monitor. Compare that to a Laptop where you are stuck with the Graphics Chip and Laptop Screen (unless your a real Tech Geek and could upgrade them).
With a Module you simply pull it and replace it with another one from the same Company OR a different Company (once an Interface Protocol is decided upon).
That leaves the Camera's User Interface (Menu) and physical container (Camera Case) to be worked on and the finicky business of reading the Sensor up to the people who actually make the Sensor (they should know what is best for their Sensor, just as WE know what we want our "Box" to look like and what it should do).
Suggestions for a "Module":
1. Aptina Q-12A65, Pixel size: 5.5 μm, Global shutter, 60dB or 90dB in HDR, 8 or 10 Bits, 66 FPS, and Interface is Camera Link.
2. Teledynedalsa FA-81-4M180, Pixel Size 6 µm, Global shutter, 8 or 10 Bits, Max. Frame Rate 168 fps, Full/Ext Mini-Camera Link
3. etc.
There are SO MANY choices and it makes the tough part easy, leaving only the other tough part (the Case). Software upgrades can change the Camera's UI making it truly open source. A few years down the road just toss the Camera Module and put in a much better one.
If you make it Super-Plug-and-Play the end-User can swap Modules On-The-Fly and take exactly the Camera they want to the shoot they are going to; how great would that be ?
That means you have one (or time to make two) Camera Bodies and a dozen (or more) choices of Sensor Module available to suit ANY need -- more people satisfied means more people PAY (which is what you need). Cut your workload and expand your Market.
References:
Aptina Q-12A65: http://www.adimec.com/en/Service_Menu/Industrial_camera_products/High_pe...
Teledynedalsa FA-81-4M180: http://www.teledynedalsa.com/imaging/products/cameras/area-scan/falcon2/...
.
Rob: http://www.youtube.com/user/LowLightVideos
the nikon 1 cameras have very
the nikon 1 cameras have very impressive video output using the burst mode and aptina sensor http://vimeo.com/62412527
Excited! Very excited!
Excited! Very excited!
Thank you for this concept
Thank you for this concept and its refreshing that you admit its a beta and that you will support purchasers with firmware upgrades. Shame on other companies that dont. You know who you are!!!
Looking forward to what and where this leads to, as I am fed up with paying a premium for a silk screen printed name on anodized aluminium. This looks like a really worthwhile project,
Super35mm over 4/3" any day
Super35mm over 4/3" any day every day.
I agree fully with this
I agree fully with this statement!!
Unless you shoot anamorphic
Unless you shoot anamorphic
Brilliant move - and good
Brilliant move - and good luck! I'm proud to have donated those 100 Euros back in Dec 31, 2012 when you were getting started. Now I can't wait to put my name down for a camera. I like the honesty that you bring to the camera industry. Also... the lack of arrogance. I'm tired of having the camera manufacturers making bad decisions for me (poor ergonomics, poor reliability, not taking full advantage of the sensor) and think it might be way more fun to play with your camera and see if I can make some mistakes of my own :)
If you are looking at other sensors, definitely consider the Altasens S35 60p (and higher) sensor that is being used in JVC's upcoming cameras, as well as pretty much anything 4K that Sony offers. Low noise, high dynamic range and good color filters are important. However the 4:3 aspect ratio, crazy high frame rate and global shutter of the CMV12000 should make things fun to play with for me for now...
Altasens is not interested in
Altasens is not interested in releasing the datasheet of their image sensors without an NDA so their products are no option for the AXIOM.
So far there are only two manufacturers that agreed to publish the full datasheet of their image sensors are Truesense and Cmosis.
For the AXIOM it's essential that the documentation of used parts is available to everybody
So far it has some amazing
So far it has some amazing specs. Very good. The only thing I question is why MicroSD?
Surely this is not by choice. I'm guessing that it just happens to be on the board?
It is done by choice and
It is done by choice and because it requires very little space. The goal is to make the Axiom Beta small and lightweight for still photography.
I'm sure this is way off beam
I'm sure this is way off beam for some, but selfishly I'd like to use Leica M glass with an axiom body. Is this a possibility?
Micro Four Thirds to Leica M
Micro Four Thirds to Leica M Adapter and you are ready to go.
I also would like to have a
I also would like to have a SD Card-Slot too.
As a still photographer i have lots of SD Cards but not a single microSD.
Maybe in the final product?
Will this Camera be an alternative for a DSLR like the Nikon D600?
I would like to have a module for a hand grip like the nikons have with a display and shutter button and an accurate auto focus system with phase detection.
Also some advanced digital manual focus systems like split image and peak would be nice to have.
http://youtu.be/7FR3l6S12JA
Hi
Hi
I'm using Linux since 2008 and admire your project.
If the price range will remain as you've mentioned and you'll have a super 35 sensor I'll put money on the project
Best regards
www.cinematographer.ro
I can't wait!! Please look at
I can't wait!! Please look at Aaton for design and formfactor (aja just did) . The cat on the shoulder idea. It makes for beter balanced shots. Not just horizontally but tilt-wise also.
Please just keep it simple. And please ad an EVF like the SI2k one .. ( it was the first oled evf i think and it was very small no interference between the camera-operator and the subject. you can use the left eye to scope the environment whilst still staying in contact with your evf for critical framing).
OKOK I'll shut up now.
Thanks!!
Ikonoskop form factor is also
Ikonoskop form factor is also very interesting...
yup
yup
I second looking at Aaton for
I second looking at Aaton for form factor, was the most comfortable 16mm camera to work with. Let's get back to building cameras for the operator. Where the camera becomes part of the body, enabling fluid shooting for documentary work. Also agree about the need for a good viewfinder, I always shoot with both eyes open!
Hear, hear...
Hear, hear...
yes Aaton beautifull cameras.
yes Aaton beautifull cameras.
Wide dynamic range (14-15
Wide dynamic range (14-15 stops, please), comfort for the operator, simplicity of set up without needing to buy complex aftermarket kit, small size. I agree that an EVF is much more useful than a monitor. When you reach mid-40s eyesight changes and monitors are not good for field work. Thank you for what you are doing. Watching with interest.
Another big thumbs up for the
Another big thumbs up for the Aaton form factor
Honesty is a good thing. I
Honesty is a good thing. I like the notion of high end stills in conjunction with 4K as it's a natural fit as many videographers, such as myself, use stills in 4/C print as well as in inserts in video. Your approach is very sound without question.
Does CMV12000 support ROI? If
Does CMV12000 support ROI? If so, what speed could be achieved for 2K resolution ( ROI, not binning)?
Regards,
denis.glinsky(at)gmail.com
Yes, but only in vertical
Yes, but only in vertical direction, speed roughly increases with the amount of skipped lines, so 3072->1080 would roughly equal a 3x speed factor. Though there are no HDMI recorders that support higher FPS.
Is there a place we can sign
Is there a place we can sign up for updates?
Yes, our newsletter in the
Yes, our newsletter in the right sidebar or here: https://apertus.org/newsletter
Is there synch for
Is there synch for stereoscopic dual-camera shooting?
We are looking into it yes.
We are looking into it yes.
I have voted for the s35
I have voted for the s35 sensor, but i'd like to add the aptina to the list. Paired with a Speedbooster can be very interesting option.
Please consider the Aaton or
Please consider the Aaton or Ikonscope form factor and the possibility of using some of the OLED EVF add on viewfinders offered in mirrorless cameras like Sony, Olympus....
I have voted for s35 sensor
I have voted for s35 sensor but I would like to see the Aptina as an option as well. Paired with something like a speedbooster, it is an amazing outcome.
It would be great using the
It would be great using the new OLED EVF viefinders offered in mirroless cameras from Sony, Olympus...
If ROI can be selected in
If ROI can be selected in direction position only, can a wide screen crop of 1:2.35 or 1:2.40 be implemented?
There are suggestions/rumours
There are suggestions/rumours that the Aptina sensor will be in the new Blackmagic Studio HD models.
I'd love to see the Aaton -
I'd love to see the Aaton - Ikonoskope form factor.
If a ND filter wheel is too
If a ND filter wheel is too difficult/expensive to incorporate, a slot for gel filters can be a better option, similar to 16mm Bolex cameras and some long telephoto lenses from Canon, Nikon...
Will under-cranking also be
Will under-cranking also be included? This in some ways, is just as important as slow motion, but overlooked by so many cameras now...
We can basically output any
We can basically output any arbitrary frame-rate but the limiting factor will likely be the HDMI recorder.
For timelapse you can capture single full resolution raw stills and save them camera internally though.
Just reading your news feels
Just reading your news feels allready like starting a new relationship: exited, keen on every further step, beeing "stoked" by the imagination what might be coming up.. Thanks for every effort till this point!
what is the price of this
what is the price of this camera, i love to study one, make videos, make software :)
How did you export the
How did you export the footage you shot '' First Moving Image from apertus° Axiom Alpha prototype '' ? Did you use the sensor Truesense KAC12040 ? And one last question, did you manage to shoot at 70fps yet ? Keep up the amazing work !!
All the Best,
The Axiom Alpha prototype
The Axiom Alpha prototype uses this image sensor: https://apertus.org/axiom_imagesensor
The bottleneck are HDMI recorders, we did not find any that can record higher than 30 FPS.
Thanks!
the odyssey 7q over hdmi only
the odyssey 7q over hdmi only record 30p, but using sdi support 4k 60p, only from canon, i think that you can speak with convergent design.
something like the atomos
something like the atomos shogun might become a way to capture the signal?!
potentially, yes
potentially, yes
It's not possible to solve
It's not possible to solve this internal with an ssd?
How about the Dynamax DYN0174
How about the Dynamax DYN0174 image sensor. Panavision sold their image sensor business, that business became Dynamax. Probably a bit more expensive though.
We have an NDA with Dynamax
We have an NDA with Dynamax Imaging and acquired the datasheet, the price is a bit cheaper than the CMV12000 but the specs are pretty similar: http://www.dynamax-imaging.com/web/docs/products/DYN0174_brief.pdf
Will the Beta release be of
Will the Beta release be of any interest to a cinematographer who isn't able to contribute much to software or hardware development? I figure the folks with those skills might benefit from being able to work with footage shot under a variety of lighting conditions.
I'll contribute regardless, but if I go all-in to purchase a camera, I'm hoping I could shoot some music videos or a short film with it.
I know it's hard to guess at this point, but I'm wondering if this initial release will mainly be for the engineers and coders.
Of course its for early
Of course its for early adopter filmmakers as well, we are looking for feedback, suggestions, ideas and testing from film makers just as much as from developers. We just want to make sure people understand its not a finished product yet.
I agree with the aptina
I agree with the aptina sensor suggestion. I don't require much in a camera. This cheap digital cinema stuff, even though I am like it's uncle, is a red herring, professional video quality is the real aim for many. Those very small 4/3rds Panasonic lens with Aptina sensor would make a small camera for $1000 if you included a 2* ssd interface to record directly etc.
Now, HDMI recording might be currently limited to 4kp30, but provably not for long. However, thunderbolt recording to 4kp60 directly to a main board of a small or portable pc might be possible, or newer Ethernet standard at a cheaper price than an initial HDMI 4kp60 recorder.
However, can we just pack Bayer p48/50/60 into 4kp30 HDMI?
So a few possible solutions.
BTW, the encoding on the site is making the pages stall when first loaded.
From my point of view, the
From my point of view, the Axiom should provide the best sensor at the best price point possible. It doesn't matter if the Axiom will cost up to 5000 EUR in total. As long as we get püerfect 4k & perfect SlowMo, this camera is worth even that price.
are their other raw options
are their other raw options in case I can't handle a 4k workflow yet, like 1080 or 2k raw?
Should be possible, we will
Should be possible, we will investigate possibilities to embed raw data into a normal image stream to an external recorder. For example you can simply fill the 24bit RGB that pretty much all recorders understand with 2x12bit raw (not taking any chroma subsampling into account to simplify the concept here), so you can basically put two pixels into one which you can utilize to record 60 FPS raw in 30 FPS RGB. The downside is that the images requires post processing to recover the original raw data again and that while shooting on the external recorder the image looks defect so you cant use it for framing...
All of it sounds good except
All of it sounds good except the MicroSD card idea.
Small? Yes of course but the size difference between SD and MicroSD is minimal on camera but major out in the field when it is being handled. I am reminded of the miniDV cassettes and their vulnerability.
There's a diminishing point of return on miniaturization and I think one reaches it at the SD size unless we're talking phones.
Please consider staying with SD. Just small enough, but not so small that a puff of wind could scatter it to oblivion.
I think it could be cheaper
I think it could be cheaper if you had a pool of a thousand people to buy a camera batch and sensor. They could use somebody like Sumix or Elphel to arrange manufacture cheaply. The sensor price is the main saving to look at. It is a shame that it could not be a batch of 10,000
Such a project might become a suitable platform for machine vision or high security cameras (such as government and casino's).
Here is another idea, if a
Here is another idea, if a simple file recorder to 2xssd or memory card were included, a simple case like the BMD pocket camera could be made by others, to carry the head with some buttons and an android device for screen control. But all we need to start is a file recorded to 2xssd/card in a module/internal The link to the android device can complete a portable test camera.
However, a big market is if this could POV record (even with a cheap sensor) but a proper cable is needed for professional production, such as action/car shots, or as a studio camera. Thunderbolt is probably the best option, as you can plug long optical cables in and direct DMA record to computer. Ethernet 10-40GB also an option, but not as broad reaching as TB in future
I'm going to put down three
I'm going to put down three suggestions for companies for sensors.
Aptina mentioned before for low cost 4k video sensors, but a lot of quality for the price, as used in the Nikon 1 series. However, even a smaller cheap 4k sensor is useful for testing and developing for the platform.
Sigma Foveon video sensor they were supposed to be working on. Sigma also needs a high end motion camera platform.
Nokia.
Looking forward to the
Looking forward to the crowdfunding campaign. Since HDMI recording may not be viable, especially for high/low frame rates, an SSD recording module is a MUST! The camera should be able to record footage directly, without requiring an outside product and the problems inherent with interfacing, powering, rigging, etc.
I could go for either the Cmosis or Truesense chips, the Truesense is cheaper and allows for choosing your priority- either dynamic range for landscape or dialog scenes, or global shutter for action sequences.
Hi Sebastian.
Hi Sebastian.
Congratulations on the Axiom development project. I am following since long time this project and I would like to enroll in the beta phase.
My contribution/suggestion would be to have a spare electronic board for sensor mounting that would connect to the main board by cables. The reason is to be able to have the optical path paralel with the main board for flexible design. One such design will be usefull for drone mounting. If the optics and the sensor will be remotely mounted the gimbal will be smaller and the drone will increase in flight time as the gimbal motors will be smaller.
Hope my idea will help the project.
Good luck!
Hi Florin, thanks for the
Hi Florin, thanks for the thumbs up.
What you suggest is not easily possible as the interface between sensor front end and processing hardware requires very high bandwidth and throughput and thus is not straightforward to route through a cable. But since the Axiom Beta is meant to be small and lightweight anyway maybe that already solves the problem.
I would love to promote this!
I would love to promote this! We need to get the ball rolling on building awareness and attract individuals that could benefit from this. I see more ingenuity and sophistication from the backyard builders than the large mainstream corporations. The more we can sell the cheaper it can get, brilliant! I think that incorporating this approach into makerspaces will breed a new community of engineers and possible pioneers. In fact, I see large potential in making a robust portable model out of carbon fiber or a modular design that incorporates a phone as the primary processing engine. The phone for the avid and the full featured for the professional.
if we purchase the beta, do
if we purchase the beta, do we have to purchase the sensor at the same time or just request the build option for the sensor we want incorporated?
You just tell us what sensor
You just tell us what sensor you want and we will build your camera.
Add new comment