Commons:Featured picture candidates/File:JohannMatthesonCFritzsch.jpg

File:JohannMatthesonCFritzsch.jpg, not featured

Voting period is over. Please don't add any new votes.Voting period ends on 24 Apr 2016 at 13:42:21 (UTC)
Visit the nomination page to add or modify image notes.

Featured picture candidates/File:JohannMatthesonCFritzsch.jpg Commons:Featured picture candidates/File:JohannMatthesonCFritzsch.jpg

• Category: Commons:Featured pictures/People
•  Info created by C. Fritzsch - uploaded and nominated by Rettinghaus -- Rettinghaus (talk) 13:42, 15 April 2016 (UTC)
•  Support -- Rettinghaus (talk) 13:42, 15 April 2016 (UTC)
•  Support - Good scan, nice print. -- Ikan Kekek (talk) 06:40, 18 April 2016 (UTC)
•  Support, this got me to read dezoomify and test it. The FPC entry was AWOL for two days. –Be..anyone 💩 07:16, 18 April 2016 (UTC)
•  Support 😄 ArionEstar 😜 (talk) 01:35, 19 April 2016 (UTC)
•  Comment No color-space metadata and no embedded color profile: Windows and Mac web browsers treat colors randomly The impact of this is probably lower for a B&W image like this. Notwithstanding, it is recommended to export as sRGB and embed the color profile in the EXIF data. -- Slaunger (talk) 20:36, 21 April 2016 (UTC)

  B&W should be linear, for RGB somebody (incl. me) could convert it to PNG and add a sRGB chunk (suggest an intent, please). –Be..anyone 💩 02:49, 22 April 2016 (UTC)

  • I have tested one of my b&w photographs (saved as a colour JPG in sRGB) and removed the profile. This does produce a small change in tonality when viewed by one of the defective browsers (i.e., all of them apart from a modified Firefox) on my wide-gamut monitor. Saving it as a greyscale JPG (i.e. only 1 colour channel rather than 3) seems to make no difference. The image above is greyscale JPG. I don't think Be..anyone is correct that it should use a linear luminance rather than one with a gamma, because that results in only 8-stops of dynamic range rather than 12. Linear is fine for intermediate files with large bit-depth (such as HDR formats, or camera RAW files), or working on computer generated images. See my comment below -- I think the subtle change caused to some viewers when a profile is missing is only really important if the image was produced with care and contained fine tonal range in a high-quality b&w photograph, which isn't the case here. -- Colin (talk) 18:32, 24 April 2016 (UTC)
    • 8 or 12 sounds bad, but 8 or 16 bits offer more than 8 steps. The TIFF specification suggests linear for grayscale. –Be..anyone 💩 18:45, 24 April 2016 (UTC)
      • Current DSLRs manage about 12-14 bits linear (i.e. 12 stops of dynamic range). So a 16-bit linear file is not unreasonable. I'm not an expert on TIFF but I believe it is just a container format and there are many possible image encodings (for example, Adobe DNG raw files are actually TIFFs, but quite different to the TIFFs one might create with Photoshop or GIMP. -- Colin (talk) 19:06, 24 April 2016 (UTC)
        • The TIFF spec. explains the issue. JPG is presumably hopeless instead of lossless, PNG works. –Be..anyone 💩 19:57, 24 April 2016 (UTC)
•  Support Daniel Case (talk) 03:48, 23 April 2016 (UTC)
•  Comment The source image (as claimed on the image description page) is a colour photo with aged yellowed paper and brown ink. This nominated image is not only converted to b&w but also the contrast increased to turn the ink fully black. I fear some subtlety is lost. It certainly doesn't look subtle. -- Colin (talk) 18:32, 24 April 2016 (UTC)

Confirmed results:

Result: 5 support, 0 oppose, 0 neutral → not featured. /Yann (talk) 22:02, 24 April 2016 (UTC)