I have seen Keith's video and can only speak from my personal experience that using 720 ppi instead of 360 does not visibly improve prints, except perhaps under a microscope or where detail of the file would otherwise be significantly sacrificed.
How about posting the full original and let folks judge for themselves.It's resampling when you upscale that makes the big improvement.
This is an iPhone image at the original 96 DPI (I think it's 96) and then 300 DPI. Obviously it's zoomed way in, and I should add that I'm not doing traditional photography.
Before upscaling
After upscaling to 300 DPI with resampling
What I posted makes the point very clearly if you view those pix full size, and that's just at screenshot resolution.
For those interested, I made some photos of the 360 PPI and 720 PPI file under a 8x zoom magnifying glass. What you are looking at is about 12x12 mm on canvas. Not the sharpest photos and there is a little blur and warping on the edges of the photo due to the lens. But it’s good enough to get an idea of the difference (or lack of) between the two.
Plus it's not a very good example.
What appears to be sharpening artifacts are definitely more prominent in the 720 dpi rendering. Otherwise not enough difference to write home about.
This might not be such a good example after all. As the photo was shot by hand on my iPhone with handheld lighting and it could very well be that the second photo was taken a little bit closer than the first resulting in the camera not focussing correctly. I need to do a comparison with a camera that can do proper macro on a tripod with consistent lighting.
Nope. The printer software takes the image and scales it to the size you requested. This can result in rounding errors. My spreadsheet tells me how to scale the image before I print, avoiding rounding errors.
How do you avoid ‘rounding errors’ if the printer, based on your paper/quality selections in printer settings, is going to ALWAYS insist on a specific PPI (for Epson 360, 720, 1440). If you do not send it what it wants, it will do its own conversion.
What are you talking about. It's a perfect example. You resample and the edges are smoother.Plus it's not a very good example.
You avoid rounding errors by making the size adjustments to give it exactly what it wants before you print. If you don't make the adjustment, then the driver's conversion can lead to rounding errors. Sometime you can't see the errors, so it makes no difference. But sometimes you can see the effect.
So, you resize to what the printer driver wants….360, 720, 1440 (for Epson).
Without seeing the entire photograph, it's impossible to judge the significance, which doesn't appear to be much. Can you produce a better example from a complete photograph? I'm not arguing for its own sake, but rather pointing out the insignificance of up-rezing beyond 360 dpi on quality. I also tried going from 360 to 720 dpi years ago and found the practical improvement to be from none at all to insignificant. I print at 12x18 to 16x24, with most native resolution at 42 mpxl and 24 mpxl.What are you talking about. It's a perfect example. You resample and the edges are smoother.Plus it's not a very good example.
I'm not talking about the export DPI, I'm talking about resampling.
It's really annoying when people argue for the sake of arguing.
Everything in the printer prints at the mechanical resolution of 1440. If you send it an image at 360, each pixel is printed by a 4x4 array of dots. The larger the array, the more effective shades you get. So it isn't what the printer wants, it boils down to how the driver is going to build the dot arrays. The driver defaults to 1440 (1:1 ratio), 720 (1:2), 360 (1:4) and 180 (1:8) ppi. Most people uses these settings, but some drivers let you pick any ppi that you want. An Epson 1440 dpi using a 3x3 array is 1440/3=480 ppi. That isn't standard, but you can do it. 5x5 is 288 ppi. 6x6 is 240 ppi. 7x7 is 205.714 ppi, which will introduce rounding errors on every pixel, so skip that.
)
