Re: Resolution and banding

[Tom Sawyer <tom@liveinbalance.com>]

This is a great thread you folks have started as its starting to take some 
of the mystery out of the Epson printing process. Am I correct that the 
printer puts one of the six inks in one spot on the page? I.E. all 
non-CcYMmK colors are actually dithered (Saurat lives)?

Rumors of earlier discussions passed on to me could be summarized as "any 
image resolution of density higher than 240ppi has no affect on the printed 
image". I'm using 720ppi wXh as these were the settings the pro-photog from 
whom I purchased the printer was using. I'm a great believer in the 
principle of inertia - if it aint movin, don't push it (I'm also lazy). Is 
this true or simplistic?

Tom Sawyer

At 01:03 PM 6/1/02 +0100, you wrote:
>In article <3CF84C65.43C1821E@bway.net>, SKID Photography <skid@bway.net> 
>writes
>>Kennedy McEwen wrote:
>>
>>>In article <ad896s+klsn@eGroups.com>, igorw2001 <iwesdorp@chello.nl>
>>>writes
>>> >Hello.
>>> >
>>> >I read somewhere that banding can occur if the image resolution is
>>> >NOT a factor of the printer resolution.
>>> >
>>> >e.g., Printing an image with a resolution of 457 dpi could cause
>>> >banding, that would dissapear if you printed at this image at 360 or
>>> >720 dpi.
>>> >
>>> >Can this be true?
>>> >
>>>Yes, it can be true.
>>>
>>>No, it isn't true - for Epson printers.
>>>--
>>>Kennedy
>>
>>It is my understanding that the Epson printer printer actually 
>>re-structures the
>>incoming resolution to an integer of 360 dpi, so setting an incoming 
>>resolution
>>of 457 (or whatever), the Epson driver just changes it (interpolates it) to
>>Epson's native resolution.  And further, that Epson does it *very* well.
>>
>>Also, in the 2 plus years that I have been on this and other printer lists, I
>>have never heard of 'banding' for odd input resolutions.
>Your understanding is not quite complete.  The Epson printers certainly 
>downsample the incoming data, but not to the closest sub-multiple of 
>360ppi.  It is downsampled to the lower of the horizontal and vertical 
>resolution limits.  So a 1440x720dpi printer has its data resampled to 
>720ppi, at 360dpi printing the data is resampled to 360dpi.
>
>Upsampling is NOT implemented directly, but is a process of the dot 
>placement algorithm which Epson uses, which redistributes the quantisation 
>noise in the image to retain as much detail as possible.
>
>Thus, in your example of a 457ppi image printed on an Epson printer, if 
>the printer is using 720ppi, it will not be directly resampled and the 
>full detail from the original resolution is retained.
>
>You can check this, as I have on several occasions, by the use of 
>synthetic test patterns and assessing exactly where the output spatial 
>frequencies are aliased to.  Providing that the input test pattern is 
>lower than the limiting resolution of the dot pattern, there is no 
>resampling and only the noise spectrum in the image is reshaped.
>--
>Kennedy
>Yes, Socrates himself is particularly missed;
>A lovely little thinker, but a bugger when he's pissed.
>Python Philosophers
>-
>Turn off HTML mail features. Keep quoted material short. Use accurate
>subject lines. http://www.leben.com/lists for list instructions.

-
Turn off HTML mail features. Keep quoted material short. Use accurate
subject lines. http://www.leben.com/lists for list instructions.