RML: I just hate dealing with cards, and the SD spec isn't an open arcitecture (it's partially open, I believe, but not all the features are implemented). It'd be possible to design it either way. 20gb is just an arbitrary number that would be about twice what I would ever need in a very long day of shooting. If someone wanted to build in more, then it'd be pretty trivial to just use bigger flash chips. But building it in also means it'll be faster and it's one less design aspect to deal with. I'm shooting for simplicity above all else.
Bluesman: I don't think that's true; there have been a ton of competitors to Leica that use their mount.
The sensor is a sticky bit at this point. I'm pretty sure that by using a sensor that's available in both b/w and color versions, we can use identical architecture and just do the colorizing in post. Which means, if a user were so inclined, they could change their sensor. No reason not to build this in to the design.
To clarify a little, I'm going to build one or two of these for myself, and put the designs out there for anyone else to copy and build one either exactly like mine or modified to their own liking. If someone wants to contribute to the work (I currently know nothing about programmable logic, which is an essential step in getting the data off the sensor and into the memory), they can also change the design I'm doing. They can then say "I'd like feature X" and I'll add it to the goals. If they go ahead and do the work for feature X, that's even better.
@sam n
No, the sensor itself wouldn't be open; unless I'm mistaken, it would be impossible for a hobbyist to fabricate them anyway. But what we can control is what ADC we use to convert the individual voltages from the photosites into digital data. And, if it turns out that it's too much computationally to have that much data moving that fast, we can scale back to 14 or 12 bits.
Speaking of sensors, nobody makes one that's exactly 12mp, it turns out. Anything 10-20 would do, since this is ground up. Everything is subject to change based on what will work and what won't. The Dalsa ftf4027 or ftf5033, the Kodak KAF-16000, and 10500, and the cypress ibis4-1400 are the ones I'm eyeing, and leaning towards the kodak, because I know I can get in single units. They all are roughly within my guidelines and I'm waiting to hear back from the sales units of all three.
No, I don't want someone else to make it. By taking control of the entire image pipeline, we're becoming more in control of the image. I also don't care about affordable. If it costs me 10k to build this, then that's what it costs. I want the camera that I want, that no one will build. And I'm sure that there are other people out there like me who'd benefit. Who think that if they could do feature x with their whatsiflex, it'd be just perfect. And like you said, there's no profit in it, so I might as well give away what I can for free. Thats all.
