Pages: [1]   Go Down

Author Topic: NEX7, M9 et al. Dumbdownsizing comparison alternative  (Read 1123 times)

rodcones

  • Newbie
  • *
  • Offline Offline
  • Posts: 20
NEX7, M9 et al. Dumbdownsizing comparison alternative
« on: December 20, 2011, 07:55:19 am »

I deliberately started a new thread as I wanted to express the idea I've had for a while which resurfaced when I read the continuing NEX7 review and the M9 comparison.

I haven't been online at all for a week and I read the articles on main page then come to the Forum, so have just caught up on the NEX7/M9 and the Forum threads, where one contributor - uaiomex - has the same idea in the "Proposed comparison between..." thread.

I can understand that the resampling/downsizing can "hint" at a reasonable comparison of differing sensor/pixel density parameters and this is the easiest way to show online in sample images. There is no insult intended by my using the alliterative "Dumbdownsizing" in the title, I just want to express that the methodology is not the best.

Mention has been made of "Real World" comparison.

Here's my idea on that.

Not just the NEX7 and M9 but any camera or capture size - MFT, APS, FF, MFDB -, what should be photographed is a scene where one _diagonal_ of the capture frame is matched exactly with 2 points in the scene.

Then you print the images such that the diagonal is the same length in each print. It would be best if the print size was at least just within A2.

This works better for matching format ratios - APS/FF are 3:2 - and I know, having done some scale pencil and paper tests, that with differing format ratios you will end up with thin triangles not quite having the same parts of the scene in comparison to "standard" matching length and/or height, as done changing a "square on" viewpoint with position or lens size.

I can envisage a large flat A0 or 2x A0 chart on a wall (approx 4' x 3', 6' x 4') as the easiest way to have the largest diagonal with surrounding variable-sized elements on it for meaningful checking of all the formats, but a suitable "Real World" scene with buildings and foliage could be found I'm sure.

The diagonal is more of a common relation to the capture area amongst the format ratios and if there are similar size elements along its length in the scene then that gives a fair run for corner through to central resolution comparison.

Part of my idea using an "A" size chart was that for an A1 or A0 chart one should print the capture image to that size and see how well the elements are resolved in comparison.

There are 2 other considerations about doing the print.

Dealing with a RAW file whether using native processing software or ACR/Lightroom.

One involves doing seperate print where the image is just "enlarged" so the diagonal is fitting inside that A2 or whatever sheet _without_ any sharpening (capture or otherwise) or any PPI change to match the printer. That is, knowing what length/width to get that fixed diagonal within the A2 example, just tell the photo editor to scale the image to the length/width and to hell with what the PPI ends up.

The other is to do appropriate sharpening and ensure the image goes to the printer with a reasonable match to the printer's native resolution - say 180/240/360ppi for Epson and 150/200/300ppi for Canon.

A slight variation in these is to use Qimage which can take the RAW file, or maybe an un-processed/un-sharpened tiff from the native processor, and do its own suitable matching for the printer's resolution.

The last consideration is how to show these results online.

The prints, or parts of the central and corner areas, would have to be scanned or maybe photographed with an MFDB at minimum and then jpegs made available for insertion in  addition to downloadable tiffs for offsite compare.
Logged
Pages: [1]   Go Up