"...46-megapixel Foveon sensor that uses 3 stacked sensors of 15.3 megapixels each to capture red, green, and blue light. The resulting images are 15.3 megapixels in size with color resolution that’s superior to traditional sensors.
The camera also has a rugged magnesium alloy build, a 3-inch LCD screen, 11 autofocus points, 98 percent and viewfinder coverage. It’ll arrive in February next year at a price that has yet to be announced. If you’re looking for a rugged DSLR and don’t care about things like Live View or video recording, then you might want to add the SD1 to your list..." (Source: PetaPixel.)
The question to ask is WHY these DSLR and photography gear giants and experts would make such a faux-pas? Actually, could this have simply been a deliberate omission? For some, it is a no-brainer and it perfectly makes sense. For one, as popular is it is amongst DSLR manufacturers today, Video is just video, and seeing that it is and has been out there for a while amongst competitors, perhaps SIGMA figured "why bother competing in such a saturated market?".
Another reason may be that SIGMA are targeting a more precise market in the Photography sector, one which they will immediately have an upper hand in.
Though, on the technical side of things, in regards to the Foveon sensor, real-time colour space conversion from Foveon to RGB is not realistic (at least not yet). Here's why; it requires immense CPU power and memory and the Foveon colour space is highly non-linear. Therefore, this "mistake" or "faux pas" was indeed utterly deliberate and even better, fully calculated.
Price:
Will it be cheap? We do not anticipate so. Its specs are simply not to be expected among the consumer-DSLR market.
Will it be popular? Who knows. One thing is for sure however, it is very unlikely that non-photography enthusiasts be exposed to this beast of a machine. Simply because the high street shops are flooded with the likes of Canon, Nikon, Fujifilm, Sony, Panasonic Lumix etc.
We look forward to have a hands-on experience with the SIGMA DS1.
2 comments:
For video I can program real time Foveon to RGB conversion in a few minutes. The Foveon sensor is at best 12 bits per color, more like 10, but for video you only need 8. So 8 bits of RGB gives you 24 bits total. A 24 bit Look Up Table (LUT) is only 16,777,216 entries. A LUT only needs to be 48 Mega Bytes to be able to do a very fast conversion, 64 MB would make programming a bit easier. When you set the white balance the camera would take a few seconds to calculate the new LUT and off you go. When you have such a limited number of possible long calculations it is standard practice to use LUTs in real time applications.
For non-programmers, a LUT is an array of pre-calculated values. In this case the raw data from the Foveon sensor would be packed into a single value and that value used to index into the array. The array would then have the RGB values. Following is an example of how a LUT for doing squares of numbers 0-9 would look:
Index Value
0 0
1 1
2 4
3 9
4 16
5 25
6 36
7 49
8 64
9 81
This one is way to simple to use in the real world, but I hope you get the idea.
I could care less for video, got the 5d2 or 7D for that. I want AA free and IR ability sensor!
I am thinking with no doubt that this will be better image quality than a 5D2?
I do wish the ISO ability was that of today's leading cameras. I think its the limiting factor in a still image SRL, as beening the major change in peoples ability to take photographs. Forget video for now, Live View would be very nice.
Post a Comment