New Project
 

So I'm thinking of starting up a new camera project similar to the RED camera.... What would you guys look for in a Digtal true S35mm camera?

Wishlist:

1) 2K resolution with 2.35:1 cinemascope guides
2) Recording to built-in swappable Hard Disks or flash (P2?). I don't want to have 17 different cables coming out the back, or be tethered to a big computer village. Handheld all the way.
3) Shoulder-mount, well balanced, not more than 18lbs.
4) Built-in rails.
5) Option to use 35mm Still lenses, as well as 16mm and 35mm Cinema lenses
6) Less than $10,000 (hey, you asked...)

Good luck!

I was thinking of #2 from the beginning because I don't like having 17 cables either. Its one of the major reason why I like the Varicam over the Cinealta.... no extra cables, variable frame rates, cost much less, ect.


And as far as #1, I was thinking about trying to push 8k and below..... 8k, 6k,4k,2k, 1080p, 720p, 480p and even NTSC and PAL (just in case). Or I could make multiple versions... where one allows 8k and below and another allows 4k and below....ect.

And I was thinking of variable frame rates but I don't want to talk about how high I was going to make it.

Anyone else?

I was thinking something like the samsung-sc-x105l

http://www.camcorderinfo.com/content...der-Review.htm

Where the lens/ccd is external and everything else you wear in a knapsack. Hopefully, it'll be light and modular. From a manufacturing point of view, it's probably simplier to do than to package everything within the space of a hand held device. I would rather the money and effort spend on better sensors and electronics than plastic mould injection and cnc milling and fancy packaging. Let the design use as much commodity hardware as possible, 35mm still lense, external lcd, common hard drives or solid state storage, basically things I don't have to buy from you ;-) A clean 2K image with some DOF and as many stops as you can squeeze from electronics nowadays is good for me. And of course I would rather you beat #6.

Simple and cheap.

There are multiple projects, going down to a couple of thousand dollars, and most price brackets might have serious competitors. But I would like to see a cheap option under $1-2K myself. A sort of B camera of movie making, but still good enough for low cost indie, and good compared to the HDV crowd.

The competition:
Under $20K Red, Infinity, Silicon Imaging Altasens model, XDCAM etc.

Colorspace, I hope it's two models are below 10K
Under $10K HVX200/HDV HDSDI, with Keith Wakeham's uncompressed HDSDI recorder.

Under $5K HDV/Sumix (if they are still working on their project) (the Sumix project being a camera of quality similar to Silicon Imaging, but without cineform or capture, price here is estimate with capture).

Under $2K, people experiment with the Elphel 323 camera (coming model a bit better)
(These are mostly complete recording setups including capture/computer).

Their are other lurkers out there, some I have heard of.

I have been thinking of a cheap camera for a very long time, accessing different technology and technique. I can say, that for a nice quality image at a cheap price, rather than a pure image, checkout www.ambarella.com. If an high enough data rate can be obtained, with some customisation, you can have near visual lossless in 10 bit bayer. This data rate should be a bit more than cineform's, and worse quality loss for editing (if an issue, import into cineform, they should get around to supporting h264 sometime). It doesn't matter what sort of case you put on this chip, as long as you can get it to do a good enough job. Short of reprogramming a PC to do it, probably one of the easiest ways.

Sensors, look for latitude (well capacity/multislope) look for sensitivity (Signal to Noise ratio). Only a handful of companies have all these. Global Shutter is also good, if you can afford it. Some companies have sensor, codec processor and control on the same chip with interface suitable for hard drive. You would be surprised at how good some of these cheap sensors can be made to be.

Ease of use. Making it easy to control, work with, and do work flow. Modes for consumer/doco, Eng, and cinema use.

Have fun.

Hold it, was that 8K I read? I understand, that is mighty big, and while there are certain cheaper processing options for that I can think of, it is going to be expensive to use, and maybe to develop. This will push the price up too. Are you sure you want to go that path?

Thanks

Wayne.

looks like we got UHD camera rain here!

For me a camera starts with a sensor and everything else follows.

Frame rate, data rate, lens quality, CODEC requirements, recording system, they are all supporting the sensor.

RED is having their own sensor fabbed, which gives them options the rest of us don't have.

Andrew,

Here is my wish list:

1. 16bit per channel 4:4:4 2K would be great! 4K or 8K is just too big (file size) and difficult to deal with (processing power) for most people who would buy this camera, but would be great to have for some.

2. lossless or atleast visually lossless codec. Image sequence might be best. I can use them native or put a quicktime wrapper around them.

3. 35mm SLR, 16mm cine, and 35mm cine lenses

4. adjustable framerate (timelapse to 240fps)

5. Intergrated 15mm rail system

6. shoulder mount with pistol grips (the layout of the movietube is nice)

7. untethered

8. 8 - 12lbs without lense

9. record to disk, flash or ram

10. SD video tap for accessories

11. HD YUV video tap for accessories (HD-SDI would be nice if it doesn't impact price)

12. use standard battery mount (IDX, Anton, etc)

13. ability to send waveform, vectorscope, and RGB parade to video taps

14. overscan to viewfinder and video taps, with safezones for every size (1.33, 1.78, 1.85. 2.35). Send 3 or 5% more to the viewfinder than will be recorded and show safezone for what will be recorded.

15. film style adjustable shutter (i.e. 1-360 degrees).

16. white balance presets and adjustable

17. advanced in-camera color correct settings like the varicam would be great if they don't impact price and performance.

18. 2k 4:4:4 HDRI would be great if it doesn't impact price performance.

19. record 2+ channels of CD quality audio in camera

20. ability to store metadata with files would be nice. camera settings, date, time, good/bad, etc

21. under $10k. If you can make this camera under $5k you won't be able to keep up with demand.

I've been back and forth between camera project and deck project for a while now. I'm solely focused on the deck now so that it can be used by most pro cameras.

As such I spent several hundred dollars to try and get preped to run a bayer filter KAI-2093 sensor (Which I still like a lot).

When I started work with FPGA's at first I was fairly new to it and I'm still not where I want to be. I rewrote the code for the drive controller a lot and keep finding new info that i never noticed in the ATA-6 documentation. I've still got Sata to tackle next year I think, thats not going to be fun.

What you will encounter is a lack of a sensor. The only one that I think you might have a chance of getting is the Altasens 357x but that won't reach production for a few months and they are phasing out the 356x series. The other option is Kodak CCDs. They don't have a 4k or 8k that could handle the frame rate but their HD 1" sensor looked great on paper (and several times i almost broke down to order it). The only other one that you can probably get is the Panavision SVI quadHD sensor. Available and true 4k, but don't want to ask about price.

If you ever were interested in using the Kodak KAI-2093 sensor I have fpga code waiting here for fixed and variable framerate along with exposure control. If I had the money I could have had it spitting out images a while ago but I moved the project in another direct. The HD-SDI recorder, which right now I'm trying to finally get things packaged together for field testing (Finally, I'm so far behind on this thing)

A word of advice though. Take you time, don't let others rush you. The only other thing I have to say is getting the parts is easy, making them work together, now thats the tricky part.

That speedcrank-controlled framerate the Kinetta had to imitate a film camera sounded pretty interesting. If you're seriously considering developing this camera, I'd suggest integrating that to separate yourself from whatever competition you might have. But then again, I don't know your price point. But that would be neat.

I knew a group that had problems getting a Altasens, there was a shortage, and they were probably just not as high priority.

Video/cinema rate UHD sensors in a cinema frame format (or any format probably) are rare. But what I would like to ask here, of Kieth and others, what good, descent priced, HD/SHD options are there apart from Altasens, Fillfactory (Cypress) and I'll include Micron as they now have multi-slope feature, though I don't know if they have improved the S/N ratio or well, capacity, or fill factor?

Well of what I know people can purchase below are the companies that I know that will sell sensors

Micron - Digikey
Fillfactory (now cypress) - Unsure
Kodak - cmos through digikey or other resellers, CCD direct from kodak
Panavision SVI - Direct from panavision SVI
Altasens - Direct from altasens
ST micro - digikey, not many
Dalsa - High end CCD that are more for still cameras directly from them

I'm not going through what exactly is available from each but suffice to say if you want improved sensor characteristics you'll need to spend a lot of money

Thats my point, out of all those virtually only the ones we already know have reasonable priced sensors with reasonable quality. I even found out ESS does sensors, but I don't know of the quality.


What is "DigiKey"?


Thanks

Wayne.

I just realized that mentioning digikey might fall into the realm of talking about places that sell stuff that isn't a dvinfo sponsor, but since they don't really deal in video equipment I don't think Chris will mind. (Chris, if you see this and you do mind, please just pull my post)

Digikey is a parts retailer/almost OEM supplier. They sell your do-dads and what-ca-ma-call-its. Like resistors, capactitors, IC's, some imaging sensors. Almost anything electronic at the smallest level.

Yes, I remember now. What is this about sponsors?

Chris Hurd, the guy running this place doesn't normally let people talk about non-sponsor companies. The sponsors for here are like B&H and EVS and stuff, thiers a link to the top right that lists them.

Anyway, I don't think Chris lurks in here to much or cares to much about mentioning part suppliers because its outside of the realm of the normal.

Do you guys think it would be crazy to have more than a 3 CCD set-up?

To what end?

Three CCD/cmos/sensor setups have a simple purpose. To seperate the incoming wavelenghts and send them to 3 different sensors, so one for red, one green and one blue, all with fairly hardline cut offs for spectral response because of the nature of the prisms.

My original understanding was the prisms were used to seperate the light, but i've heard that some 3 sensor setups use single colour filters which would reduce light even more, either way their is light loss with prisms and it is apparently an issue for larger sized sensors.

How are you thinking about spliting it, are you just concerned with increasing sensor size by using multiple smaller sensors, or pixel shift or what?

their was someone that connected two DV cameras together to make one larger sensor here a while ago. It was interesting and sucessful, needed some tweaking in after effects to fix the seam but it was neat.

A 3 sensor setup using cheap cmos sounds possible without to much trouble but you need a fairly complex setup to align the sensors and mount them onto the prisms. Its to bad you can't get a sensor with only a blue red pattern and use a 2 sensor setup with a monochrome to make 4:2:2. Much easier to deal with large scale data this way.

Unless you want to so some extreme pixel shift stuff, or do special stuff to use two extra sensors, one for IR and one for UV, it is a bit of over kill. I have considered it myself previously. What Keith says is ok, but look on the web for home made/DIY vacuum Chambers, prisms, and alignment stuff, I might have seen one specifically used on a sensor.

Keith that dual camera setup, sounds interesting, you know that you can source prisms to do that so the seam problem is less of a problem. Olympus did it for their SHD camera. If you had three cameras on their sides you would have close to 2.35:1 720p 5:1 Dv through a prism. Ohh, driving everybody crazy again ;).