I've always found the Foveon sensor a bit weird, and felt the sensor's development was always limited by Foveon themselves, and their limited resources. Each pixel gets 3 times more light than the standard pixel on the typical Bayer sensor (assuming they were the same physical size), and yet its high ISO performance is pathetic. Pretty crazy, really. They also didn't manage to increase the number of pixels to something like 8 MP (or 24 MP (8 x 3), by their definition), which would put them on par with 12 MP DSLR bayer sensors because they don't need to use surrounding pixels to interpolate. Or I guess you could think about it the other way around, that a typical 12 MP sensor only gives around 8 MP of actual detail. Either way, Foveon can only match 8 MP camera sensors right now, and that hurt them a lot.
Err... somebody did a comparison to a FULL FRAME Canon EOS 5D (Mark I then), which was rather the pinnacle of extremely expensive digital cameras. The SD14 lost a little in monochrome resolution, but convincingly beat it in terms of capturing subtlety and resolving *color* details.
I'm pretty convinced the X3 sensor's shortcomings are in 1) The sensor size (still slightly smaller than your average camera) 2) No real in-camera noise processing. Well, everything you see after ISO800 on normal cameras have been processed, especially for the Nikons and Canons. (Sony just leaves it there and gets called noisy as well)
1 is more of a minor niggle, 2 is supposedly being helped out in the DP2 and SD15 cameras with a new image processor.
But I disgress. Most people not doing sports or action tracking would hardly touch 1600 and above. Plus, it's not like normal consumer cams still rock at 1600. I've seen ISO1600 images from the 350/400D and they were frankly, awful.
Actually, at best (pure white light), each Foveon pixel *in theory, at 100% efficiency* can get only the same amount of light as a bayer pattern pixel, it just splits it three ways through each layer. With any other color of light, you have varying degrees of light loss through leakage and imperfect efficiency in each layer, meaning that a Foveon pixel receives less light than a bayer pattern pixel, which gets 100% of the photons hitting it for any given dyed surface. So a red bayer pixel gets close to 100% of the red hitting it because nothing else obstructs it. However, if a Foveon pixel is designed with red at the lowest level, it may lose 5% through the green layer and 5% through the blue layer, giving it only 90% of the red light striking the top surface of the sensor.
This correlates with the decreased high-ISO performance from the X3 sensor, which is too bad. It's still a good idea, but as mentioned, they were too small to keep up with the R&D and manufacturing resources of Sony or Canon.
"I'm moving to a small studio and for some reason the cable connection is in an awkward place and I need a way to transmit HD quality video and audio no more than 20 feet away. What is the best wireless HDMI transmitter / receiver for this situation? Thanks!"
The most commented posts on Engadget over the past 24 hours.
Now that we've thrown 'em off the trail, use the form below to get in touch with the people at Engadget. Please fill in all of the required fields because they're required.
I've always found the Foveon sensor a bit weird, and felt the sensor's development was always limited by Foveon themselves, and their limited resources. Each pixel gets 3 times more light than the standard pixel on the typical Bayer sensor (assuming they were the same physical size), and yet its high ISO performance is pathetic. Pretty crazy, really. They also didn't manage to increase the number of pixels to something like 8 MP (or 24 MP (8 x 3), by their definition), which would put them on par with 12 MP DSLR bayer sensors because they don't need to use surrounding pixels to interpolate. Or I guess you could think about it the other way around, that a typical 12 MP sensor only gives around 8 MP of actual detail. Either way, Foveon can only match 8 MP camera sensors right now, and that hurt them a lot.
Err... somebody did a comparison to a FULL FRAME Canon EOS 5D (Mark I then), which was rather the pinnacle of extremely expensive digital cameras. The SD14 lost a little in monochrome resolution, but convincingly beat it in terms of capturing subtlety and resolving *color* details.
I'm pretty convinced the X3 sensor's shortcomings are in 1) The sensor size (still slightly smaller than your average camera) 2) No real in-camera noise processing. Well, everything you see after ISO800 on normal cameras have been processed, especially for the Nikons and Canons. (Sony just leaves it there and gets called noisy as well)
1 is more of a minor niggle, 2 is supposedly being helped out in the DP2 and SD15 cameras with a new image processor.
But I disgress. Most people not doing sports or action tracking would hardly touch 1600 and above. Plus, it's not like normal consumer cams still rock at 1600. I've seen ISO1600 images from the 350/400D and they were frankly, awful.
Actually, at best (pure white light), each Foveon pixel *in theory, at 100% efficiency* can get only the same amount of light as a bayer pattern pixel, it just splits it three ways through each layer. With any other color of light, you have varying degrees of light loss through leakage and imperfect efficiency in each layer, meaning that a Foveon pixel receives less light than a bayer pattern pixel, which gets 100% of the photons hitting it for any given dyed surface. So a red bayer pixel gets close to 100% of the red hitting it because nothing else obstructs it. However, if a Foveon pixel is designed with red at the lowest level, it may lose 5% through the green layer and 5% through the blue layer, giving it only 90% of the red light striking the top surface of the sensor.
This correlates with the decreased high-ISO performance from the X3 sensor, which is too bad. It's still a good idea, but as mentioned, they were too small to keep up with the R&D and manufacturing resources of Sony or Canon.