D300 vs D200 compressed RAW

On an newly formatted 8GB card, I just noticed selecting Compressed Raw on a

D300 gives me an initial estimate of 541 shots vs 480 on the D200.

 

Doesn't make sense to me since the D300 is 12mp vs the D200's 10mp. Lossless

Raw Compression on the D300 gives an estimate of 394 shots - I would expect that

to be less.

 

Either the D300 is more aggressive with its compression or the way it estimates

is different. I've always felt the D200 was conservative when estimating # of

photos that will fit in the available space on a card.

On the D200, an 8G CF card should give you approximately 480 UNcompressed RAW files. Apparently the D200's remaining frame estimate is so conservative that it initally assumes no compression is achievable at all.

i've found what the manual states and what happens in real life are two different things (which is probably why there is no "human" owner's manual :)

 

much depends too on the filesystem formatting, more specifically the cluster size. to format with an inefficient cluster size will result in wasted space on the card (or hard disk in the case of a computer).

Since there is a range of file sizes whenever compression is used, it is tricky to have the camera guess how many images will be left, and best if the camera's estimate is conservative.

 

Checking some folders full of recent D200 and D300 images, I see the following:

 

D200 compressed = 7.1 to 10.9 MB (about 8 MB average)

 

D300 12 bit lossless compressed = 11.2 to 14.2 MB (about 12 MB average)

 

D300 14 bit lossless compressed = 15.2 to 16.8 MB (about 15 MB average)

 

Normally I do not apply compression to the D200 files or D300 files, but shooting my niece's wedding it made sense to use compressed NEF to avoid going through too many CF cards and I am comfortable with Nikon's assurance that lossless compressed on the D300 is just that -- lossless. For day to day shooting, I am happy that my D300 12 bit lossless compressed files are smaller than my D200 uncompressed files and my D300 14 bit lossless compressed files are about the same size as the D200 uncompressed files.

Is the D300's lossless RAW compression really lossless? If so, what are the advantages, if any, of using the larger uncompressed format?

Lossless means lossless, no loss at all. Therefore, there is indeed no technical reason to use uncompressed RAW on the D3 and D300.

But as we have discussed recently, for example, some lawyers may exploit this issue in court; therefore, in some situations, you may want to not compress anything just to avoid those silly arguments.

Thank you Shun. Those pesky lawyers!

There's one other compressed/uncompressed issue. If, somehow, there's a single bit (or single byte) error in a compressed file, it can render the entire portion of the file following the error as unreadable. The same error on an uncompressed file will produce a single "hot" pixel.

 

Then again, in this day and age, single bit or single byte errors are comparatively rare. It's much more common to lose whole files ;)

"If, somehow, there's a single bit (or single byte) error in a compressed file, it can render the entire portion of the file following the error as unreadable."

 

Interesting. I have lost one file so far and it was a 14 bit lossless compressed one. It represents too small a percentage to become alarmed yet and is most likely a card issue; but this makes me wonder if compression might have played a role, either by itself or in conjunction with an admittedly older card that I had in the camera.

 

On the other hand, the advantage of compression on the D300 is that in addition to using less card and archival space, it allows for faster processing of files from the buffer; on the D200 compression made no difference to the buffer when shooting RAW.

Some comments -

 

There are fewer bits to go bad if compression is used. So compression could make your data safer.

 

One bad bit, in the wrong place, could potentially cause an entire card to be lost, let alone just one file. For this sort of error, there is a high probability that data recovery software can save the images on the card.

 

One bit errors are actually unlikely. With error check summing techniques, it's more likely an entire sector is lost due to hardware failure. As simple a device as they appear, memory cards are incredibly complicated.

 

I'd bet fewer images as a percentage of total shots taken get lost with digital than with film.

 

If you really want to be paranoid, get a dual card camera. In the Nikon world, this means the D3.

 

I'm curious about what kind of circumstances caused Anthony's single file loss. Maybe the bad luck of a sector on the card failing and automatically remapping, but with the data loss?