downsizing large images

[wmchauncey]

I currently use a 1Ds3 combined with a 180 macro lens and do a lot of stacking and photo-merging using PS CC,
often winding up with 3-5' images at 300ppi.
I'm finding that when I downsize the resulting big images...they seem to be ultra sharp in spite of the fact that the only
sharpening being done is applied to the RAW images in LR (set to 100).  Is this normal and why?     

[Jim Kasson]

I currently use a 1Ds3 combined with a 180 macro lens and do a lot of stacking and photo-merging using PS CC,
often winding up with 3-5' images at 300ppi.
I'm finding that when I downsize the resulting big images...they seem to be ultra sharp in spite of the fact that the only
sharpening being done is applied to the RAW images in LR (set to 100).  Is this normal and why?     

Can be. Depends on the downsizing algorithm. What are you using?

Jim

[wmchauncey]

Image size in PS CC...

[Jim Kasson]

Image size in PS CC...

There are six choices for the resizing algorithm in CC. If you have if set to Automatic, it's probably choosing BiCubic Sharper. Try BiCubic or Bilinear. If you use Bilinear, you might want to sharpen a bit afterwards.

Jim

[wmchauncey]

That tidbit of information never occurred to me...thank you.       

[Jim Kasson]

That tidbit of information never occurred to me...thank you.       

You're welcome. See if one of those works for you. If not, there are levels and levels of complexity and quality for us to try. If you're on a PC, you can use QImage. Let us know if the standard routines don't do it for you.

Nicolas Robidoux and Bart van der Wolf hang out here, and they are true experts on the subject, if you get interested in the technology and the tricks of resizing.

Jim

[Wayne Fox]

I do mostly panos with mfdb, and end up with pretty large files (25-50k pixels wide).  When downsizing for web, I always do so in 3 50% increments. (I’ve built it all into some actions).  The first step asks for a starting size which is my final desired size x 8 using “Fit Image”  (so for a 1000px web jpeg I put in 8000).  I then down size 50% using bicubic, 50% using bicubic sharper, then 50% with bicubic again.  The final image usually gets a very low smart sharpen (91% at 0.3 pixel threshold), but this is decided visually with the image at 100% (200% on my retina display) and can be tweaked or skipped.

The idea came from Jack Flesher, and he tested quite a few different ways (when to use bicubic vs bicubic sharper).  whether this gives me any better image than single downsize I don’t know, but seems to.  Interestingly enough, even if the first fit image requires the file to be uprezzed the quality is still very good and maybe better than a single resize.

[Jack Hogan]

I currently use a 1Ds3 combined with a 180 macro lens and do a lot of stacking and photo-merging using PS CC,
often winding up with 3-5' images at 300ppi.
I'm finding that when I downsize the resulting big images...they seem to be ultra sharp in spite of the fact that the only
sharpening being done is applied to the RAW images in LR (set to 100).  Is this normal and why?     

Good answers above. To complement them, and assuming that you are viewing the same image at two different sizes on your screen, the smaller image will often look sharper (and cleaner and more saturated - but we digress) than the larger one in part because of the physiology of the human visual system.

Granger showed empirically in the 1970s that the spatial frequencies that are most closely related to our perception of sharpness vary with the size and viewing distance of the displayed image. He found that the spatial frequencies that are most representative of the perception of sharpness are those that project detail at between 3 and 12 cycles/degree on the retina.  When you downsize an image you are squeezing detail (cycles) closer together, effectively shifting some of the lower spatial frequencies (what we normally call 'contrast') into the more sensitive 3-12 cycles/degree range on the retina of the viewer.

As you found out, this contrast-compressing effect can be quite surprising and lead to some interesting conclusions, one of which is that for 'standard' applications one should take a close look at his/her lens' performance around MTF90 (!). You can read a bit more on this subject here .

Jack