Digital BW, The Print

Michael King wrote:
> Txs Alan as I was scratching my head trying to figure out what was going on.
>
> Also Sarah you can download a resolution test image from this link that will
> show you what resolution you are really printing at.
>
> http://www.ddisoftware.com/qimage/quality/
>
> Mike
>   
I think there must be a slight difference between the 2400 and the 4800 
drivers -- I couldn't get that particular option to appear whatever I 
did. However, I reproduced my printing procedure with the Qimage test 
image as you suggested, and it seems that the r2400 was printing at 
720ppi after all. You can clearly see the difference between the 720 and 
the 360ppi tests, particularly in column D, so I think it's pretty much 
proven. One thing that I would say, though, is that in column A, you 
can't actually resolve adjacent lines, so whilst the ppi is clearly 720, 
I have my suspicions that the real optical resolution as seen in the 
final print is probably closer to 360, but if you're printing from an 
image that's come from a Bayer pattern sensor, printing at 720 will 
match this pretty much exactly. Since my images aren't 
Bayer-interpolated at all, it might make more sense for me to be 
resampling the image exactly 2:1 in each axis, so the Nyquist limit of 
the image will match the Nyquist limit of the printer. Hmmm... I shall 
have to try that...

Actually, I've probably answered my own question here -- adjacent black 
and white lines are actually *exactly* on the Nyquist frequency, so 
clever DSP algorthms responsible for the dithering are probably going to 
brick wall below that frequency in order to avoid aliasing (i.e. visible 
pixellation). Since the frequency limit of an image sourced from a Bayer 
sensor is actually f/4 for red and blue and a bit more for green, this 
should all be comfortably under the true Nyquist limit.

[s]