New DIY HD Cinema Camera Project
I'm new here but I've been a lurker for a long time. Specially since I started with my little project.
First of all let me tell you I was planning to make my first post when I could actually show you my first images, but I came to a point where I need professional advice.
Now, let's start with the basics: what's my project all about?
I'm planning to build a very simple low cost DIY HD Cinema Camera. It'll be able to shoot Academy 2K (1828 × 1332 / 1.37:1), Digital Cinema 2K (2048 × 858 2.39:1) and 1920x1080 real progressive @ 24fps / 48Hz using 5MP Micron sensors (just like the Elphel 353). Of course, lower resolutions (720P) will lead to higher frame rates.
Let me tell you I already know this is not a perfect solution. All I want is to be able to shoot my own footage with my own camera. I don't want to spend $4k-$8k on something that doesn't give the image quality and motion feeling I really want, which is basically the closest to film I can possibly get. It'll probably have image artifacts due to the rolling shutter and it won't be able to shoot very long takes cause I'll be capturing raw bayer to the laptop but hey, I'll have real progresive, filmlike motion blur, shallow DOF, 4:2:2 compression and an image closer to film than any HDV camera and everything for about $2k.
The basic structure will be very simple: Optics - ground glass - achromat - Micron 5MP Demo Board and my laptop capturing 10bit raw bayer footage with Micron software. I've been talking with Micron and they include several debayering filters within their software.
I already have a couple (just in case) of nikon D focusing screens, I'll buy the achromat and GG holder from http://www.jetsetmodels.info/ and the demo board from Micron will arrive within the first days of July.
Now this is it. Those are the basics. I'll start a new thread talking about my first problem: lenses.
Of course if anyone wants to know more about my project, don't hesitate to ask.
First choice: LENSES
I know lenses are the most important choice for a project like this but unfortunately I'll be spending most of my budget in the demo board so that leaves little money for the lenses, that's why I want to make the right choice. What's good about that is that once I choose a mount, I'll be able to upgrade my lenses when I have more money.
Choosing a lens for this camera is basically the same as choosing a lens for a DOF adapter, so I want a fast fixed focal lenght manual focus lens. In fact I was thinking about buying three of them: 28mm, 50mm and 135mm.
The real problem is what kind of lens to buy. Nikon? Canon EF? Canon FD? I know FD and Nikon (manual focus ones) have manual iris control which would be the way to go but I also was thinking about using an old second hand EOS body as part of my camera so I could autofocus and use the electronic iris control. EF would be the perfect lenses for that purpose but I'm not too sure I want autofocus and I really like old fixed focal lenght manual canon lenses better than the new ones.
I have seen three FD canon lenses at KEH.COM (28mm, 50mm and 135mm) for $125. They're not new and they may not be the best choice but they're fast (one 1.8F and two 2.8F) and may be a good starting point. I could also buy an old canon body and I'd have the actual FD mount system (one of the bests ever) and also a prism which would be useful to get a correct image before capturing. Anyway I'm not too sure about the FD mount (because of the lack of compatibility with others) and if you say it'll be way better to spend like $150 on a single very high quality 50mm Nikon lens, I'll buy it.
What do you think?
Just something I forgot...
I like to think of this camera as something that's "alive" and will remain alive once it's built, cause it'll be upgradeable from the ground up. If Micron or another company develops a low cost global shutter 12MP sensor in about two years that's able to deliver 4k@24fps and I have enough money to buy it, I'll do it. The camera will also have better lenses and ground glasses in time. Computers will be faster and will be able to capture larger frames at higher rates. It'll all come in time, but the best of all is that nowadays it's possible to build a very simple yet powerful camera which will be the base for everything else.
I'm very interested in reading more about your project. I have 2 questions.
How much does the demo board cost?
What knowledge and skill is required to do what you are attempting? Do I need to be a programmer or electrial engineer?
Thanks and good luck.
The demo board costs 900 euros, that's about $1200 and it comes ready to plug, install the software and start capturing. It has an USB 2.0 interface so you don't have to be an electrical engineer to plug it to your computer. Of course it depends on your computer speed if you can or cannot capture 1080p at 24fps.
Now, this' not a simple camera. You don't hit "record" and that's all. You must configure everything before you can start. Configurations can be saved but you must determine resolution, fps, shutter speed, gain, binning (if needed) and so on... and after that you get a raw bayer sequence which can be filtered and converted to any standard format.
I say this because in the first post everything sounds very simple and easy. I'd suggest everyone thinking about going for this particular solution to wait for me to start testing everything before buying because there're lots of things that can go wrong. I've emailed other companies like Omnivision and Cypress asking for demo boards just in case the Micron sensor doesn't give me what I want. That's exactly why I wanted to start posting after getting the first images. Nowadays the project is at very early stages of development and I can't really know if something will go wrong, but I really needed advice with the lenses. I need all the optics to be ready when the sensor arrives.
Thanks for posting, Paul. I'll keep you (and everyone who's interested) updated with the project.
Thanks Jose, I'm really interested in your progress and appreciate your posts.
Is is too simplistic to say that with the demo chip, computer, control/calibration software, a housing and lens, plus optionally a 35mm adapter, you've got yourself a 1080p/24 camera?
Is USB inherently fast enough for the transfer needed?
Is this the same chip that is made for cellphone cameras? If so, why is the demo board so expensive?
Why the Micron? Is it better than other alternatives?
I'm very non-technical so I can't really help much at all in that respect, but I'm fascinated with these type of projects
I do have a question, is there a way to do it without doing the whole adapter thing? so that you dont' have to have all that light loss etc? like instead of having a c mount like the elphel you have something that will directly work with a nikon or canon lens?
First question: That's the main idea, yes.
Second: I asked the same to Micron. They told me it's possible. Anyway I've got two weeks to test it myself. If it doesn't work I can get my money back.
Third: These chips are used for still cameras and HD security cameras and one single sensor is just about $30. The demo board is able to deliver all the sensor features directly to the computer through USB, software included, and it's made specifically for Micron megapixel sensors so I guess that's why it's so expensive (if you consider 12bit 1080p@24fps for $1200 expensive).
And fourth: Till the day, Micron is the cheapest and most balanced (quality/price) 1080p solution I've found. I'm very interested in Omnivision chips as well, like I said. I'm waiting for an answer from them. Of course Altasens chips are better, but they don't support small projects.
Of course if anyone knows of a better solution...
John: Of course it's possible to use the board "out of the box". I wouldn't recomend using 35mm lenses without a ground glass though simply because of the sensor size. 1/2.5" is way too small for 35mm.
Re: if you consider 12bit 1080p@24fps for $1200 expensive.
LOL, I guess if you put it that way. I was just thinking that it is just a $30 chip. Nonetheless, it's worth it. Maybe we should form a cosortium to research and build these.
You could eliminate the DOF adapter, but I guess if you need 1080p/24 you want the depth of field control too.
Well, I said it in the first post. I know it's not perfect but I want something very close to film. As close as possible.
Appart from that, if you think about it, without the adapter, you get a very digital image (you can see some 5MP samples at www.micron.com) and I think the ground glass plus 35mm lenses will soften the picture, giving it a much more filmic look.
In fact I see the ground glass and achromat more like part of the whole camera than an adaptor. They're there because the sensor's not larger.
I might have some lens adapter parts (or even a unit) that I could loan to you for research, if the timing worked out, and we could make the neccesary arrangements.
I'm looking forward to your test results.
Thanks for posting
I'll contact you if I need something but at the time I'm focused on trying to decide what kind of lens to use. By the way, you wouldn't happen to have lens mounts to add to my project would you? I'm trying to decide between Nikon F and Canon FD.
By the way, I know of www.keh.com and I also know eBay is a no-go for lenses but, does anyone know of other sites to buy used photography equipment?
Jose -- great to see another DIY HD camera thread.
Iím experimenting with two Sumix cameras: an M73 (3 mgp) and M72 (2 mgp). These cameras record an uncompressed 8-bit Bayer clip to a laptopís RAM, before automatically saving it out to the HDD. Like you I wanted to do more work before starting a thread here; everyone naturally wants you to "show us the pictures" (Iím glad I didnít post my early stuff because I wasnít using the exposure controls as well as I am now!). The two cameras have the same size sensor, which means the M72 has better sensitivity (bigger pixels) and in using the whole sensor width (a respectable 1600 pixels), the field of view I get with C-mount lenses isnít too bad. I currently have a 10mm Schneider and 15mm Angenieux and usually shoot 1.85:1 or 2.39:1 in various frame widths (between 1600 and 1400) at 25 fps. Depending on frame size and aspect ratio, the recorded clips (using a laptop with 2 GB of RAM) last between 25 and 45 seconds each. Using an 35mm adaptor is beyond me for this project, though the mature look of the uncompressed image is sufficently different from interlaced video for my purposes.
I started using the camera either on a tripod or with a pistol grip (which needs to be held smoothly to avoid obvious rolling shutter), though I never really got used to looking to one side to view the laptop in order to see what I was shooting. So my current project is making a shoulder rig to combine laptop and camera, including a quick release plate to put it on a tripod.
I think I'll go for Canon FD. Nikon is too expensive for me now mainly because F lenses can still be mounted on new DSLR cameras but prices for FDs are really low. I just found 3 (or 4) lenses at www.keh.com. They're VERY cheap but they're fast and may make a good starting point.
The lenses are:
∑28mm F2.8 FD (52)
∑50mm F1.8 FD (52)
∑85mm F1.8 FD (52) (I may not buy this one)*
∑135mm F2.8 FD (52)
Everything for just $145. I know it's VERY cheap but they're used lenses ("bargain" at keh.com) with the obsolete FD mount so I guess that's why they have such a low price.
Anyway I'll buy better lenses in time or I'll change the mount. This is just a start.
(*) The 85mm lens is the only one with "UG" (Ugly) quality. "Bargain" quality is acceptable but according to KEH, "UG" quality lenses may have marks which may affect picture. I may change that one for a 35-105 zoom (3.5-4.5F) to have a bit of everything.
What do you think?
what i do not understand about building the camera, as with the Ephel, it will be not sync on camera sound, which in most cases is good, you know film is shot like that, blah dee dah, but would there be a possible way to hookup some sort of sync sound with it?
|All times are GMT -6. The time now is 11:22 PM.|
DV Info Net -- Real Names, Real People, Real Info!
1998-2017 The Digital Video Information Network