Originally posted by rawr Some firmware suggestions re RAW format saving options for the K-1 ...
In some Nikons like the D810, there are options for saving RAW files in different bit depths - specifically 12/14 bit. 12 bit RAW can shrink files, boost the fps a bit, and boost the amount stored in the RAW buffer by ~20-50%.
12 bit RAW is used in some Pentaxes already (eg K-S2), so it would not be a radical change to see that feature offered in the K-1.
Additionally, some Nikons have the option to save RAW's (12 or 14 bit) with different compression options - either lossless-compressed; lossy compressed; or un-compressed RAW, which allows a range of file size and image quality options.
Both Nikon and Canon also offer options to save RAW files in various image dimensions - eg Large (ie maximum resolution), Medium or Small, which adds more useful options to reduce file sizes, boost buffer and fps, and reduce storage.
I agree that some further options for RAW file quality and size would be desirable - especially with such a high resolution sensor. Personally I would never want to shoot in JPEG, but do sometimes want to reduce the amount of storage space required for non-critical work. Both reduced size and lossy compressed RAW options would be useful for me. I have used Canon's M-Raw format in the past and found it useful, though it's not that efficient (look at the numbers - less than half the pixels but two thirds of the file size).
Having worked with the compressed DNG format available through Lightroom, DNG converter etc, I find this offers a very high quality-to-file-size ratio which is far preferable to JPEG compression, maintaining much higher dynamic range and post-processing flexibility and of course being free from sharpening, noise reduction and other in-camera-processing artifacts. Of course you can convert files in post, but this takes time and you don't get the additional card space while shooting. Perhaps there would be a performance hit by using this kind of compression, but if it's just an option, I don't see the problem.
Either way, yes - size, bit-depth and lossy/lossless options could all be welcome if implemented well (Nikon's long list of combinations is a bit confusing and probably not the best way to do it). If there's a slight performance hit to deal with the compression then I personally wouldn't mind. I would have thought these could be implemented via firmware, though not being a camera engineer, I'm not sure.