FS7: Buying Shogun Inferno; is the XDCA extension really necessary?

[Mitch Gross]

My right rack went so far south, it ended up north.

But, if you have a few seconds to humor me - is it not possible to convert an 8-bit footage into a floating "fake" 10-bit? I believed I read a post of yours when you referred to something in that vein as a "9-bit". Or is it something usually done, if it's done at all, neither in camera or a recorder but in post, with plenty of processing power and time?

You can do this but only when going from a higher resolution to a lower resolution. And we offer it on the Odyssey7Q+ -- the ONLY recorder in the world to do so (why on earth would anyone choose something else ;-P). Take an UHD 8-bit signal and reduce it to HD. Simplistic math says that four pixels become one pixel (you actually do a much more complicated process for the best image). But don't just think resolution, think about the color. There are four references for color info. Again, simple math just averages them and creates a new color number. But if you want to utilize that color information with greater precision, you average it out using more decimal places to get a more descriptive answer. That's increasing bit depth.

Think of it in this overly simplified way. I did all the math in Base 10 because I didn't want to go there for those who have not dealt with other systems since high school. Anyway, four original pixels in 8-bit color:

R100 G100 B100
R110 G110 B110
R120 G090 B105
R115 G095 B102

Sums of color channels:
R445 G395 B417

Averaged pixel in 8-bit (divide by 4, 3 numerals):
R111 G099 B104

Averaged pixel in 10-bit (divide by 4, 4 numerals):
R111.3 G098.8 B104.3

As you can see, the 10-bit numbers are more precise, and that precision is not "invented" but in fact comes from the data of the original higher resolution 8-bit numbers. A lot of 8-bit data can yield a smaller amount of 10-bit data. Is it the same as having an HD sensor that captured data directly in 10-bit? Not quite, but it is certainly more precise than just capturing in 8-bit, which is why I refer to it as "pseudo 10-bit."

In a camera such as an A6300 or an A7S, the processing architecture is completely 8-bit. There's only so much room in these little cameras and so they can't utilize anything other than 8-bit for the math. So when they sample an UHD image down to HD, they use 8-bit color info and through the rest of the data away. The Odyssey7Q+ has greater processing space available for this operation, so it can do the conversion instead and capture the additional info. That's why we refer to it as SuperSampling, because you get more from it than a simple downconversion. The Odyssey7Q+ offers SuperSampling in two modes: in FS RAW going from a 4K RAW signal to either a 2K or HD ProRes recording, and when taking a DCI4K/UHD video signal over HDMI and recording it as 2K/HD ProRes.