These clips were shot over the course of around 18 months of development. Some of the periodic developments improved image quality eg. by reducing noise, but some tweaks didn't make the image noticeably better to any significant degree so please be aware that this collection of clips demonstrate technical snapshots relating to software and hardware development only. Nothing was shot to highlight specific aspects of the camera's development, rather, we wanted to test and evaluate its operating processes in various environments that happened to be available to us during particular times. The footage is all 1080p - some later clips were shot in UHD mode but processed at 1080p.
Lenses
In the beginning we were borrowing lenses, then we started buying old Nikon photography primes for shooting with the AXIOM Alpha prototype, but over the years we have acquired a selection of old Contax Yashica, M42, Canon and Nikon lenses like the the Zeiss 50mm F1.4 or Distagon 35mm F1.4 and 85mm F1.4. We haven't used any real cinema lens with the AXIOM Beta up to this point as we didn't feel that our test-shots would justify their outlay.
Note that all AXIOM Beta developer kits currently use no optical low pass filter and have an infra-red cut-off filter screwed onto the lens. However, our integrating an infra-red filter into the lens mount in the future is planned. In addition, a new optical low pass filter module option will be offered but as yet we're not sure when.
The image sensor datasheet declares 10 f-stops native dynamic range but a so called Piecewise Linear Response (PLR) High Dynamic Range (HDR) mode is also mentioned. We've done initial tests with this mode and found that it allowed us to get quite a few additional stops dynamic range out of the image sensor, which is very promising. PLR can make individual sensels (pixels on an image sensor) less light sensitive based on their saturation value during each exposure cycle. In other words, highlights can be tuned to contain a predefined additional number of f-stops. Trust is a Shakespearean short film directed by Juliana Neuhuber and the below clip was the first footage shot with the AXIOM Beta's PLR HDR mode turned on. While the settings might not have been 100% optimal at this point, and while we felt there may still be some room for improvement, the results were better than expected. A single response curve knee-point was used for the time being - in addition the image sensor supports the incorporation of a second knee-point, so, dynamic range increases in the highlights can be tuned for further improvement.
You downloaded the ungraded clips we provided here and put them into a new order, added your own music and regraded the "raw" footage? We would love to see and share the results. If you have such a clips please upload it and let us know.
Downloading clips is not working...
This is the error message I get:
Not Found
The requested URL /AXIOM-Beta/samples/Trust-Ungraded-DNxHD-HQ-1080p-V01.zip was not found on this server.
Apache/2.2.22 (Debian) Server at files.apertus.org Port 80
really impressive camera and the footage looks really good. If the dynamic range of the camera can get to around 13 stops this will be a serious contester against some of the the more established cinema cameras out there! keep up the good work!
The amount of color information in the files and the way they behave in DaVince is amazing! even do they are DNxHD files.
The PLR settings sound promising. especially since I shoot a lot of HDR.
I also shoot a lot of video (well until 2 of my 3 CCD's went out... lol) will this camera suffer from the same issues as DSLR's when shooting video while tracking a moving object ? (jello)
I also shoot high speed video, typically using a Casio EX-F1 Pro at 600 fps and 1200 fps video speeds. (Any chance that the AXIOM will have any High Speed Video Potential in the Future ? ) I ask because there is a new very affordable high speed camera that is also a good Video Camera (that I hear... I have yet to test it or review that high speed and Video Camera.
The image sensor has a global shutter - so no "jello" effects and supports up to 300 FPS at 4K resolution, reducing the number of lines increases the number of frames - we still have to find solutions how to store this amount of data somewhere though
That's great you finally share footage from actual Beta (s?) !
Could you also share (or make if you don't have any) some handeld, walking, driving in a car, fast pan... footage so we can enjoy some Global Shutter magic ?
There is one shot in the Danube Clip shot from a moving train. You can see that buildings are not bending over due to the global shutter. We can also shoot some more hectic footage in the future but I am not sure we need to proof that the global shutter is actually global :)
Yes I saw that and I was so happy ! Like some camera porn to me ;)
I don't need proof it works, I trust you about it. But modern sensors with global shutter DON'T exist ! Almost every camera has a rolling shutter and I strongly believe DOP's, directors and more widely film audience can really feel the difference on handeld footage, pans and car shots. I also believe the handeld camera work has disappear on TV and movies because of the Jello effect you get with a rolling shutter camera. And with the Beta you can claim to bring that back on top of the "raw" thing other cameras have :)
Anyway, it's a minor thing but I do believe it's a fantastic thing to bring back and to praise on...
Inspiring footages.
The page https://www.apertus.org/axiom-beta-status has not been updated for 4 months. It's not clear how fast the process moves. Are there any deadlines, how many years do you plan to reach the usable mature version?
I understand people are eagerly waiting to get their hands on the cameras but its difficult to impossible for us to provide accurate estimates. Rather we decided to share what is still left to be done to arrived at this more mature version: With the developer kits orders being close to being all fulfilled soon we will shift focus again towards software development and that means improving usability for users. On the enclosure front things are moving slowly unfortunately as the original CAD designer is currently unable to work on the AXIOM Beta design as his day job occupies him completely - CAD designers to finish the enclosure wanted - help would be much appreciated!
Hey! I do CAD work for my own camera projects and film friends. I'm an NYC area filmmaker and work with RED, BMP, and Arri cameras on set weekly, and am a member of a hackerspace with full access and ability to use Shopbot CNC router, SLA 3D printers, and most recently: 50W C02 Epilogue laser and a Bridgeport CNC, among various other toys. I would sell my soul to prototype and improve the enclosure with in-field testing and user feedback. This project is exactly what I've been looking for, and I wish you and your team nothing but the best. I get what you're trying to do and I love it. The film-like sample footage brought me close to tears.
If there's anyway that I can help, please don't hesitate to reach out to me.
Please show us some more footage (however ugly) of how well the camera handles mixed lighting in a shot to show the light sensitivity of the camera/sensor. This is a good reference video of the kind of shot I mean: https://www.youtube.com/watch?v=eOjmYDQmhok
I think developers dont want to show light sensitivity tests until final presentation as they results can varies at significant amount with implementations. Just compare x-t2 vs a6500 and keep in mind X-trans technology pattern don't helping to superior x-t2 at all. Also keep in mind sony just doubled amount of pixels in a9 with pretty the same efficiency for low-light. Im just saying iso performance is very tricky subject as it heavily depends on algorithms. Right guys?
We indeed believe by now that most (if not all) other cameras "raw" data isn't raw at all and that a lot of noise processing to increase dynamic range and light sensitivity is happening but nobody knows what their secret algorithms really do...
19 Comments
Downloading clips is not
Downloading clips is not working...
This is the error message I get:
Not Found
The requested URL /AXIOM-Beta/samples/Trust-Ungraded-DNxHD-HQ-1080p-V01.zip was not found on this server.
Apache/2.2.22 (Debian) Server at files.apertus.org Port 80
Thanks for the notice. Fixed!
Thanks for the notice. Fixed!
Hei guys!
Hei guys!
really impressive camera and the footage looks really good. If the dynamic range of the camera can get to around 13 stops this will be a serious contester against some of the the more established cinema cameras out there! keep up the good work!
The amount of color information in the files and the way they behave in DaVince is amazing! even do they are DNxHD files.
Looking forward to se more from this camera!
Thanks! I think finding a
Thanks! I think finding a good balance with the PLR settings should allow us to go to 13 fstops easily.
The PLR settings sound
The PLR settings sound promising. especially since I shoot a lot of HDR.
I also shoot a lot of video (well until 2 of my 3 CCD's went out... lol) will this camera suffer from the same issues as DSLR's when shooting video while tracking a moving object ? (jello)
I also shoot high speed video, typically using a Casio EX-F1 Pro at 600 fps and 1200 fps video speeds. (Any chance that the AXIOM will have any High Speed Video Potential in the Future ? ) I ask because there is a new very affordable high speed camera that is also a good Video Camera (that I hear... I have yet to test it or review that high speed and Video Camera.
Best Regards,
Weaponeer
Disabled Veteran
gary@weaponeer.net
The image sensor has a global
The image sensor has a global shutter - so no "jello" effects and supports up to 300 FPS at 4K resolution, reducing the number of lines increases the number of frames - we still have to find solutions how to store this amount of data somewhere though
That's great you finally
That's great you finally share footage from actual Beta (s?) !
Could you also share (or make if you don't have any) some handeld, walking, driving in a car, fast pan... footage so we can enjoy some Global Shutter magic ?
Anyway that's great progess, keep them coming !
There is one shot in the
There is one shot in the Danube Clip shot from a moving train. You can see that buildings are not bending over due to the global shutter. We can also shoot some more hectic footage in the future but I am not sure we need to proof that the global shutter is actually global :)
Yes I saw that and I was so
Yes I saw that and I was so happy ! Like some camera porn to me ;)
I don't need proof it works, I trust you about it. But modern sensors with global shutter DON'T exist ! Almost every camera has a rolling shutter and I strongly believe DOP's, directors and more widely film audience can really feel the difference on handeld footage, pans and car shots. I also believe the handeld camera work has disappear on TV and movies because of the Jello effect you get with a rolling shutter camera. And with the Beta you can claim to bring that back on top of the "raw" thing other cameras have :)
Anyway, it's a minor thing but I do believe it's a fantastic thing to bring back and to praise on...
Good point! We will see what
Good point! We will see what we can do.
Inspiring footages.
Inspiring footages.
The page https://www.apertus.org/axiom-beta-status has not been updated for 4 months. It's not clear how fast the process moves. Are there any deadlines, how many years do you plan to reach the usable mature version?
Ah thanks for the notice,
Ah thanks for the notice, updated.
Hi Sebastian,
Hi Sebastian,
Lovely images!
Care to comment on Ilu's question?
'Are there any deadlines, how many years do you plan to reach the usable mature version?'
I understand people are
I understand people are eagerly waiting to get their hands on the cameras but its difficult to impossible for us to provide accurate estimates. Rather we decided to share what is still left to be done to arrived at this more mature version: With the developer kits orders being close to being all fulfilled soon we will shift focus again towards software development and that means improving usability for users. On the enclosure front things are moving slowly unfortunately as the original CAD designer is currently unable to work on the AXIOM Beta design as his day job occupies him completely - CAD designers to finish the enclosure wanted - help would be much appreciated!
Hey! I do CAD work for my own
Hey! I do CAD work for my own camera projects and film friends. I'm an NYC area filmmaker and work with RED, BMP, and Arri cameras on set weekly, and am a member of a hackerspace with full access and ability to use Shopbot CNC router, SLA 3D printers, and most recently: 50W C02 Epilogue laser and a Bridgeport CNC, among various other toys. I would sell my soul to prototype and improve the enclosure with in-field testing and user feedback. This project is exactly what I've been looking for, and I wish you and your team nothing but the best. I get what you're trying to do and I love it. The film-like sample footage brought me close to tears.
If there's anyway that I can help, please don't hesitate to reach out to me.
Best,
David
Please get in touch with us:
Please get in touch with us: https://www.apertus.org/contact :D
Please show us some more
Please show us some more footage (however ugly) of how well the camera handles mixed lighting in a shot to show the light sensitivity of the camera/sensor. This is a good reference video of the kind of shot I mean: https://www.youtube.com/watch?v=eOjmYDQmhok
I think developers dont want
I think developers dont want to show light sensitivity tests until final presentation as they results can varies at significant amount with implementations. Just compare x-t2 vs a6500 and keep in mind X-trans technology pattern don't helping to superior x-t2 at all. Also keep in mind sony just doubled amount of pixels in a9 with pretty the same efficiency for low-light. Im just saying iso performance is very tricky subject as it heavily depends on algorithms. Right guys?
We indeed believe by now that
We indeed believe by now that most (if not all) other cameras "raw" data isn't raw at all and that a lot of noise processing to increase dynamic range and light sensitivity is happening but nobody knows what their secret algorithms really do...
Add new comment