16 bit vs 8 bit with banning printing problems

[donna stew]

I sometimes have banning problems on my ipf8400 canon printer. I am currently editing in 8 bit. My question is should I be editing in 16 bit to solve the banning problems? If so do I then have to print in 16 bit or can I print in 8 bit?

[William Michael]

Probably a typo and you mean "Banding" (i.e. "Colour Banding").

This issue is not to do with the Post Production editing and changing the bits will not assist.

 

The issue is to do with the Printer - specifically the Printer's Ink Heads. You need to check the flow.

 

Click on this sentence to go to a link to the Canon Reference Sheet specific to this issue and this Printer.

 

WW

[digitaldog]

Yes, edit in high bit (what some call and often isn't "16-bits").

http://digitaldog.net/files/TheHighBitdepthDebate.pdf

Most printer drivers only accept and need 8-bits of good data.

[Ed_Ingold]

The more common type of banding is caused by the print head. Clogged or partially clogged print heads cause banding comparable in width to the print head. It is more likely to occur in "economy" print modes than high resolution mode (where the head passes overlap).

 

<br><br>Another kind of banding occurs in areas with a very slight gradient, such as blue sky, where you see the transition from one level of ink density to another. This is more common in monitors than printers, which generally have a greater bit depth than screens. It would be unusual to attribute banding to the limits of an 8-bit/channel image, but anything is possible.

 

<br><br>A high resolution (bit-wise) image may produce out-of-gamut areas when translated to a lower bit resolution device. Banding near the OOG region would be a possible consequence.

[Didier Lamy]

If you can, whatever the printer, always edit in 16 bits. Just look at an histogram after adjusting light parameters. No comb in 16 bits

[DeanJones]

Bit is a PC expression for information stockpiling. It can just contain two qualities, regularly 0 or 1. 8-bit basically implies the information lump is 8 bits altogether (or 2 to the energy of 8, as each piece can be either "1" or '0'). This takes into account numeric qualities running from 0 to 255.

[Didier Lamy]

(a) is a 5600x3750 px 16 bits greyscale 4000 ppi scan. (b) is the same photo + processing for contrast in/out = 75/50 and 185/200. © is the same photo, turned into 8 bits, and + the same processing.