Digital BW, The Print

Hi Bob,

> Surely Nearest Neighbour
> interpolation just duplicates pixels (or should I say
> quadruplicates?).

Nearest neighbor takes the nearest neighbors, adds them up, then divides by
the number, and basically gives an average.  It certainly isn't an optimal
interpolation algorithm, and in fact, is the most basic interpolation
algorithm I can think of.

> Which is why QImage claims to produce better prints by
> interpolating up to 720 ppi at print time using Lanczos
> interpolation, thus
> avoiding the nearest neighbour interpolation in the Epson driver.

Run some images and let me know if you see a difference.

I'd use that by doing as I've suggested, let the PPI fall where it may, not
letting PS interpolate, then sending it to what ever your favorite
interpolating program is.

Now, realize that is only for the Epson driver, which quite a few people
don't use when printing B&W.  The Cone driver is entirely different.
Whether the Cone driver would benefit at all from this, I doubt.

I'm still not quite convinced of this claim, as it seems entirely
unnecessary...when using random/stochastic dither algorithms, like the Epson
supposedly does.  These dither algorithms don't use fixed cell size, so I'm
not clear of what the benefit of resizing the image is.

Regards,

Austin