8bit, 16bit, 16bit linear...

[svhardesty]

Can someone give me a breakdown of the differences between 8bit,

16bit, and 16bit linear?

I am scanning B/W negs on my new Dimage 5400 II, and I am thinking

that 8bit is sufficient...

16bit linear produces a negative image...whats that all about?

Thanks for the help...

[fk319]

Bit is your color depth. 1 bit, is Black/White, 2 bit is Black/DarkGrey/LightGrey/White. 8 bit would give you 256 shades of grey. This would be ok if you had a very busy photo, liket trees, but you would see blocks of grey on a persons face. If memory serves me correctly, the eye can distinguish 10-bit. Often 12-bit is used, 4096 colors. For larger prints where you can adjust your eye for a darker and lighter areas, 16 bit is used. (This is the same way your pupil changes from looking at the bright sky to looking at the land.)

[jpursley]

This is a bit confusing in the scanner software, as there are two different elements tied to that one drop-down list.

<br><br>

Issue #1: Bit Depth - You can think of 8-bit images as the lowest common denominator of the digital photography world. Printers and jpegs use 8-bit images, so after you're all done with your digital darkroom work, your output will likely be an 8-bit image.

<br><br>

The more you manipulate the levels, curves, hue, etc of your image, the more likely you'll benefit from working with a 16-bit file. With 16 bits of data, there are a lot less rounding errors with each adjustment, which can really add up and cause posterization and loss of fine transitions in 8-bit images. Once you're done editing, you'd then output an 8-bit copy of your 16-bit master file for printing or web posting. Not all functions are available when using 16-bit images, and the file size is double that of an 8-bit image.

 

<br><br><br>

 

Issue #2: Scan Mode - 8-bit and 16-bit mode result in the scanning software adjusting the exposure and color balance to look decent in most photo editing applications. The other mode, linear, does not adjust the image, but rather just writes the unmanipulated data collected to disk. Linear scans will look horribly dark until you assign the 'Minolta 5400 Posi (Linear)' profile to them, then convert to your working profile.

<br><br>

I use 16-bit linear because it yeilds consistent results. I got frustrated with 8-bit and 16-bit modes because I'd get different exposure and color results across multiple scans of similar scenes. All of the above is based on my experience in color (mostly slides) only. I've found the extra steps of linear scanning to be worth it, but others have reported other results.

<br><br>

<a href=http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=009mEH>http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=009mEH</a> has links to more workflows for this scanner, including a negative workflow.

[mendel_leisk]

Jeremy, assuming you're talking about Minolta Scan Utility, I don't think 16 bit linear output is by definition locking exposure. Unless of course you are setting exposure manually. If you use auto exposure and 16 bit linear output with a number of frames of the same scene, there will still be variations.

 

BTW, is my usual output as well.

[david-m]

Can I just ask then, if you have converted and edited (in PS) a colour photo as a 16 bit

image, is it best to then reconvert this back to 8 bit before either sending to your own

inkjet printer or giving the file to a commercial printer?

 

Thanks

 

David

[jpursley]

Whoops...you are correct, Mendel - I forgot I also have autoexposure turned off in the Minolta Scan software.