Resolution measurements - TZ30 (ZS20) - Many images

Started Apr 3, 2013 | Discussions thread
J C Brown
Senior MemberPosts: 1,388
Like?
Re: Resolution measurements - TZ30 (ZS20) - For kkardster
In reply to kkardster, Apr 4, 2013

kkardster wrote:

Regarding your yellow woes, perhaps part of the reason it measures poorly is its yellow-on-white color combination and the difficulty that that low contrast combination presents to the human eye.  Perhaps half of the yellow Es should be displayed against a background with better contrast to increase the chances of identifying its minimum clarity.  (Of course, having to print both the foreground and background rather than just the foreground can lead to bleeding or encroachment - depending on how good your printing process is.)

-- hide signature --

Bruce
You learn something new every time you press the shutter

Bruce

Thanks for taking the time to consider the results of my tests and for passing on your thoughts about the much lower resolution values for the yellow Es and your suggestions in relation to increasing the contrast for these letters by using a grey background.

Having endeavoured to take a purely objective approach to such problems throughout my career I don’t regard these results for the yellow Es as “woes” but simply as limitations in need of a logical explanation.

Having been aware for several years of very similar behaviour in several different camera models I have given a good deal of thought to the various factors which could influence the lower values for the cyan and yellow Es.

In addition to the difficulty which you describe of the low contrast that the yellow on white combination presents to the human eye I have considered several other factors. These include the accuracy of the printed image which in addition the colour accuracy of all the letters including yellow also involves the whiteness of the paper on which the chart has been printed.

To achieve that with the highest possible accuracy when constructing the test chart I used combinations of the Zero and 255 settings for Red, Green and Blue in Photoshop Elements to set the three primary and the three secondary colours. As I don’t have access to professionally calibrated analytical instruments I have no means of measuring the accuracy of the printed colours and have had to rely on my own visual assessments.

The accuracy of the colours in the recorded image may also be affected by the properties of the light used to illuminate the chart, the accuracy with which the white balance has been set and the accuracy with which the grey level has been adjusted by means of the EV setting.

Assuming that as a result of the care taken with all of these factors and settings the image projected onto the sensor is reasonably accurate the accuracy of the recorded image may be affected very significantly by the properties of the sensor, the position of the individual Es in relation to the arrangement of the colour filters of the Bayer Matrix and the accuracy of the demozaicing software used to derive the recorded image.

As yellow is a 50:50 combination of red and green and cyan is a 50:50 combination of blue and green, the resolution with which the cyan and yellow letters is recorded is likely to be adversely affected by their positions in relation to the positions of the red, green and blue filters on the sensor.

In spite of these limitations I believe that although my new colour resolution test cannot provide absolute values it does provide a useful means of making quantitative rather than qualitative comparisons of cameras and camera lens combinations.

Thanks again for your interest and for your helpful comments.

Jimmy

-- hide signature --

J C Brown

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow