Need optimal RAW settings explained

I always shoot in raw to have the most available data for post processing. The bit of confusion comes when selecting the optimum or best raw setting to shoot in, MY options are:

NEF (raw)

Compressed 12-bit: 31 MB

Compressed 14-bit: 38.1 MB

Uncompressed 14-bit: 77.3 MB.

 

right off the bat I assume that Compressed 14 bit gives a more detailed data that Compressed 12 bit and the storage is only slightly more in MB's.

Uncompressed 14 bit requires twice the storage that compressed 14 bit.

 

How much of a trade off is there between each. Is twice the storage worth it or is just 12 it compressed enough?

For me, 14 bit, uncompressed RAW is the way to go. Storage is cheap. Trying to do it over again is expensive or impossible. If you need something to put out immediately, shoot JPEG in parallel with the raw files. Some modes (e.g., silent shutter in a Sony A7/A9) are truncated to 12 bits, but still available uncompressed.

 

Image data is mostly random, so 50% compression has to be lossy. What kind of artifacts depends on the method of compression and subject.

so 50% compression has to be lossy

Nikon offers lossless compressed on some cameras (something Sony needs to implement sooner rather than later) and for the D850, the file sizes are 92MB for 14-bit uncompressed and 51.6MB for 14-bit lossless compressed; 14-bit compressed is 43.8MB. Very little reason to choose compressed over lossless compressed and I personally have never seen a difference between lossless compressed and uncompressed, so no need to waste storage space.

 

To answer the OP's question: uncompressed is the safe option. 12-bit vs 14-bit compressed isn't enough difference in file size to matter; and in most cases 12-bit vs 14 bit isn't something that can be distinguished anyway. What to choose pretty much depends on what you are shooting, how much processing is going to be involved and what the images are used for. For many applications 12-bit compressed may not be limiting at all.

As far as bit depth, I would take more over less. Whether it makes a practical difference will depend on a number of factors, including the scene, ISO, and what processing you do. You may find there's more useful data in shadows with a greater bit depth. As you point out, it's a relatively small file size difference for the extra two bits.

 

For compression, I don't know whether your camera does lossy or lossless compression. I'm a Canon shooter, and Canon's RAW files use lossless compression. If your camera uses lossy compression, well, it's up to you, but to me, lossy compression would erode one of the benefits of RAW files (that none of the data will be thrown away between when you take the shot and when you start editing it), so I personally wouldn't use it.

For compression, I don't know whether your camera does lossy or lossless compression.

The OP mentions NEF which indicates Nikon. If his camera (I assume it's a consumer model) doesn't state "lossless" than the compression is indeed lossy. Never understood why Nikon doesn't offer all options on all cameras.

I have never observed any differences between Sony compressed and non-compressed RAW images, but there have been examples published on DPReview and other places. While the uncompressed RAW files are large (42 MP, 85 MB), I still get nearly 700 images on a 64 GB memory card. It's scary when you think it takes 3 BD discs to backup each card, but I continue to do that after suffering several drive failures. RAW files are my digital negatives, and I still keep the real negatives too.

Somewhere in internet was once study of nikon uncompressed nef, loslessly compressed nef, lossy compressed nef. The raw compression seemed to operate mainly in dark shadows and lost higlights. There was shown extremely processed images that managed to show some advantage to uncompressed nef sourced images. IIRC the losslessly compressed sourced image was practically as good as uncompressed sourced image.

 

Older cameras usually have 12-bit raw and I usually shoot 12-bit uncompressed nef with them. New cameras from consumer midrange upwards seem to have also 14-bit raw and I usually shoot 14-bit uncompressed nef with them. With 16Mpix camera this results 36MB per image. 36MB is about half of size of color film scan and image quality is arguably better.

 

I have lurking interest of shooting incamera jpeg fine small, but since modern computers are so good and I shoot so few frames, there has been no practical reason to do so.

 

I have also tried inkjet printing and that process seems to like the more image information is available.

Any data word larger than 8 bits is stored as 16 bits, with some bits blank. Words are sized in whole byte (8-bit) increments.