Timeline for Will scaling down incrementally hurt quality?
Current License: CC BY-SA 3.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 10, 2011 at 20:21 | comment | added | Jari Keinänen | I tested some incremental resizing with ImageMagick: first (1 * 25%) vs (50% * 2) then (1 * 25,24%) vs (99% * 137). Once vs twice — there was no much of a difference; but once vs 137 — the effects were emphasized. (Original and all the intermediate files were 16-bit TIFF images; converted to JPEG as a last step.) I know, ImageMagick's settings could be tweaked a lot (better), but the idea persists: few times won't hurt and one probably shouldn't really worry about it, but in the long run downsizing is best done in one step. | |
| Dec 10, 2011 at 20:17 | comment | added | Jari Keinänen | Good guess is a good guess; so incremental downsizing is a good guess of a good guess of a good guess ad nauseam. Good point is that "the usual bicubic algorithm often benefits from a little bit of sharpening" which conversely mean the usual "resizer" will do some sharpening/blurring. Incremental resizing then means sharpening the sharpening or blurring the blurring. Also, a good sanity check on quality is always "worth a reputation". | |
| Dec 10, 2011 at 19:59 | history | bounty ended | Jari Keinänen | ||
| Dec 9, 2011 at 1:35 | comment | added | Farray | +1 Lots of good bits: "Quality is a measure of how convincing and pleasing the end result is, not how "accurate" it is. ", "accurate ... is not what you're looking for unless you're a scientist", " the slight difference between "degraded image quality" and "slightly less degraded image quality" has no practical value" | |
| Dec 4, 2011 at 8:37 | comment | added | Alan Gilbertson | @muntoo: "If you charge by the hour..." is a mild sarcasm that's common in discussions about design workflows, to emphasize that inefficient and wasteful uses of work time are both inefficient and wasteful. | |
| Dec 4, 2011 at 8:33 | comment | added | Alan Gilbertson | Bicubic doesn't blur by intention, but a slight softening is a common artifact. When it occurs, a modest sharpen fixes it. | |
| Dec 4, 2011 at 8:33 | history | edited | Alan Gilbertson | CC BY-SA 3.0 |
added 1117 characters in body
|
| Dec 4, 2011 at 7:13 | comment | added | Mateen Ulhaq | BTW, isn't the perceived 'blurring' done by downscaling algorithms on purpose? | |
| Dec 4, 2011 at 7:11 | comment | added | Mateen Ulhaq | "If you charge by the hour, by all means go 1% at a time." ...Unorthodox? :) | |
| Dec 4, 2011 at 0:37 | history | answered | Alan Gilbertson | CC BY-SA 3.0 |